Tropical Rainforest Research Current Iss

TROPICAL RAINFOREST RESEARCH - CURRENT ISSUES
MONOGRAPHIAE
BIOLOGICAE
VOLUME 74
Series Editors
H.J. Dumont and M.J .A. Werger
Tropical
Rainforest Research -
Current Issues
Proceedings of the Conference
held in Bandar Seri Begawan, April 1993
Edited by
D. S. EDWARDS
W.E.BOOTH
and
S.C.CHOY
University of Brunei Darussalam
KLUWER ACADEMIC PUBLISHERS

DORDRECHT / BOSTON / LONDON
Library of Congress Cataloging-in-Publication Data
Tropical raInforest research current issues / edited by D.S. Edwards

and W.E. Booth and S.C. Choy.
p. cm. -- (Monographlae biologicae ; v. 74)
Includes bib I iographical references and index.
1. RaIn forests. 2. Rain forest ecology. 3. Rain forest

conservation. 4. Rain forests--Economic aspects. 5. Biological
diversity--Troplcs. I. Edwards. David S. II. Booth. W. E. (Webber
E.) III. Choy. S. C. (Satish C.) IV. Series.
OH86.T76 1996
574.5'2642'0913--dc20 96-14529
ISBN-13: 978-94-010-7255-7 e-ISBN-13: 978-94-009-1685-2

DOl: 10.1007/978-94-009-1685-2
Published by Kluwer Academic Publishers,

P.O. Box 17,3300 AA Dordrecht, The Netherlands.
Kluwer Academic Publishers incorporates

the publishing programmes of
D. Reidel, Martinus Nijhoff, Dr W. Junk and MTP Press.
Sold and distributed in the U.S,A. and Canada

by Kluwer Academic Publishers,
101 Philip Drive, Norwell, MA 02061, U.S.A.
In all other countries, sold and distributed

by Kluwer Academic Publishers Group,
P.O. Box 322, 3300 AH Dordrecht, The Netherlands.
Printed on acid-free paper
All Rights Reserved

© 1996 Kluwer Academic Publishers and copyright holders
as specified on appropriate pages within.
Softcover reprint of the hardcover 1st edition 1996
No part of the material protected by this copyright notice may be reproduced or

utilized in any form or by any means, electronic or mechanical,
including photocopying, recording or by any information storage and
retrieval system, without written permission from the copyright owner.
CONTENTS
Preface ix
BIODIVERSITY
N.E. Stork: Tropical forest dynamics: the faunal components
R. L. Kitching and J. Zalucki: The biodiversity of arthropods from Australian
rain forest canopies: some results on the role of the tree species 21
Gaden S. Robinson and Kevin R. Tuck: Describing and comparing high
invertebrate diversity in tropical forest - a case study of small moths in Borneo 29
Axel Dalberg Poulsen: The herbaceous ground flora of the Batu Apoi
Forest Reserve, Brunei Darussalam 43
David W. Roubik: Wild bees of Brunei Darussalam 59
D.M.J.S. Bowman and J.C.Z. Woinarski: Australian monsoon rainforest
mammals and possible implications for the conservation of tropical
rainforest biodiversity 67
Douglas Brandon-Jones: Presbytis species sympatry in Borneo versus allopatry
in Sumatra: an interpretation 71
Klaus Riede: Diversity of sound-producing insects in a Bornean lowland rain forest 77
Damir Kovac and Bruneo Streit: The arthropod community of bamboo
internodes in Peninsular Malaysia: microzonatlOn and trophic structure 85
Clive F. Mann: The avifauna of the Belalong forest, Brunei Darussalam 101
Maryati Mohamed, Azizah Hussein and Arbain Kadri: Terrestrial Ants
(Hymenoptera: Formicidae) of Poring, Kinabalu Park, Sabah 117
A. G. Orr and C. L. Haeuser: Temporal and spatial patterns of butterfly
diversity in a lowland tropical rainforest 125
Axel Dalberg Poulsen, Ivan C. Nielsen, Sylvester Tan and Henrik Balslev:
A quantitative inventory of trees in one hectare of mixed dipterocarp forest in
Temburong, Brunei Darussalam 139
Jiro Kikkawa: Forest bird diversity in Oceania 151
Dionysius Sharma, Hydeir Kamarudin, Halijah Ibrahim and On~
Hean Chooi: The fauna and flora of a semi-deciduous forest in PerilS,
Peninsular Malaysia 153
M. J. E. Coode, J. Dransfield and D. W. Kirkup: Brunei and
biodiversity - the Kew-Brunei checklist project 163
Dietrich Burkhardt and Ingrid De La Motte: Stalk-eyed Flies:
Caprices of Evolution? 169
Joseph K. Charles: Small mammal diversity in riparian and dipterocarp
habitats in Belalong forest, Brunei Darussalam 175
Patrick Grootaert: Species radiation of predatory empidoids (Insecta, Diptera)
in lowland rainforest in northern Papua New Guinea 183
Nick Mawdsley: The theory and practice of estimating regional species
richness from local samples 193
vi
RAINFOREST DYNAMICS
Peter J. Grubb: Rainforest dynamics: the need for new paradigms 215
Arsenio B. Ella: The interaction of rainfall, tree diameter and sulphuric-acid
treatments on resin yield of Palosapis
(Anisoptera thurifera (Blanco) Blume ssp. thurifera) 235
C. A. Pendry and J. Proctor: Preliminary studies of altitudinal zonation and
nitrogen supply on Bukit Belalong, Brunei Darussalam 241
S. M. Ross and A. Dykes: Soil conditions, erosion and nutrient loss on
steep slopes under mixed dipterocarp forest in Brunei Darussalam 259
Alexander K. Sayok and Nalau Bingeding: Construction of a low-cost runoff
and sediment monitoring plot 271
Lok Eng Hai, Hashim Md. Noor and Fauzidah Ahmad: The growth performance
of four dipterocarp species under different terrain and shade conditions 277
George D. Weiblen: Variation in reproductive success of gynodioecious figs
(Ficus spp., Moraceae) and their pollinators (Hymenoptera: Agaonidae)
in New Guinea 283
Konrad Fiedler: Interactions between Iycaenid butterflies
and ants in Peninsular Malaysia 291
David T. Jones: A quantitative survey of the termite assemblage and its
consumption of food in lowland mixed dipterocarp forest of Brunei Darussalam 297
Satish C. Choy, Salwana A. Latif and Yong N. Yung: Resource use in a
freshwater fish community of a tropical rainforest stream in northern Borneo 307
Indraneil Das: Spatio-temporal resource utilization by a Bornean rainforest
herpetofauna: Preliminary results 315
Eustaquio G. Aragones: The altitudinal distribution of forest tree
families on the western slopes of Mount Banahao, Luzon, Philippines 325
Ruth Levy: Interspecific colony dispersion and niche relations of three
large tropical rain forest ant species 331
Zainuddin Dahaban, M. Nordin and Elizabeth L. Bennett: Immediate effects
on wildlife of selective logging in a hill dipterocarp forest in Sarawak: mammals 341
Abd. Rahman, K., Shahruizaman, I. and Weinland, G.: Growth
performance of a naturally regenerated stand of Kapur
(Dryobalanops aromatica Gaertn. f.) 37 years after a heavy low thinning treatment 347
Arthur Y. C. Chung and Maryati Mohamed: A Comparative Study of the
Ant Fauna in a Primary and Secondary Forest in Sabah, Malaysia 357
Webber E. Booth: Comparative morphology and physiology of gap and
understorey seedling leaves from a mixed dipterocarp forest 367
Satish C. Choy: Distributional ecology of freshwater fishes in tropical
rainforest streams of Borneo 377
A. P. Dykes: Analysis of factors contributing to the stability of steep
hillslopes in the tropical rainforest of Temburong, Brunei Darussalam 387
Herujono Hadisuparto: The effects of timber harvesting and forest
conversion on peat swamp forest dynamics and environment in West Kalimantan 411
vii
SOCIO-ECONOMICS
Peter Eaton: Socio-economic values and tropical rain forest conservation 417
F. Goeltenboth and K. H. Timotius: Impact of rainforest destruction -
the Siberut Island case, Sumatra, Indonesia 425
Jay H. Bernstein: Higher-order categories in Brunei Dusun
Ethnobotany: the Folk-Classification of Rainforest Plants 435
Mubyarto: The Socio-Economic Benefits of Indonesian Forest 451
Roy Ellen: Putting plants in their place: Anthropological approaches to
understanding the ethnobotanical knowledge of rainforest populations 457
David H. Bennett: Valuing a tree: the ethics of environmental evaluation 467
Robert A. Voeks: Extraction and Tropical Rain Forest Conservation
in Eastern Brazil 477
Eberhard F. Bruenig and Sonja Csomos: Geographic information systems
in rainforest ecosystem research: the Sarawak pilot model 489
John T. Wills: The creation of a geographical information system for
the Kuala Belalong Field Studies Centre 495
W. Erdelen: Tropical rain forests in Sri Lanka: characteristics, history of
human impact, and the protected area system 503
Waranoot Tungittiplakom and Bengt-Erik Bengtsson: A Conflict over
Natural Resources between Highland and Lowland Populations in Thailand 513
M. C. Stockdale and H. L. Wright: Rattan inventory:
determining plot shape and size 523
Glyn Davies: Conservation of biodiversity in Kenya's indigenous forests:
biological currencies used in management planning 535
Index to genera and species 547
General Index 555
Preface
The Conference on Tropical Rainforest Research: Current Issues was organised by the
University of Brunei Darussalam and The Royal Geographical Society, London, and held in
Bandar Seri Begawan, Brunei Darussalam, over 8 days in April 1993. Over 160 participants
from 22 countries attended the Conference, which was opened on the 9th April by the Brunei
Darussalam Minister for Home Affairs, Yang Berhormat Pehin Orang Kaya Laila Setia Bakti
Diraja Dato Laila Utama Haji Awang Isa bin Datu Perdana Menteri Dato Laila Utama Haji
Awang Ibrahim.
The conference was initially intended to provide a forum to present the results of the
expedition into the lowland mixed dipterocarp forest of the Temburong District of Brunei
Darussalam, which had been jointly organised by the University of Brunei Darussalam and
The Royal Geographical Society, London. The 15-month expedition, lasting from January
1991 to March 1992, was based at the then newly-completed Kuala Belalong Field Studies
Centre, a research and education facility set up by the University with funds provided by the
Brunei Government and Brunei-Shell Petroleum Sdn Bhd. The expedition, with over 70
scientists taking part, received financial support from a wide range of sponsors and Corporate
Patrons, including Royal Brunei Airlines, the Baring Foundation, Daiwa-Dicam, Greencard
Trust, the Hongkong Bank, Morgan Grenfell and Nomura-Nimco. The conference itself was
supported by donations from Brunei-Shell Sdn Bhd, Royal Brunei Airlines and Standard
Chartered Bank.
In the event, although papers resulting from work at the Field Centre formed a substantial
part of the presentations, the theme of the conference was widened to include other aspects
of research into tropical forests. Three main themes were identified - Biodiversity, Forest
Dynamics and Socio-economic Aspects, with each theme being introduced by a keynote
speaker; the proceedings have been organised in a similar manner. A total of 85 papers were
presented over the 4 days of lecture sessions; the present volume contains the texts of 53 of
these. Apart from the formal presentations at the Conference, a session was also devoted to
discussing the future development of the Kuala Belalong Field Studies Centre, and the
integration of its activities into the wider context of research into tropical forests within the
south-east asian region as a whole; the resolutions drawn up to guide the University have not
been included in these Proceedings.
The editors would like to thank the authors for their patience and understanding through
what has been a long gestation period, as well as the many anonymous reviewers for their
comments on the original manuscripts. The conference itself benefitted from the dedication
and commitment of many of the staff of the University, far too numerous to mention here;
the organisational abilities of Madam Pang Yoke Kiew (University of Brunei Darussalam)
and Ms Catriona Prebble (Royal Geographical Society) were admirably demonstrated through
its success. Publication of these Proceedings has been made possible through the generosity
of grants from the Government of Brunei Darussalam and Brunei-Shell Petroleum Sdn Bhd.
ix
x
Madam Pang Yoke Kiew and Mrs C.Fereday were most helpful in assisting in the
preparation of the typescripts; the editors are also appreciative of the efforts of the staff at
Kluwer Academic Publishers, particularly Mr. Wil Peters and Ir. Hendrik Prins, whose
patience in the face of interminable delays was endless. We hope that this collection of
papers will serve as a fitting epilogue to this phase of the productive collaboration between
the University of Brunei Darussalam and The Royal Geographical Society, and will give an
indication of the current state of tropical rain forest research in south-east Asia.
Finally, we would like to thank The Honourable Minister of Education, Yang Berhormat
Pehin Orang Kaya Laila Wijaya Dato Seri Setia Awang Haji Abdul Aziz bin Begawan Pehin
Udana Khatib Dato Seri Paduka Awang Haji Umar, and the Vice-Chancellor of Universiti
Brunei Darussalam, Yang Dimuliakan Pehin Orang Kaya Putera Maharaja Dato Seri Laila
Jasa Haji Abu Bakar bin Haji Apong, for the-r enthusiastic and wholehearted support for our
efforts in developing the Kuala Belalong Field Studies Centre and in the preparation of these
Proceedings.
David S. Edwards
Webber E.Booth
Satish C. Choy
Bandar Seri Begawan,

Brunei Darussalam.
May 1996
Tropical forest dynamics: the faunal components
N.E. STORK
Biodiversity Division, Department of Entomology, The Natural History Museum, Cromwell

Road, London SW75BD, United Kingdom
ABSTRACT. The form, diversity and dynamics of tropical forests are influenced by both abiotic factors,
such as the physical envirorunent and climate, and biotic factors, such as interactions between organisms.
To most people forests are principally comprised of trees, large mammals and birds. But, tropical forests
are comprised of many other smaller life forms including arthropods, fungi, nematodes and bacteria, and
these too have an important part to play in forest dynamics. The relative importance of different groups
of animals in tropical forests, as measured through abundance, biomass and species richness, is presented
here. Because of the sheer diversity of the life forms in tropical forests, only a few of the complex webs
of interactions between organisms of the different kingdoms have been determined with any accuracy.
In spite of this fact, some general patterns have been recognised and the major components of trophic
systems in tropical forests have been determined. Many detailed studies of large mammals and birds have
shown the significance of their roles in the dynamics of tropical forests. What is less clear is the precise
contribution of many groups of invertebrates such as arthropods and nematodes. In part this is because
the abundance, species richness and sometimes biology of these groups is poorly understood. Data are
presented on the (i) distribution, abundance and biomass of arthropod groups in the main habitats of
lowland rain forest in Seram, and (ii) the species richness of different arthropod groups in lowland forest
in Sulawesi. Arthropods are particularly important in two contrasting strata of tropical forests, the canopy
and soil, where herbivory and decomposition, respectively, are the principal trophic processes. The
importance of arthropods in these forest strata and to forest dynamics are discussed through examples
of two groups of arthropods, ants and termites. An understanding of tropical forest dynamics may help
in determining how biodiversity and sustainability may sit more easily together. This is discussed with
suggestions as to how we can best use available information toward this end, what new information is
needed, and which ideas require examining.
KEY WORDS: diversity, biomass, biotic factors, abiotic factors, arthropods, ants, termites, Seram,
Sulawesi.
INTRODUCTION
For obvious reasons much of the attention within studies of tropical forest dynamics has been
on trees. The natural process of tree growth, death, gap creation and new tree growth has
been well documented (Brokaw 1874, Grubb 1977. Whitmore 1984). The role of nutrient
cycling in forest dynamics is also well documented but again largely focuses on botanical
elements (DeAngelis 1992). However, tropical forests are not just comprised of trees or even
just of plants. Other groups of organisms, such as insects and other invertebrates, are also
essential to tropical forests, as they are to all ecosystems, and are often highly abundant and
speciose. Prokaryote organisms (e.g. bacteria and viruses) and eukaryote plants (e.g. fungi)
are clearly essential to ecosystem function but are beyond the scope of this paper and are not
discussed here. The first aim of this paper is to provide a brief review of the eukaryote
faunal components of tropical forests and their contribution to forest dynamics.
In reviewing faunal dynamics in tropical forests it is perhaps appropriate to consider this
subject in the light of current scientific interest in both environmental change and
biodiversity. Biodiversity, a contraction of 'biological diversity' (Wilson 1988), is the term
used by many individuals, as diverse in interests as biologists and economists, to cover a
very wide-ranging field of science. Most recently it has come to prominence through the
D. S. Edwards et at. (eds.), Tropical Rainforest Research - Current Issues, 1-20.

© 1996 Kluwer Academic Publishers.
2 N.E. STORK
Biodiversity Convention agreed at the UNCED meeting in Rio de Janeiro in 1992 and signed
by more than 160 countries. To most scientists biodiversity represents the diversity of life
on Earth as expressed typically by genetic diversity, species diversity and ecosystem diversity
(IUCN, UNEP & WRI 1990). In reality it involves many other aspects of nature including
landscape diversity, biome diversity, and historical biogeography (Holloway & Stork 1991).
Recent upward revisions of estimates of the number of species on Earth to around 5-15
million species (Gaston 1991, Hammond 1992, in press (a,b) May 1992, Stork 1993) and
beyond have implicitly suggested that perhaps (i) more than 80 % of species are undescribed,
(ii) more than 80% of species of all living eukaryote organisms are insects, and (iii) perhaps
90% of all species are in tropical forests (although some marine biologists might argue this
latter point - see Grassle & Maciolek 1992). The major faunal component of tropical forests
is the arthropods, and their significance is vastly underestimated.
The second aim of this paper, therefore, is to examine more closely the contribution of
arthropods to tropical forests. First, the abundance and biomass of different arthropod
groups, and their distribution in the different habitats (canopy, soil etc.) are discussed using
an example of a study in Seram. Second, the species richness of arthropods in tropical forests
is examined using the example of an intensive study in Sulawesi. Although many arthropods
are found in aquatic systems these are not considered here.
The third aim of this paper is to demonstrate the significance of arthropods to tropical
forest dynamics. What we know of the contribution of arthropods to the dynamic processes
in tropical forests is extremely variable. A few groups are better known tha~ others and here
I use two examples, ants and termites, to show the extent and complexity of their role in
forest dynamics.
Scientists are faced with the task of providing answers to many pressing international
problems. With limited budgets it is essential that efforts are coordinated to produce best
value for money. Some answers are required more rapidly than others and therefore priorities
need to be made if we are to contribute to limiting the effects of problems such as global
climate change, forest destruction and pollution. Particularly important is the implementation
of policies advocating sustainable utilisation of resources such as tropical forests.
Understanding the natural processes of tropical forests and the changes in these processes
when forests are affected by humans is essential. In the final section of this paper I briefly
review some ideas that are either being discussed by international organisations or are being
implemented.
CONCEPTS OF ECOSYSTEM DYNAMICS
What are the forces that drive tropical forest systems? The study of the dynamics of tropical
forest ecosystems is extremely complex and no unifying theory exists. Numerous concepts
overlap and sometimes conflict and no attempt is made in this paper to discuss these.
However, perhaps the concept most fundamental to an understanding of tropical forest
dynamics is that of scale and the balance between local and regional forces (see Figure 1,
after di Castri 1991, and Delcourt & Delcourt 1987, and Figure 2 after Cornell & Lawton
1992). As humans we perceive nature in a limited range of scale and this has created bias
in the way we examine it. As a result, large more obvious organisms such as birds, large
mammals and trees have received more of our attention than smaller ones (Gaston & May
1992). Similarly, we have difficulty in looking at changes with respect to time when the
timescale is either very short (less than a few seconds) or very long (more than a few years)
with respect to the life-span of human. Cornell & Lawton (1992) noted that two groups of
processes they identify, ecological and biogeographical, operate at the local and regional
spatial scales, respectively. This is also true of the temporal scale: ecological processes
Tropical forest dynamics 3
DISTURBANCE REGIME BIOTIC RESPONSE ECOSYSTEM STEPS
-
Evolution of
the Biota
r- Plate- Tectonics
lO' I-
10' I-
~
b
~
Evolution of ecosystems
10' r-
Climate Cycles
'"
~0 IO' - Soil developme<nt
'"::E - Adaption
"'.... 10'
J02 - Hurric:mes Succession of communities
10' :-- Drought Cycles Popublioo

Dynamics
Fire Cycles Population Dyn:tmics
Wet/Dry Season
Phenology
I Yr ~ Rain Storms
,M r- Le3fPosi{ion ;. •••••• -----..:
'0 I- ~ D'Yr'gh'{Yclej I I I I I I: I I I I I I I I I I I
10' 10' 10" 10" 10' 10' 1010 10' 10' 1010
SPATIAL SCALE (m')
Figure I. Spatial and temporal scales of disturbance forces and biotic responses (after di Castri 1991 and simplified
from Delcourt & Delcourt 1987).
LOCAL FORCES REGIONAL FORCES

population dynamics long-distance dispersal
parasitism speciation
predation species extinction
predation fluctuation in species distribution
abiotic fluctuations
or disturbance
'ecological forces' = 'biogeographical forces'
Figure 2. The balance between local and regional forces and how these forces are
represented in the biology of organisms.
operating more in the short-term and biogeographical processes operating in the long-term
(Figure 2).
Of the abiotic forces affecting the balance of tropical forests, particularly important are
the flows of water, carbon, nitrogen and other elements such as phosphorus. Changes in the
balance of the cycling of these can considerably affect the nature and stability of the
ecosystem (e.g. see Furley et at. 1992). Also extremely important is the stability of local and
regional climatic conditions; hurricanes, for example creating immense natural damage to
some forest systems (e.g. Weaver 1989). These abiotic forces can rapidly become controlling
in some circumstances of forest change leading to erosion and desertification. The organisms
that combine to form tropical forests are each physiologically adapted to a limited range of
tolerances with respect to water and nutrient availability, sunshine, temperature, and relative
4 N.E. STORK
humidity. Some species have sufficient genetic variability to be able to adapt to large and
rapid changes in the environment whereas others are replaced by more adaptable species. The
level of fluctuation in such abiotic factors will determine the degree of change in species
composition, population structure, and biological disturbance, and the rapidity of recovery.
THE FAUNAL COMPONENTS OF TROPICAL FORESTS
What groups of animals are found in tropical forests? What are their abundances, species
richness, biomass, and what are their roles in the dynamics of tropical forests? In practice,
abundances and species richness of different organisms vary in time and space and no sites
have been adequately sampled to present even a model view of their dynamics. Table 1
provides an overview of several measures of the faunal component of tropical forests and is
discussed below. No attempt has been made to give a complete list of the higher taxa within
the phyla, only those being listed which have a recognised role in one way or other in
tropical forests.
Many phyla are entirely restricted to aquatic environments and only a few, such as
Arthropoda, Chordata, Mollusca and Annelida have groups that have adapted to the rigours
of life out of water. Many groups of small organisms found in soil, such as Protozoa,
Rotifera and some Nemata, are still aquatic, surviving in the interstitial water between soil
particles. Data for abundance, biomass, and species richness for these and many of the other
groups figured in Table 1 are scarce. Furthermore, the term 'tropical forest' covers a wide
range of forest types and hence data from one source are rarely directly comparable with
those from another. To these extents, Table 1 is a conglomerate and provides a simplistic
view of the faunal contribution to tropical forest dynamics. The actual contribution of
different groups such as earthworms and termites in different forest types or parts of the
world may vary considerably.
Nematodes are known from numerous studies to be superabundant in agricultural
ecosystems but we are uncertain how diverse or how numerous they are in tropical forests.
Preliminary results from studies in Brunei and Cameroon (Hodda et aI., in press) indicate
they are very abundant and contribute large biomass in moist tropical forests. Evidence is so
far lacking on the species richness of this group in tropical forests because of the enormous
numbers of individuals per m2 • Lepidoptera can be very abundant particularly on short time
scales and, as most are herbivores, they can have a dramatic affect on temperate forests of
low tree species richness. Whether they have the same affect on similarly species poor
tropical systems such as Brachystegia forests, or in more diverse tropical forests is uncertain.
In Australian eucalyptus forests detritivorous Lepidoptera of the family Oecophoridae,
numbering some 2,300 species, are essential in the nutrient cycle as they are major
decomposers of eucalypt leaf litter (Common 1990).
It is evident from Table I that in terms of biomass, abundance and species richness,
Arthropoda is the most important group of animals with respect to tropical forests. Many
groups of arthropods have key roles in ecosystem function and the dynamic processes of
tropical forests and examples of these are discussed below.
Site inventories
There is no single terrestrial site in the world, tropical or temperate, where we know all
the living organisms and therefore not surprisingly, our understanding of the biological
components of tropical forests is poor. Botanists have tackled this problem in a more
systematic way (e.g. Dallmeier 1992) and lists of species for trees above lO cm dbh are now
available for many hectare units around the world. More intensive studies of the plants are
Table 1. Summary of the main parameters of the terrestrial faunal components of tropical forests. Arthropod data
compiled from sources quoted in the text and from unpublished data. Vertebrate data compiled from (1) Inger 1979,
(2) Inger 1979, (3) Davidson 1991, Johns 1992 and Woinarski 1991, (4) Chivers & Davis 1979, Curtin & Chivers
1978, Johns 1992, Medway 1983, Payne 1979 and Payne & Mahedi 1991. Groups in italics are exclusively marine.
Classification following Parker (1982). Habitats: T - terrestrial (ie. not aquatic), G - ground, C - canopy; Key (key
groups): P - predators, D - decomposers, H - herbivores.
Key to symbols:
*** ** *
Abundance (individuals) > 106lha > 104 /ha > 1(J"Iha < l<J"/ha
Biomass (gm) > 104lha > 10'/ha > lO'/ha < l(J"/ha
Sp. richness (species) > IO'lha > 102 /ha > lO/ha < IO/ha
Abundance Biomass Sp. rich. Habitat Key

Subkingdom Protozoa ? ? ? T:G
Subkingdom Phagocytellozoa
Subkingdom Parazoa
Subkingdom Eumetazoa
Phylum Cnidaria
Phylum Ctenofhhora
Phylum Platy elminthes T:G
Phylum Nemertea
Phylum Gnathostomulida
Phylum Mesozoa
Phylum Gastrotricha
Phylum Roti/era
Phylum Kinorhyncha
Phylum Nemata *** ** ** T:G P,D,H
Phylum Nematomorpha
Phylum Acanthocephala
Phylum Pria~Uld
Phylum Mol usca T:G
Phylum Annelida *** *** * T:G D
Phylum Po~nphra
Phylum Ec iura
Phylum S/f.uncula
Phylum nychophora * T:G
Phylum Tardigrada
Phylum Phoronida
Phylum Bryozoa
Phylum Entoprocta
Phylum Brachiopoda
Phylum Chaetognatha
Phylum Echinodermata
Phylum Hemichordata
Phtlum Arthrol?oda
lass Arachnida
Order Scorpiones * T:G
Order Araneae ** ** *** T:G,C
Order Pseudoscorpiones ** * * T:G,C
Order Opiliones ** * * T:G,C
Order Acari *** * *** T:G,C P
Class Crustacea
Order Isopoda * T:G,C
Class Hexa~d
Order Co lembola *** * ** T:G,C D
Class Insecta
Order Tlzsanura * T:G
Order 0 onata * ** F
Order Isoptera *** *** * T:G D
Order Orthoptera ** ** * T:G,C
Order Dermaptera ** * T:G,C
Order Hemiptera ** * ** T:G,C
Order Homoptera ** * ** T:G,C H
Order Neuroptera ** * T:C
Order Coleoptera *** ** *** T:G,C D,H,P
Order Diptera *** ** *** T:G,C D
Order Lepidoptera *** *** *** T:G,C H
Order Hymenoptera *** ** *** T:G,C P
Phylum Chordata
Subphylum Vertebrata
Class Osteichthyes
Class Amphibl3, * ** * T:G P
Class Reptilia, * ** * T:G P,H
Class Aves. * ** * T:G,C P,H
Class Mammalia. * *** * T:G,C P,H
6 N.E. STORK
far fewer although some have managed to identify and measure all trees above 1 cm DBH
for 50 hectare plots (Foster & Hubbell 1990). Detailed censuses are also available for large
mammals, birds, reptiles and amphibians. However, for the vast majority of groups such as
the ferns, mosses and lichens on the plant side, and for arthropods, annelids, and protozoa
on the animal side, we have very little idea of their abundance let alone their richness in
species. Furthermore, the data that are available are from a variety of different sites and it
is not always possible to make valid comparisons.
Two attempts, both in Indonesia, have been made specifically to document arthropod
diversity in restricted areas of tropical forest, loosely comparable with the botanists' hectare
and 50 hectare plots, and these are discussed here. The first examines arthropod abundance
and biomass and the second largely examines species richness.
(i) Arthropod abundance and biomass in Seram
The relative abundance and biomass of the arthropods were estimated in a hectare of
lowland tropical forest in Seram, Indonesia through collection of direct quantitative samples
of arthropods from what were deemed to be the five most important habitats - canopy, tree
trunks, ground herb-layer, leaf litter and soil (Stork 1988, Stork & Brendell 1993). Samples
were collected using a variety of sampling methods (canopy - knockdown insecticide fogging;
tree trunks - knockdown insecticide spraying; ground herb layer - DVac; leaf litter and soil -
Tullgren funnels). Samples were sorted to orders and counted. In order to estimate biomass
the samples were further sorted to log. size classes and the following body sizelbiomass
relationship used:
B = N(0.0305U62) (Rogers et al. 1976)
where B is biomass in mg, N is number of individuals, and L is body length in mm. The
geometric mean of the two limits for each size class was used as the average body length of
each size class in these calculations.
The number or biomass of arthropods from each of the five habitats in a hectare of Seram
forest was calculated on the expected number of duplicate samples in a hectare. For example,
since a 1 m2 sampling unit was used for the canopy and a hectare is 10,000 m2, the number
of arthropods in the canopy in a hectare = 10,000 x M (where M = the mean number of
arthropods 1m2). Multiples for the other habitats are: leaf litter - 10,000, herb-layer
vegetation - 10,000/16, tree trunks - 800, soil - 650,000 (the figure for tree trunks is based
on a guess of 800 'tree trunk equivalents' in a hectare, there being no data available for a
more accurate assessment).
The results of this study indicated that in a hectare of forest (i) there are more than 42
million arthropods, equivalent to more than 30 kg dry weight, (ii) soil, canopy and leaf litter
contribute the greatest abundance and biomass of arthropods, (iii) Collembola, Formicidae
and Acarina are the most abundant groups, (iv) non-insect arthropods (primarily Diplopoda)
comprise more than 50% of the biomass (Figure 3 and see Figures 7-8 in Stork 1988,
Figures 1-3 in Stork & Brendell 1993).
The biases in the various sampling methods used in this study are well known and
discussed in full by Stork & Brendell (1993). These and the results of other studies (Ad is
1988, Adis & Schubart 1985, Collins 1983, Andre et al. 1992, 1994, Eggleton et ai. in
press, Eggleton & Bignell, in press) would suggest that these estimates for total arthropod
abundance and biomass are low. Those for soil arthropods are particularly suspect and the
abundance of Collembola and Acarina may be much higher. Termites are almost certainly
a large component of the fauna of most tropical forest soils (see below) even if they were
of genuinely low abundance or absent from the Seram study site. In spite of these anomalies,
it seems probable that the relative significance of the five habitats studied, in terms of
arthropod abundance and biomass, is correctly identified. For all five habitats examined
a)
4
3
1 Soi l
2 Lear liner
3 Herb layer
4 Tree trunks
5 Canopy 3 2
b) 1314
d)
2
11 3
14 4
4
5
5 6
6
7
8 7
9
9 121 110
1 Formicidae 8 Thysanoptera
2 Othcr Hymenoptera 9 Collembola
3 Diptcra 10 Other insects
4 Coleoptera 11 Araneac & Opilioncs
5 Hemiptcra 12 Acari
6 Psocoptcrn 13 Pseudoscorpiones
7 Orthoptcroids 14 Othcr arthropods
Figure 3. Pie chilrts showing the relative contribution of arthropod abundance with respect to a) habitats, and b)
taxonomic groups, and of biomass with respect c) habitats, and d) taxonomic groups (note: Thysanoptera and
Pseudoscorpiones biomass too low to include in Fig. 3d), in lowland tropical forest in Seram, Indonesia (after Stork
& Brendell 1993).
arthropod abundance decreases and biomass increases with body size (Stork & Blackburn
1993), although the average body size increases vertically with habitat from soil to canopy.
The modal body size for arthropod abundance is about 1.2 mm for above ground habitats and
less than 0.5 mm for below ground habitats (soil and leaf Jitter).
Although some 45,000 arthropods were sampled in the Seram study this was insufficient
to carry out an analysis of the species richness of the arthropods in a hectare of forest and
the relative species contribution and overlap for each of the five habitats. A more intensive
8 N.E. STORK
sampling programme, such as that described below, is required to determine arthropod

species richness of tropical forest.
(ii) Arthropod species richness and composition in N.Sulawesi
The insect fauna of the Dumoga-Bone region of N.Sulawesi was intensively studied by
over 100 entomologists during 1985 as part of the Royal Entomological Society's Project
Wallace. The main results of this project are summarised in Knight and Holloway (1990).
The largest single study was that of the insect fauna of lowland forest by a team from the
Natural History Museum (London, UK). Their intensive sampling programme (see Hammond
1990, Stork & BrendellI990), which involved Malaise, flight interception, light, yellow pan,
and pitfall (baited and unbaited) trapping, knockdown insecticide canopy fogging, and litter
sampling, was supplemented by additional sampling and hand collecting by others (Barlow
& Woiwod 1990, Hodkinson & Casson 1987, 1991, Holloway et al. 1990, Noyes 1989a,b).
The most extensively sampled group were Coleoptera with over 1.1 million collected
specimens being examined (Hammond 1990, Stork & Brendell, 1990). The vast majority of
these were from a lowland forest area of about 500 hectares. Hammond identified 5,649
morphospecies from the whole region and 3,488 from the lowland forest area. He estimated
that with additional unsorted materials the totals for the latter would be over 6,000. From
his detailed analyses of these samples a number of observations can be made: (1) the most
species rich groups in the samples are also those groups with the largest numbers of
described species (see Gaston 1991); Staphylinoidea (26% of all beetle species),
Curculionoidea (18.7%), Tenebrionoidea (12.9%), (2) only 17.4% of the beetle species feed
on non-woody living higher plant materials (ie leaves, pollen etc.); 29% are predators and
23.4% fungivores (Figure 4). Hammond's (1990) cumulative species curves for both ground
and aerial Malaise trap weekly catches throughout the year show no obvious asymptote.
However, since none of the traps for which cumulative totals were presented collected more
than about 400 species (ie less than 10% of the minimum number of species known from the
area) this is not surprising. What would be interesting is to see a cumulative catch week by
week for all trap methods combined. However, difficulties in cross-referencing species
between samples and databasing this information probably preclude such a task for the
moment.
Species richness data are available for several other species rich groups of Sulawesi
insects. Noyes' (1989a,b) staggering collection of 4,109 species of Hymenoptera was from
samples he and others collected in a three month period (April - June). Hodkinson and
Casson's studies of the Hemiptera (Casson & Hodkinson 1991, Hodkinson & Casson 1987,
a) b)
Par~itods Fresh Water

0.9% 5.0%
Saprophages
12.7%
Wood-
Xylophages associated
16.0% 32.7%
Dung, carrion
etc.
6.8%
Non-woody
parts of
higher plants
28.3%
Figure 4. Pie charts showing a) the feeding guild associations and b) the habitat associations of 5,469 beetle species
collected from N. Sulawesi, Indonesia (see text) (after Hammond 1990).
Table 2. Species richness of different arthropod groups for areas within Sulawesi: (i) for the lowland forest area of
about 500 hectares including three hectare plots A-C at 200 m, 300 m and 400 m altitude, (ii) the Dumoga valley
and surrounding regions, including upland sites up to 1760 m and over a geographical distance of 150 krn, (iii) the
whole of the island of Sulawesi.
Plots A-C N. Sulawesi Sulawesi Author

500 ha 15 krn2 150,000 krn2
Coleoptera 3500 6,250 (10,000) ? Hammond (1990)
Diptera
Hemiptera 1,690 (2,500) ? Hodkinson & Casson (1991)
Hymenoptera 4,100 (10,000) ? Noyes (1989a, b)
Lepidoptera
Butterflies 300 350 470 R. Vane-Wright (pers. comrn.)
Macro-Lepidoptera 1,600 Holloway et al. (1990) and J.
Micro-Lepidoptera Holloway & G. Robinson (pers.
comrn.)
Tineidae 114
Tortricidae 285
Orthoptera
Thysanoptera
Psocoptera ? ?
Estimated Totals (30,000) (50,000) (75-100,000)
1991) were also based on limited samples; most of their 1690 species were collected by
knockdown insecticide fogging from trees (Stork & Brendell 1990).
In Table 2 an attempt has been made to compile very rough and preliminary species
richness estimates for the insect fauna of the Dumoga-Bone area at three geographic scales:
(i) the insect fauna of the lowland forests area of about 500 hectares including three hectare
plots A-C at 200m, 300m and 400m altitude, respectively, (ii) the Dumoga valley and
surrounding regions, including upland sites up to 1760m and over a geographical distance
of 150 km, (iii) the whole of the island of Sulawesi. The Sulawesi project gives an
approximate idea, perhaps for the first time, of the number of species in an area of moist
tropical forest. It also indicates that unless levels of local endemism of species have been
seriously underestimated, estimated local and regional species richness differ by less than
an order of magnitude.
TROPICAL FOREST DYNAMICS: TWO EXAMPLES OF KEY GROUPS

OF ARTHROPODS
The contribution of ants to tropical forest dynamics

(1) Ant abundance, diversity and biomass
A number of studies have shown that ants form a major part of the faunal component
of tropical forests in terms of both abundance and biomass. Fittkau & Klinge (1973)
suggested that ants, termites, bees and wasps comrise 75 % of the insect biomass in
Amazonian forest. In many knockdown insecticide studies of the canopy arthropod fauna of
tropical forests in countries such as Brazil (Adis et al. 1984), Peru (Erwin 1988), Venezuela
(Stork & Davies unpublished), and Seram (Stork 1988, Stork & Brendell 1993), ants
comprise 50% of the individuals. Ants are low in abundance in only a few tropical forests,
such as in Sulawesi and Brunei (Stork 1991, Stork & Brendell 1990). In the Seram study
discussed above, ants comprised an estimated 17% of all arthropods in lowland tropical
forest.
10 N.E. STORK
Table 3. Abundances of the eight species (from 98 in total) of ant with more than 1()() individuals on ten trees
sampled by canopy fogging in Borneo (Stork 1991, identifications provided by B. Bolton).
Trees
Species 2 3 4 5 6 7 8 9 10 Total
Crematogaster sp. B 0 15 522 0 0 0 214 38 0 0 789
Crematogaster sp. H 0 0 119 3 0 0 0 0 0 0 122
Crematogaster sp. I 6 0 21 0 75 0 17 0 5 0 124
Crematogaster sp. L 317 0 0 0 0 0 0 0 0 0 317
Myrmicaria sp. A 45 0 94 159 62 6 0 0 109 13 488
Myrmicaria sp. B 0 458 0 0 0 0 799 53 0 0 1310
Dolichoderus sp. D 0 9 0 52 59 0 0 0 0 0 120
Camponotus sp. C 0 0 0 0 0 0 0 0 0 199 199
In species terms, ants represent a relatively high diversity group. B. Bolton (personal
communication) collected almost 400 species of ants in a hectare of lowland tropical forest
in Sarawak. Htilldobler & Wilson (1990) quote other examples with lower species numbers
from other ant workers collecting intensively in different tropical forest sites.
(2) Patterns of distribution
Ants have a broad geographical distribution being found throughout the world in temperate
to tropical ecosystems. The well known social structure of ant colonies is discussed by
Htilldobler & Wilson (1990) and elsewhere. Some species have interconnected nests forming
colonies that may cover several km2 • One of the most unusual patterns that has been
described for ants is the 'mosaic' distribution pattern. First described for ant species in cocoa
plantations in Ghana by several authors (Leston 1973a,b, Majer 1976a,b,c, Room 1971), the
mosaic pattern of distribution is now also recognised as a phenomenon of arboreal ants in
natural forests (Htilldobler 1983, Majer 1993). Dominant species occupy large areas of single
or several trees to the exclusion of other competing species. Some sub-dominant species
appear to co-exist successfully in the territories of dominant species. In Table 3 the numbers
of individuals are shown for the eight commonest species of the 98 collected from ten trees
in Brunei by knockdown insecticide fogging (Stork 1991). This table demonstrates that
individuals of the different species are not evenly distributed across all trees. The majority
of individuals of a species are found on just one or two trees. The two commonest species
of ant (both Myrmicaria spp) appear to be mutually exclusive. This would appear to
demonstrate the dominance pattern typical of the ant-mosaic model.
(3) Contribution to forest dynamics
The importance of ants to tropical forest dynamics has been lauded by many including
Wilson (1971, Htilldobler & Wilson 1990). Some of the more well recognised roles of ants
are discussed here.
(i) Controlling levels of herbivory
Most species of ants have a significant role in herbivory either as predators of herbivores,
tenders of, and hence protectors of, herbivores, or as herbivores themselves. The two main
groups of herbivores, 'chewers' and 'suckers', are affected in different ways by some species
of ants. Most sucking insects are Homoptera, are largely sedentary, and feed on plant
phloem. Some species of ants feed on the honeydew excreted by Homoptera, and thus could
be termed herbivores. These ants provide protection to Homoptera from predators and
parasitoids, and by depredation of leaf-chewing herbivores (e.g. caterpillars) reduce
herbivorous competition. Some species of Monomorium are predators of batches of
Lepidoptera eggs thus reducing herbivory at the earliest stage (Way et al. 1989). Some
reduce levels of herbivory through predation of leaf mining larvae (Memmott et at., in
press). Leaf-cutter ants of the genus Atta have been estimated to account for the removal of
12-17% of total leaf production in Neotropical forests (Cherrett 1986, 1989). Although these
ants do not eat the leaves they cut (see below), they are nevertheless responsible for the leaf
damage and hence can be considered herbivores. Their harvesting of leaves from trees is not
arbitrary but follows clear patterns such that individual trees are not over-cropped. They
maintain the correct degree of sunlight and shade over their extensive nests in the ground by
pruning of the surrounding trees. Hence, these ants to some extent determine levels of plant
growth in neotropical forests. Several studies have also shown how species of ant closely
associated with a particular species of tree will prune other trees or destroy seedlings growing
in the vicinity of their host tree (eg Pseudomyrmex with Acacia, Janzen 1967).
(ii) Control of other organisms by predation
Most ants are predominantly predatory and although many are generalist predators some are
more specialised. Below are listed a few examples of such specialisms (B. Bolton, personal
communication) :
a) Amhlyopone, Leptanilla - geophilomorph centipedes,
b) Leptogenys - Dermaptera, land isopods, termites, flat millipedes,
c) Discothyrea, Proceratium - arthropod eggs in leaf litter,
d) Plectroctena, Gnamptogenys - Diplopoda (eggs and adults),
e) Dacetini - mostly Collembola,
t) Many ants (e.g. Dorylus, Centromyrmex, Decamorium, Carebara) - termites,
g) Cerpachys, some Tetramorium - other ants,
h) Myopopone, Aphaenogaster - beetle larvae,
i) Nomamyrmex - leaf-cutter ants,
j) Psalidomyrmex - earthworms.
Insects which have been parasitised by other insects are often more prone to attack by ants
as they sometimes lose some of their mobility and defensive escape mechanisms.
(iii) Plant adaptations
The benefits of reduced herbivory from hosting large populations of predatory ants, have
resulted in many plants evolving morphological and physiological adaptations to promote the
establishment of ant populations. Some have developed extra-floral nectaries which produce
a sugary sap on which the ants feed. Removal of the nectaries results in marked increases
in herbivory (see Holldobler & Wilson 1990). In one study 60% oftree species in clearings
and 30-40% of tree species in canopy samples showed mutualistic ant-plant associations
(McKey 1992). Some plants produce structures called domatia which serve to protect ant
colonies. These structures include hollow thorns, stems, or expanded leaf bases. Some
domatia provide food bodies on which the ants feed (Holldobler & Wilson 1990). One of the
best known ant-plant associations is that between Cercropia trees and Azteca ants (see
Davidson & Fisher 1991, Longino 1989). The ants vigorously defend the trees from
herbivores, leaf-cutter ants and even vines. In return the plant provides shelter in the form
of hollow stems and nutrition from food bodies at the bases of leaf petioles. If a Cercropia
stem is cut, ants pour out and aggressively attack the damage-causing organisms. One study
showed that young Cercropia trees have a reduced rate of survival when their Azteca colonies
are removed (Schupp 1986).
One of the most unusual forms of domatia is that of the hugely expanded stems of
epiphytic plants of the Myrmecodia (Rubiaceae). These domatia provide a honeycomb-like
nest for the ant lridomyrmex cordatus (Huxley 1978, 1980).
(iv) Seed dispersal
Ants play an important role in seed dispersal and seed growth in tropical forests, perhaps
more so than any other group of animals. Many species of plants produce seeds bearing
chemical attractants for ants and appendages which ease the task of carrying them (Beattie
1985). Ants harvesting seed falls do not manage to successfully carryall seeds back to their
nests. Some seeds are dropped on the way, thus aiding seed dispersal. Some ants also provide
12 N.E. STORK
a service to seedling growth by carrying seeds underground into the nutrient rich environment
of ant nests. In this way they protect the seeds from predators and from the possible hazards
of fire in dry forests (Holldobler & Wilson 1990).
(v) Ants as a food source/centres of communities
With ants comprising such a high proportion of insect biomass in most tropical forests
they naturally provide a rich food source for other animals. Well known examples of
vertebrates associated with ants are many species of birds. Some 50 species of neotropical
birds feed on ants or are associated with species of ants. Many of these prey on the insects
flushed out by army ant colonies. Some parasitic species of insects, such as phorid flies,
specialise on ants and have a rich fauna (Disney 1990). Many other groups of arthropods
(including spiders, beetles, mites, flies, isopods, Collembola, Orthoptera, mites, bugs, and
Lepidoptera) take advantage of the large food resource that ant nests represent and Holldobler
& Wilson (1990) list some of these and their relationships with ants. In terms of species,
beetles predominate, particularly staphylinoid beetles. Some use the ant brood as a food
source, many species developing structures and chemical secretions mimicking those of the
ants to do so. The evolution of species rich groups, such as Paussidae and Pselaphidae
(Coleoptera) and Lycaenidae (Lepidoptera), are strongly linked with those of the ants on
which they depend (e.g. Pierce 1983, 1987). Many species of arthropods utilise the rich
nutrient source of ant waste products.
Many plants also utilise the nutrient rich environment of ant nests for development,
notably those species that develop in so-called ant gardens (VIe 1902). Seeds of epiphytes
brought in by the ants develop in the ant carton nests which they build above ground in trees
or bushes.
(vi) Nutrient cycling
In most tropical forests ants play an important role in soil creation and turnover, although
probably to a lesser extent than termites, except perhaps in some neotropical forests. Here,
leaf-cutter ants, Atta species, create huge nests many metres in diameter and depth. These
ants carry leaves underground where they are decomposed by fungi cultured by the ants in
specialised chambers. The ants feed on the fungi rather than the leaves. Thus the ants are
instrumental in the decomposition process. The by-products from the decomposition of leaves
are stored in dumping chambers and contribute considerably to nutrient cycling in neotropical
forests.
The underground chambers and galleries of ants probably play an important but
unmeasured role in the passage of rain through the soil, hence reducing water-logging,
subsequent run-off, and soil erosion.
The contribution of termites to tropical forest dynamics

(1) Termite abundance, diversity and biomass
Termites are far less visible than ants and it is only through their nests and galleries that
most people view their contribution to tropical forests. However, although not particularly
speciose, they are one of the most abundant groups of organisms in this environment (Table
4). Estimating abundance and biomass is not easy and hence some of the figures in Table 4
may be gross underestimates. In dry tropical forests termites represent as much as 65 % of
the soil faunal biomass (Goffinet 1973). Some nests or galleries reach many metres into the
soil and during the time taken to reach them by digging, many of the termites have long
since disappeared. Similarly estimating termite abundance in dead wood is difficult. Numbers
of species of termites recorded from tropical forest sites range from 25 to 100 (Constantino
1992, Eggleton et al., in press, Eggleton & Bignell, in press, Wood et al. 1982).
(2) Patterns of distribution
Termites have a tropical/sub tropical distribution and few are found in temperate zones.
Most termites are found under ground, in fallen timber on the ground, or in the lower parts
Table 4. Abundance and biomass of termites in tropical forests (after Wood et al. 1982).
Localit~ Numbers m2 Biomass E m·2 Author

W. Africa
Riparian forest, Rabba, Nigeria 2646 6.9 Wood et al. 1982
Semi-decid. forests, Ibadan, Nigeria 3163 8.0 Wood et. al. 1982
Riparian forest, Yangambi, Zaire 1000 11.0 Maldague 1964
s. & C. America
Rain forest, Trinidad, W. Indies 4450 Strickland 1944
S.E. Asia
Rain forest, Pasoh, Malaya 3280 10.9 Abe 1975
Sarawak:
Lowland rain forest
Mixed dipterocarp forest 1125 Collins 1979
Alluvial 254 Collins 1979
Kerangas 1405 Collins 1979
Lower montane forest 38 Collins 1980
Upper montane (1310-1860 m) 99-295 Collins 1980
Upper montane (1970-2376 m) 0 Collins 1980
of trees. Few are truly arboreal. They form large or small nests above or below the ground
and some are connected to each other or to food sources by galleries in the soil or in dead
and living wood. As Collins (1979, 1980) has shown, termites are predominantly a lowland
forest phenomenon, their abundance and diversity decreasing with altitude.
(3) Contribution to forest dynamics
The role of termites in tropical forests is discussed in some depth by several authors (Lee
1983, Lee & Wood 1971, Wood 1988) and a summary is provided here.
(i) The decomposition process and nutrient cycling
Termites are an essential component of many tropical forests representing one of the major
groups concerned with the decomposition of plant tissues aided by symbiotic microflora and
microfauna (Collins 1981, Coventry et al. 1988). In breaking down these tissues they play
a key role in nutrient cycling and hence maintenance of forest stability (Lee 1983, Lee &
Wood 1971).
(ii) Soil generation, turnover and fertility
The construction of vast termite nests and complex galleries involves the movement of
enormous quantities of soil and leaf litter (Collins 1983). Such termite constructions are
composed of a mixture of termite faeces, saliva and soil particles of a particular size range.
Thus the texture of the soil is affected by termites. Termitaria and galleries provide a rich
source of nutrients for plant growth. C. Martius (personal communication) has shown that
old arboreal termitaria falling to the ground provide nutrients for trees in varzea forest of
Amazonia. In some dry tropical forests termites are responsible for the deposition of large
amounts of soil on the surface over long periods of time (Bagine 1984).
(iii) Termitophilous communities
Many other organisms, particularly insects, live with termites. These associations are less
well documented than those between ants and other organisms.
(iv) Termites as a food source
Because termites comprise such a large portion of the standing faunal biomass they offer
an important food source to other organisms such as insects, birds and mammals. In E.
Africa, even the mounds of termites are eaten by primates (G. Davies, personal
communication).
(v) Gap creation and cycles of tree growth
Many tree falls are the result of termite attack and thus termites have a crucial role in gap
14 N.E. STORK
creation and cycles of tree growth. Precisely how important this role is uncertain.
(vii) Hydrology
The galleries and chambers constructed by termites offer ready made channels for the
through flow of rain and therefore termites, as many other insects, are important in the
hydrology of tropical forests. Elkins et al. (1986) found that low vegetation cover areas
treated with insecticide had significantly lower water infiltration rates than untreated areas.
They suggested that the galleries of termites enhance water infiltration and, through reduced
runoff, aid retention of topsoil.
(vi) The global carbon cycle
We are only just beginning to recognise the important role that termites may play in the
global carbon cycle (Higashi et al. 1992, Jones 1990). Termites are believed to produce up
to twice the global CO2 generated by fossil fuel combustion and up to 40% of the global
atmospheric CH 4 released each year (see Jones 1990, and references therein).
In the sections above we can see how at least two groups of insects, ants and termites, are
essential to tropical forest dynamics. But what of other groups? The Sulawesi and Seram
studies described above and Table 1 show that other arthropod groups such as Collembola,
Acarina, Coleoptera, parasitic Hymenoptera, Lepidoptera, Diptera, and Hemiptera, are all
either very abundant, have great biomass or are hyper-diverse groups with 10 times or more
species per unit area than either termites or ants. The roles of some of these groups in
tropical forest dynamics have received much less attention than ants and termites and are far
less clear. One can see how the loss of some species of ants and termites could affect an
ecosystem, but the effects of loss of many species of beetle, moth or fly are less apparent.
The chains of interactions of these organisms with each other and with plants, fungi and
microorganisms are extremely complex. Few such food webs have been examined for
tropical forest systems and as a result the ecosystem function and value of many species of
invertebrates are unclear. Separate studies of the biology of individual species or groups of
species indicate that many are surprisingly essential for the survival of other organisms.
However, there has been a real bias towards studies of the interactions of insects and living
plants and of the processes of herbivory and insect parasitism. Far less attention has been
paid towards insects and fungi and the decomposition process, and towards the role of
animals in nutrient cycling.
THE WAY FORWARD
The purpose of this paper has been to highlight the importance of the faunal component and,
in particular, the arthropod component of tropical forest dynamics and to clarify those areas
where new research will provide a better understanding. Why we need this better
understanding is discussed in the conclusions.
That some groups of organisms have received a disproportionate amount of attention or
conversely, hick of attention, is clear. May (1988) showed that the number of scientific
papers published per species for mammals, for example, is 100 times more than that for
insects and 1000 times that for nematodes. The possible reasons for that are several. First,
humans recognise features in mammals and birds that are also found in man. Human
attributes, such as eyes, nose, mouth and ears, or gentleness, anger and kindness, are not
always quite so readily identifiable in nematodes, mites, springtails or slugs! Second, some
groups, such as butterflies, giant pandas and so on, appear to be more 'attractive' to the
human eye than others (see Kellert 1993, Kim 1993). Third, we readily see and identify
only a small size range of organisms. Many invertebrates and all microorganisms are beyond
the vision of the human eye and optical aids are necessary for their study. Fourth, groups
that are taxonomically intractable and species rich have been poorly examined or ignored.
The vast numbers of individuals and species of insects and other invertebrates, and the lack
of synthesising taxonomic studies deter many from the study of these organisms. Finally, the
important contribution that many of these groups of smaller organisms make to ecosystem
function is not known or is poorly publicised within the scientific and public communities.
The continuing disproportionate distribution of scientific resources maintains our level of
ignorance; a greater proportion of resources needs to be applied to studies of microorganisms
and invertebrates and their role is ecosystems.
These points are not simply of academic interest but show how, as often as not, selection
of groups for study is not based on sound scientific criteria. The same is also true of what
we know about the distribution and biology of those organisms that have received attention.
The goal of most early naturalists and taxonomists was to maximise the richness of their
collections. Far less attention was paid to the examination of geographical or biological
patterns. Thus much of the data provided by museum specimens, although important, is
difficult to use until such biases have been determined (Stork 1994).
It is clear from examination of present rates of description of species (circa 10,000 per
year) and new estimates of 5-15 million species on earth, that unless remarkably fundamental
changes occur in the way we describe species, the task is not going to be completed within
the foreseeable future. Given the international concern for the environment, as exemplified
the Biodiversity Convention and other international agreements signed at the United Nations
Conference on the Environment and Development in Rio de Janeiro in 1992, drastic scientific
measures are needed to assess the extent of biodiversity and the threats to it. Measures to
inventory and monitor changes to the world's fauna and flora are, and have been, considered
by international and national organisations. Some aspects are discussed briefly below, and
in more detail elsewhere (di Castri et at. 1992, Stork 1994).
1) Selection of indicator groups and key species based on sound rather than historical
criteria: It is impossible to fully inventory all organisms in the foreseeable future even with
a massive injection of resources and yet it is important to have a better understanding of the
distribution of life on Earth both for purely scientific reasons and for practical reasons
concerned with the conservation of biodiversity. Similarly, it is impossible to monitor the
effect of environmental change on all organisms. It is therefore necessary to select species
and groups of species that act as indicators (sensu lato) for a wider group of organisms. In
this respect no group can truly be an indicator for others as all species and groups of species
have their own unique characteristics, but in practice compromises have to be made. A broad
and biologically representative range of indicator groups need to be studied, their selection
being based on sound ecological criteria such as species richness, body size, trophic group,
and geographical distribution. There is a particularly urgent need for indicators in tropical
regions and the proposed 'Tropical Areas Bioindicators System' (TABS)(Hawksworth &
Ritchie, in press) should provide a method for early detection of environmental threats to
forest disturbance as well as air pollution and water pollution in tropical forest ecosystems.
2) Preparation of manuals/handbooks and 'sampling packages' for indicator groups:
Sampling of indicator groups needs to be evaluated and standardised to allow comparison
between sites (e.g. see Hammond 1990). Where a range of indicator groups are to be
examined, then the use of standard 'sampling packages' will be useful (Gadagkar et at. 1990,
Stork 1994). The handbook for inventorying and monitoring changes in amphibian
populations and species distributions is perhaps a model for other groups (Heyer et at. 1994).
3) Selection of intensive and extensive study sites: No single tropical or temperate site in
the world has been completely inventoried. An intensive study of all the taxa at a single site -
a so-called 'all-taxa biological inventory' (Janzen, personal communication) - could provide
measured standards for less intensive studies at other sites, and a better understanding of
ecosystem function (di Castri & Younes 1990). A full inventory of a single site using
standardised sampling methods raises many experimental and procedural problems, such as
16 N.E. STORK
the problem of scale (e.g. compare the temporal and spatial problems of sampling bacteria
and vertebrates). One good example of the benefits of an intensive study of a single site is
the result of the long-term sampling of the components of the beech woodland soil fauna in
Germany by Schaefer & Schauermann (1990). Through their studies they were able to give
a complete or near-complete analysis of the trophic and energetic systems of the soil.
At a large number of 'extensive' study sites the species richness and abundance of species
of the selected indicator groups should be determined using standardised sampling methods
and sampling packages. These data could be used to determine local, regional and global
patterns in diversity and to provide biological measures of the consequences of environmental
change.
4) Improvement of taxonomic and ecological resources in developing countries through
training and development of regional and country centres: The global species inventory task
so far has largely failed because it has relied almost entirely on human resources in the
developed world. If this process is to change, then the taxonomic community needs to be
strengthened in developing countries. INBio in Costa Rica is seen by some as one model
for developing countries (see Gamez 1991).
5) Integration of international and national programmes for inventories and monitoring
change: Much of the funding for environmental and projects has been directed towards single
countries, either through bilateral agreements or through international organisations. There
has been a lack of coordination of research between countries with resultant duplication of
effort and lack of information flow. There also has been no recognition of the need for
funding of international programmes of research.
6) Improve availability of information through networking: There have been enormous
advances in information technology such that networking and exchange of data has been
made much easier and relatively cheap. Development of such resources in tropical countries
will enhance the ability of tropical ecologists and taxonomists to make maximum use of data.
CONCLUDING REMARKS
In a recent editorial Andre et af. (1992) made a strong plea for more 'biodiversity data'
particularly with respect to the African continent. I strongly endorse this call but with a
warning. If we are to form a better understanding of the patterns of distribution of life on
earth, the value of individual faunal and floral components to ecosystem function, and most
important of all, develop methods of truly sustainable use of the environment, then we need
to examine carefully what we really know about life on earth. We need to determine the
biases and correct for them in future research programmes. Undoubtedly, one of the results
such actions will be a better understanding of the faunal components of tropical forest
dynamics.
ACKNOWLEDGEMENTS
I am grateful to the organisers of the Brunei Rain Forest Symposium for the opportunity to
present this paper. I thank Victoria Best for assistance in compiling Table 1 and Martin
Brendell and Victora for preparing the figures. I am particularly grateful to my colleagues
Barry Bolton, Gaden Robinson and Dick Vane-Wright for provision of unpublished data.
Constructive criticisms of the manuscript were provided by Barry Bolton, Paul Eggleton and
Kevin Gaston.
LITERATURE CITED
ABE, T. 1975. Studies on the distribution and ecological role of termites in a lowland rainforest of West Malaysia.
Doctoral Thesis, Kyoto University, Japan.
ADIS, J. 1988. On the abundance and density of terrestrial arthropods in central amazonian dryland forests. Journal
of Tropical Ecology 4: 19-24.
ADIS, I., LUBIN, Y. D. & MONTGOMERY, G. G. 1984. Arthropods from the canopy of inundated and terra
firme forests near Manaus, Brazil, with critical considerations of the Pyrethrum-fogging technique. Studies of
the Neotropical Fauna & Environment 19:223-236.
ADIS, I. & SCHUBART, H. O. R. 1985. Ecological research on arthropods in central amazonian forest ecosystems
with recommendations for study procedures. Pp. 111-144 in Cooley, J.H. & Golley, F.B. (eds). Trends in
Ecological Research for the 1980's. Nato Conference Series, Series 1: Ecology. Plenum Press, New York,
London.
ANDRE, H.M., LEBRUN, P. & NOT!, M. 1992. Editorial. Biodiversity in Africa: a plea for more data. Journal
of African Znology 106:3-15.
ANDRE, H.M., LEBRUN, P. & NOTI, M. 1994. The soil fauna: the other last biotic frontier. Biodiversity and
Conservation 3: 45-56.
BAGINE, R.K.N. 1984. Soil translocation by termites of the genus Odontotennes (Holmgren) (Isoptera:
Macrotermitinae) in an arid area of Northern Kenya. Oecologia (Berlin) 64:263-266.
BARLOW, H.S. & WOIWOOD, LP. 1990. Seasonality and diversity of macrolepidoptera in two lowland sites in
the Dumoga-Bone National Park, Sulawesi Utara. Pp. 167-172 in Knight, W.J. & Holloway, J.D. (eds). Insects
and the Rain Forests of South East Asia (Wallacea). Royal Entomological Society of London, United Kingdom.
BEATTIE, A.I. 1985. The Evolutionary Ecology of Ant-Plant Mutualisms. Cambridge University Press, New York.
BROKAW, N.V.L. 1984. Treefalls, regrowth, and community structure in tropical forests. In Pickett, S.T.A. &
White, P.S. (eds). Natural disturbance: the patch dynamics perspective. Academic Press, New York.
CASSON, D.S. & HODKINSON, LD. 1991. The Hemiptera (Insecta) communities of tropical rain forest in
Sualwesi. Znological Journal of the Linnean Society 102:253-275.
di CASTRI, F. 1991. Ecosystem Evolution and Global Change. Pp. 189-217 in Solbrig, O.T. & Nicolis G. (eds.).
Perspectives on Biological Complexity. ruBS Monograph Series No 6, Paris.
di CASTRI, F. & YOUNeS, T. 1990. Ecosystem Function of Biological Diversity. Biology International, Special
Issue 22.
di CASTRI, F., ROBERTSON VERNHES, J. & YOUNeS, T. 1992. Inventorying and Monitoring Biodiversity.
Biology International, Special Issue 27, 28 pages.
CHERRETT, J.M. 1986. History of the leaf-cutting ant problem. Pp. 10-17 in Lofgren, C.S. & Van der Meer, R.K.
(eds). Fire Ants and Leaf-Cutting Ants; Biology and Management. Westview Press, Boulder, Colorado.
CHERRETT, I.M. 1989. Leaf-cutting ants. Pp. 473-488 in Lieth, H. & Werger, M.l.A. (eds). Tropical Rain Forest
Ecosystems: Biogeographical and Ecological Studies. Ecosystems of the World 14B. Elsevier, Amsterdam.
CHIVERS, D.l. & DAVIS, G. 1979. Abundance of primates in the Krau Game Reserve, Peninsular Malaysia. Pp.
9-32 in Marshall, A.G. (ed.). Transactions of the Sixth Aberdeen-Hull SYmPosium on Malesian Ecology. Dept.
of Geography, University of Hull.
COLLINS, N.M. 1979. Comparison of the soil macrofauna of three lowland forest types. Sarawak Museum Journal
4, 27:267-282.
COLLINS, N.M. 1980. The distribution of the soil macrofauna on the west ridge of Gunung (Mount) Mulu,
Sarawak. Oecologia (Berlin) 44:263-275.
COLLINS, N.M. 1981. The role of termites in the decomposition of wood and leaf litter in the southern Guinea
savanna of Nigeria. Oecologia (Berlin) 51:389-399.
COLLINS, N.M. 1983. Termite populations and their role in litter removal in Malaysian rain forests. Pp. 311-325
in Sutton, S.L., Whitmore, T.C. & Chadwick A.C. (eds). Tropical Rainforest: Ecology and Management.
Blackwell Scientific Publications, Oxford.
COMMON, LF.B. 1990. Moths of Australia. Melbourne. vi + 535pp.
CONSTANTINO, R. 1992. Abundance and diversity of termites (Insecta: Isoptera) in two sites of primary rain
forest in Brazilian Amazonia. Biotropica 24:420-430.
CORNELL, H.V. & LAWTON, I.H. 1992. Species interactions, local and regional processes, and limits to the
richness of ecological communities: a theoretical perspective. Journal of Animal Ecology 61: 1-12.
COVENTRY, R.I., HOLT, J.A. & SINCLAIR, F. 1988. Nutrient cycling by mound building termites in low
fertility soils of semi-arid tropical Australia. Australian Journal of Soil Research 26:375-390.
CURTIN, S .H. & CHIVERS, DJ. 1978. Leaf-eating primates of Peninsular Malaysia. Pp. 441-464 in Montgomery,
G.G. (ed.). Ecology of Arboreal Folivores. Smithsonian Institution, Washington D.C ..
DALLMEIER, F. 1992. Long-term monitoring of biological diversity in tropical forest areas: methods for
establishment and inventory of permanent plots. MAB Digest II. UNESCO, Paris. 72pp.
18 N.E. STORK
DAVIDSON, D.W. & FISHER, B.L. 1991. Symbiosis of ants with Cercropia as a function of light regime. pp. 289-
309 in Huxley, C.R. & Cutler, D.F. (eds). Ant-Plant Interactions. Oxford University Press, Oxford.
DAVIDSON, G.W.H. 1991. Birds. pp. 135-142 in Kiew, R. (ed.). The State of Nature Conservation in Malaysia.
Malayan Nature Society, Malaysia.
DEANGELIS, D.L. 1992. Dynamics of nutrient cycling and food webs. Population and community biology series,
9. Chapman & Hall, London, 27Opp.
DELCOURT, H.R. & DELCOURT, P.A. 1987. Quaternary EcolOgy. A Paleoecological Perspective. Chapman &
Hall, London. 242pp.
DISNEY, R.H.L. 1990. Scuttle flies (Diptera: Phoridae) associated with ants and termites in Toraut Forest,
Sulawesi. Pp. 301-304 in Knight, W.J. & Holloway, J.D. (eds). Insects and the Rain Forests of South East Asia
(Wallacea). Royal Entomological Society of London, United Kingdom.
EGGLETON, P. & BIGNELL, D.E. in press. Monitoring the response of tropical insects to changes in the
environment: troubles with termites. In Harrington, R. & Stork, N.E. (eds). Insects in a changing environment.
Academic Press, London.
EGGLETON, P., BIGNELL, D.E., SANDS, W.I., WAITE, B., WOOD, T.G. & LAWTON, J.H. in press. The
species richness of termites (Isoptera) under differing levels of forest disturbance in the Mbalmayo Forest
Reserve, southern Cameroon. Journal of Tropical Ecology.
ELKINS, N.Z., SABOL, G.V. WARD, T.J. & WHITFORD, W.G. 1986. The influence of subterranean termites
on the hydrological characteristics of a Chihuahuan desert ecosystem. Decologia (Berlin) 68:521-528.
ERWIN, T.L. 1988. Canopy arthropod biodiversity: a chronology of sampling techniques and results. Revista
Reruana de Entomologia 32:71-77.
FITTKAU, E.J. & KLINGE, H. 1973. On biomass and trophic structure of the central Amazonian rain forest
ecosystem. Biotropica 5:2-14.
FOSTER, R.B. & HUBBELL, S.P. 1990. The floristic composition of the Barro Colorado Island Forest. pp. 85-98
in Gentry, A.H. (ed.). Four Neotropical Forests. Yale University Press.
FURLEY, P.A., PROCTOR, J. & RATTER, J.A. (eds) 1992. Nature and Dynamics of Forest-Savanna Boundaries.
Chapman & Hall, London.
GADAGKAR, R., CHANDRASHEKARA, K. & NAIR, P. 1990. Insect species diversity in the tropics: sampling
methods and a case study. Journal of the Bombay Natural History Society 87:337-53.
GAMEZ, R. 1991. Biodiversity conservation through facilitation of its sustainable use: Costa Rica's National
Biodiversity Institute. Trends in Ecology and Evolution 6:377-8.
GASTON, K.J. 1991. The magnitude of global insect species richness. Conservation Biology 5:283-296.
GASTON, KJ. & MAY, R.M. 1992. Taxonomy of taxonomists. Nature 356:281-282.
GOFFINET, G. 1973. Contribution a l'etude de l'ecosysteme foret claire (miombo). Note 11. Etude comparative
des effectifs de quelques groupes arthropodiens du sol intercalique de quatre biotopes Katangais (Zaire). Annuaire
Universitaire Abidjan Serie E 6:25 I -256.
GRASSLE, J.F. & MACIOLEK, N.J. 1992. Deep-sea richness: regional and local diversity estimates from
quantitative bottom samples. American Naturalist 139:313-41.
GRUBB, P.J. 1977. The maintenance of species-richness in plant communities: the importance of the regeneration
niche. Biological Reviews 52:107-145.
HAMMOND, P. 1990. Insect abundance and diversity in the Dumoga-Bone National Park, N. Sulawesi, with
special reference to the beetle fauna of lowland rain forest in the Toraut region. pp. 197-254 in Knight, W.J.
& Holloway, J.D. (eds). Insects and the Rain Forests of South East Asia (Wallacea). Royal Entomological
Society of London, United Kingdom.
HAMMOND, P.M. 1992. Species Inventory. Pp. 17-39 in Groombridge, B. (ed.). GlobalDiversity. Status of the
Earth's Living Resources. World Conservation Monitoring Centre, Chapman & Hall, London.
HAMMOND, P.M. in press. Practical approaches to the estimation of the extent of biodiversity in speciose groups.
Philosophical Transactions of the Royal society, Series B.
HAMMOND, P:M. in press. Described and estimated species numbers: an objective assessment of current
knowledge. Philosophical Transactions of the Royal society, Series B.
HAWKSWORTH, D.L. & RITCHIE, M. in press. Tropical Areas Bioindicators System (TABS). CAB International,
Wallingford.
HEYER, W.R., DONELLY, M.A., McDIARMID, R.W., HAYEK, L.-A.C. & FOSTER, M.S. (eds) 1994.
Measuring and monitoring biological diversity. Standard methods for amphibians. Smithsonian Institution,
Washington D.C. 364pp.
HIGASHI, M., ABE, T. & BURNS, T.P. 1992. Carbon-nitrogen balance and termite ecology. Proceedings of the
Royal Society of London, Series Biological Sciences 249:303-308.
HODDA, M., BLOEMERS, G.F., WANLESS, F.R. & LAMBSHEAD, P.J.D. in press. Nematodes and carbon
flux in a tropical forest in Cameroon: preliminary results. In Nilsson, L.-O. & Mathy, P. (eds). Nutrient Uptake
and Cycling in Forest Ecosystems. C.E.C., Brussels.
HODKINSON, I.D. & CASSON, D. 1987. A survey of the food plant utilization by Hemiptera (Insecta) in the
understorey of primary lowland rain forest in Sulawesi, Indonesia. Journal of Tropical Ecology 3:75-85.
HODKINSON, LD. & CASSON, D. 1991. A lesser predilection for bugs: Hemiptera (Insecta) diversity in tropical
rain forests. Biological Journal of the Linnean Society 43:101-109.
HOLLDOBLER, B. 1983. Territorial behaviour in the green tree ant (Oecophylla smaragdina). Biotropica 15:241-
250.
HOLLDOBLER, B, & WILSON, E.O. 1990. The ants. Belknap Press of Harvard University Press, Cambridge,
Massachusetts.
HOLLOW A Y, J.D., ROBINSON, G.S. & TUCK, K.R. 1990. Zonation in the Lepidoptera of northern Sulawesi.
pp. 153-166 in Knight, W.L & Holloway, J.D. (eds). Insects and the Rain Forests of South East Asia
(Wallacea). Royal Entomological Society of London, United Kingdom.
HOLLOWAY, J.D. & STORK, N.E. 1991. The dimensions of biodiversity: the use of invertebrates as indicators
of human impact. pp. 37-62 in Hawksworth, D.L. (ed.). The Biodiversity of Microorganisms and Invertebrates:
Its role in Sustainable Agriculture. CAB International, Wallingford.
HUXLEY, C.R. 1978. The ant-plants Myrmecodia and Hydnophytum (Rubiaceae), and the relationships between
their morphology, ant occupants, physiology and ecology. New Phylologist 80:231-268.
HUXLEY, C.R. 1980. Symbiosis between ants and epiphytes. Biological Reviews 55:321-340.
INGER, R.F. 1979. Abundances of amphibians and reptiles in tropical forests of South-East Asia. pp. 93-110 in
Marshall, A.G. (ed.). Transactions of the Sixth Aberdeen-Hull Symposium on Malesian Ecology. Department
of Geography, University of Hull.
IUCN, UNEP & WRI 1990. The Biodiversity Conservation Strategy Programme, Mobilizing Worldwide action to
Sustain the Living Resources of our Planet. Gland, Switzerland.
JANZEN, D.H. 1967. Interaction of the bull's horn acacia (Acacia comigera L.) with an ant inhabitant
(Pseudomyrmex ferruginea F. Smith) in eastern Mexico. University of Kansas Science Bulletin 47: 315-558.
JOHNS, A.D. 1992. Vertebrate responses to selective logging: implications for the design of logging systems.
Philosophical Transactions: Biological Sciences of the Royal Society of London 335(1275):443-457.
JONES, LA. 1990. Termites, soil fertility and carbon cycling in dry tropical Africa: a hypothesis. Journal of
Tropical Ecology 6:291-305.
KELLERT, S.R. 1993. Values and perspections of invertebrates. Conservation Biology 7:845-855.
KIM, K. C. 1993. Biodiversity, conservation and inventory: why insects matter. Biodiversity and Conservation
2:191-214.
KNIGHT, W. & HOLLOWAY, J.D. (eds) 1990. Insects and the Rain Forests of South East Asia (Wallacea). Royal
Entomological Society of London, United Kingdom.
LEE, K.E. 1983. The influence of earthworms and termites on soil nitrogen cycling. Pp. 35-48 in Lebrun, P.,
Andre, H.M., de Medts, A., Gregoire-Wibo, C. & Wauthy, G. (eds). New Trends in Soil Biology. Dieu-
Brichart, Louvain-Ia-Neuve.
LEE, K.E. & WOOD, T. G. 1971. Termites and Soil. Academic Press, London.
LESTON, D. 1973a. Ecological consequences of the tropical ant mosaic. Proceedings of the Seventh International
Congress of the International Union for the Study of Social Insects. London 19:311-341.
LESTON, D. 1973b. Ants and tropical tree crops. Proceedings of the Royal Entomological Society of London, series
C 38:1.
LONGINO, LT. 1989. Geographic variation and community structure in an ant-plant mutualism: Azteca and
Cercropia in Costa Rica. Biotropica 21: 126-132.
MAJER, LD. 1976a. The maintenance of the ant mosaic in Ghana cocoa farms. Journal of Applied Ecology 13:123-
144.
MAJER, LD. 1976b. The ant mosaic in Ghana cocoa farms: further structural considerations. Journal of Applied
Ecology 13:145-155.
MAJER, LD. 1976c. The influence of ants and ant manipulation on the cocoa farn1 fauna. Journal of Applied
Ecology 13: 157-175.
MAJER, J.D. 1993. Comparison of the arboreal ant mosaic in Ghana, Brazil, Papua New Guinea and Australia -
its structure and influence on arthropod diversity. Pp. 115-141 in LaSalle, 1. & Gauld, LD. (eds). Hymenoptera
and Biodiversity. CAB International, Wallingford, United Kingdom.
MALDAGUE, M.E. 1964. Importance des populations de termites dans les sols equitoriaux. Transactions of the 8th
International Congress of Soil Science, Bucharest 3:743-751.
MAY, R.M. 1988. How many species are there on earth? Science 241:1441-9.
MAY, R.M. 1992. How many species inhabit the earth? Scientific American, October, 18-24.
McKEY, D. 1992. Interactions between ants and plants: comparison of canopy, understorey and clearing
environments. In: Biologie d'une Canopee de foret Equatoriale - II Rapport de Mission: Radeau des Cimes
Octobre Novembre 1991, Reserve de Campo, Cameroun.
MEDWAY, Lord, 1983. The Wild Animals of Malaya (Peninsular Malysia and Singapore. 2nd ed. Oxford
Univ.Press,Kuala Lumpur.
20 N.E. STORK
MEMMOTT, J., GODFRAY, H.C.J. & BOLTON, B. 1993. Predation and parasitism in a tropical herbivore
community. Ecological Entomology 18:348-352.
NOYES, J.S. 1989a. The diversity of Hymenoptera in the tropics with special reference to Parasitica in Sulawesi.
Ecological Entomology 14:197-207.
NOYES, J.S. 1989b. A study of five methods of sampling Hymenoptera (Insecta) in a tropical rainforest, with
special reference to the Parasitica. Journal of Natural History 23:285-298.
PARKER, S.P. (ed.) 1982. Synopsis and c/ossijication of living organisms. 2 Volumes. McGraw-Hill, New York.
PAYNE, J.B. 1979. Abundance of Diurnal Squirrels at the Kuala Lompat Post of the Krau Game Reserve, Pahang,
Peninsular Malaysia. pp. 37-53 in Marshall, A.G. (ed.). Transactions of the Sixth Aberdeen-Hull Symposium
on Malesian Ecology. Department of Geography, University of Hull.
PAYNE, J. & MAHEDI, A. 1991. Large Mammals in Sabah. Pp. 177-183 in Kiew, R. (ed.). The State of Nature
Conservation in Malaysia. Malayan Nature Society, Malaysia.
PIERCE, N. E. 1983. The ecology and evolution of symbioses between lycaenid butteiflies and ants. Ph.D.
dissertation. Harvard University x + 274pp.
PIERCE, N. E. 1987. The evolution and biogeography of associations between lycaenid butterflies and ants. Pp.
89-116 in Harvey, P.H. & Partridge, L. (eds). Oxford surveys in evolutionary biology, volume 4. Oxford
University Press, Oxford.
ROGERS, L.E., HINDS, W. T. & BUSCHBOM, R.L. 1976. A general weight vs length relationships for insects.
Annals of the Entomological Society of America 69:387-389.
ROOM, P.M. 1971. The relative distributions of ant species in Ghana's cocoa farms. Journal of Animal Ecology
40:735-751.
SCHAEFER, M. & SCHAUERMANN, J. 1990. The soil fauna of beech forests: comparison between a mull and
a moder soiL Pedobiologia 34:299-314.
SCH UPP, E. W. 1986. Azteca protection of Cercropia: ant occupation benefits juvenile trees. Oecologia (Berlin)
70:379-385.
STORK, N.E. 1988. Insect diversity: facts, fiction and speculation. Biological Journal of the Linnean Society
35:321-337.
STORK, N .E. 1991. The composition of the arthropod fauna of Bornean lowland forest trees. Journal of Tropical
Ecology 7: 161-80.
STORK, N.E. 1993. How many species are there? Biodiversity and Conservation 2: 215-232.
STORK, N.B. 1994. Inventories of biodiversity: more than a question of numbers. Pp. 81-100 in Forey, P.,
Humphries, C.J. & Vane-Wright, R.1. (eds). Systematics and Conservation Evaluation. Systematics Association
Special Volume no 50, Oxford University Press.
STORK, N.E. & BLACKBURN, T.M. 1993. Abundance, body size and biomass of arthropods in tropical forest.
Oikos 67: 483-489.
STORK, N .E. & BRENDELL, M.J.D. 1990. Variation in the insect fauna of Sulawesi trees with season, altitude
and forest type. Pp. 173-190 in Knight, W.J. & Holloway, J.D. (eds). Insects and the rainforests of South East
Asia (Wallacea). Royal Entomological Society of London, London.
STORK, N .E. & BRENDELL, M.J .D. 1993. Arthropod diversity studies in lowland rainforest ofSeram, Indonesia.
Pp in The Natural History ofSeram. Edwards, I., Macdonald, A.A. & Proctor, J. (eds). Intercept, Andover.
STRICKLAND, A.H. 1944. The arthropod fauna of some soils. Tropical Agriculture (University of the West Indies,
Trinidad & Tobago) 21:107-114.
ULE, E. 1902. Ameisengarten im Amazonasgebiet. Botanische Jahrbucher /iir Systematik, ptlanzengeschichte und
ptlanzengeographien 30:45-52.
WAY, M.J., CAMMELL, M.E., BOLTON, B. &KANAGARATNAM, P. 1989. Ants (Hymenoptera: Formicidae)
as egg predators of coconut pests, especially in relation to biological control of the coconut caterpillar, Opisina
arenosella Walker (Lepidoptera: Xyloryctidae), in Sri Lanka. Bulletin of Entomological Research 79:219-233.
WEAVER, P.L. 1989. Forest changes after hurricanes in Puerto Rico's Luquillo Mountains. Interciencia 14:181-
192.
WHITMORE, T.C. 1984. Tropicalforests of the Far East. Clarendon Press, Oxford, 2nd Edition.
WILSON, E.O. 1971. The Insect Societies. Belknap Press of Harvard University Press, Cambridge, Massachusetts.
548pp.
WILSON, E.O. 1988. The current state of biological diversity. Pp 3-18 in Wilson, B. O. & Peter, F.M. (eds).
Biodiversity, National Academy Press, Washington D.C.
WOINARSKI, J.C.Z. & TIDEMANN, S.C. 1991. The bird fauna of a deciduous woodland in the wet-dry tropics
of northern Australia. Wildlife Research 18:469-500
WOOD, T.G. 1988. Termites and the soil environment. Biology and Fertilily of Soils 6:228-236.
WOOD, T.G., JOHNSON, R.A., BACCHUS, S., SHITU, M.O., & ANDERSON, J.M. 1982. Abundance and
distribution of termites (lsoptera) in a riparian forest in the southern Guinea savanna vegetation zone of Nigeria.
Biotropica 14:25-39.
The biodiversity of arthropods from Australian rain forest canopies: some
results on the role of the tree species
R. L. KITCHING and 1. ZALUCKI
Australian School o/Environmental Studies, Griffith University, Nathan, Brisbane, Qld 4111,
Australia
ABSTRACT. Australia's modest tropical rainforest resources have been recognised both by measures
taken to conserve them in recent years and by the establishment of a major Cooperative Research Centre
to investigate their biodiversity and dynamics to underpin proper management as well as to generate
biological understanding. Investigations of the arthropod biodiversity of the canopy strata have been
underway for several years. Over the last two years the focus of this work has moved to include sites
in the lowland forests of Cape Tribulation. These canopy studies take a 'top-down' approach to
biodiversity definition generating, in sequence, ordinal, familial and specific profiles as well as guild
structures for particular sites, samples or target trees. Recent work, following earlier hypotheses of Terry
Erwin, has included investigations of the tree-species specificity of arthropod assemblages focussing in
particular on the Coleoptera. Early results of these studies are presented, in particular relating to the
canopy faunas of the very local endemic tree Noahdendron nicho/asii and the more widespread Ryparosa
javanicum. Difficulties associated with sampling, analysis and the identification of specificity at the level
of the assemblage are discussed. Other work in progress and extension of the tree-specificity work under
the aegis of the newly proposed Tropical Rainforest Research Centre is foreshadowed.
KEY WORDS: biodiversity, arthropods, Australia, Coleoptera, canopy sampling, tropical rainforest.
INTRODUCTION
Rainforest covers only 5 % of the Australian continent but, in terms of biodiversity, public
interest and global importance, this charismatic ecosystem looms much larger in the public
consciousness than its area would suggest. The Australian biological community and the
political forces at work in the country are keenly aware of the importance of our rainforests
and the biological diversity they contain. The recent announcement of federal government
support, to the tune of 13 million Australian dollars, for the establishment of a Cooperative
Research Centre for Tropical Rainforest Ecology and Management, is perhaps the most
tangible evidence of this awareness among scientists and their paymasters. The proposal for
and approval of this research centre, scheduled to open in mid-1993, coincides with the
development of a National Strategy for Biodiversity Conservation, the draft of which was
released in late 1991 and the final version of which is due for final release shortly.
At the Kuala Lumpur conference in 1991 attention was drawn to the rainforest resources
of Australia (Kitching 1992). At that stage the Government of Queensland was in dispute
with the federal government concerning the future declaration of a substantial area of
remaining tropical rainforest area as a World Heritage property. Between the presentation
of that paper and the appearance of the printed version there was an unexpected change of
government at the state level, the World Heritage Dispute was resolved and, in December
1990, the Wet Tropics of Queensland was inscribed as a World Heritage property. Since
then a statutory body, the Wet Tropics Management Authority, has been established with
both federal and state support and is now designing and supporting a coordinated programme
of research to develop management strategies for the region in line with Australia's
responsibilities nationally and internationally. This programme includes an extensive
programme of biological inventory across all taxa.
21
D. S. Edwards et al. (eds.), Tropical Rainforest Research - Current Issues, 21-28.
22 R.L.KITCHING AND J. ZALUCKI
Canopy Arthropod Studies

It is now generally accepted that the arthropods in the canopies of rainforest trees
represent 'the heart of biodiversity' to paraphrase E. O. Wilson. Following pioneering
studies by Erwin (1982, 1983) in the Neotropics, Stork and his co-workers carried out
seminal studies in Brunei, in Sulawesi and in Seram (Stork 1987a, 1987b, Stork & Brendell,
in press). These workers developed pyrethrum-knockdown techniques to obtain large
samples from high in the forest canopy. These samples, duly sorted, have revealed a
massive speciosity unrivalled in the animal world. Erwin (1982) in particular was so
impressed by the results he obtained in Panama that he erected a series of testable hypotheses
which, if upheld, lead to upward revision of our estimates of global arthropod biodiversity
by more than an order of magnitude.
Parenthetically these new estimates have caused such a furore in some circles as to have
engendered a whole secondary industry of comment and re-analysis which continues more
than ten years after their original publication (Erwin 1992, Gaston 1991, Hodkinson &
Casson 1988, May 1988, Stork 1988). What has happened, but rarely, is any attempt to test
the sequential hypotheses on which Erwin's estimates rest. It is into this last partial vacuum
that our present contribution falls!
METHODS
Canopy Studies in Australia

Since 1988 we have been using comparable pyrethrum-knockdown techniques to obtain
data on the arthropod assemblages from Australian rainforest canopies. Our work initially
focussed on the cool temperate and subtropical forests of northern New South Wales and
south-east Queensland but, since 1991, has expanded to include lowland tropical sites in
Australia's north. Our basic approach, methods, field sites and ordinal level results are
presented by Kitching et al. (1993).
To date we have taken two separate approaches to the study: the collection of 'baseline'
data from a set of 1Oxl0m plots within which we carry out parallel surveys of woody plants;
and, studies of the canopy faunas of individual trees of known species. The basic methods
and some results of the baseline studies will appear elsewhere ( Kitching & Arthur 1993,
Kitching et al., in press).
The tree species specificity project is designed to examine the first of Erwin's sequential
hypotheses: that is that the average host specificity of Coleoptera on tropical rain forest trees
is l3.5 %. This is the first of the multipliers used in Erwin's calculation and is one to which
the overall calculation is very sensitive: a 10% figure, for instance, will produce a global
estimate 25 % smaller than that arrived at by Erwin; a 20 % figure one that is 50 % greater.
We hasten to add that our aim is not to produce better global estimates of arthropod
biodiversity, which must be regarded, at best, as an elusive goal, and at worst an
unachievable diversion from the real business in hand! There can be no doubt, however, that
the estimation of tree species specificity, which is fundamental to an understanding of how
canopy faunas are assembled, presents formidable technical difficulties, but is nevertheless
an attribute vital to any biodiversity conservation activities.
Tree-specificity Studies
Our work on tree specificity remains at an early stage. We have selected a number of
understorey trees the growth forms of which facilitate sampling. Each of these is between
six and 12 metres in height, growing more or less isolated (or easily isolatable!) from its
neighbours. In each case a set of 0.5 m2 collecting hoops are disposed around the base of
the subject tree and a painters' scaffold erected adjacent to it. The canopy of the tree is then
Arthropod biodiversity in Australian rainforest canopies 23
sprayed with a commercially available mixture of pyrethrins (see Kitching et at., in press,
for details of composition, dilution etc) for about six minutes using a backpack dispenser with
the operative on the scaffold. All samples are taken in the morning to reduce variance
associated with diel activity patterns. This somewhat simple approach allows a carefully
controlled and thorough sampling of the particular canopy of interest but is less effective in
dense forest where canopies are interdigitated.
Arthropod samples are collected three to four hours after the spraying event. Samples are
sorted to Order at field laboratories and returned to the laboratory in ethanol. Subsequently
particular Orders are further curated and sorted to family and morphospecies. We chose to
focus further on the Coleoptera, an Order in which we are developing particular expertise
and one which allows for direct comparison with the results of workers elsewhere in the
world. Following removal from liquid storage specimens are pointed and separately labelled
before further sorting. Other work, not discussed further here, examines the Collembola,
Psocoptera, Lepidoptera and Araneida from our samples.
The results presented below represent the partial outcome from our 1992 field season
during which, inter alia, we sampled six trees each of two known species: Ryparosa
javanica, a widespread species in Australia and elsewhere, and Noahdendron nicholasii, a
very restricted endemic found only in the Noah Creek and adjacent catchments (see below).
STUDY SITE
The results to be described all originate from sites located in the catchment of Noah Creek,
a permanent waterway some 11 km south of Cape Tribulation within the greater Daintree
area of the Wet Tropics World Heritage area. The vegetation of the region has been
described by Tracey (1982) and species lists of the trees and climbing plants of Cape
Tribulation have been published by Jessup and Guymer (1985) and McKenzie (1985). The
vegetation of the actual sites sampled has been described (McIntyre, Kitching & Jessup, in
press). The forest is a complex mesophyll vine forest following the structural-physiognomic
classification of Webb (1959) and Webb et at. (1976). Palms are an important element of
the vegetation, with Calamus spp. being abundant on all sites; Licuala ramseyi, Linospadix
spp. and Normanbya normanbyi were also present although in lesser numbers. Numbers of
tree and shrub species in each 10 x 10 m plot ranged from about 12 to 22.
The climate of the neighbouring village of Cape Tribulation is described by Kitching et
al. (in press). Although there may be considerable falls of rain throughout the year, the
area has a distinct wet season centred in February and March, with monthly falls of over 600
mm. The Noah Creek catchment is substantially wetter than this although not separately
instrumented. Single rainfall events of over 225 mm are not unknown. There is a small
annual temperature range with mean maxima ranging from 27°C in July to 34°C in January,
and mean minima ranging from 21°C in July to 26°C in January.
This largely undisturbed catchment is one of a number that are widely recognised as
containing relictual fauna and flora and almost certainly were refugia during the last
glaciation when broad-leaved rainforest in Australia retreated to a number of such locations
along the north-east coast. There is substantial interest in this and neighbouring catchments
both as conservation objects and natural laboratories for the study of rainforest processes.
It is one of the sites designated by the Wet Tropics Management Authority for which the only
designated use is scientific research.
24 R.L.KITCHING AND 1. ZALUCKI
RESULTS
Figure 1 presents the ordinal profiles for insects and non-insectan arthropods from the two
species of tree. The insects profile shows some key differences that warrant further
attention: Collembola and Psocoptera are far more abundant on Ryparosa than Noahdendron
whereas the reverse is true for the Orthoptera, Homoptera and Hymenoptera. The relative
abundance of some Orders (eg. Blattodea, Coleoptera, Thysanoptera) appears unaffected by
the tree species. For the non-insects all three groups present were somewhat more abundant
on Noahdendron but the differences were not large.
Focussing on the Coleoptera, 135 individuals sorted to 59 species, 38 on Noahdendron
and 35 on Ryparosa (14 being shared). These 59 species represent 22 families (Figure 2)
dominated for both tree species by the Curculionidae, Coccinellidae and Corylophidae.
The species profiles for each species are presented separately in Figure 3. Essentially
both assemblages show the highly skewed distributions characteristic of such data from
virtually all ecosystems. Only a few species are present in any numbers, the greater part of
the species diversity being contributed by 'rare' species. For Ryparosa, of 35 species only
3 occurred in the samples more than three times, 27 species occurred as singletons. For
Noahdendron comparable figures are 6 and 29 respectively, out of 38 species.
4 INSECTS
....<(I)'"...
0
=1
~ g 3
~:
Q "
zoo!
: g 2
ON
~ E
...,1Il
Q:l Q 1
=.
I: ..
z,9
-< -< -< -<

-'
o '"co0 ~ ~
c:I
I: I- '" l-'"e.
I-
e.
'"
-' I-
-< eo 0
z
3-' -'
c:I e.
'" I:
'"
-' ~
NON-INSECTS
c:J Noahdendron
• Ryparosa
o
ARANEIDA ACARINA ISOPODA
Figure 1. Distribution of insects and non-insects in samples from the canopies of Ryparosa javanica and
Noahdendron nicholasii from Cape Tribulation, North Queensland.
30
.;
..." 20
'>
:;;
.=
.
':;
......e 10
z"
o Noahdendron
• R~porO'8
Figure 2. Distribution of Coleoptera into families in samples from the canopies of Ryparosa javanica and
It is when these data are plotted together (Figure 4) that the overlap between the two
assemblages become apparent and the question of 'specificity' can be addressed. Of the six
'commonest' species, only one is not shared by both tree species. That 'exclusive' species,
a coccinellid, occurred on Ryparosa. Two species were almost 'exclusive' with just single
individuals breaking the pattern. A coccinellid primarily on Ryparosa and a galerucine
chrysomelid primarily on Noahdendron made up this set. More exclusivity was to be found
among the rarer species, as indicated by asterisks in Figure 4, but the real contribution to
beta diversity is made, by definition, by the singletons.
DISCUSSION
Top-down Approaches
The results we have presented can do little more than act as a vehicle for the posing of
further questions at this stage. Attempts to define a figure for host tree specificity using the
top-down approach we have taken can never do more than produce an indication of sample
overlap - that is they can never be reliable guides as to whether or not any particular species
is monophagous or polyphagous. Only detailed autecological studies can do that and they
are not compatible, in general, with the sort of assemblage-level studies we have set
ourselves to do. Nevertheless, in terms of defining beta, and subsequently gamma, diversity
it is reliable measures of this overlap that are required.
30
R'plrll.Ul ./1It'lIl1icllllt
20
.
..."
E
z"
10
0
10 20 30
S~il$Rk
30 I A'#IhtleRtlrlllllicMlnii I
..
...."
20
E
z"
10
10 20 30
S~il$Rk
Figure 3. Numbers of individuals of species of Coleoptera in samples from the canopies of Ryparosajavanica and
30
., 25
0 Noahdendron
'"
=
"C
:~"C
20 • Ryparosa
...= 15
..
C>
10
.<:I
S
z= 5
0
5 10 15 20 25 30 35 40 45 50 55
Species Number
Figure 4. Overlap between the beetle fauna in samples from the canopies of Ryparosajavanica and Noahdendron
nicholasii, from Cape Tribulation, North Queensland.
The Importance of Oligophagy

Erwin's specificity-multiplier is naive in the sense that it assumes that a beetle species is
either tree species specific or not - that is, either monophagous or polyphagous. It does not
allow for the various degrees of oligophagy - and yet these are probably more the rule than
the exception. Tree species exhibit various sorts of phylogenetic relationship and it is
entirely reasonable that two species of tree of the same genus, for instance, might be
expected to share more species than two less closely related ones. The continuation and
extension of our work in Australian forests will eventually allow quantification of the role
of such factors in determining this level of overlap in diversity of the canopy faunas of
co-occurring trees.
Current data shortcomings and Solow's analysis

Further analysis is required to give confidence that the numbers of specialists and
generalists identified within our current data set is other than an artifact of small sample sizes
and unequal distributions of species across replicates. However, M. P. Zalucki kindly
carried out for us an analysis following that of Solow (1993) in which the distributions of
beetle species and abundances within the assemblages from each tree species are compared
with one constructed randomly. This suggested that the observed assemblages on Ryparosa
and Noahdendron are significantly different from the one constructed on the null hypothesis
that each is but a different random sample of a common assemblage.
The epiphyte problem

Problems inherent in our method will always restrict the power of biodiversity analyses
of this kind. In particular, our selection for technical reasons of smaller, less mature trees
means that these seldom have heavy loads of macroepiphytes. Less selective samples on high
canopy trees inevitably include samples of the fauna of the rich epiphyte flora of these strata.
These epiphytes undoubtedly playa large part in generating the very rich arthropod speciosity
observed in canopy samples. At this stage separate sampling of epiphytes and their host trees
on a sufficient scale to generate useful statistical estimates presents formidable technical
difficulties with no obvious, affordable solutions.
Future Work
In July 1993 we are scheduled to spend an extended period in North Queensland
continuing the tree specificity study. We shall repeat our studies of Noahdendron nicholasii
and Ryparosa javanica, increasing replication and adding additional species including the
endemic Gymnostoma and the widespread Lindsayomyrtus racemoides. We shall continue
our 'baseline' studies obtaining dry season information for the first time and carry out light
trap surveys of canopy Lepidoptera.
Our involvement in the Tropical Rainforest Cooperative Research Centre will also allow
us to extend this and other work along an altitudinal base coordinated with extensive
vegetation surveys and studies of the determinants of biodiversity in other animal groups,
through the application of a centrally maintained database and associated geographical
information systems facilities. The large databases on a few arthropod assemblages which
we are accruing will also provide the context within which to define predictor-subsets of taxa
which can be used statistically to predict the species richness of the whole set. Such
techniques of Rapid Biodiversity Assessment, backed by adequate baseline data, will provide
an essential and powerful tool for the assessment of impact and the efficacy of restoration,
on key components of biological diversity. This aspect of our work may ultimately be its
greatest significance.
ACKNOWLEDGEMENTS
The field work and subsequent analyses reported in this paper have been generously funded
by the Wet Tropics Management Authority, the Ian Potter Foundation and Earthwatch
(Australia). We are grateful to these bodies for their patient and generous support. We
thank Bill Upjohn for his provision of able technical support in the field, Heather Mitchell
for sorting, mounting and generally assisting in the handling of very large samples and
Graham Floater for assistance in the preparation of material for this paper. We are also
grateful to John Lawrence, coleopterist with the Australian National Insect Collection,
CSIRO Canberra, who has given freely of his time to advise and guide us during the sorting
of our beetle specimens. Earlier work by Peter Hammond upon our collections has created
a resource for which we remain grateful.
LITERATURE CITED
ERWIN, T. L. 1982. Tropical forests: their richness in Coleoptera and other species. Coleopterists' Bulletin
36:74-75.
ERWIN, T. L. 1983. Beetles and other arthropods of the tropical forest canopies at Manaos, Brasil, sampled with
insecticidal fogging techniques. pp. 59-75. in Sutton, S.L., Whitmore, T.C. & Chadwick, A.C. (eds). Tropical
Rain Forests: Ecology and Management. Blackwell Scientific Publications, Oxford.
ERWIN, T. L. 1991. How many species are there?: revisited. Conservation Biology 5:330-333.
GASTON, K. J. 1991. The magnitude of global insect species richness. Conservation Biology 5:283-296.
HODKINSON, I. D. & CASSON, D. 1991. A lesser predilection for bugs: Hemiptera (Insecta) diversity in tropical
JESSUP, L. W. & GUYMER, G. P. 1985. Vascular plants recorded from the Cape Tribulation area. Queensland
Naturalist 26:2-19.
MAY, R. M. 1988. How many species are there on Earth? Science 241:1441-1449.
MAY, R. M. 1990. How many species? Philosophical Transactions of the Royal Society of London B330:293-304.
KITCHING, R. L. 1992. Biodiversity, research and conservation in Australian rainforests. Malayan Nature Journal
45:31-54.
KITCHING, R. L., BERGElSON, J. M., LOWMAN, M. D., MCINTYRE, S. & CARRUTHERS, G., in press.
The biodiversity of arthropods from Australian rain forest canopies: general introduction, methods, sites and
ordinal results. Australian Journal of Ecology 18.
KITCHING, R. L. & ARTHUR, M. 1993. The biodiversity of arthropods from Australian rain forest canopies:
summary and the impact of drought. Selbyana 14:29-35.
MCINTYRE, S., KITCHING, R. L. & JESSUP, L. W., in press. Structure and composition of rain forest plots
at Cape Tribulation, Queensland. Proceedings of the Royal Society of Queensland.
MCKENZIE, F. 1985. Climbing plants from 'Pilgrim Sands', Cape Tribulation. Queensland Naturalist 26:20-25.
SOLOW, A. R. 1993. A simple test for change in community structure. Journal of Animal Ecology 62:191-193.
STORK, N. E. 1987a. Guild structure of arthropods from Bornean rain forest trees. Ecological Entomology
12:69-80.
STORK, N. E. 1987b. Arthropod faunal similarity of Bornean rain forest trees. Ecological Entomology
12:219-226.
STORK, N. E. 1988. Insect diversity: facts, fiction and speculation. Biological Journal of the Linnean Society
35:321-337.
STORK, N. E. & BRENDELL, M. J. D., in press. Arthropod diversity studies in lowland rain forest of Seram,
Indonesia. In Proctor, J. & Edwards, I. (eds). The Natural History of Seram.
TRACEY, J. G. 1982. The Vegetation of the Humid Tropical Region of North Queensland. CSIRO Division of
Plant Industry, Melbourne.
WEBB, L. J. 1959. A physiognomic classification of Australian rain forests. Journal of Ecology 47:551-570.
WEBB, L. J., TRACEY, J. G. & WILLIAMS, W. T. 1976. The value of structural features in tropical forest
typology. Australian Journal of Ecology 1:3-28.
Describing and comparing high invertebrate diversity in tropical forest -
a case study of small moths in Borneo
GADEN S. ROBINSON and KEVIN R. TUCK
Biodiversity Division, Department of Entomology, The Natural History Museum, Cromwell

Road, London SW75BD, United Kingdom
ABSTRACT. The description and comparison of invertebrate diversity in tropical habitats poses
enormous problems. Diversity indices, of which there is a plethora, are often seen as ends in themselves;
this has led to much sterile phenomenology which is opaque to most naturalists. We mimic field samples
by bootstrap sampling from computer-generated populations based on logarithmic series. This provides
a simple way of comparing samples in terms of their shared species, defining confidence limits and
describing diversity in easily understood terms. Field samples of Microlepidoptera from sites in Borneo,
including two sites 1 Ian apart in the Batu Apoi Forest Reserve, are used to demonstrate how simple
questions such as Are the moth faunas of these habitats the same?", OOHow diverse are they?" or OOHow
00
different are they?OO may be answered. Extrapolations from these field samples to minimum estimates of
the total number of species of a habitat or territory are attempted, and some of the problems and
limitations encountered in sampling are examined.
KEY WORDS; Microlepidoptera, biodiversity, tropical rainforest, Brunei Darussalam.
INTRODUCTION
In this paper we discuss the diversity of a group of invertebrate animals in lowland Bornean
tropical rainforest; specifically, we deal with Microlepidoptera - small moths - in the valley
of the Sungai Belalong in the Batu Apoi Forest Reserve in Brunei.
The Lepidoptera, moths and butterflies, have traditionally been divided into three groups -
the Macrolepidoptera, large moths, the Microlepidoptera, small moths, and the butterflies.
Studies of the diversity of 'Lepidoptera' have almost invariably concentrated on the large
moths. To complement previous studies we have focussed on the Microlepidoptera as they
comprise more than 50% of the known species of moths and are a significant but neglected
component of the invertebrate biodiversity of tropical forest. We include the Pyraloidea in
our definition of Microlepidoptera. Although many pyraloids are robust and powerful flyers
and are as large as many Macrolepidoptera, they have been omitted from most studies of
'Lepidoptera' diversity as being 'Microlepidoptera'. A noteworthy exception to this is the
work of Barlow & Woiwod (1989). For further discussion of this separation of the
Lepidoptera in studies of diversity see Robinson & Tuck (1993).
Diversity has come to mean many things to many people. Its measurement and
significance has become an area of considerable debate. We do not intend contributing to
this debate. Our approach is based on the principle that "if it ain't broke, don't fix it",
tempered with common sense derived from considerable experience of working with
Lepidoptera in the tropical forests of South-East Asia.
The diversity that we describe here falls in the category of 'alpha-diversity'. It is a
measure of the species-richness of a particular group pertaining in a particular habitat or
locality at a particular time. There are two ways to measure such species-richness. One is
to enumerate all the species present in, say, the Belalong valley. The second is to measure
the diversity of a sample or series of samples, compare these and extrapolate from them to
29
D. S. Edwards et al. (eds.). Tropical Rainforest Research - Current Issues. 29-42.
30 GAD EN S. ROBINSON AND KEVIN R. TUCK
40
40 ~
~
." 20
«"
.a
$
u
; Port Vila, 24.vi.1971:
~ 20 N=175. 5=32
~ ~_- _ _L -_ _ _ _ _ _ _ _ ~
50 Rank 100
Kuala Belalong site 2,

1S.vii.1991: N=174, 5 =146
50 Rank 100 150
Figures \ & 2 (inset). Rank abundance curve of (\) a high diversity sample of Microlepidoptera from lowland
dipterocarp forest in Borneo, and (2) a low diversity sample of Macrolepidoptera from grassland and gardens in
Vanuatu.
provide meaningful biological observations.

The first method is time-consuming: it necessitates the capture or at least recording of all
the species present of the group of organisms under discussion. A full enumeration of
Britain's Lepidoptera fauna (some 2500 species) has taken an army of naturalists some 200
years to achieve. It is impractical in the much more diverse tropical systems to contemplate
such enumeration in the short term except with a few small groups of large vertebrates.
The sampling approach, while subject to the biases to which any sampling system is
prone, has been applied widely in studies of tropical Lepidoptera with varying degrees of
success. The measurement of diversity, or at least the method that we shall use, relies upon
knowledge of the distribution of individual organisms among particular groups - in this case
species. We thus need to sort our samples into their constituent species and count the
numbers of each.
Measuring diversity
It is often obvious by inspection what is a high-diversity sample and what is a low-
diversity sample (Figures 1 & 2). The first sample is of Microlepidoptera collected at light
on 16 July 1991 on the eastern slope of the Belalong valley: it comprises 174 individuals
representing 146 species. Figure 1 is a linear plot showing the number of individuals of each
of these species in descending order of abundance. Only three species are represented by
more than two individuals and most species are represented by just a single specimen; the
resulting species-abundance curve is very long and flat.
The second sample is of Macrolepidoptera trapped on 24 June 1971 from the edge of a
field in Port Vila, Vanuatu. It is of almost exactly the same size as the Belalong sample,
consisting of 175 individuals, but contains only 32 species. Plotted in similar fashion to
Figure 1, it can be seen that the rank-abundance curve (Figure 2) is very differently shaped;
there are several abundant species and the curve, while hollow like that in Figure 1, is
considerably steeper. These samples, in the context of tropical Lepidoptera, are respectively
of high and low diversity.
The diversity of small moths in Borneo 31
Table L The number of species represented by different numbers of individuals in the entire (pooled) sample of
Microlepidoptera from Kuala Belalong, with the number predicted from the log-series.
SPECIES WITH: Observed Expected (log series)

I individual 320 309
2 individuals 116 116
3 individuals 49 58
4 individuals 33 32
5 individuals 15 20
6 individuals 19 12
7 individuals 6 8
8 individuals 4 5
9 individuals 2 3
10 individuals 0 3
RESIDUAL SPECIES (> 10) 7 6
At its simplest, for a sample consisting of a given number of individuals, the diversity will
be higher the greater the number of groups (species) in the sample.
In order to be able to understand, compare and manipulate samples we need a measure
of the species-richness expressed in each that is independent of sample size. A species-
abundance curve (such as that in Figure 1 or 2) may be summarised in terms of the number
of species contributing to each abundance category starting with the rarest. As illustration,
we take here the total sample of Microlepidoptera from the Belalong valley collected in 1991.
It consists of 1230 individuals comprising 571 species (Table 1). The decreasing frequency
of species in each class of (increasing) abundance is regular and predictable: it has long been
known that such frequency distributions conform to a mathematical series known as the
logarithmic series. The logarithmic series has been found to apply to an enormously wide
range of samples of different organisms of different sizes taken over different periods of time
(Fisher, Corbet & Williams 1943, Williams 1944, 1947, 1964).
Given the crucial elements N, the number of individuals, and S, the number of species in
the sample, the applicable log-series can be calculated. In the case of the Kuala Belalong
sample it will be seen by inspection that the expected (calculated) series matches closely the
series observed in the field sample (Table 1). The log-series has particular properties that
make it especially useful as a descriptor of samples:
1. It is relatively simple to generate log-series and manipulate them using a computer.
Populations (i.e. mixed-species 'assemblages') of different sizes can be generated using
parameters derived from samples so that samples can be 'magnified' or 'rarefied' and
hypotheses explored.
2. One of the parameters of the logarithmic series, alpha (a), relates Nand S directly and
is a diversity index. Any particular log-series is defined by the three parameters N, Sand
a, or by a combination of any two. The three are related by the formula:
S=a{loge(1 + (N/a»)
Alpha is most easily computed, given Nand S, by iteration and substitution in this
formula. The number of species represented by 1, 2, 3 (etc.) individuals is given by:
a(N/(N +a», a«N/(N +a»2J2), a«N/(N +a»3/3), etc.

32 GADEN S. ROBINSON AND KEVIN R. TUCK
100
.
7l.
.
a.
50
..
'C
~
-;;
~
0 0L-- - - - - - 5 0 L O - - - - - 1 - ' 0 0 0
Sample size
FIGURE 3: Shared species in pairs of similarly sized samples from the same mixed-species assemblage: variation
in percentage with the different sample size and diversity. [Data from bootstrap sampling procedure from an
assemblage of 107 individuals]
3. Random samples of a log-series are of themselves a log-series with the same ex as the
'parent' assemblage (although, of course, Nand S will be different).
The converse of the third point is that any sample must be from a putative mixed-species
assemblage that has the same diversity (ex) as the sample.
There are numerous other indices of diversity available; the reader is referred to Wolda
(1981, 1983), Magurran (1988) and Taylor (1978) for further discussion. These authors
provide additional justification for the use of Williams' ex.
Comparing samples
Our next requirement is for a means to compare samples in terms of their species
composition - do they represent the same or different assemblages?
One problem encountered in examining samples from high-diversity species-assemblages
is that pairs of samples even from the same assemblage may have only a very small
proportion of their species in common (Figure 3). Most indices of sample similarity or
dissimilarity such as Preston's coefficient (Preston 1962) assume exhaustive sampling of
species, rarely possible except in the case of a few small groups of large vertebrates. Wolda
(1981) has provided a 'consumer guide' to similarity indices. Preston's coefficient was not
included in his survey. We have used this index in the past (Holloway, Robinson & Tuck,
1990), and it is perhaps appropriate therefore that we demonstrate here its limitations in this
context. Like many of the indices tested by Wolda, Preston's coefficient is sample-size and
diversity-dependent (Figure 4) and is therefore not a good sample comparator.
This is not to condemn all such indices; Wolda (1981) found that modified versions of
Morisita's index (Magurran 1988, Morisita 1959) for example, did not suffer defects of
sample size- and diversity-dependence. However, computation of the Morisita index is
difficult and we find it impossible to conceptualise what it is actually measuring.
Faced with the problem of comparing samples, we turned to Wolda's (1981, 1983)
technique for randomly sampling computer-generated assemblages derived from the log-
series. The concept of comparison that appeared to us to be statistically valid while at the
same time remaining understandable was that of how many species are shared between our
samples and how many we would expect to be shared.
1.0
"h ~
·~-t __~.'<o ____ ~ --i
1000 2000 3000

SIz.e or nth tamp'.
Figure 4. Variation in the value of Preston's coefficient of similarity between sample pairs (of the same size) from
the same master population with different sample size and diversity. [Data derived as in Figure 3]
Given a log-series population, this can be sampled such that x individuals are randomly
selected and the number and identity of species they represent can be recorded. Another
sample can then be taken and the species common to the two compared. Replication of
samples provides the means to evaluate the variance and confidence limits of the comparison.
Translated back into the field problem, we can therefore mimic the observed field situation
by building a gigantic population (mixed-species assemblage) within the computer, the
population of the same order of magnitude as that which might be present in, say, the
Belalong valley. That population is generated at the same diversity as observed in the field
samples. Pairs of samples of identical size to the field samples are then taken from the
computer population and the number of species shared between sample pairs is compared
with the number observed. The field samples are tested against the null hypothesis that they
come from a common population. In practice we generate a population of 106 individuals;
the diversity figure used is that of the pair of field samples combined, but there are other
alternatives. Possibil ities are the higher of the two diversities, the mean diversity, or the
lower diversity (Robinson & Tuck 1993), but there are sundry other possibilities. Ten pairs
of replicate samples are taken (with replacement) from the computer-generated log-series
population; the species common to pairs of samples are enumerated, and means, variances
and 2 x SD confidence limits calculated. The observed/expected ratio of shared species in
the field samples is taken as a measure of similarity, with 1.00 indicating common origin -
a match of observed to expected. Copies of the software are available from the authors.
The questions
Our sampling of Microlepidoptera in the forest adjacent to the Kuala Belalong Field
Studies Centre addressed five questions:
1. How species-diverse is the forest of the Belalong valley?
2. How does the scale of this diversity compare with that of other habitats in Borneo or with
temperate regions?
3. How many species are there likely to be in the area immediately adjacent to the Kuala
Belalong Field Studies Centre - a minimum estimate?
4. Is diversity uniform across the Belalong valley?
5. Is species composition similar across the valley or between the Belalong valley and other
sites in Borneo?
METHODS
Moths were attracted to a white sheet hung vertically and illuminated by a 125 watt
mercury-vapour lamp suspended one metre from the sheet surface. Lamps were run from
a portable generator.
Small specimens (wingspan < 40 mm) were collected individually into corked glass tubes,
the tubes kept cool overnight and the specimens killed, pinned and the wings spread (as
described by Robinson 1984) the following day. Large specimens (>40 mm) were killed
on site using conventional killing-bottles and pinned and stored the following morning.
Collecting began at dusk and usually ceased by about 2300 hrs when flight activity had
become minimal. The capture of all Microlepidoptera and Pyraloidea on the sheet was
attempted; on the few occasions when this proved logistically impossible a subsample was
taken and it was attempted to make this random.
Pinned specimens were stored with a few crystals of chlorocresol to prevent growth of
mOUld, dried, and returned to the laboratory. There they were labelled and sorted by
appropriate specialists to family and to species. Individuals, species and higher taxonomic
units (genera/subfamilies/families/superfamilies) were tabulated.
The localities studied

Kuala Belalong, Temburong, Brunei, altitude 125 m, sites 1 & 2 (KBl, KB2). These sites
were on the sides of the valley of the Sungai (River) Belalong above the Kuala Belalong Field
Studies Centre. Both were on steep slopes with closed-canopy primary lowland mixed
dipterocarp forest. Site 1 was on the western side of the valley and site 2 diagonally opposite
on the eastern slope about 1000 m from site 1. Collections were made from site 1 on 4, 6-8
and 10-11 July 1991 (6 nights) and from site 2 on 13-16 July (4 nights). Nightly samples
were kept separate and specimens from each labelled with the precise night of collection.
Comparisons were made with two other localities in northern Borneo:
Crocker Range, Sabah, altitude 1500 m (CR). This site was in closed-canopy montane
forest on a trail along a broad ridge crest at S.4SN, 116.19E, and appears to be in the
transitional zone between lower and upper montane forest (Anderson & Chai 1983).
Collections were made on the 6 nights from 7-12 August 1991. Nightly samples were
pooled.
Kampong Kapok, Muara, Brunei, altitude 0 m (MRI and MR2). This site was a
white-painted house wall facing into undisturbed mangrove (Rhizophora mucronata) forest.
Cleared land with sparse weed growth and a few ornamentals stretches 6-700 metres behind
the house to the edge of disturbed secondary forest. The mangrove forest extends for more
than 0.5 km before open water is reached. A more detailed description of this site is given
by Robinson & Tuck (in press). Collections were made from 19-21 July 1991 (3 nights) and
samples were pooled (MR1). Further samples were taken between 20 January and 10
February 1992 (ca. 14 nights' collections) and these were also pooled (MR2).
RESULTS
Numbers and diversity

The numbers of individuals and species and the diversity of each of the samples is
summarised in Tables 2, 3 and 4. A more detailed summary is given by Robinson & Tuck
(1993).
Values of alpha for the total sample at each of the two Kuala Belalong sites are very high
- 355 and 413 (Table 4). In practical terms, samples of 10,000 specimens with the same
diversities would contain 1197 and 1333 species respectively. A typical figure for
TABLE 2: Nwnbers of individuals of Microlepidoptera (sensu stricto - excluding Pyraloidea (MICRO), Pyraloidea
(pYRAL) and Microlepidoptera including Pyraloidea (ALL) collected at thee sites in Borneo.
SITE MICRO PYRAL ALL

KBI 372 377 749
KB2 166 315 481
KB1 + 2 538 692 1230
CR 386 127 513
MRI 128 57 185
MR2 323 362 685
MR1 + 2 451 419 870
TABLE 3: Nwnbers of species of Microlepidoptera (sensu stricto - excluding Pyraloidea) (MICRO), Pyraloidea
(pYRAL) and Microlepidoptera including Pyraloidea (ALL) collected at three sites in Borneo.

KBI 194 209 403
KB2 124 195 319
KBI + 2 270 301 571
CR 198 70 268
MRI 40 35 75
MR2 99 113 212
MRI + 2 118 135 253
TABLE 4: Diversity (Williams' a) of species of Microlepidoptera (sensu stricto - excluding Pyraloidea) (MICRO),
Pyraloidea (pYRAL) and Microlepidoptera including Pyraloidea (ALL) collected at three sites in Borneo

KBl 164±30 193±35 355±46
KB2 222±77 218±46 413±74
KBI + 2 216±32 203±25 414±39
CR 163±29 64±20 226±35
MRI 20±6 39±19 47±1l
MR2 49±9 56±9 105± 13
MRI + 2 52±8 69±11 120±13
Macrolepidoptera in Britain would be a diversity of 40, i.e. 221 species for 10,000
specimens.
The individual elements - Micro-Iepidoptera and Pyraloidea - of the two Kuala Belalong
samples have broadly comparable diversities at the two sites as do the total samples.
However, the diversity (a) of Microlepidoptera at KB1 is notably lower than at KB2. The
KB2 sample is small, however, and KB 1 diversity falls within the 95 % confidence limits of
KB2 though not vice versa.
The total sample from the Crocker Range has a lower diversity than that from Kuala
Belalong, and the sample confidence limits of the two do not overlap. This holds true if
Microlepidoptera and Pyraloidea are considered separately.
500
i?:
~ 250
2:
c
o~-
o 500 1000 1250
CumulaUve sample size
FIGURE 5: Trend in the index of diversity (Williams' a, with 95% confidence range) with the accumulation of
individuals (N) of Microlepidoptera and Pyraloidea over ten days at Kuala Belalong, Brunei. Note that on 12 July
(seventh point) the collecting site was changed from KBI to KB2.
The mangrove forest samples from Muara are clearly of much lower diversity than those
from the other sites and are little more diverse than samples from a mixed area of secondary
vegetation in a temperate country.
If we enumerate the daily samples from Kuala Belalong and measure the diversity of the
sample accumulated day by day it will be seen that there is a steady upward drift of the
diversity index (Figure 5).
Comparison of the sites

The species content of the samples from KB, CR and MR was radically different. Kuala
Belalong and Muara samples had more species in common than any other pair of sites, just
37 out of 787 (Table 5). However, expressed as a ratio of observed to expected similarity,
this figure is not much different from that between Kuala Belalong and Crocker Range
although much higher than the similarity between Muara and Crocker Range.
Comparison of the samples from the two Kuala Belalong sites, about one kilometre apart,
shows that the total samples share species to within 15 % of the figure expected, were the
samples from a common assemblage. Examined separately, however, the Pyraloidea match
the expected figure almost precisely whereas the Microlepidoptera do not (32 % divergence).
There is a remarkable disparity between the two Muara samples (MRI and MR2) which were
taken six months apart although the degree of difference between each of these and the
samples from KB and CR is similar.
Commonness, rarity and taxonomic composition of the samples

Neither rainforest sample (KB or CR) contained markedly abundant species. Only seven
species were represented in the KB sample by 10 or more individuals, the most abundant
having 27. In the CR sample only three species had 10 or more individuals, the most
abundant having 19. These "dominant" species comprise less than 3.75% of the total
samples.
The pooled MR samples contained four numerous species with 30 or more individuals,
the most abundant with 95, the four together comprising 28% of the sample; dominance in
these samples is much more apparent.
TABLE 5: Differences between samples of Microlepidoptera (sensu stricto - excluding Pyraloidea) (MICRO),
Pyraloidea (PYRAL) and Microlepidoptera including Pyraloidea (ALL) collected at three sites in Borneo in tenns
of tbeir observed and expected shared species (expected figure derived by simulation metbod - see text).
SHARED SPP:
SITES COMPARED OBS. EXP. OlE
KBI x KB2 MICRO 48 71±9 0.68
KBI x KB2 PYRAL 103 100± 13 1.03
KBI x KB2 ALL 151 175±20 0.86
MRI x MR2 MICRO 21 49±7 0.43
MRI x MR2 PYRAL 13 32±6 0.40
MRI xMR2 ALL 34 87±1O 0.39
KB1+2 x CR ALL 27 223±13 0.12
KBI +2 x MRI ALL 13 109±7 0.12
KBl+2 x MR2 ALL 28 266±22 0.11
KBl+2 x MR1+2 ALL 37 299± 17 0.12
MRI xCR 2 86+6 0.02
Further details, with notes on the taxonomic composition of the samples, are given
elsewhere (Robinson & Tuck 1993).
DISCUSSION
How diverse?
The vegetation of the Belalong valley is mixed dipterocarp forest growing on slopes from
the tame to the precipitous with a variety of aspects. This forest is a mosaic in which
different trees and herbs dominate or are clumped in particular small areas and in which
floristic diversity is remarkably high.
Most moth caterpillars feed on or in the leaves of plants and are to a greater or lesser
extent host-specific. Thus moth diversity in a habitat reflects surrounding botanical diversity,
and the high diversity of lowland dipterocarp forest in the Belalong valley is matched by high
moth diversity. The mosaic nature of the forest is also matched by the moths but tempered
by their ability to disperse by flight and thereby 'smooth out' the heterogeneity of the habitat.
We have taken a 'snapshot' of diversity of the smaller moths (Microlepidoptera and
Pyraloidea) at Kuala Belalong. The a-value recorded, 414, is of a similar order of
magnitude to that of about 300 for Macrolepidoptera found by Barlow & Woiwod (1989) in
a mature secondary forest habitat in W. Malaysia and by Holloway (1984) in lowland
dipterocarp forest in the Gunung Mulu National Park, Sarawak. Our diversity measure is
probably an underestimate - see Figure 5, discussion below, and Robinson & Tuck (1993).
The scale of diversity ?

This diversity is spectacularly high; mean a in the UK for Pyraloidea is 2.9; at Genting,
W. Malaysia, it is 91 and at Kuala Belalong it is 203. We do not have data for the diversity
of light-trapped Microlepidoptera in Britain but we suggest that the a value at anyone site
would be unlikely to exceed significantly the typical Macrolepidoptera value of 40 cited
above.
The sample of Microlepidoptera described here from montane forest in the Crocker Range
is less diverse than that from lowland forest. This matches the pattern found in
Macrolepidoptera in SE Asia by Holloway (1984, 1987) and, in Sulawesi, in Macro- and
Microlepidoptera by Holloway et al. (1990). These authors note that there is a peak of
diversity at about 1000 meters. Holloway's (1984) site 11 at 150 metres altitude in mixed
dipterocarp forest is broadly comparable to KB: his sample of Macrolepidoptera had an ex
of 309. His sites 1-5 in upper montane forest may be comparable to CR; ex for
Macrolepidoptera at these sites ranged from 73 to 118 whilst ex for CR Microlepidoptera is
163.
Spectacularly high diversity is not universal in Bornean habitats, however. Floristically
depauperate habitats such as mangrove forest, dry heath forest (kerangas), and Shorea albida
peat-swamp forest have much lower diversities than mixed dipterocarp or montane forest.
Species collected from the Muara site are from adjacent mangrove forest, cleared land with
regenerating weeds and planted ornamentals, and (distant) secondary forest. Here the
diversity (ex) of the larger sample just exceeds 100. Our subjective assessment of diversity
in dry kerangas and in peat-swamp forest, based on collecting in 1982 and 1988, is that these
are no more diverse than mangrove forest.
How many species in the lowland forest?

There is no such thing as an absolute number of species present in a particular habitat
except at a particular instant. Mobile organisms pass through, establish temporary
populations, become locally extinct. Collecting and/or inventorying invertebrates in a
lowland forest habitat such as the Belalong valley is an open-ended enterprise. However, the
number of 'new' species yielded by each successive unit of collecting effort (usually
measured as a number of individuals) becomes steadily less. The log-series formula can be
used to predict the number of species obtained with progressively larger samples. This
presupposes that the value of ex remains constant. In the early stages of a sampling
programme it is evident that it does not. Both at Kuala Belalong (Figure 5) and elsewhere
(Barlow & Woiwod 1989) ex rises with sampling effort. This phenomenon is also seen on
'bad' and 'good' collecting nights. A small sample invariably has a lower ex than a large one
(Robinson 1975). The effect is caused by heterogeneity of the surrounding habitat. On
single good nights with abundant insect activity flight-ranges are considerable and the insects
attracted to the light represent, metaphorically speaking, numerous tiles from the
environmental mosaic. On a bad (perhaps cold and windy) night flight-ranges are restricted
and the tiles sampled are fewer. Similarly, successive sampling will result in varied catches
due to climatic factors and the random nature of some insect movement:the heterogeneity of
the sampled habitat will increase steadily. Eventually some stability may develop, but
superimposed now will be effects of seasonality and floristic phenology as well as longer-
term environmental changes.
The values of ex obtained by 'snapshot sampling' are therefore underestimates so that any
prediction of species numbers from these will also be underestimates.
The basis for a 'total species estimate' can be anything that we want it to be. Here we
propose the number of species that could be collected in two years. Likely values of ex (from
the figures above) for Macrolepidoptera, Pyraloidea and Microlepidoptera respectively are
310, 170 and 210 for lowland forest. The combination of a Rothamsted trap and MV light
and sheet would allow a collector to collect on a nightly basis perhaps 100 Macrolepidoptera,
60 Pyraloidea and 60 Microlepidoptera. Extrapolating to a two-year catch and computing
the number of species of each group from the proposed diversity value we get:
Macrolepidoptera, 73000 individuals and 1694 species; Pyraloidea, 43800 individuals and
944 species; Microlepidoptera, 43800 individuals and 1122 species. The combined total is
160600 individuals and 3760 species. These represent no more than informed guesses of the
number of Lepidoptera species that may be present at a single site in Bornean rainforest. We
The diversity of small nwths in Borneo 39
stress that they are likely to be underestimates.

Janzen (1988) has published inventory figures for the Lepidoptera of the Santa Rosa
National Park, Costa Rica. This is a tropical dry forest and not necessarily comparable with
the forest type described here. From 10 years' observations he recorded 2797 moth and 345
butterfly species; of the moths 1130 are Microlepidoptera and Pyraloidea and 1667 are
Macrolepidoptera.
Elsewhere (Robinson & Tuck 1993) we have extended this method to derive a minimum
figure for Lepidoptera in Borneo (excluding butterflies) of more than 8600. Holloway (1987)
estimated that there are 4500 Macrolepidoptera species in Borneo; Robinson, Tuck & Shaffer
(in preparation) have seen more than 6500 species of Microlepidoptera plus Pyraloidea from
SE Asia (effectively Thailand + W. Malaysia + Borneo). These figures suggest to us that
our extrapolation is of at least the right order of magnitude.
The question remains to what degree our 'snapshot' figures are underestimates of the
'true' diversity. This phenomenon requires further investigation to determine after what
period of collection or trapping the diversity value stabilises if, indeed, it ever does. One
approach that has been adopted to circumvent this problem is to define 'faunal elements' -
associations of species characteristic of particular forest types - and to examine the diversities
of these independently (Holloway 1989, Robinson 1975). However, large samples from a
large number of sites are needed for this technique.
Uniform diversity?
The data presented above (Table 4) suggest that there is no significant difference in
diversity between the two sites on either side of the Belalong valley. However, it must be
stressed that the samples are small.
Similar species composition?

The extrapolation above assumes a uniformity in lowland Bornean forest types that may
be illusory. We are able to speculate about habitat homogeneity in lowland dipterocarp forest
only on the basis of the two KB sites, one kilometre (and a deep valley) apart. While the
Pyralidae at the two sites appeared to be drawn from the same assemblage, the
Microlepidoptera did not (Table 5). Many species in the latter group have limited dispersal
powers. The mosaic nature of tropical forest may mean that Microlepidoptera appear more
'patchy' than the more strongly dispersive Pyralidae; further collecting of Microlepidoptera
at the two sites should result in a convergence of the shared species with the number
expected if that is the case. Further testing of habitat homogeneity of, particularly, lowland
mixed dipterocarp forest is needed.
On a larger scale it is apparent that there are great faunistic differences between lowland
and montane forest in Borneo (Table 5); few species are common to the two habitats.
Elsewhere we describe differences in the proportion of families and superfamilies and
indicate groups that are almost entirely restricted to one or other forest type (Robinson &
Tuck 1993). The difference between the lowland forest samples and those from the edge of
mangrove forest are of comparable magnitude. here is almost complete faunal dissimilarity
between the Muara sample and that from the Crocker Range.
The two samples from Muara bear closer examination. Taken six months apart, the
diversity of the Microlepidoptera is significantly different in the two as is the diversity of
Microlepidoptera and Pyraloidea combined but not Pyraloidea alone. Sample comparison in
terms of shared species (Table 5) indicates a much greater degree of difference in the
temporally separated Muara samples than between the two spatially separated samples from
Kuala Belalong. Combining the two Muara samples produces an increase in the index of
diversity, suggesting that the two are drawn from putative parent assemblages that are
different. There are several explanations.
It is possible that there is faunal change with season that is affecting the samples. Barlow
& Woiwod (1989, 1990) have demonstrated temporal heterogeneity of samples of
Macrolepidoptera from W. Malaysia and Sulawesi.
The samples may show the effects of succession. MRI was taken shortly after completion
of a block of houses built on bare infill soil. By the time MR2 was taken there was
considerable growth of weeds, grass and ornamentals in the housing block which could add
a particular specialised fauna to the second sample.
The samples may differ because of collector-bias; the two samples were taken by different
entomologists.
The samples may differ because of their size. The mosaic nature of the habitat around
the Muara site is obvious: it is at the intersection of mangrove forest and developing gardens,
and mature secondary forest begins some 700 metres away. From samples taken elsewhere
and from the work of Murphy (1990), we know that the majority of the 14 species with 10
or more individuals in the combined (MRI + MR2) sample are mangrove specialists and
have been found only in association with this habitat. Current work has identified other
species that are also part of this specialised fauna. The remaining species in the Muara
samples are characteristic of disturbed and open habitats and of secondary forest. We have
given evidence above that heterogeneity within the sampled habitat contributes to an increase
in diversity. It may be that in the case of the Muara site this heterogeneity is so acute as to
affect faunal composition in samples of differing size to the extent that we have observed.
CONCLUSIONS
The rich lowland dipterocarp forest of the Batu Apoi Forest Reserve harbours a fauna of
small moths that is spectacularly diverse. This moth fauna is quite different in terms of its
species composition from that of montane forest and from a mixed secondary habitat with
mangrove. It is probably no more diverse than the fauna of any other pristine tract of
undulating lowland rainforest in northern Borneo, but such habitats are becoming increasingly
rare. Whether they are uniform in terms of their species composition beyond a particular
level of sampling remains to be investigated. However, first indications from Kuala Belalong
suggest that over a distance of one kilometre there is apparent uniformity in a strongly-
flighted group of moths but not in a poorly-dispersive group when comparatively small
samples are used.
Projections from 'snapshot' measures of diversity suggest that such a forest contains
upwards of 3750 Lepidoptera species excluding butterflies. Further extrapolations and a
limited examination of small-scale turnover suggest that we are not looking at astronomical
species numbers. However, the effect of turnover in increasing species diversity across a
greater distance in this habitat now requires urgent investigation.
Apparently spectacular turnover figures for moths have been quoted from unpublished data
by Erwin (1991) from two sites 500 km apart in the Amazon Basin. However, it is not made
clear whether they are from the same forest type, and the total sample size is not given.
Erwin recorded 3.2 % of species common to the two sites; here 3.3 % of species are common
to our Kuala Belalong (lowland) and Crocker Range (montane) samples from sites 180 km
apart.
Ultimately, figures such as these are intended to demonstrate that to inventory life on earth
by naming and describing the planet's species may appear to be a Herculean task but is by
no means impossible. As taxonomists we need to plan strategies to inventory biodiversity;
strategies cannot be devised unless the scale of the problem is estimated. If the moth fauna
of Borneo is 'only' four times the size of that of the United Kingdom there is no reason to
fear that we cannot provide at least a top-down approach to describing the common, the
The diversity of smaLL moths in Borneo 41
conspicuous and the biologically important elements of the biodiversity that we talk so much
about and seem to do so little to illuminate. Indeed, that is precisely the task that we as
individuals are now engaged upon (Robinson, Tuck & Shaffer, in preparation).
ACKNOWLEDGEMENTS
We thank the many individuals and organisations who made our fieldwork in Borneo
possible: work by Gaden Robinson at Kuala Belalong was undertaken as part of the Brunei
Rainforest Project 1991-92 organised by Universiti Brunei Darussalam and the Royal
Geographical Society. Kevin Tuck sampled in the Crocker Range with Operation Raleigh's
Expedition Perwira, with project approval by the Social and Economic Research Unit of the
Prime Minister's Department, Malaysia (SERU).
Fieldwork at Muara was made possible through the courtesy of Glenn and Maxine Ping;
we are indebted to them and to Eric Classey for taking sample MR2.
Financial support for this project was provided by the Royal Society South-east Asian
Research Committee, the Godman Fund and the Loke Wan Tho Memorial Foundation to
whom we are most grateful.
We are indebted to our colleagues: Jeremy Holloway and Klaus Sattler reviewed early
drafts of this manuscript; Pat Haynes provided painstaking technical support; Michael Shaffer
gave us extensive help in sorting and identifying Pyraloidea; Peter Barnard gave unstinting
assistance in overcoming serious software problems.
LITERATURE CITED
ANDERSON, J.A.R. & CHA!, P.P.K. 1983. Vegetation. pp. 195-223 in Jerrny, A.C. & Kavanagh, K.P. (eds).
Gunung Mulu National Park, Sarawak. Sarawak Musewn Journal Special Issue 2 (part 1) 30.
BARLOW, H.S. & WOIWOD, I.P. 1989. Moth diversity of a tropical forest in Peninsular Malaysia. Journal of
BARLOW, H.S. & WOIWOD, I.P. 1990. Seasonality and diversity of Macrolepidoptera in two lowland sites in the
Dwnoga-Bone National Park, Sulawesi Utara. Pp. 167-172 in Knight, W.J. & Holloway, J.D. (eds). Insects
and the rain forests of South East Asia (Wallacea). Royal Entomological Society, London. iv + 343 pp.
ERWIN, T. L. 1991. How many species are there?: revisited. Conservation Biology 5:330-333.
FISHER, R.A., CORBET, A.S. & WILLIAMS, C.B. 1943. The relation between the nwnber of species and the
nwnber of individuals in a random sample of an animal population. Journal of Animal Ecology 12:42-58.
HOLLOWAY, J.D. 1984. The larger moths of the Gunung Mulu National Park; a preliminary assessment of their
distribution, ecology, and potential as environmental indicators. Sarawak Museum Journal 30:149-190.
HOLLOWAY, LD. 1987. Macrolepidoptera diversity in the Indo-Australian tropics: geographic, biotopic and
taxonomic variations. Biological Journal of the Linnean Society 30:325-341.
HOLLOWAY, LD. 1989. Moths. In: Lieth, H. & Werger, M.LE. (eds). Tropical Rain Forest Ecosystems.
Ecosystems of the World. 14B. Elsevier, Amsterdam. xvii + 713 pp.
HOLLOWAY, J.D., ROBINSON, G.S. & TUCK, K.R. 1990. Zonation in the Lepidoptera of northern Sulawesi.
Pp. 153-166 in Knight, W.L & Holloway, LD. (eds). Insects and the rainforests of South East Asia (Wallacea).
Royal Entomological Society, London. iv + 343 pp.
JANZEN, D.H. 1988. Ecological characterization of a Costa Rican dry forest caterpillar fauna. Biotropica
20:120-135.
MAGURRAN, A.E. 1988. Ecological diversity and its measurement. Croom Helm, London. 179 pp.
MORISITA, M. 1959. Measuring of interspecific association and similarity between communities. Memoirs of the
Faculty of Science of Kyushu University, Series B, Biology 3:65-80.
MURPHY, D.H. 1990. The natural history of insect herbivory on mangrove trees in and near Singapore. Raffles
Bulletin of Zoology 38: 119-203.
PRESTON, F.W. 1962. The canonical distribution of commonness and rarity. Ecology 43:185-215, 410-432.
ROBINSON, G.S. 1975. Macrolepidoptera of Fiji and Rotuma: a taxonomic and biogeographic study. Faringdon.
vii + 363 pp.
42 GAD EN S. ROBINSON AND KEVIN R. TUCK
ROBINSON, G.S. 1984. Microlepidoptera in Brunei. The fourth VIu Temburong Expedition. Brunei Museum
Journal 5(4):146-177.
ROBINSON, G.S. & TUCK, K.R. 1993. Diversity and faunistics of small moths (Microlepidoptera) in Bornean
rainforest. Ecological Entorrwlogy 18:385-393.
TAYLOR, L.R. 1978. Bates, Williams, Hutchinson - a variety of diversities. In: Mound, L.A. & Waloff, N. (eds).
Diversity of Insect Faunas. Royal Entomological Society, Oxford. x + 204 pp.
WILLIAMS, C.B. 1944. Some applications of the logarithmic series and the index of diversity to ecological
problems. Journal of Ecology 32: 1-44.
WILLIAMS, C.B. 1947. The logarithmic series and its application to biological problems. Journal of Ecology
34:253-272.
WILLIAMS, C.B. 1964. Patterns in the balance of nature, and related problems of quantitative ecology. Academic
Press, London. 324 pp.
WOLDA, H. 1981. Similarity indices, sample size and diversity. Oecologia (Berlin.) 50:296-302.
WOLDA, H. 1983. Diversity, diversity indices and tropical cockroaches. Oecologia (Berlin) 58:290-298.
The herbaceous ground flora of the Batu Apoi Forest Reserve, Brunei
Darussalam
AXEL DALBERG POULSEN
Department of Systematic Botany, Institute of Biological Sciences, Aarhus University,

Nordlandsvej 68, DK-8240 Risskov, Denmark
ABSTRACT. The herbaceous ground flora was sampled in different habitats of the Batu Apoi Forest
Reserve, Brunei, and analysed in more detail in two one-hectare plots on ridge slopes at about 250 m
elevation in primary mixed dipterocarp forest. A total of 261 species were recorded, representing 26
families of angiosperms and 14 families of pteridophytes. Most of the species were proto-terrestrials
(93 %). The largest families of ground herbs in both the Forest Reserve generally and in the sampled
plots were the Zingiberaceae and Araceae, although the collective contribution of all fern families is
similar or greater. Grasses are not a significant component of ground herbs inside the forest; a major
difference compared with results from a similar study in the Amazon. The species richness of vascular
herbs was 92 in one plot (total of 6264 individuals with 191 m2 cover) and 68 in the other plot (total of
2479 individuals with 72 m2 cover). It is suggested that the rugged landscape and the dynamics of
landslides may be important factors for the high species richness of ground herbs in the area.
KEY WORDS: diversity, ground herbs, inventory, mixed dipterocarp forest, Borneo.
INTRODUCTION
As early as in 1907 Raunkiaer (1934) noted that a vast number of more or less herbaceous
plants of the phanerophytic type are found as ground cover, especially in forest regions with
an almost constantly humid tropical climate. Their stems are less lignified than those of the
trees and they are typically found under a closed tree canopy where the most favourable
conditions prevail. Most inventories of tropical rain forests have concentrated on large trees
and information of herbaceous plants is often not included. General descriptions of the
ground herb flora of tropical lowland rain forests in Borneo mention most common plant
families represented: Araceae, Begoniaceae, Cyperaceae, Gesneriaceae, Melastomataceae,
and Zingiberaceae (Richards 1952), but information is often limited to local floras or
checklists (Kiew 1978, 1987), and very few attempts have been made to measure quantitative
parameters.
The general notion has been that ground herbs contribute insignificantly to the total
diversity in the forest. However, Gentry & Dodson (1987) stressed the importance of non-
woody species in contributing to total species richness in South America. The present paper
includes a checklist of all ground herb species registered by the author in the scientific area
of the Batu Apoi Forest Reserve. Other results were obtained from inventories of the ground
herb vegetation in selected plots.
The ground herb life form is subordinate to trees and shrubs. The conditions on the
ground in a closed-canopy forest are characterized by low light intensities and high humidity.
Vegetative spread is important in ground herb reproduction (Burtt 1977) resulting in
regenerative capabilities and mobility not found in trees. In some cases ground herbs depend
on gaps to promote seed production (Smith 1987). Not surprisingly, ground herbs are often
found to have a patchy distribution and large areas of the forest floor are often devoid of
herbaceous cover (Poulsen & Balslev 1991).
43
44 AXEL DALBERG POULSEN
THE STUDY AREA
The study was conducted in the Batu Apoi Forest Reserve which covers 45,000 hectares of
Temburong District, Brunei Darussalam (115°10' E; 4°32' N). The topography is generally
very hilly with steep slopes. The elevation ranges from 65 m at Kuala Belalong up to 913
m (2995 ft) at the highest peak in the area, Bukit (Bt.) Belalong. The climate is generally
hot and humid. The total precipitation the first year of measurement was 4135 mm with less
than 200 mm per month in January, February, July, and August. Periods with no rain for
ten days occurred in these drier months. All soils are residuals from erosion of the shale
rock of the Miocene Setap shale formation (Ashton 1964). The slopes are dominated by
mixed dipterocarp forest. Human impact on the vegetation is low; Iban tribes have used the
area for pig hunting exclusively. Bryophytes are rare on the forest floor in the lowland
mixed dipterocarp forest as opposed to heath forest (kerangas).
One sample plot was established on the ridge west of the Kuala Belalong Field Studies
Centre by scientists from the University of Aarhus and will be referred to as Plot A.
Another plot, the second plot established by the Universiti Brunei Darussalam, on the ridge
between Kuala Belalong and Bt. Belalong, will be referred to as Plot B. Each plot was of
one hectare projected in primary mixed dipterocarp forest at approximately 250 m elevation
on either side of the river Sungai Belalong (Figure 1). A flood stream was present in each
plot but only after heavy rain. No major recent gaps existed in either of the plots. Plot A
was much steeper than B - the differences between the highest and lowest points were about
60 and 30 m respectively (Figure 2). Because of its steepness, Plot A had 17.3% more
surface area than a perfectly flat plot and Plot B had 6.4 % additional surface area.
MATERIAL AND METHODS
Fieldwork was undertaken from April to June 1991 and from January to April 1992, a total
of six months. Ground herb species were collected throughout the research area of the
Reserve and notes were made of habitat and elevation. The ground herb vegetation was
SOUTH CHINA SEA
BORNEO
I. _ ._~
,
-50km-
I I I I I ,
'.......... ,
BRUNEI DARUSSALAM
Figure 1. Map showing the position of the two plots in the Batu Apoi Forest Reserve, Brunei, Borneo.
Herbs of the Batu Apoi Forest Reserve 45
Figure 2. Topography, distribution of tree falls, and distributions of all herbaceous individuals on the ground in
Plots A and B. The contours of Plot A are shown with 5 m intervals; Plot B with 2 m intervals. The position of
tloodstreams are shown on contour maps . The asterix indicates the position of the reference point (0;0).
studied in detail in the two quadratic one-hectare plots, A and B. Plot A was, at the same
time, used for an inventory of trees (Poulsen et at., this volume). The plots were subdivided
into 400 5 x 5 m subplots with nyfon strings . In each subplot the following data were
recorded for all ground herb species: number of individuals, coordinates, and cover. Cover
was estimated as the vertical projection of shoot area on a horizontal plane. Voucher
specimens were coIlected for all species in the plots. The coIlections are deposited at the
herbarium of Brunei (BRUN) , with additional sets at Kew Herbarium (K), and Aarhus
University (AAU).
Much influenced by botanical traditions at higher latitudes, a herb has been defined as a
plant having stems which are not secondarily thickened and lignified and which die down
annually (Lincoln et al. 1982). Herbaceous plants in tropical lowland rain forests are
perennials and do not die back every year during the unfavourable season. Therefore tropical
herbs may develop secondary lignified tissues in the basal part even if they are herbaceously
derived. In the Batu Apoi Reserve this phenomenon is common in the Acanthaceae,
Gesneriaceae, and Rubiaceae. The use of terms such as shrub let, undershrub, subshrub,
suffrutex, and woody herb, noted from herbarium labels, indicates the difficulties that
collectors experience when trying to categorize members of this transition group. It is not
possible to come up with a clear cut definition of herbaceous plants, even after studies of
anatomical cross sections of borderline species. Kiew (1988) stated that herbs may still be
recognized as such because they have soft leaves, in contrast to shrubs and trees with
typically leathery leaves.
In this study only a relatively small proportion of the species belongs to the transition
group. The checklist includes all the Gesneriaceae although some are quite woody at the
base (Cynandra, Didymocarpus). Thottea penitilobata Ding Hou (Aristolochiaceae) is an
erect shrub according to Flora Malesiana (Ding 1984) and is excluded. Anatomical sections
of Neckia serrata Korth (Ochnaceae) and Myrioneuron cyaneum Hallier (Rubiaceae) reveal
well developed lignification (Figure 3). Both were excluded in this study despite the fact that
the former occurs in a checklist of herbs from Peninsular Malaysia (Kiew 1987). Bamboos
and tree ferns were not also included.
Life form stratification definitions in this paper follow Poulsen & Balslev (1991). Ground
herbs are species that in certain developmental stages have stem parts at ground level rooted
in the ground. Obligate terrestrial species are found on ground level only and include proto-
terrestrials, saprophytes, and scandents. Facultative terrestrial species include climbers and
epiphytes which do not complete their life cycle on the forest floor.
An important component of the ground layer not included in this study is that of woody
plants like palms, shrubs, and the seedlings and saplings of trees.
Figure 3. Anatorrucalcross·sections of A: Neckia serrata (Ochnaceae) and B: Myrioneuron cyaneum (Rubiaceae),

at the apical (a) and basal part (b) of the stem showing secondary xylem 00. The material of Ba is compressed
because it derives from a herbarium specimen.
RESULTS AND DISCUSSION
Species richness of the Reserve

Appendix 1 lists 261 species of ground herbs encountered by the author in the Batu Apoi
Forest Reserve. It should be noted that this is not a complete account of the herbaceous
ground flora. Other important collections of the ground herb life form in the area are those
made by Argent & Mitchell, Boyce and Wong. The checklist nevertheless gives a good
indication of the landscape-level (g) diversity in the Reserve. Kiew (1978) listed 105 species
of the ground flora from the nearby Gunung Mulu National Park, including some woody
species, but stressed that this figure probably is an underestimate.
The most species-rich family in the Reserve was the Zingiberaceae with 46 species,
followed by Araceae (35), Gesneriaceae (16), Orchidaceae (16), and Cyperaceae (10); 21
families were found with less than 10 species (Figure 4). The group comprising the 14
families of pteridophytes (ferns and fern allies) is, however, more species-rich (72 species)
than any of the angiosperm families.
The genera most rich in species were Cyrtandra (Gesneriaceae) and SelagineUa
(Selaginellaceae) each with 12 species, followed by Diplazium (10, Dryopteridaceae),
Tectaria (10, Dryopteridaceae), Amomum (9, Zingiberaceae) and Etlingera (8,
Zingiberaceae). Forty eight genera were only represented by one species.
Life form strategies

By far the most important life form strategy of ground herbs is the proto-terrestrial one
(93%). The only scandent species found was the fern Lygodium circinnatum. Ten species
of saprophytes were found in the area, with the genera Epirixanthes and Sciaphila being most
PTERIOOPHYTES
Zingiberaceae
Araceae
Gesneriaceae
Orchidaceae
Cyperaceae
Rubiaceae
Begoniaceae
Acanthaceae
Maranthaceae
Melastomataceae
Costaceae
Urticaceae
Musaceae
Polygalaceae
Poaceae
Commelinaceae
o Climbers and fallen epiphytes
Hanguanaceae El Scan dent
Hypoxidaceae
Triuridaceae til Saprophytes
Campanulaceae
Liliaceae • Proto-terrestrials
Myrsinaceae
Taccaceae
Petrosaviaceae
Burmanniaceae
Piperaceae
20 30 40 50 60 70 80
no. of species
Figure 4. Number of species per family represented in the ground herb flora of the Batu Apoi Forest Reserve. The
14 families of pteridophytes are shown as one group.
common. Only 19 species of climbers and three epiphytes were rooted on the ground. Two
of the most abundant species of the 13 aroid climbers were Amydrium medium and
Epipremnum falcifolium. The genus Rhaphidophora was also found to have many juveniles
contributing to the ground flora, and it is likely that more aroid species germinate on the
forest floor than have been observed so far (e.g., Pothos). Even if climbers were omitted,
Araceae would still be the second largest family in the area.
Habitats
In the nearby Gunung Mulu National Park, Kiew (1978) noted the association of ground
layer species with different habitat types which she roughly divided into those associated with
open canopy and others associated with closed canopy. Such habitat preferences were also
observed in the Batu Apoi Forest Reserve.
About a quarter of the species were restricted to moist habitats near the rivers. Most
strictly confined are the periodically flooded rheophytic herbs on shale rocks (e.g.,
Bucephalandra motleyana, Piptospatha burbidgei, Tectaria lobU var. lobi!). On the river
banks Etlingera punicea and E. fimbribracteata frequently occurred in dense stands in alluvial
deposits. Phacelophrynium maximum was also abundant and attained a large size in the light
open habitats near the river, but was also found on slopes in closed canopy forest. However,
in the latter habitat no flowering was observed. This could be an example of the so called
mass-effect (Shmida & Wilson 1985).
In areas with large gaps along the river created by landslides or major tree falls the
herbaceous ground vegetation was dominated by giant herbs like bananas (Musa) , and
Alocasia spp. Costus speciosus was also typical of this habitat.
Alocasia lowii, A. guttata, Cyrtandra pendulijlora, and many ferns grew on the moister
slopes near the river but not in open gaps. Along small streams species of Elatostema and
Tectaria were particularly common.
Within the closed canopy forest there was a difference between the ridge tops and the
slopes. The ridge top community of ground herbs was less species-rich and characterized
by Didymocarpus lanceolatus and Diplazium crenatoserratum. The proto-terrestrial orchid,
Apostasia elliptica, of which only four individuals were found, was only observed on dry
ridge tops. The ginger Alpinia glabra was also very abundant on dry ridge top areas - a fact
that is in confl ict with Kiew (1978), who states that this species dom inates in areas of
frequent waterlogging.
The saprophytic species were generally found in dark and moist habitats (Sciaphila spp.
and Epirixanthes spp.).
On the steep slopes with exposed mineral soils, ground herbs like Mapania monostachya,
Ptyssiglottis psychotriifolia, Trichomanes singaporeanum, Begonia cyanescens ined. were
very abundant. On litter-covered slopes the gingers Elettariopsis kerbyi and Boesenbergia
burttiana and the sedge Mapania palustris were more common.
The composition of ground herbs also changes with elevation. Near Bukit Belalong five
species of Sonerila were found in habitats that were also very rich in Selaginella and ferns.
These species of Sonerila, and 45 other species, were never observed at elevations lower than
600 m. Sixty-six species were never found above 100 m elevation. Other species had a
distribution ranging from near the Sungai Belalong all the way to the top of Bukit Belalong,
but never occurring on the dry ridge at intermediate elevations (e.g., Etlingera brevilabris,
Plagiostachys strobilifera).
Vegetation studies, species richness, density and cover

In each plot seven pteridophyte and 13 angiosperm families were present: Plot A had one
pteridophyte family and three angiosperm families which were not found in Plot B, and vice
versa. In both plots the Zingiberaceae were represented by the highest number of species
(A:19; B:19), followed by Araceae (A: 17; B:lO). The remaining families were represented
by less than 10 species. The pteridophyte families combined have more species than any
angiosperm family in Plot A, and rank second in Plot B. This pattern of family contribution
to species diversity based on small scale studies on habitat level (ex-diversity) is similar to the
results of the checklist (Figure 4) collected in a large area pooling various community types
of the landscape (g-diversity).
Plot A had 92 species and Plot B 68 species of ground herbs. A total of 117 species were
registered in the two plots and 38 % were shared. Even when the larger surface area of the
steep Plot A is compensated for, this plot has markedly more species. More microhabitats
undoubtedly exist on the steep slope, thus enabling more species to coexist. Certainly a
longer gradient from local ridge top and down slopes is represented in Plot A.
The number of tree species with dbh ::?: 10 cm in Plot A was 231 (Poulsen et at. this
volume) and additional woody species not recorded (with dbh less than 10 cm) are present.
Even though the total species richness in the plot is unknown, the species contribution of
ground herbs is still significant.
Most of the individuals belong to a few species. In both plots 50% of the individuals are
accounted for by 5-6 % of the most common species of ground herbs. The pattern that a few
species are common and many species rare is more pronounced for the herbs than for trees
in the same plot (Figure 5).
Plot A had 6264 individuals of ground herbs whereas Plot B had 2479 individuals (Figure
2). It is noteworthy that ground herbs have a much more clustered distribution pattern than,
for example, trees. Particularly in Plot A there is an evident density gradient correlated with
topography and underlying edaphic factors. Plot B has a much lower overall density of
ground herbs probably because the drier rich top constitutes a bigger part in this plot.
Despite the high precipitation, a dry period of two weeks had a pronounced effect during
which many ground herbs were observed to wilt.
.
.
x
X
;/
x
.'
.'
.'
~o x
. x
60 x
x
x
·x •
40
Trees, Plot A
Herbs, Plot A
20 .' x Herbs, Plot B
~
o 20 40 60 80 100
% species
Figure 5. Species-abundance curve illustrating how a few of the species with highest density constitute the majority
of the individuals.
Five species are found on both the lists of the 10 species with highest density for each
plot: Mapania monostachya, Mesophlebion falcatilobum, Elettariopsis kerbyi, Tn'chomanes
singaporeanum, and Boesenbergia burttiana. The density of these five species was in all
cases more than 100 individuals per plot.
The total cover of ground herbs was 191.2 m2 and 72.4 m2 in Plot A and B respectively.
The explanation for the markedly lower cover in Plot B may be the same as that for the
density difference. In both plots, SO % of the cover is accounted for by 2-4 % of the most
common species of ground herbs.
Comparing the 10 species with highest cover in each plot, four species are the same:
Mapania monostachya, Mesophlebionfalcatilobum, Elettariopsis kerbyi, Mapania palustris.
These species all had a cover greater than 2 m 2 • In Plot A, Stachyphrynium cf. jagorianum
had the greatest cover of almost 79 m2 • This species was found in a large, dense cluster, a
feature probably explained by an historical event affecting the vegetation. The cluster was
probably formed in a gap-phase situation, where vegetative spread as well as seed production
for this species was optimal (Smith 1987). Established plants are then able to survive for
some time after the canopy has closed. This is an example of a gap-phase species as opposed
to the shade-loving species (Richards 1952), but the species may already have been present
on the site before the gap situation.
Comparison with a Neotropical study

In a similar inventory of ground herbs in a one-hectare plot in lowland rain forest in the
Amazon (Poulsen & Balslev 1991), the total number of species was 96, comprising 10,960
individuals and 2.5 % cover. The Amazonian plot was not as steep as any of the Bornean
ones; the additional surface area was only 2 % more than a perfectly flat plot. The species
richness of the Amazonian plot is, however, higher due to the larger contribution of
facultative species, which have a high diversity in this region (Gentry 1982). If only proto-
terrestrials are compared, the Amazonean plot has about the same number as the flattest
Bornean Plot B (50/54). Despite the great variation in individual numbers associated with
the plots, species richness of the Bornean and Amazonian plots is of the same magnitude,
reflecting the similarities in conditions for herbs on the ground of tropical lowland rain
forests in general.
The major floristic difference between the two studies is that grasses are a very prominent
component of the Amazonian study. In contrast, in all inventories, only one individual of
a grass (Scrotochloa urceolata) was found in the Bornean study. In both investigations the
Araceae and pteridophytes contribute massively to the ground herb species richness, whereas
Zingiberaceae plays a much greater role in the Bornean study. In fact, the Zingiberaceae
seems to have a stable representation in the vegetation studies in terms of the number of
species. The two most abundant ginger species in each Belalong plot were the same and
their density and cover in each plot were also approximately equal.
CONCLUSIONS
The species diversity of ground herbs in the Batu Apoi Forest Reserve is relatively high in
spite of the fact that the substrate is all clayish derivatives of Setap shale. Several habitats
with a characteristic ground herb flora may be observed in the area. The 10 richest ground
herb families were the Zingiberaceae, Araceae, Gesneriaceae, Orchidaceae, Cyperaceae,
Rubiaceae, Begoniaceae, Acanthaceae, Marantaceae and Melastomataceae. These families
were also among the highest ranking in respect of density and cover in vegetation studies in
closed canopy forest. Pteridophytes, taken as a single group, contribute more to ground herb
diversity than any of the angiosperm families. One of the reasons for the high number of
species in the Batu Apoi Forest Reserve may very well be the diversity of habitats provided
by the rugged landscape and the dynamics of landslides.
ACKNOWLEDGEMENTS
The author is indebted to P. Boyce, B. L. Burtt, 1. Dransfield, D. S. Edwards, P.

Goetghebeur, B. Hansen, K. U. Kramer, K. van der Meijden, M. Sands, G. Seidenfaden,
D. Simpson R. M. Smith, E. de Vogel, B. 0llgaard for help with identification, and to the
Royal Geographical Society and the Universiti Brunei Darussalam for permission to
participate in the Brunei Rain Forest Project 1991-92. The study was conducted during a
scholarship from the Faculty of Science, the University of Aarhus, and financed by the
Danish Natural Science Research Council (Grant no. 11-9018).
LITERATURE CITED
ASHTON, P. S. 1964. Ecological studies in the mixed Dipterocarp forests of Brunei State. Oxford Forestry Memoirs
25:1-75.
BRUMMITT, R. K & POWELL, C.E. (eds) 1992. Authors of plant names. Royal Botanic Gardens, Kew.
BURTT, B. L. 1977. Notes on rain-forest herbs. Gardens' Bulletin Singapore 29: 73-80.
DING, HOU. 1984. Aristolochiaceae. Pp. 53-108 in Flora Malesiana, Ser. I, Vol. 10.
GENTRY, A. 1982. Neotropical floristic diversity: phytogeographical connections between Central and South
America, pleistocene climatic fluctuations, or an accident of the Andean orogeny? Annals of the Missouri
Botanical Garden 69:557-593.
GENTRY, A. & DODSON, c. 1987. Contribution of non-trees to the species richness of a tropical rain forest.
KIEW, R. 1978. Floristic components of the ground flora ofa tropical lowland rain forest at Gunung Mulu National
Park, Sarawak. Pertanika 1: 112-119.
KIEW, R. 1987. The herbaceous flora of the Ulu Endau, Johore-Pahang, Malaysia, including taxonomic notes and
descriptions of new species. Malayan Nature fournaI41:201-234.
KIEW, R. 1988. Herbaceous flowering plants. Pp. 56-76 in Cranbrook, Earl of(ed.). Malaysia, key environments.
Pergamon Press, Oxford.
LINCOLN, R. 1., BOXSHALL, G. A. & CLARK, P. F. 1982. A dictionary of ecology, evolution and systematics.
Cambridge University Press, Cambridge.
POULSEN, A. D. & BALSLEV, H. 1991. Abundance and cover of ground herbs in an Amazonian rain forest.
foumal of Vegetation Science 2:315-322.
RICHARDS, P. W. 1952. The tropical rainforest. Cambridge university Press, England.
RAUNKIAER, C. 1934. The life-forms ofplants and statistical plant geography. Oxford University Press, London.
SHMIDA, A. & WILSON, M. V. 1985. Biological determinants of species diversity in temperate and tropical
forests. American Naturalist 126:760-776.
SMITH, A. P. 1987. Respustas de hierbas del so to bosque tropical a claros ocasiondos por la caida de arboles.
Revista de Biologia Tropical 35 (Supl.l):111-118.
Appendix. Checklist of ground herb species collected or registered by the author in the scientific area of Batu Apoi
Forest Reserve during the Universiti Brunei Darussalam and Royal Geographical Society's Rain Forest Project
1991-92, with notes on occurrence and habitat. Abbreviations: KBFSC = Kuala Belalong Field Studies Centre;
K. = Kuala; Bt. = Bukit; S. = Sungai. Plot A was established by Aarhus University (AAU); plot B was the second
plot established by Universiti Brunei Darussalam (UBD); C plots were establiShed by Colin Pendry with specified
occurrence as low (c 200 m elev.), middle (c 500 m elev.) or high (c 800 m elev.) in brackets. Voucher number
is given in round brackets. Specimens are deposited at BRUN, with additional sets at K and AAU. Author names
according to Brummitt & Powell 1992.
Acantbaceae
Cosmianlhemum oblusijolium Bremek. Proto-terrestrial; S. Belalong, on bank above Hoodlevel; 70 m; (Poulsen 117).
linarianlha bicolor B.L.Burtt & R.M.Sm. Proto-terrestrial; small creeper along the ground, on trail along S. Belalong; 65 m; (Poulsen
256).
Ptyssiglollis psycholriijolia (Stapt) B.Hansen. Proto-terrestrial, rather woody at base; common on slopes particularly on exposed mineral
soils; plot A, C (low); 150-250 m; (Poulsen 40, 157, 158).
Ptyssiglol1is staminodifera B. Hansen. Proto-terrestrial, rather woody at base; S. Belaiong, near river in dark, moist habitat; 65 m;
(Poulsen 67).
Staurogyne jaherii Bremek. Proto-terrestrial; S. Belalong, near river; plot A; 60 m; (Poulsen 28).
Staurogyne spp. Proto-terrestrial; BI. Belalong, steep rock near Pondok Belalong; 850 m; (Poulsen 311).
Araceae
Aglaonemo simplex Blume Proto-terrestrial; ridge W. of KBFSC; plot A; 250 m; (Poulsen 212).
Alocasia denudala Eng!. Proto-terrestrial; ridge W. ofKBFSC; plot A, B, C (low, high); 150-850 m; (Poulsen 79).
Alocasia guttala N.E.Br.Proto-terrestrial; S. Baki, S. Belalong, and S. Temburong, on shale bank near river; 65-70 m; (Poulsen 108,
235).
Alocasia lowii Hook. f. Proto-terrestrial; S. Belalong and S. Enkiang, near river; 60 m; (Poulsen 18, 264).
Alocasia peltala M.Hotta Proto-terrestrial; ridge W. of KBFSC; plot A; 250 m; (Poulsen 213).
Alocasia sp. Proto-terrestrial, giant herb; in disturbed areas along S. Belalong; 65 m; (Poulsen 43).
Alocasia sp. Proto-terrestrial; ridge slope W. ofKBFSC, gully between A and S. Belalong; 150 m; (Poulsen 155).
Alocasia sp. Proto-terrestrial; S. Esu, on vertical rock near creek; 80 m; (Poulsen 296).
Amydrium medium (Zoll. & Mor.) Nicolson Climber, juveniles germinate on the forest floor, flower about 2 m up a tree trunk;
common; plot A, B, C (low, midlle, high); 150-850 m; (Poulsen 173, 236, 364).
Bucephalandra molleyana Schott. Rheophyte; on shale rocks in river; 70 m; (Poulsen 109).
Epipremnumfalcijolium Eng\. Climber, germinating on the forest floor; common; plot A, C (low, high); 65-850 m; (Poulsen 172, 233,
363).
Homolomena cordala (Houtt.) Schott. Proto-terrestrial; common; plot A, C (low, high); 250-850 m; (Poulsen 25, 234).
Hornalomena geniculala M.Hotta Proto-terrestrial; BI. Belalong; C (high); 850 m; (Poulsen 322).
Hornalomena humilis Hook.f. Proto-terrestrial; S. Belalong, on rocks near river; 65 m; (PouL"n 190).
Hom%mena ovata Eng\. Proto-terrestrial; steep soil slopes near river and in forest; plot A, C (low); 70-250 m; (Poulsen 168, 268).
Homolomena vagans P.C.Boyce ined. Rheophyte; S. Belalong, on steep river banks close to K. Belalong; 65 m; (Poulsen 47, 273).
Homolomena sp. Proto-terrestrial; ridge slope W. ofKBFSC; 150 m; (Poulsen 181).
PiplOspatha burbidgei (N.E.Br.) M.Hotta Rheophyte; common on shale rocks in river; 60 m; (Poulsen 19).
Rhaphidophora korthalsii Schott. Climber, juvenile on ground; ridge W. of KBFSC; plot A. B; 250 m; (Poulsen 211).
Rhaphidaphora minor Hook.f. Climber; ridge slope W. of KBFSC; plot A, B; 200 m; (Poulsen 153).
Rhaphidpphora sp. Climber; ridge slope W. ofKBFSC; plot B; 180 m; (Poulsen 152).
Rhaphidophora sp. Climber, fallen or juvenile on ground; ridge slope W. ofKBFSC; plot A, B, C (low); 150-250 m; (Poulsen 165).
Rhaphidophora sp. Climber, juvenile on ground; ridge W. ofKBPSC; plot A, B, C (low, high); 250-850 m; (Poulsen 207).
Rhaphidophora sp. Climber, juvenile on ground; ridge W. ofKBFSC; plot A, C (high); 250-850 m; (Poulsen 232).
Rhaphidophora sp. Climber, juvenile on ground; on ridge between K. Belalong and BI. Belalong; plot B; 250 m; (Poulsen 362).
Rhaphidophora sp. Climber, juvenile on ground; BI. Belalong; plot C (high); 850 m; (Poulsen 366).
Rhaphidaphora sp. Climber, juvenile on ground; plot C(low); 850 m; (Poulsen 367).
Schismologlollis beccarianus Eng\. Proto-terrestrial; BI. Belalong; plot C (high); 850 m; (Poulsen 365).
Schismatogloltis brevicuspis Hook.f. Proto-terrestrial; S. Baki, on steep rock slope near stream; 70 m; (Poulsen 139).
SchisrnalOglottis calyptrala Zoll. & Mor. Proto-terrestrial; S. Belalong, on bank of small stream E. of S. Belalong; plot A; 65-250 m;
(Poulsen 192, 238, 285).
Schismologlottis modesta Schott Proto-terrestrial; ridge slope W. ofKBFSC; plot A; 150-250 m; (Poulsen 177, 237).
SchismolOglollis sp. aff. monoplacenla M.Hotta Proto-terrestrial; ridge W. of KBFSC; plot A; 250 m; (Poulsen 86, 239).
Schismologlollis sp. Proto-terrestrial; S. Belalong, near KBFSC, bank above floodlevel; 70 m; (Poulsen 143).
Scindapsus longipes Eng!. Climber, fallen or juvenile on ground; ridge between K. Belalong and BI. Belalong; plot A, B, C (low); 150
m; (Poulsen 313).
Scindapsus pictus Hassk.Climber, juvenile on ground; ridge W. ofKBFSC; plot A, B, C (Iow,high); 250-850 m; (Poulsen 206).
Begoniaceae
Begonia baramensis Merr. Proto-terrestrial; DI. Belalong; plot C (high); 850 m; (Poulsen 316). I ~ Martin Sands sp. # IJ
Begonia (cyanescens ined.) Proto-terrestrial; steep slope behind KBFSC; plot A, B, C (low, high); 70-850 m; (Poulsen 6, 30). I~
Martin Sands #13J
Begonia sp. Proto-terrestrial; ridge W. of KBFSC; plot A; 250 m; (Poulsen 31). I ~ Martin Sands sp. # 9]
Begonia sp. Proto-terrestrial; S. Belalong, near waterfall; 65 m; (Poulsen 45). I ~ Martin Sands sp # II]
Begonia sp. Proto-terrestrial; S. Temburong, on steep rocky bank near river; 70 m; (Poulsen 58). I ~ Martin Sands sp. # 10J
Begonia sp. Proto-terrestrial; S. Temburong, riverbank, in wet soil; 70 m; (Poulsen 105). I ~ Martin Sands sp. #18]
Begonia sp. Proto-terrestrial; trail along S. Belalong, above floodlevel, damp area; 65 m; (Poulsen 258). I ~ Martin Sands sp. # 4]
Begoniaceae (continued)
Begonia sp. Proto-terrestrial; Bt. Belalong; plot C (high); 850 m; (Poulsen 317). [ = Martin Sands sp. # 17J
Burmanniaceae
BUl7llJl1U!ia coelestis Don Saprophyte; ridge between Pondok Busiri and Bt. Belalong; 700 m; (Poulsen 101).
CampanoIaceae
Pentaphragma prostratum Kiew Proto-terrestrial; trail along S. Belalong, above f1oodlevel. damp area; plot C (low, high); 65-800 m;
(Poulsen 254).
Commelinaceae
Amisehotolype mollissima (BI.) Hassk. Proto-terrestrial; river bank and gully between A and S. Belalong; 65-150 m; (Poulsen 156,
186).
Pollia pumila Hallier Proto-terrestrial; S. Belalong, river bank; 65 m; (Poulsen 187).
Costaceae
Costus globosus Blume s. lat. sp. I Proto-terrestrial; ridge W. of KBFSC; plot C (low); 150m; (Poulsen 167).
Costus globosus Blume s. lat. sp. 2 Proto-terrestrial; ridge slope W. of KBFSC; 150 m; (Poulsen 175).
Costus paradoxus K.Schum. Proto-terrestrial; S. Temburong and S. Belaiong, river bank; 65 m; (no voucher).
Cosrus speciosus (Koenig) Sm. Proto-terrestrial; S. Temburong, near landslide in open and disturbed, rocky habitat; 75 m; (Poulsen
54).
Cyperaceae
Cyperus diffusus Vahl Proto-terrestrial; S. Temburong, in landslide area near river; 70 m; (Poulsen 66).
Mapania euspidflta (Miq.) Uittien Proto-terrestrial; S. Temburong, S. Belalong, above river, on bank of small flood stream near B;
75 m; (Poulsen 135, 300, 369).
Mapania graminea Uittien Proto-terrestrial; ridge W. of KBFSC; plot A, C (low. middle); 250 m; (Poulsen 41).
Mapania longiflora C.B.Clarke Proto-terrestrial; ridge between K. Belalong and Bt. Belalong; plot B; 250 m; (Poulsen 318).
Mapania monostaehya Uittien Proto-terrestrial; common on ridge slopes particularly on steep, exposed soil; plot A. B, C (low, middle,
high); 220-900 m; (Poulsen 32, 39, 284).
Mapaniapaluslris (Hassk. ex Steud.) Fern.-Vill. Proto-terrestrial; common on ridge slopes; plot A, B, C (low, middle, high); 180-850
m; (Poulsen 201).
Mapania riehardsii Uittien Proto-terrestrial; ridge W. of KBFSC; 200 m; (Poulsen 36).
Paramapania parvibraeteata (C.B.Clarke) Uittien Proto-terrestrial; ridge W. ofKBFSC; plot A, B, C (low, high); 250-850 m; (Poulsen
53).
Seleria motleyi C.B.Clarke Proto-terrestrial; Bt. Belalong, in clearing on the summit; plot C (high); 850-900 m; (Poulsen 97).
Scleria purpurascens Steud. Proto-terrestrial; S. Beialong, near river; 60 m; (Poulsen 16).
Gesneriaceae
Cyrtandra basiflora C.B.Clarke Proto-terrestrial, rather woody at base, about I m; on banks of flood stream; plot A, C (low, high);
75-750 m; (Poulsen 288).
Cyrtandra digitaliJlora B.L.Burtt Proto-terrestrial; Bt. Belalong; plot C (high); 850 m; (Poulsen 381).
Cyrtandra pendulijlora Kraenzl. Proto-terrestrial; on steep slope behind KBFSC; 75 m; (Poulsen 49).
Cynandra phoenicoiasia Lauterb. Proto-terrestrial; S. Beiaiong, in rocky area near little stream; plot A, D, C (low) 70 Ill; (Poulsen
188).
Cyrtandra sarawakensis C.B.Cl. Proto-terrestrial, ascending woody stem with fleshy outer layer; S. Esu, wet bank above river; 70
m; (Poulsen 297).
Cyrtandra sp. aff. lambirense Proto-terrestrial; steep slopes on ridge top W. of KBFSC, near A; 250 m; (Poulsen 76).
Cyrtandra sp. Proto-terrestrial; ridge slope W. of KBFSC, gully between A and S. Belalong; 150 m; (Poulsen 161).
Cyrtandra sp. Proto-terrestrial; ridge W. of KBFSC; plot A, B, C (low); 250 m; (Poulsen 119).
Cyrtandra sp. Proto-terrestrial, rather woody at base; ridge slope W. of KBFSC; plot A, B, C (middle, high); 250-800 m; (Poulsen
120).
Cyrtandra sp. Proto-terrestrial, rather woody base; ridge W. of KBFSC; plot A; 250 m; (Poulsen 220).
Cyrtandra sp. Proto-terrestrial, rather woody base; ridge W. of KBFSC; plot A, B, C (low); 250 m; (Poulsen 221).
Cyrtandra sp. Proto-terrestrial; Bt. Belalong; plot C (high); 850 m; (Poulsen 302).
Didymocarpus amoenus C.B.Clarke s. lat. Proto-terrestrial; from river banks of S. Belalong to Bt. Belaiong, below summit; 65-850
m; (Poulsen 189. 283).
Didymocarpus gracilipes C.B.Clarke s. lat. Proto-terrestrial; Dt. Bcialong, steep rock ncar Pondok Belalong; 850 m; (Poulsen 310).
Didymoearpus laneeolalus C.B.Clarke Proto-terrestrial, woody base; ridge top W. ofKBFSC; plot A, C (low); 280 m; (Poulsen 85).
Loxoearpus periolaris (C.B.Clarke) B.L.Burtt Proto-terrestrial; Bt. Belalong; plot C (high); 850 m; (Poulsen 321).
Hanguanaceae
Hanguana malayana (Jack) Merr. Proto-terrestrial; short stem rather woody, leaves linear; S. Belalong, on river bank; 70 m; (Poulsen
121).
Hanguana sp. I Proto-terrestrial, stem rather woody, broad leaves; S. Belalong, on slope near river; plot B, C (high); 60-850 m;
(Poulsen 33).
Hypoxidaceae
Cureu/igo lalifolia Dryand. Proto-terrestrial; BI. Belalong, S. Belalong, steep slope behind KBFSC; plot A, B. C (low, middle, high);
80-850 m; (Poulsen 90; 295).
Curcu/igo racemosa Ridl. Proto-terrestrial, occasional epiphyte; from S. Babi. steep bank above river. to ridge slopes near Bt.
Belalong; plot C (high); 70-850 m; (Poulsen 38. 60, 278).
Liliaceae
Peliosanthes tela Andr. ssp. humilis (Andr.) Jessop. Proto-terrestrial; Bt. Belalong; plot C (high); 850 m; (Poulsen 368).
Marantaceae
Donax cannaeformis (G. Forst.) K. Scum. Proto-terrestrial, woody 4 m tall; on shale in gully between A and S. Belalong; 150 m;
(Poulsen 166).
PluJcelophrynium maximum (Blume) K.Schum. Proto-terrestrial; S. Belalong river bank and on steep slope behind KBFSC; plot A,
C (low) 65-150 m; (Poulsen 2, 44, 61).
Phrynium cf. placemarium (Laur.) Merr. Proto-terrestrial; on bank near river; 65 m; (Poulsen 191).
Phrynium sp. aff. injlarum Merr. Proto-terrestrial; ridge W. of KBFSC; plot A, B, 250 m; (Poulsen 209).
Slachyphrynium bomeense Rid!. Proto-terrestrial; ridge W. of KBFSC, ridgetop; plot A; 250 m; (Poulsen 75).
Slachyphrynium cf. jagorianum (K.Koch) K.Schum. Proto-terrestrial; ridge between K. Belalong and Bt. Belalong and ridge W. of
KBFSC; plot A; 250 m; (Poulsen 282).
Melastomataceae
Sonerila begoniefolia Blume. Proto-terrestrial, very low with woody base; Bt. Belalong; plot C (high); 880 m; (Poulsen 308).
Sonerila bomeensis Cogn. var. selulosa Stapf. Proto-terrestrial, occasionally epiphyte on mossy trunks; Bt. Belalong; plot C (high);
850 m; (Poulsen 320).
Sonerila helerophylla Jack. Proto-terrestrial; Bt. Belalong; plot C (high); 8S0 m; (Poulsen 319).
Sonerila obovala Schwartz. Proto-terrestrial; Bt. Belalong; plot C (high); 880 m; (Poulsen 94).
Sonerila pulchella Stapf. Proto-terrestrial; Bt. Belalong; plot C (high); 900 m; (Poulsen 100).
Musaceae
Musa 3 spp. Proto-terrestrial, giant herb; S. Belalong and S. Temburong, in large gaps near river; 65 m; (no voucher, Argent pers.
comm.).
Myrsinaceae
Labisia pumila (Blume) Fer.-Vill. Proto-terrestrial, with woody base; ridge W. of KBFSC; plot B, C (low, high); ISO-8S0 m;
(Poulsen 77).
Orcbidaceae
Aphyl/orchis pal/ida Blume. Saprophyte, less than 20 cm; ridge between K. Belalong and Bt. Belalong; plot C (middle); 2S0-400 m;
(Poulsen 281).
Apos/aSia elliplica J.I.Sm. Proto-terrestrial; ridge between K. Belalong and Bt. Belalong, near Pondok Busiri; plot B; 250-S00 m;
(Poulsen 89).
Aposlasia wallichii R. Br. in Wall. Proto-terrestrial, 50 cm; ridge W. ofKBFSC and Bt. Bel.long; plot A, B, C (low, high); 250-800
m; (Poulsen 122, 312).
Claderia viridiflora Hook.f. Climber, on ground; ridge between K. Belalong to Bt. Belalong; plot B; 250-300 m; (Poulsen 219, 371).
Cyslorchis variegala Blume. Proto-terrestrial; in moist, dark places in forest; plot B, C (high); 2S0-800 m; (Poulsen 323).
Dendrobium sp. Epiphyte, probably fallen, rooted in ground; ridge W. ofKBFSC; plot A; 2S0 m; (Poulsen 215).
Dipodium sp. Climber, juvenile on ground; on ridge between K. Belalong and Bt. Belalong; plot B; 2S0 m; (Poulsen 374).
Eria sp. Epiphyte, probably fallen, rooted in ground; on ridge between K. Belalong and Bt. Belalong; plot B; 2S0 m; (Poulsen 375).
Lecanorchis mulliflora J.J.Sm. Saprophyte, SO cm; ridge between K. Belalong and Bt. Belalong; plot A, C (low, middle); 350 m;
(Poulsen 88).
liparis wrayi Hook.f. Proto-terrestrial; trail from K. Belalong to Bt. Belalong, on big, old well-decomposed log; plot C (low); ISO-
-275 m; (Poulsen 200).
Malaxis sp. Proto-terrestrial; on trail along S. Relalong, above floodlevel, damp area; 65 m; (Poulsen 255).
Mischobulbum scapigerum (Hook. f.) Schltr. Proto-terrestrial; Bt. Belalong; plot C (high); 850 m; (Poulsen 373).
Ploeoglollis sp. Proto-terrestrial; S. Belalong; 70 m; (Poulsen 218).
Tainia paueifolia (Breda) J.J.Sm. Proto-terrestrial; ridge W. of KBFSC; plot A; 250 m; (Poulsen 214).
Tropidia graminea Blume. Proto-terrestrial; ridge W. of KBFSC and Bt. Belalong; plot A, C (high); 2S0-850 m; (Poulsen 216).
Tropidia saprophyliea J.J .Sm. Saprophyte; ridge between K. Relalong and Bt. Belalong, Rt. Belalong; plot B; 2S0-7S0 m; (Poulsen
303).
Petrosaviaceae
Pelrosavia slellaris Becc. Saprophyte; damp flat area below Bt. Belalong; 8S0 m; (Poulsen 305).
Piperaceae
Piper cf. abbreviatum Opiz. Climber, juvenile on ground, ridge W. of KBFSC; plot A; 2S0 m; (Poulsen 241).
Poaceae
Paspalum conjugalum Berg. Proto-terrestrial; S. Belalong, near river; 60 m; (Poulsen 17).
Seroloehloa ureeolala (Roxb.) Judz. Proto-terrestrial; valley West of KBFSC; plot B; 110-2S0 m; (Poulsen 293).
Thysanolaena maxima Kuntze Proto-terrestrial, 3-4 m long stems hanging a few meters above river; S. Temburong, steep rocky river
bank in light gap; 70 m; (Poulsen 274).
PolygaIaceae
Epirixanlhes rylindrica Blume. Saprophyte; ridge between K. Belalong and Bt. Belalong, damp flat area below Bt. Belalong; plot C
(high); 800 m; (Poulsen 304).
Epirixamhes elongala Blume. Saprophyte; ridge W. of KBFSC; plot A, B; 2S0 m; (Poulsen 224).
Epirixanlhes pallida Wendt. Saprophyte; ridge W. of KBFSC; 80 m; (Poulsen 3).
Rubiaceae
Aeranlhera sp. aff.jruleseens Valeton. Proto-terrestrial, with woody base; ridge slope W. ofKBFSC; plot A, C (low, high); 250-800
m; (Poulsen 52).
Acranlhera sp. aff. involuerala Valeton. Proto-terrestrial, with woody base; ridge slope W. of KBFSC; plot A, C (low); 250-800
(Poulsen 51).
Acranlhera sp. aff. longipeliolala Bremek. Proto-terrestrial, woody at base; valley W. ofKBFSC, near river; 100 m; (Poulsen 286).
Rubiaceae (continued)
Argostemma hameliaefolium Wemham in Gibbs. Proto-terrestrial; S. Beiaiong, river bank above tloodlevel, damp area; 65 m; (Poulsen
12,253).
Argoslemma moultonii Rid!. Proto-terrestrial; Bt. Belalong; plot C (high); 850 m; (Poulsen 309).
Argoslemma subfalcifolium Bakh.f. Proto-terrestrial; Bt. Belalong; plot C (low, high); 850 m; (Poulsen 325).
Hedyolis sp. Proto-terrestrial; ridge W. of KBFSC; plot A; 250 m; (Poulsen 222).
Ophiorrhiza sp. Proto-terrestrial; ridge W. of KBFSC; plot A; 250 m; (Poulsen 223).
Psycholria sp. Climber, genninating on ground, small plant; ridge W. of KBFSC; plot A, B, C (low. middle, high); 250-850 m;
(Poulsen 118).
Taccaceae
Tacca integrifolia Kef Gawl. Proto-terrestrial; S. Temburong, bank above river; 65 m; (Poulsen 240).
Triuridaceae
Sciaphike densiflora Schltr. Saprophyte; ridge W. ofKBFSC; plot A, C (low); 250 m; (Poulsen 27).
Sciaphike secundiflora Thw. ex Benth. Saprophyte; ridge W. ofKBFSC, between K. Belalong and Bt. Belalong, and damp ftat area
below Bt. Belalong; plot A, C (high); 250-800 m; (Poulsen 306, 307).
Urticaceae
EkelOslema sp. aff. kabayense (Gibbs) Schroter Proto-terrestrial. long creeping, also epiphytic(l93); on trail along S. Belalong, above
ftoodlevel, damp area; 65 m; (Poulsen 193, 252).
Eketoslerna sp. aff. sesquifolium (BI.) lIassk. Proto-terrestrial; ridge slope W. ofKBfSC, gully between A and S. Belalong; 150 m;
(Poulsen 159).
Ekeloslerna sp. Proto-terrestrial; S. Belalong and S. Temburong, river bank in wet soil; 60 m; (Poulsen 10, 107).
Elatoslema sp. Proto-terrestrial; S. Baki. on shale rocks in river; 70 m; (PouLsen J 11).
Ziugiberaceae
Alpinia giabra Rid!. Proto-terrestrial; particularly common on dry ridges; plot B, C (middle, high); 180-800 m; (Poulsen 29, 279).
A/pinia ligulata K.Schum. Proto-terrestrial; creek adjoining S. Belalong and along trail to S. Esu, bank; 60-65 m; (Poulsen 13.329).
Alpinia sp. aff. capilelkela Jack. Proto-terrestrial; ridge W. of KBFSC; Ruth Levy's plot, A. B; 250 m; (Poulsen 226).
Alpinia sp. sect. Cenolophon. Proto-terrestrial. big dark green ginger over 2m; ridge slopes; plot B, C (high); 250-850 m; (PouLsen
327).
Alpin;a sp. sect. Allughas. Proto-terrestrial; Bt. Belalong; plot C (high); 850 m; (Poulsen 328).
Amomum bicomicukelum K.Schum. Proto-terrestrial; ridge slopes; plot B. C (low. high); 250-850 m; (Poulsen 326).
Amomum ('oriaceum R.M.Sm. Proto-terrestrial; common from river banks to Bt. Belalong: plot A. B. C (low, high): 70-800 m;
(Poulsen 21; 277).
Amomum Jaxisquarrosum K.Schum. vel aff. Proto-terrestriaL S. Belalong, steep soil; 70 m; (Poulsen 267).
Amomum sp. atf. oliganthum K.Schum. Proto-terrestrial; on bank of small stream from A adjoining S. Belalong; 70 m; (Poulsen 261).
Amomum pungens R.M.Sm. vel afT. Proto-terrestrial; S. Babi. on wet soil bank in shade; 70 m; (Poulsen 276).
Amomum sp. Proto-terrestrial; ridge W. of KBFSC; plot A; 250 m; (Poulsen 20).
Amomum sp. aff. ridleyi R.M. Sm. Proto-terrestrial; ridge between K. Belalong and Bt. Belalong; plot B, C (high); 30~85 m;
(Poulsen 87).
Amomum sp. afr. flavoalbum R.M. Sm. Prolo-terrestrial; very common on ridge slopes; plot A, B, C (low, middle, high); 250-800
m; (Poulsen 280).
Amomum sp. Proto-terrestrial; ridge W. of KBFSC; plot A; 250 m; (Poulsen 24).
Boesenbergia belalongensis A.D.Poulsen. Proto-terrestrial: from S. Belaiong near river to Bt. Belalong. steep shale bank; pial C
(middle, high); 65~80 m; (Poulsell 182).
Boesenbergia jlavoalba R.M.Sm. Proto-terrestrial; common on ridge slopes in dosed forest; plot A, B. C (low, high); 250-800 m:
(Poulsen 50).
Boesenbergia gracilipes (K. Schum.) R.M. Sm. Proto-terrestrial; S. Temburong, on flat bank 50 m from river; 75 m; (PouL'ten 137).
Boesenbergia orbiculata R.M.Sm. Proto-terrestrial, creeping; on trail along S. Belalong, above floodlevel, damp area; 65 m; (Poulsen
257).
Boesenbergia urceoligena A.D. Poulsen .. Proto-terrestrial; on trail along S. Belalong, above floodlevel. damp area: 65 m; (Poulsen
251).
Elettariupsis kerbyi R.M.Sm. Proto-terrestrial; common species on ridges; plot A, B, C (low. middle. high): 250-850 m: (Poulsen
74).
Elellan'opsis sp. aff. curtisii Baker Proto-terrestrial; on slopes; plot A, B, C (low); 250 111; (Poulsen 227).
Elellariopsis sp. Proto-terrestrial; S. Esu, wet bank near creek; plot B: 65-250 m; (PouLsen 330).
Ellingera breviiabris (Valeton) R.M.Sm. Proto-terrestrial; from S. Belalong, river bank to Bt. Belalong; plot A, B. C (Iuw. high);
65-800 m; (Poulsen 170).
Etlingera fimbriobracteata (K. Schum.) R.M.Sm. Proto-terrestriaL giant ginger confined to river banks of S. Temburong and S.
Belalong; 65 m; (Poulsen 56; 376).
Etlingera punicea (Roxb.) R.M.Sm. Proto-terrestrial. giant ginger: confined to river banks of S. Temburong and S. Belalong; iban
name: tapas; 65 m; (Poulsen 46).
Ellingera velutina (Rid!.) R.M.Sm. Proto-terrestrial; S. Baki, in wet area near creck, and on ridge slopes W. of KBFSC; plot A;
60-250 m; (Poulsen 83).
Ellingera sp. nov. Proto-terrestrial; ridge slopes; plot A, B, C (high); 80-800 m; (Poulsen 35).
Ellingera sp. nov. Proto-terrcstrial; ridge W. of KBFSC; plot A, C (high) 220-800 m; (Poulsen 130).
Etlingera sp. afT. muluensis R.M.Sm. Proto-terrestrial; S. Belalong, on river bank; common in B; 65-250 m; (Poulsen 169).
Etlingera sp. Proto-terrestrial; ridge slopes; plot A. C (low. middle); 250-500 m; (Poulsell 225).
Globba atrosanguinea Teijsm. & Binn. Proto-terrestrial; S. Belalong; in moist areas near river; 65 111; (Poulsen 171. 331).
Globba tn'color RidL var. gibbsiae (RidL) R.M.Sm. vel. aff. Proto-terrestrial: S. Enkiang, S. Babi, common in moister places on
slopes; plot A, C (low, high); 80-800 Ill; (Poulsen 116, 263, 275).
Zingiberaceae (continued)
Hedyehium muluense R.M.Sm. Epiphyte, on fallen debris; S. Belalong, near KBFSC; 70 m; (Poulsen 287).
Homs/emia scyphifera (Koenig) Steud. Proto-terrestrial; S. Temburong and S. Belalong, river bank; 70 m; (Poulsen 55, 64).
Homs/emia IOmen/osa (Bl.) Bakh. f. Proto-terrestrial; Bt. Belalong; plot C (high); 850 m; (Poulsen 332).
Plagioslaehys albiflora Ridl. Proto-terrestrial; very common on ridge slopes; plot A, B, C (middle, high); 150-800 m; (Poulsen 22,
34, 176).
Plagios/aehys s/robilifera (Baker) Ridl. Proto-terrestrial; very common species on ridge slopes, Bt. Belalong; Iban name: banjang; plot
A, B, C (low, high); 180-800 m; (Poulsen 4, 23, 95).
Zingiber grijJi/hii Baker. Proto-terrestrial; on steep slope behind KBFSC, Bt. Belalong; plot C (high) 70-800 m; (Poulsen 202).
Zingiber pseudopungens R.M.Sm. Proto-terrestrial; S. Baki, valley W. of KBFSC, in wet places; 70-110 m; (Poulsen 110, 289).
Zingiber cf. albiflorum R.M.Sm. Proto-terrestrial; ridge W. of KBFSC; plot A, B, C (high); 250-800 m; (Poulsen 228).
Zingiber sp. Proto-terrestrial; ridge slopes; plot A, B, 250 m; (Poulsen 337).
Genus indet. Geoeharis or Alpinia. Proto-terrestrial; ridge slopes; plot A, B, C (low, middle); 250 m; (Poulsen 229).
Genus indet. Proto-terrestrial; Bt. Belalong; plot C (high); 850 m; (Poulsen 333).
Genus indet. Proto-terrestrial; .Bt. Belalong; plot C (high); 850 m; (Poulsen 334).
PTERIDOPHYTES
Blechnaceae
Blechnumfinlaysonianum Wall. ex Hook. & Grev. Proto-terrestrial; S. Temburong, on bank in dark area; plot A; 70 m; (Poulsen
136).
Cyatbeaceae
Cya/hea molueeana R.Br. Proto-terrestrial; S. Temburong, river bank; 65 m; (Poulsen 69).
Dennstaedtiaceae
Lindsaea bomeensis Hook. ex Baker. Proto-terrestrial; Bt. Belalong; plot C (middle, high); 500-850 m; (Poulsen 93).
Lindsaea cui/rata (Willd.) Sw. Proto-terrestrial; ridge slope W. ofKBFSC ; plot A, B, C (low, high); 250-850 m; (Poulsen 128, 343).
Lindsaea divergens Hook. & Grev. [= Isoloma divergens]. Proto-terrestrial; ridge between K. Belalong and Bt. Belalong, in small
valley near B; 200 m; (Poulsen 271).
Lindsaea doryphora Kramer. Proto-terrestrial; on ridge between K. Belalong and BI. Belalong; plot B; 250 m; (Poulsen 341).
Lindsaea integra Holttum. Proto-terrestrial; Bt. Belalong; plot B, C (high); 250-850 m; (Poulsen 344).
Lindsaea lucida Blume ssp. lucida. Proto-terrestrial; S. Temburong, muddy river bank; 65 m; (Poulsen 72).
Lindsaea parallelogram"", A1derw. Proto-terrestrial; common from S. Temburong, muddy river bank to Bt. Belalong; plot A, B, C
(high); 65-850 m; (Poulsen 71, 342).
Tapainidium pinnalum (Cav.) C.Chr. Proto-terrestrial; S. Temburong, muddy bank near river; 65 m: (Poulsen 70).
Tapeinidium luzonieum (Hook.) Kramer. Proto-terrestrial; BI. Belalong; plot C (high); 850 m; (Poulsen 350).
Dipteridaceae
Dip/eris lobbiana (Hook.) Moore. Rheophyte; S. Temburong, close to river; 70 m; (Poulsen 63).
Dryopteridaceae
Araehniodes cf. /ripinnala (Goldm.) Sledge agg. Proto-terrestrial; BI. Belalong; plot C (high); 850 m; (Poulsen 353).
Cye/opel/is erenala (Fee) C.Chr. Proto-terrestrial; ridge slope W. of KBFSC; ISO m; (PouL"n 178).
Diplazium asperum Blume. Proto-terrestrial, up to 2.S m; S. Belalong, small stream adjoining S. Belalong; 65 01; (Poulsen 184).
Diplazium cordifolium Blume. Proto-terrestrial, often with irridescent leaves; common on moist slopes; plot A. B, C (low. high);
250-850 m; (Poulsen 203).
Diplazium crenaloserra/um (Blume) Moore. Proto-terrestrial; on higher slopes near ridge tops; plot A, B, C (low, middle, high);
250-850 m; (Poulsen 26, 73, 91).
Diplazium esculenlum (Retz.) Sw. Proto-terrestrial; young leaves eaten as vegetable; iban name: pakoh; K. Belaiong, light open
riverbank on accumulated organic material; 65 m; (Poulsen 144).
Diplazium rnallaeense C.Presl. Proto-terrestrial; Bt. Belalong; plot C (high); 850 m; (Poulsell 348).
Diplazium pal/idum (Blume) Moore. Proto-terrestrial; leaves often irridescent; BI. !lelalong; plot C (high); 850 m; (Poulsen 346).
Diplazium porphyroraehis (Baker) Diels. Proto-terrestrial; common on steep soil slopes with exposed soil; plot A, B, C (low, high);
220-850 m; (Poulsen 80).
Diplazium riparium Holttum. Proto-terrestrial; near flood steram in forest; plot A, C (low); 250 m; (Poulsen 123).
Diplazium xiphophyllum (Baker) C.Chr. Proto-terrestrial; S. Baki, on clayish soil slope near river; 70 m; (Poulsen 114).
Dip/azium sp. Proto-terrestrial; Bt. Belalong; plot C (high); 850 m; (Poulsen 347).
Pleocnemia irregulan·s (Presl) Holttum. Proto-terrestrial; valley and ridge slope W. ofKBFSC, in wet area near pig walo~ 110 m;
(Poulsen 162, 294).
Tee/aria angula/a (WiUd.) Copel. Proto-terrestrial; ridge slope W. of KBFSC; plot A; 150-250 m; (Poulsen 154, 163).
Tee/aria barberi (Hook.) Copel. Proto-terrestrial; ridge slopes in moist places; plot A, C (high); 250-850 m; (Poulsen 125).
Teclaria decurrens (Presl) Copel. Prot-esial~ S. Baki, on soil slope near stream; 70 m; (Poulsen 140).
Tee/aria inopinata Holttum. Proto-terrestrial; Bt. Belalong, on rocks; plot C (high); 850 m; (Poulsen 354).
Tee/aria lobii (Hook.) Copel var. lobii. Rheophyte; S. Belalong, river bank shale; 70 m; (Poulsen 299).
Tee/aria /emala (Baker) Copel. Rheophyte; S. Baki, on shale rocks in river; plot A, C (low, high); 70-850 m; (Poulsen 112, 132).
Tee/aria vas/a (Blume) Copel. Proto-terrestrial; S. Temburong, soil slope on river bank, in shade; 70 m; (Poulsen 138).
Tee/aria sp. Proto-terrestrial; ridge slope W. of KBFSC, between A and S. Belalong, gully; 150 m; (Poulsen 164).
Tee/aria sp. Proto-terrestrial; Bt. Belalong; plot C (high); 850 m; (Poulsen 349).
Tee/aria sp. Proto-terrestrial; Bt. Belalong on rocks; plot C (high); 850 m; (Poulsell 355).
Gleicheniaceae
Dicranopleris linean's (Burm.f.) Underw. ssp. subpeclinala. Proto-terrestrial, becoming scandent with age; ridge W. of KBFSC. in
light gap; plot C (high); 250-850 m; (Poulsen 82).
GLeichenia lruncala Willd. Proto-terrestrial; from S. Temburong, overhanging river to Bt. Belalong, in clearing on the summit; 70-913
m; (Poulsen 96. 104).
HymenopbyUaceae
Tn'chomanes obscurum Blume. Proto-terrestriaL on ridge between K. Beiaiong and Bt. Belalong; plot B. C (high); 250-850 m;
(Poulsen 340).
Tn'chomanes singaporeanum (Bosch) Alderw. Proto-terrestrial; common on ridge slopes, abundant on steep slopes with exposed soil;
plot A. B. C (low. middle. high); 250-850 m; (Poulsen 204).
Lomariopsidaceae
Lomariopsis lineala (Presl) Holttum. Climber; juvenile on ground; plot A; 250 m; (Poulsen 126).
Teralophyllum rOlundifolialum (Bonap.) Holttum. Climber. juvenile with bathyphylls on ground; ridge slope W. of KBFSC; plot A;
250 m; (Poulsen 131).
Lycopodiaceae
Lycopodium cernuum L. Proto-terrestrial; BI. Belalong; plot C (high); 850 m; (PouLsen 356).
Marattiaceae
Angiopleris sp. Proto-terrestrial; Ill. Ilelalong; plot C (high); 850 m; (Poulsen 345).
Chn"slensenia aesculifolia (Blume) Maxon. Prot~esial; S. Baki. soil slope near creek; 70 m; (Poubien J 13).
Polypodiaceac
Asplenium Ienerum Forst. var. relUsum. Proto-terrestrial; S. Esu, on rock; 70 m; (Poulsen 259).
Pteridaceae
Syngramma alismifoLia (Presl) I.Sm. Proto-terrestrial; on ridgebetween K. Belalong and Ill. Ilelalong; plot IJ. C (high); 250-850 m;
(Poulsen 338).
Taeni,is blechnoides (Willd.) Sw. Proto-terrestrial; on ridge between K. Belalong and BI. Relalong; plot R. C (low. high); 250-850
m; (Poulsen 339).
Schizaeaceae
Lygodium circinallum (Burm. f.) Sw. Scandent; ridge slope W. of KBFSC; plot A; 70-250 m; (Poulsen 133).
Schizaea dicholoma (L.) Sm. Proto-terrestrial; ridge slope W. of KBFSC; plot A. B; 250 m; (Poulsen 127).
Schizaea digilala (L.) Sw. Proto-terrestrial; on ridge near Pondok Busiri; 500 m; (PouLsen 360).
Selaginellaceae
Selaginella bluuensis Alderw. Proto-terrestrial; common on ridge slopes; plot A. B. C (low. high); 70-850 m; (Poulsen 7).
Selaginella dielsii Hieron. Proto-terrestrial; BI. Belalong; in clearing on the summit; plot C (high); 800-900 m; (Poulse" 99).
SelagineLla cf. boschai Hieron. Proto-terrestrial; on ridge W. of KBFSC; A ; 250 m; (Poulsen 230).
SelagineLla subserpenlina Alderw. Proto-terrestrial; BI. Belalong. in clearing near the summit; plot C (high); 800-900 m; (PouL"" 98).
Selaginella cf. brevipes A.Br. Proto-terrestrial; on steep slope behind KBFSC; 70 m; (Poulsen 8).
Selaginella sp. Proto-terrestrial; S. Temburong. river bank; 70 m; (PouLsell 231).
SelagineLla sp. Proto-terrestrial; ridge W. of KBFSC; 250 m; (Poulsen 242).
Selaginella sp. Proto-terrestrial; BI. Belalong; plot C (high); 850 m; (PouLsen 357).
SelagineLla sp. Proto-terrestrial; Ill. Belalong; plot C (high); 850 m; (Poulsen 358).
Selaginella sp. Proto-terrestrial; BI. Belalong; plot C (high); 850 m; (Poulsen 378).
SelagineLla sp. Proto-terrestrial; ridge slopes to BI. Belalong; plot B. C (high); 250-800 m; (PouLsen 379).
Selaginella sp. Proto-terrestrial; BI. Belalong; plot A. C (high); 850 m; (Poulsen 380).
Thelypteridaceae
Mesophlebion beccarianum (Ces.) Holttum. Proto-terrestrial; B1. Belalong; plot C (middle. high); 850 m; (Poulsen 92).
Mesophlebionfalcalilobum Holttum var.faLcalilobum. Proto-terrestrial; one of the most common fems in the area; plot A. B. C (low.
middle); 250-800 m; (PouLsen 134).
Mesophlebion molleyanum (Hook.) Holttum. Proto-terrestrial. juvenile from ecological study lalternatively M. dulitense Holtt.l; Ill.
Belalong; plot C (high); 850 m; (Poulsen 351).
Pronephrium nilidum Holttum. Proto-terrestrial; ridge slope W. ofKBFSC; plot A; 250 m; (PouL"n 129).
Pronephrium cf. menisciicarpon (Blume) Holtturn. Proto-terrestriat valley W. ofKBFSC. in wet flood stream; 110 m; (Poulsen 291).
Pronephrium sp. Proto-terrestrial; BI. Belalong; plot C (high); 850 m; (Poulsen 352).
Sphaeroslephanos plerosporus (Alderw.) Hottturn var. crenalus Holttum. Proto-terrestrial; ridge slope W. of KBFSC. on rocks in
floodstream; plot A; 220 m; (Poulsen 124).
Sphaeroslephanos sp. Proto-terrestrial; K. Belalong, light open riverbank on accumulated organic material; 65 m; (Poulsen 145).
Sphaeroslephanos sp. Proto-terrestrial, valley West of KBFSC, in wet flood stream; 110m; (Poulsen 290).
Sphaeroslephanos sp. aff. lrichoclamys Holttum. Proto-terrestrial; on ridge NW of A. near floodstream; 250 m; (PouLsen 361).
Wild bees of Brunei Darussalam
DA YID W. ROUBIK
Smithsonian Tropical Research Institute, Unit 0948, APO AA 34002-0948, USA.
ABSTRACT. The wild bee fauna of Borneo and Brunei is singularly depauperate compared with other
tropical areas but very rich in the two bee groups at the zenith of social insects. In the forests of
Belalong I found 4 Apis and 22 stingless bee species; the latter, mostly Trigona, included 76% of all
stingless bee species known from Borneo, and all three endemics. Other bee species probably number
about 80 if the proportions of the highly social bees found in Indonesia apply to Belalong. A 'rapid
assessment' of the highly social bee fauna was possible in Brunei, even during rainy weather in
November, by spraying sugar and salt baits on vegetation. Virtually every species of Trigona, Apis and
Pariotrigona came to these baits during a few days. Assays of various salt solutions revealed the
following preferences by the stingless bees (predominantly Trigona melina) and honey bees
(predominantly Apis koschevnikovi and A. dorsata) : Na > Mg > K > NH. > PO.. Preference for
sodium, magnesium, and potassium is possibly linked to neural transmission requirements, osmotic water
balance and low levels of these elements in soil, expressed directly in the nectar and pollen that bees
consume. Apis koschevnikovi was found to be similar to African and other tropical cavity-nesting Apis
in showing the highest profitability (measured as energy harvest rate of sucrose solutions) at 45 % sugar
solutions, compared with solutions as low as 15% and as high as 65% sugar. Ground-level surveys of
bee nests located 33 colonies of stingless bees along 12 Ian of trails. I calculate the density of stingless
bee colonies to be approximately 3 per hectare - lower than the 4 to 6 colonies in each hectare found in
some Neotropical forests. The sugar-salt baiting technique also revealed the first known association
between an orchid and the genus Apis. Apis koschevnikovi was found with a small orchid pollinarium
attached to its scutellum, thus it may have a pollinator role often played by Trigona in the neotropics for
a few orchid genera. Other noteworthly bees were the large, nocturnal Xylocopa myops, megachilids and
long-tongued Amegilla, all likely principal pollinators of obligately outcrossing tree species.
KEY WORDS: bee biodiversity, Trigona, Pariotrigona, Apis, feeding behaviour, sugar-salt baiting,
Borneo.
INTRODUCTION
Recent treatments of animal or plant biodiversity place too much emphasis on species
number, rather than on an actual assessment of their biological diversity, which includes
ecological roles, behaviour, and many characteristics of living things that are seldom
uncovered by comparative morphology and museum specimens (see also Williams 1993).
In work described here I give results of a short survey of the bees at Belalong field station
in Brunei, and an effort to characterize some of their feeding behavior. These results are
compared to other equatorial and some temperate areas. In an attempt to make a rapid
assessment of the richness of the socially-foraging colonial bees at Belalong, I used baiting
techniques that generally provide results in any of the tall equatorial forests. By attracting
bees to feeding stations, other studies can be made with the trained bees. At Belalong these
were aimed at answering two questions: 1) What are some of the important non-floral
resources that bees obtain by visiting such odd substrates as chil i sauce containers, laundry,
human skin, campfire ashes, and seepage areas along stream banks? 2) What kinds of nectar
sources are colonial bees most likely to prefer, to judge from energy intake rate at different
sugar solutions? (cf. Roubik & Buchmann 1984).
59
D. S. Edwards et al. (eds.). Tropical Rainforest Research - Current Issues, 59-66.
60 DAVID ROUBIK
METHODS
Between 28 October until 25 November, 1991, when very little flowering took place in the
forest understory, I set up monitoring stations for the social bees by spraying combinations
of honey, sugar and table salt solutions with banana flavouring on small bushes 1-3 m high,
near the laboratory clearing at Belalong. The amount of flavouring used was one or two
drops per litre solution. Using a hand-held refractometer, I prepared solutions with sugar
concentration of 28 % by weight. A hand-held sprayer was used to make bait applications
(after Wille 1962). General procedure was to spray 200-300 ml of solution at each site,
beginning at 0640, and replenishing the solutions as they evaporated in direct sunlight or
were depleted by foraging bees. A fine mist spray was used to enhance odour dispersal,
while a coarser spray was used to cover leaf surfaces with the bait. After bees began to find
the solutions and bring nest mates there, within the first morning in which weather was fair,
bees were periodically collected with an aerial net. Additional survey and collecting were
carried out by walking the forest trails searching for social bee nests in living trees. I
covered about 12 km of hilly forest on either side of the Belalong River, monitoring a swath
of forest 10 m wide (5 m on either side of the trail). Thus I surveyed roughly 12 ha of the
forest for bee nests.
Baiting experiments were performed on 23-25 November at the laboratory clearing, under
the corner of a building which provided shelter during intermittent rains. Bees which had
not been given baits for more than two weeks immediately found baits when these were
placed in their former positions on the afternoon of 23 November. Sugar solution feeders
made of inverted, paired petri dishes were placed on one bush used for baiting, where bees
visiting them were taken to the new site used for the feeding assays. This was done by
carefully moving a group of bees on the dish they fed upon. Salt and sugar solutions for
behavioural assays were made with distilled water as follows: sugar solutions of 15, 30, 45
and 65 % were made with table sugar, their concentration determined using an Atago
hand-held refractometer corrected for temperature. Five 0.5 M salt solutions were prepared
of NaCl, MgCI 2 , Na3 P04 , NH4 Cl, and KCl, with the equivalent solute quantity per litre of
5.8 g, 20.3 g, 38 g, 5.4 g, and 7.5 g, respectively. Feeders for the salt solutions consisted
of petri plate lids each with a moistened yellow paper square in the middle, and also of open
vials 9 cm long by 2 cm in diameter, 1/3 filled with each solution.
The five salt solutions feeders were placed on a tray in a four-cornered arrangement with
one feeder in the centre. Relative positions of feeders were changed periodically, and the
tray was rotated 1800 each 15 min. to prevent a tendency of forager site constancy from
determining feeding choice. The total number of each bee species at the salt feeders was
recorded each 15 min. during 5-min. periods between 0700 and 0930 on 23 and 24
November. During the latter part of the day on the 24th and all day on the 25th, the five
open vials containing the solutions were left in the feeding arena. Vials served as traps, not
allowing bees to leave once they had entered. Total bees captured on each day were
combined with observed numbers feeding at the moistened paper patches. A feeding bout
was scored when a bee extended its proboscis and fed > 10 seconds.
Sugar solution feeding assays were made following Roubik & Buchmann (1984). Two
sugar-solution reservoirs of 45 % sugar were maintained in the feeding arena, where field
bees fed freely during the study. Two smaller feeding platforms were maintained in the area
for imbibement assays; a platform was carefully removed from the foraging area, thus
avoiding disturbance by other bees, when an individual bee arrived at one. A 25 ilL drop
was measured using a calibrated micropipette and placed in a circular depression (the
inverted base of a plastic film canister) for measurement of imbibement rate. Individual bees
were timed at the depression from the start of feeding at the droplet until the end of a feeding
bout, when the entire droplet had been removed. I terminated the stopwatch recording as
Wild bees of Brunei Darussalam 61
soon as the bee started to move the proboscis laterally, as if seeking additional solution in
the feeding depression. Imbibement assays were conducted on 25 November from 0930 until
1630, at ambient temperature of approximately 28°C, and 100% relative humidity. The
lowest concentration of separate sets of the different solutions were presented first,
progressing to the highest. At least 20 bees were recorded feeding at each type of resource,
and no bee was monitored in more than one feeding assay at a particular concentration.
RESULTS
Species Richness and Nesting

The social bees recorded at feeders included all those found nesting in the area, and many
whose nests were never located (Table 1). Some bees have Iban names that were provided
by Awang Mergi (Table 1), others are mentioned by Beccari (1904). The stingless
honey-making bees (Meliponinae, or 'c1ulut') were the richest group, having 22 species
ranging in size from the minute Pariotrigona pendleburyi and Trigona (Heterotrigona)
fuscobalteata (2.5 mm) to the relatively large Trigona (Heterotrigona) thoracica and T.
(Homotrigona) Jimbriata (8 mm). These constitute 76% of stingless bee species recorded for
the island of Borneo (Sakagami et at. 1990). Further, the males of Pariotrigona pendleburyi
were previously unknown (Michener 1990) but were found exiting from their nests at
Belalong. There are four species of honeybees at Belalong, the most common being the giant
Table I. Highly social bees and nests encountered near the Belalong field station.
Species Nest (trees, No.) Nest height (m)

Apis koschevnikovi 'ranyuan'
A. dorsata 'manyi' or 'Ianyeh' Koompasia, 1 40
A. cerana 'ranyuan tikung'
A. andreniformis 'nuang'
Pariotrigona pendleburyi Shorea, Dryobalanops, 8 0.04-0.4
Trigona (Lepidotrigona) nitidiventris
T. (L.) terminata 4
T. (L.) ventralis 5.5
T. (Heterotrigona) melina 2 ground level
T. (H.) haematoptera 8
T. (H.) hobbyi
T. (H.) collina 11 ground level
T. (H.) thoracica
T. (H.) apicalis 3 1.5-3
T. (H.) drescheri 1 ground level
T. (H.) melanocephala 3 0.1-0.9
T. (H.) reepeni
T. (H.) itama
T. (H.) erythrogastra
T. (H.) canifrons
T. (H.) juscobalteata
T. (H.) fuscibasis
T. (H.) rufibasalis
T. (H.) geissleri 2 0.2-0.5
T. (H.) laeviceps
T. (Homotrigona) fimbriata
62 DAVID ROUBIK
honeybee Apis dorsata ('manyi' or 'lanyeh') and the Malesian honeybee Apis koschevnikovi
('ranyuan'), with the small Apis andreniformis ('nuang'), a bee about 8 mm long, and Apis
cerana ('ranyuan tikung') seemingly less common.
Nests of the honeybees proved elusive, and only one small nest of Apis dorsata was seen
on an individual of Koompassia (Fabaceae). Virtually all bee nests except those of
Lepidotrigona were in or under the bases of large trees, 90 to 180 cm diameter at breast
height (mean = 130 cm). The 33 nests of stingless bees listed in Table 1 were mostly within
a few metres of ground level, and all were associated with living trees. All species but T.
apicalis produced a resin and wax entrance tube that projected from a few cm to a few dozen
cm from the substrate. Three species nested in the ground under large trees: T. collina, T.
drescheri and T. melina. A nest of the last species was in the same tree base as a colony of
the carpenter ant Camponotus gigas. Pariotrigona pendleburyi and Trigona collina were the
most common colonies, amounting to over half of those seen. The former frequently had
two nests separated by a few dozen centimetres in the same tree, and those of the latter were
frequently found in groups of two to four, within root systems. The results of predator
attacks (presumably civets and the sun bear, Selenarctos) were seen at the terrestrial nests
of T. drescheri and T. melanocephala, while unmistakable claw marks of Selenarctos were
seen going up the trunks of five trees, and a cavity in one tree had been chewed open,
through 6 cm of living wood, at 3 m aboveground, exposing a hole 30 x 12 cm - very likely
a former nest of Apis or Trigona.
Salt Selection
Three species were the primary foragers at salt solution feeders - Apis dorsata, A.
koschenikovi and Trigona melina. Sodium salts appeared the most attractive to bees, with
magnesium and potassium also sought to greater extent than phosphate or ammonium (Figure
1). The Trigona were substantially more numerous at magnesium and sodium solutions than
at other mixtures, but there were only slight differences between potassium, ammonium and
sodium for the two honeybees. Individual bees infrequently took full loads of these
solutions, with the exception of A. dorsata, which repeatedly imbibed full loads of sodium
solution. In one instance a worker of Trigona melina fought with one of T. itama by locking
mandibles in a venter-to-venter position at a sodium bait. Combining all of the bee counts
at feeders suggests that social bees were primarily interested in sodium salt (Figure 1).
Nectar Selection
The Malesian honeybee Apis koschevnikovi imbibed the dilute solutions more rapidly than
viscous solutions, but the energy return rate was greatest at the 45 % sugar solution and
possibly at higher concentrations near 60% (see Roubik & Buchmann 1984), as it is for other
honeybees of its size (Figure 2). For comparative purposes, I provide preliminary data,
taken at ambient conditions of 26-31°C and 80% humidity, for the other 'hive bees' of the
world. Three other tropical varieties of Apis mellifera, two from the Yucatan peninsula in
Mexico, where a tropically-adapted honeybee of European descent existed before colonization
by Africanized honeybees, and Japanese Apis cerana - show nearly the same profile of
energy harvest from sugar solutions. The most viscous solutions, 60-65 % sugar, which
provide 3.04 and 3.35 calories per microlitre, respectively, are nonetheless imbibed more
slowly and are less energetically rewarding. The coefficients of variation at different
solutions were between 10 and 26% for the 22 replicates of different solutions (N = 925
bees). However, A. koschevnikovi, with the smallest sample size (see legend for Figure 2),
also had the smallest coefficient of variation, ranging between 11 and 14% for the time
required to imbibe 25 microlitres of each of the four different solutions. Apis
koschevnikovi was unusual in readily taking dilute solutions of 15% sugar, as do African
Potassium
Figure 1. Salt Solution choice by highly social bees of the forest near Belalong field station.
Api3 me/lifera ligustica
Api3 m. tAfrican~ed'
Sugar Concentration
Figure 2. Sugar solution imbibement performance, shown as relative energy (calories) taken per second, by thl
multiple-comb or cavity-nesting Apis. Sample sizes at each concentration were usually equal for a given bee, N =
82, 142, 202, 249 and 250 for Apis koscMvnikovi, A. mellifera scutellata, A. cerana, A. mellifera ligustica x A. m
scutellala, and A. m. ligustica, respectively.
honeybees, A. mellifera scutellata (Figure 2). The other honeybees rarely completed feedin
bouts at these solutions or rejected the solution after close approach.
64 DAVID ROUBIK
DISCUSSION
In terms of biological diversity, temperate and tropical bee communities are difficult to
compare, owing to the predominance of social species in the tropics and their near absence
in the temperate zone (Michener 1979, Roubik 1979, 1989, 1990, 1992). For French Guiana
and Sumatra, in which the total bee fauna in a given forest is approximately 250 and 125
species, respectively (Inoue et al. 1990, Roubik 1990), about 25% of the bees are stingless
bees and honeybees, and 50% are to some degree social. Compared to Mexico and the
United States, where there are 1600 and 3745 recorded bee species, respectively, only 12 %
and 2 % are social, and in Mexico alone there are native highly eusocial species, comprising
3% of the fauna (Ayala 1992, Ayala et al. 1993). The number of genera, however, is 153
in Mexico and 122 in the United States (Ayala et al. 1993). There is thus higher generic and
presumably biological diversity in areas such as Mexico that share tropical and temperate
components. The ecological diversity of the stingless bees and honeybees, which are by
definition perennial colonies, must be greater since they are active every day and have the
opportunity to interact with far more plant species and other resources and natural enemies.
If the lowland forests near Belalong have the same component of honeybees and stingless
bees as those of west Sumatra, approximately 25 %, then the total number of forest bees
should be somewhat higher than 100 species, exactly as it is in west Malesia (Inoue et al.
1990). The number of nests of 3 per hectare was perhaps low compared with forest of
similar stature in the Neotropics, where between four and six nests might be found (Roubik
1989 & in press). The few other prominent bees that I observed in Belalong were the
nocturnal carpenter bee, Xylocopa (Nyctomelitta) myops, Amegilla (Glossamegilla) sp., a
large bee having among the longest tongues of any bee (Kato et al. 1991), and large
megach il ids , which along with Amegilla occasionally visited the salt solutions and laundry
at the Belalong camp. Considering the stingless bees, all three of the species known only
from Borneo (the other 26 are generally distributed from Thailand to Java [Sakagami et al.
1990]) were relatively common in the Belalong forest. These are T. melanocephala, T.
hobbyi and T. haematoptera. The large complement of stingless bees and honeybees suggests
that the bee fauna is intact and rich, and further study of solitary and other native bees would
repay the effort for uncovering the composition and ecology of a lowland tropical forest bee
assemblage.
Of all the species that recruited to baits at Belalong, none was notably aggressive toward
other species and one in particular, Apis koschevnikovi, recruited very rapidly and quickly
depleted baits. One Apis koschevnikovi carried an unidentified orchid pollinarium on its
scutellum. This is the first known instance of an Apis potentially pollinating a tropical forest
orchid, although some neotropical Trigona s. str. regularly visit orchid flowers (Maxillaria
and Xylobium) and carry pollinaria in the same position as seen on Apis (Dodson 1967,
Roubik, unpublished data).
The stingless bees and honeybee that visited the five salt solutions seemed to seek sodium,
magnesium and potassium salts. Sodium, potassium and other ions are physiological
requirements for water retention and neurotransmission (Nicholson 1990), but the role of
magnesium is unclear. One or more of these elements may be deficient in the soil, because
the pollen and nectar ordinarily forming bee diets often reflect the local soil conditions and
nutrients (Kearns & Inouye 1993, Stanley & Linskens 1974). Previous studies with halictid
bees, sometimes called 'sweat bees', had also shown a tendency to take sodium in solution,
rather than pure water (Barrows 1974). The natural sources of such salts in the forest are
both inorganic, where water leaches them from the subsoil or bedrock at river margins, or
from animal feces and urine, particularly of carnivores. Whether any of the chemicals
collected from these non-floral sources are critical in maintaining bee populations or species
richness would be an interesting comparative study with no precedent.
The sugar contained in floral nectars used by bees ranges between 20 and 70 % by weight,
while that of fruit sap utilized by some bees, like the African honeybee in the neotropics,
falls between 10 and 15 % sugar (Roubik, unpublished data). Sugar composition of honey
stored in nests of highly eusocial bees is generally between 70 and 80%, yet the most
common nectars foraged by bees are usually less than 60% sugar, and often half of that
(Roubik 1989). By demonstrating a peak efficiency at nectar of 45% sugar concentration,
the cavity-nesting Apis (Figure 2) colonies would be expected to dominate these resources
and prefer them to others, all other factors being equal. Given the predominance of these
bees in their natural habitats, one might expect resource partitioning to encourage the solitary
bees to seek nectars significantly different from 45 % sugar, and to visit flowers presenting
45 % sugar nectar only when floral morphology, abundance and distribution discourages
recruitment by Apis. It is generally unknown if other components of nectar are favoured or
rejected by some bees but not others. A study of trees and other flowering plants dependent
on bees for outcrossing, which appears to be the predominant breeding system in tropical
trees (Bawa 1990, 1992), might readily reveal whether or not they utilize these three Apis
species. A plant providing 45% nectar, especially if nectar sugar increases during the day
(Roubik & Buchmann 1984) would display a behaviour that would encourage visitation by
Apis mellifera, cerana or koschevnikovi. Other patterns and concentrations of nectar would
likely favour other pollinators, although the comparative study of energy harvest rate by the
non-social bees that form the bulk of tropical species still awaits investigation.
ACKNOWLEDGEMENTS
I thank Marina Wong, Tamiji Inoue, Peter Becker, Gathorne Cranbrook, Catriona Prebble,
David Edwards and collaborators at Universiti Brunei Darussalam and Royal Geographic
Society in general for making this work possible. Special thanks are given to T. Inoue and
R. Villanueva, and for funding from Smithsonian Scholarly Studies grants for assistance in
studies of nectar use by Apis.
LITERATURE CITED
AYALA, R. 1992. Revision de las abejas sin aguijon de Mexico (Hymenoptera: Apidae: Me/iponinae). Thesis.
National Autonomous University of Mexico, Mexico City. 73 pp.
AYALA, R., GRISWOLD, T.L. & BILLOCK, S.H. 1993. The native bees of Mexico. pp. 179-228 in
Ramamoorthy, T.P., Bye, R., Lot, A. & Fa, H. (eds). Biological diversity in Mexico: Origins and distribution.
Oxford University Press, New York.
BARROWS, E. M. 1974. Aggregation behavior and response to sodium chloride in females of a solitary bee,
Augochlora pura (Hymenoptera: Halictidae). Florida Entonwlogist 57:189-193.
BAWA, K. S. 1990. Plant-pollinatior interactions in tropical rain forests. Annual Review of Ecology and Systematics
21 :399-422.
BA WA, K. S. 1992. Mating systems, genetic differentiation and speciation in tropical rain forest plants. Biotropica
24:250-255.
BECCARI, O. 1904. Wanderings in the great forests of Borneo. Archibald and Constable and Co., London
[reprinted by Oxford University Press, Singapore, 1989). 346 pp.
DODSON, C. H. 1967. Relationships between pollinators and orchid flowers. Atlas do Simposio Biota Amazonica
5 (Zoologia):I-n.
INOUE, T., SALMAH, S., SAKAGAMI, S.F., YAMANE, S.K. & KATO, M. 1990. An analysis of antho philous
insects in central Sumatra. Pp. 201-218 in Sakagami, S.F., Ohgushi, R. & Roubik, D.W. (eds). Natural history
of social bees and wasps in equatorial Sumatra. Hokkaido University Press, Sapporo, Japan.
KATO, M., ITINO, T., HOTTA, & INOUE, T. 1991. Pollination of four Sumatran Impatiens by hawkmoths and
bees. Tropics 1:59-73.
KEARNS, C. A. & INOUE, D.W. 1993. Techniques for pollination biologists. University Press of Colorado,
Boulder. 568 pp.
66 DA YID ROUBIK
MICHENER, C. D. 1979. Biogeography of the bees. Annals of the Missouri Botanical Garden 66:277-347.
MICHENER, C. D. 1990. Classification of the Apidae. University of Kansas Science Bulletin 54:75-164.
NICHOLSON, S. W. 1990. Osmoregulation in a nectar-feeding insect, the carpenter bee Xylocopa capitata: water
excess and ion concentration. Physiological Entomology 15:433-440.
ROUBIK, D. W. 1979. Africanized honey bees, stingless bees, and the structure of tropical plant-pollinator
communities. pp. 403-417 in Caron, D. (ed.). Proceedings N lntemational Symposium on Pollination, Marylond.
Agric. Exp. Sta. Misc. Publ. 1.
ROUBIK, D. W. 1989. Ecology and natural history of tropical bees. Cambridge University Press, New York. 514
pp.
ROUBIK, D., W. 1990. Niche pre-emption in tropical bee communities: A comparison of neotropical and malesian
faunas. Pp. 245-257 in Sakagami, S.F., Ohgushi, R. & Roubik, D.W. (eds). Natural history of social bees and
wasps in equatorial Sumatra. Hokkaido University Press, Sapporo, Japan.
ROUBIK, D. W. 1992. Loose niches in tropical communities: Why are there so many trees and so few bees? pp.
327-353 in Hunter, M.D., Ohgushi, T. & Price, P.W. (eds). Resource Distribution and Animal-Piont
lnteractions. Academic Press, San Diego.
ROUBIK, D. W. 1993. Direct costs of forest reproduction: bee-cycling and the efficiency of pollination modes.
Journal of Biosciences 18:1-16.
ROUBIK, D. W. & BUCHMANN, S.L. 1984. Nectar selection by Melipona and Apis mellifera (Hymenoptera:
Apidae) and the ecology of nectar intake by bee colonies in a tropical forest. Oecologia (Berlin) 61:1-10.
SAKAGAMI, S. F., INOUE, T. & SALMAH, S. 1990. Stingless bees of central Sumatra. pp. 125-137 in
Sakagami, S.F., Ohgushi, R. & Roubik, D.W. (eds). Natural history of social bees and wasps in equatorial
Sumatra. Hokkaido University Press, Sapporo, Japan.
STANLEY, R. G. & LINSKENS, H.F. 1974. Pollen: Biology, biochemistry, management. Springer-Verlag, New
York. 307 pp.
WILLE, A. 1962. A technique for collecting stingless bees under jungle conditions. lnsectes Sociaux 9:291-293.
WILLIAMS, P. H. 1993. Choosing conservation areas: using taxonomy to measure more of biodiversity. Pp.
194-227 in Moon T.-Y. (ed.). Colloquium on Biodiversity. Korean Entomological Institute, Seoul.
Australian monsoon rainforest mammals and possible implications for the
conservation of tropical rainforest biodiversity
D.M.I.S. BOWMAN and I.C.Z. WOINARSKI
Conservation Commission of the Northern Territory, PO Box 496, Palmerston, Northern

Territory, 0831 Australia
ABSTRACT. Monsoon rainforests form an archipelago of small habitat fragments throughout the wet-dry
tropics of northern Australia. According to one definition the current monsoon rainforest mammal
assemblage contains only one rainforest specialist mammal species (restricted to Cape York Peninsula),
and is dominated by eutherian habitat generalists (murids and bats) that mostly occur in surrounding
savanna habitats. The mammal assemblages in monsoon rainforests across northern Australia (Cape York
Peninsula, Northern Territory and the Kimberley) are essentially regional subsets of the local savanna and
mangrove mammal assemblages, and consequently share only a limited number of species in common
(most of which are bats). The lack of rainforest specialists in north western Australia is believed to be
due to lack of large tracts (> I ()()() ha) of monsoon rainforest habitat, the possible substantial contraction
of these habitats in the past and the limited extent of gallery rainforests: gallery rainforests in South
American savannas are important habitats for rainforest mammals. Unfortunately it is not possible to
identify the threshold of habitat area required to maintain populations of rainforest specialist mammal
species because of an impoverished fossil record pertaining to the past spatial distribution of monsoon
rainforests. The implications of the lack of Australian monsoon rainforest specialist mammal fauna for
the future of heavily fragmented tropical rainforests elsewhere in the world is briefly discussed. It is
concluded that the analogy of habitat fragments to true islands is weak, that rainforest plant species are
less vulnerable to local extinction than mammals, and that the loss of mammal rainforest specialists may
not result in a dramatic loss of plant species.
KEY WORDS: biogeography, conservation, fragmentation, mammals, northern Australia, rainforests.
INTRODUCTION
It is widely accepted that the destruction of extensive rainforests will result in an archipelago
of rainforest 'islands' within a 'sea' of anthropogenically modified vegetation. The
consequences for biota specialized to the rainforest environment are uncertain (Simberloff
1990). Meave et al. (1991) have recently addressed this problem by analyzing tree species
diversity in naturally created rainforest habitat islands that occur in large tracts of neotropical
savanna.
Australian monsoon rainforests form an archipelago of small (typically less than 5 ha in
extent) habitat fragments throughout the wet-dry tropics of northern Australia (Russell-Smith
1991). Given recent zoogeographic research these 'habitat islands' provide an opportunity
to consider the impact of anthropogenic clearance upon rainforest mammal specialists.
ECOLOGICAL BIOGEOGRAPHY OF AUSTRALIAN MONSOON

RAINFOREST MAMMALS
The mammal assemblages across northern Australia (Cape York Peninsula, Northern
Territory and the Kimberley) are dominated by eutherian habitat generalists (murids and bats)
and are essentially regional subsets of the local savanna and mangrove mammal assemblages
67
68 D.M.J.S. BOWMAN AND J.C.Z. WOINARSKI
(Friend et al. 1991, Menkhorst & Woinarski 1992). There are only a few species (most of
which are bats) that occur throughout northern Australia. The current monsoon rainforest
mammal assemblage contains only one mammal species (Rattus leucopus) with exclusive
dependence on rainforest habitats (Winter 1988, Winter & Atherton 1985).
Menkhorst & Woinarski (1992) suggest that the following features have prohibited
colonization of north west Australian monsoon rainforests by non-volant mammal species that
occur in the monsoon rainforests on Cape York Peninsula: the small area of monsoon
rainforest and the typically large distances between patches; the extreme microclimatic
differences between savanna and rainforest habitats (particularly in the dry season); the
scarcity of gallery rainforests; and the existence of a major biogeographic barrier (the Gulf
of Carpentaria and the Barkly Tablelands) that separates north western Australia from large
tracts of tropical rainforest in eastern Australia.
In marked contrast to South American tropical rainforest mammal assemblages the
majority of the Australian monsoon rainforest mammal assemblage have body weights below
0.1 kg (Redford 1992). In the South American seasonally dry tropics there are few mammals
adapted to savanna conditions even though savanna dominates over 90% of the landscape
(Mares et al. 1985, Redford & Da Fonseca 1986). Most of the mammals are restricted to
gallery rainforests, which act as habitat corridors between Amazon and Atlantic rainforests
(Redford & Da Fonseca 1986). Gallery rainforests are very restricted in northern Australia
because the characteristically Australian genus Melaleuca dominates virtually all riverine
habitats.
HISTORICAL BIOGEOGRAPHY OF AUSTRALIAN MONSOON

RAINFOREST MAMMALS
Until the mid-Tertiary northern Australia supported a rich marsupial rainforest fauna (Archer
et al. 1989). Currently Australian mammal species adapted to rainforest habitats have
populations restricted to large tracts of rainforest in north eastern Australia (Winter 1988).
It is assumed that the lack of rainforest specialists in south eastern Australia and north
western Australia is due to reduction in the area of rainforest habitat in the past (Winter
1988). Localized extinction following recent European habitat destruction in the core area
of Australian tropical rainforest mammals demonstrates the vulnerability of rainforest
specialists to habitat fragmentation (Lawrence 1990). Patches less than 3000 ha in extent
appear to be too small to support rainforest specialists, although mammal populations have
probably not equilibrated to this latest episode of rainforest fragmentation. In this context
it is significant that the largest tracts of monsoon forest in northern Australia are less than
1000 ha.
IMPLICATIONS FOR THE CONSERVATION OF TROPICAL

RAINFOREST BIODIVERSITY
Given an extremely depauperate paleao-environmental record it is impossible to evaluate the

hypothesis that past patch areas have fallen below a critical threshold to support monsoon
rainforest specialists (cf. Winter 1988), or indeed if specialist mammals ever evolved to
exploit large tracts of monsoon rainforest. Similarly, although both inter-patch distance and
distance from core refugia appear to be important factors controlling the richness of monsoon
rainforest mammal assemblages, uncertainty concerning the nature of past landscapes,
including the extent of gallery forests, prohibits simple quantification of these variables.
The high diversity of monsoon rainforest plants and the low diversity of specialist
rainforest mammals (following their mass extinction in the mid-Tertiary?) suggests that
specialist mammals are not necessarily critical for the functioning of all rainforest ecosystems
Australian rainforest mammals 69
(cf. Da Fonseca & Robinson 1990, Dirzo & Miranda 1990, Howe 1984, Terborgh 1992).
It appears that the lack of rainforest mammal specialists is compensated for by vagile winged
vertebrates that exploit fruits and flowers from rainforest trees. Birds, frugivorous and
nectarivorous bat species apparently play the most ecologically important role in the
maintenance of the diversity of woody plants through their dispersal of pollen and seeds
between isolated populations of plant species (Russell-Smith & Lee 1992).
The lack of rainforest specialists does not mean that the monsoon rainforests are of little
significance to the regional mammal assemblage. The monsoon rainforests are
opportunistically utilized by savanna adapted mammals in order to avoid climatic extremes,
predators, and ubiquitous savanna fires. The use of monsoon rainforests by savanna animals
supports Janzen's (1983) notion that "no park is an island" because small patches of
rainforest will be exploited by biota in the surrounding habitat. North Australian monsoon
rainforest biogeography also supports the notion that rainforest specialist vertebrates are far
more vulnerable to extinction than plant species (Meave et al. 1991, Redford 1992). In this
context the "empty" (sensu Redford 1992) north Australian monsoon rainforest fragments
may be a window into the future of the world's tropical forests.
LITERATURE CITED
ARCHER, M., GODHELP, H., HAND, S.L & MEGIRIAN, D. 1989. Fossil mammals of Riversleigh,
Northwestern Queensland: Preliminary overview of biostratigraphy, correlation and environmental change.
Australian Zoologist 25:29-65.
DA FONSECA, G.B.A. & ROBINSON, J.G. 1990. Forest size and structure: competitive and predatory effects
on small mammal communities. Biological Conservation 53:265-294.
DIRZO, R. & MIRANDA, A. 1990. Contemporary Neotropical defaunation and forest structure, function, and
diversity - A sequel to John Terborgh. Conservation Biology 4:444-447.
FRIEND, G.R., MORRIS, K.D., & McKENZIE, N.L. 1991. The mammal fauna of Kimberley rainforests. Pp.
393-412 in McKenzie, N.L., Johnson, R.B. & Kendrick P.G. (eds). Kimberley Rainforests of Australia. Surrey
Beattie and Sons, Sydney.
HOWE, H.F. 1984. hnplications of seed dispersal by animals for tropical reserve management. Biological
Conservation 30:261-281.
JANZEN, D.H. 1983. No park is an island: increase in interference from outside as park size decreases. Oikos
41 :402-410.
LAURANCE, W.F. 1990. Comparative responses of five arboreal marsupials to tropical forest fragmentation.
Journal of MamJlUllogy 71:641-653.
MARES, M.A., WILLIG, M.R. & LACHER, T.E. 1985. The Brazilian Caatinga in South American
zoogeography: tropical mammals in a dry region. Journal of Biogeography 12:57-69.
MEAVE, J., KELLMAN, M., McDOUGALL, A. & ROSALES, J. 1991. Riparian habitats as tropical forest
refugia. Global Ecology and Biogeography Letters 1:69-76.
MENKHORST, K.A. & WOINARSKI, LC.Z. 1992. Distribution of mammals in monsoon rainforest of the
Northern Territory. Wildlife Research 19:295-316.
REDFORD, K.H. 1992. The empty forest. BioScience 42:412-422.
REDFORD, K.H. & Da FONSECA, G.A.B. 1986. The role of gallery forests in the zoogeography of Cerrado's
non-volant mammalian fauna. Biotropica 18:126-135.
RUSSELL-SMITH, J. 1991. Classification, species richness, and environmental relations of monsoon rainforest
vegetation in the Northern Territory, Australia. Journal of Vegetation Science 2:259-278.
RUSSELL-SMITH, J. & LEE, A. 1992. Plant populations and monsoon rain forest in the Northern Territory,
Australia. Biotropica 24:471-487.
SIMBERLOFF, D. 1990. Species-area relationships, fragmentation and extinction in tropical forests. In Kheong,
S. & Win L. (eds). In Harmony with Nature. Proceedings of the International Conference on Tropical
Biodiversity. Malayan Nature Society, Kuala Lumpur.
TERBORGH, J. 1992. Maintenance of diversity in tropical forests. Biotropica 24:283-292.
WINTER, LW. 1988. Ecological specialization of mammals in Australian tropical and sub-tropical rainforest:
refugial or ecological determinism. Proceedings of the Ecological Society of Australia 15:127-138.
WINTER, J.W. & ATHERTON, R.G. 1985. Survey of the lnamJlUlls and other vertebrates of the Weipa region.
Cape York Peninsula. Queensland National Parks and Wildlife Service, Brisbane.
Presbytis species sympatry in Borneo versus allopatry in Sumatra: an
interpretation.
DOUGLAS BRANDON-JONES
11 York Road, London SW11 3PX, United Kingdom
ABSTRACT. A review of the diversity of opinion on the nwnber of Swnatran species recognizable for
the genus Presbytis sensu stricto, indicates that an over-reliance on sympatry as a criterion for specific
status, produces an underestimation of the species nwnber in Swnatra, compared with Borneo. It is
advocated that, for reasons of consistency, geographic variation in pelage colour must retain senior
taxonomic status over geographic variation in adult male loud vocalization, but that the latter character
corroborates the conspecificity rather than convergence of the Borneanand Swnatrangrey-backed species,
Presbytis comata. It is inferred that Presbytis potenziani was the sole species of its genus to survive an
early stadial of the most recent glaciation, and that Presbytis comata derived from it during the
subsequent interstadial and was approximately restricted to its modern distribution by a milder second
stadia!. The remaining species of the genus have derived from Presbytis comata since the glaciation.
Allopatry in Swnatra is attributed to the northern part of the island forming a more active post-glacial
dispersal centre than the comparable area in north Borneo. Only one Bornean species, Presbytis frontata
has derived from Presbytis comata in Borneo. The other two, Presbytis femoralis and Presbytis
rubicunda, had their origins in Swnatra.
KEY WORDS: dispersal, geographic variation, glaciation, pelage colour, Presbytis melalophos species
group, vocalization.
The Asian representatives of the Old World monkey subfamily, Colobinae, consist of the
proboscis monkey, the snub-nosed monkeys, the douc monkeys and twenty or so simple-
nosed species (Brandon-Jones 1984). Most authors refer the simple-nosed species to a single
genus, Presbytis, rejecting the opinion of Pocock (1935) that they constitute three genera.
A growing body of evidence indicates that at least a bipartite generic division is warranted,
and that the genus, Presbytis sensu stricto should include the Mentawai Island species,
Presbytis potenziani (Brandon-Jones 1993). Bornean populations of this genus overlap each
other in geographic distribution. Evidence of sympatry in Sumatra derives solely from
speculation prompted by the absence of a river boundary between two distinct populations
in the central part of the island (Borner 1974). This paper seeks to highlight the resulting
imbalance in the number of species recognized for the two islands, and to propose an
explanation for the zoogeographic contrast.
Miller (1934), dealing only with the "Presbytisjemoralis group", recognized four species
from Sumatra, Presbytis batuana, P. catemana, P. percura and P. sumatrana. Pocock
(1935) segregated "Presbytis thomasi", but treated the remaining populations as respectively
"blackish races", a "reddish race", or "whitish mutants" of "Presbytis melalophus". Chasen
(1940), although acknowledging its earlier authorship, idiosyncratically treated Simia
melalophos Raffles 1821, as a subspecies of "Pithecus jemoralis". His recognition of
"Pithecus jemoralis jusco-murinus" [=Presbytis melalophos mitrata}, and his description of
a new subspecies, "Pithecus jemoralis jZuviatilis", implied rejection of Pocock's (1935)
opinion that they were whitish mutants of no taxonomic standing. Chasen (1940)
innovatively recognized "Pithecus aygula thomasi" as conspecific with populations in Jawa
and north Borneo, and showed remarkable insight in suggesting that, "taking an extremely
broad view," Presbytis potenziani might "be regarded as a very distinct race" of the same
71
72 DOUGLAS BRANDON-JONES
species which, for nomenclatural reasons, must now be referred to as "Presbytis comata"
(Napier 1985, Napier & Groves 1983, Weitzel & Groves 1985). This relationship,
appreciated only by Thomas (1895) in papers predating Chasen's, is corroborated by more
recent research (Brandon-Jones 1993).
Hooijer (1962) and Kuhn (1967) recognized two Sumatran species, Presbytis aygula and
P. melalophos. Medway (1970) considered Pocock's (1935) more conservative
classification, implying convergence between the grey-backed populations in northwest
Sumatra, Java and north and east Borneo, as more acceptable than Chasen's (1940) indication
that Presbytis comata is disjunct in its present distribution and represents the initial colonizer
of the Malay archipelago. Medway (1970) offered three alternative classifications. The
most radical would recognize a single species (Presby tis melalophos) inhabiting Java,
Sumatra, the Malay peninsula and west Borneo. The more orthodox would exclude Java
from this distribution. The most cautious would also recognize P. thomasi as a distinct
species in north Sumatra. Thorington & Groves (1970) opted for his most cautious
interpretation.
Their extensive field survey of the primates of Sumatra led W.L. Wilson & C.C. Wilson
(1975) to conclude that geographic variation in the adult male loud vocalization in Sumatran
Presbytis (sensu stricto), differentiated it into three species, P. melalophos, P. femoralis and
P. thomasi. C.C. Wilson & W.L. Wilson (1977) reported that the adult male three-part call
of Presbytis potenziani "is as similar to that of P. thomasi as are the calls of the P.
melalophos subspecies to one another. In fact, the Mentawai Island langur may prove to be
a subspecies of P. thomasi" (P. potenziani is actually the earlier name). Their separate
report (C.C. Wilson & W.L. Wilson 1975) that, based on vocalization, "Presby tis aygula"
in east Kalimantan may have "Po thomasi" as its "nearest relative in Sumatra", corroborates
Chasen's (1940) opinion that the distribution of Presbytis comata is disjunct, with P.
potenziani possibly "as a very distinct race". Convergence both in pelage colour and
vocalization between P. potenziani and two geographically isolated subspecies of P. comata
would seem improbable.
Kawamura (1984) and Aimi et al. (1986) conducted further field surveys of Sumatran
Presbytis, and endorsed the Wilsons' conclusion that there are three species. Aimi & Bakar
(1992) claimed that their taxonomic arrangement was substantiated by the cranial morphology
of specimens they collected in 1981 and 1983.
The classification adopted by Brandon-Jones (1984) was based on the appreciation that,
for reasons of consistency, geographic variation in vocalization must remain subordinate in
taxonomic status to geographic variation in pelage colour, the principal character on which
Presbytis specific status has been allocated. Brandon-Jones (1984) reinterpreted as a natural
group the cluster of pale-thighed populations in the southern Malay peninsula, Natuna Besar,
the Riau archipelago and east central Sumatra, interpreted by Medway (1970) as convergent.
The greyish white pelage colour of the thigh is exceptionally well developed in the east
Sumatran subspecies, Pithecus femoraUs paenulatus Chasen, 1940, which occurs in close
proximity to surrounding dark-thighed populations. Unlike comparable populations in the
Malay peninsula, there is no indication of subspecific pelage colour intergradation. It is this
subspecies which is subject to speculation as to its sympatry with the adjacent species.
Brandon-Jones (1984) accordingly recognized four species from Sumatra: the dorsally
grizzled-grey species, Presbytis comata; the blackish brown species, Presbytisfemoralis; the
pale-thighed species, Presbytis siamensis; and the reddish and whitish species, P.
melalophos.
These Sumatran popUlations, which Medway's (1970) radical and more orthodox
classifications would reduce to one species, exceed all four unanimously recognized Bornean
species together in pelage colour variation. The dorsally grizzled-grey populations in north
and east Borneo are conspecific with, or convergent to, the similar populations in north
Presby tis distribution in Borneo and Sumatra 73
P. siamensis 0
P. potenziani
P. comata
P.e. sabana
*6o P.me/a/ophos nobi/is
P. m. me/a/ophos
\1
•
••
P. femoralis P. m. mitrata and <)
P.fredericae P. m. f/uviatilis
P. frontata ... P. rubieunda
P. r. chrysea
1:3:
*
Figure 1. Geographic distribution of Presbytis (sensu stricto).
Sumatra. Even Pocock (1935) who granted it specific status, acknowledged that "Presby tis
chrysomelas" , inhabiting the northwest coastal region of Borneo, was "perhaps only racially
distinct from P. jemoralis". This view is now generally accepted. P. jemoralis and the
species, Presbytis jrontata, endemic to east, southeast and central Borneo, are superficially
so similar in pelage colour that museum skins have been subject to mistaken identity (e.g.
Jentink 1898). P. rubicunda, which inhabits Karimata island and peripheral Borneo other
than most of the area occupied by P. jemoralis, is the counterpart of P. melalophos both in
reddish pelage colour and, according to C.C. Wilson & W.L. Wilson (1975), in some
aspects of behaviour and vocalizations. There is no Bornean equivalent of P. siamensis, nor
of the southeast Sumatran whitish subspecies. Conversely, there is no Sumatran counterpart
of the black, red and white pelage-coloured morph which has a high incidence in the eastern
populations of Bornean P. jemoralis.
Its geographic distribution leaves little doubt that P. jemoralis originated in Sumatra and
migrated to Borneo. Similarities in external characters and vocalization, and evidence of
intergradation indicate that P. melalophos derived from P. jemoralis. If it is possible for a
predominantly reddish species to derive from a predominantly brownish species in Sumatra,
then it is equally plausible for a similar species to have evolved in transit to Borneo. The
similarities in vocalization between P. rubicunda and P. melalophos reported by the Wilsons
have already been mentioned, but similarities in head hair disposition suggest that P.
rubicunda actually originated from east Sumatran populations of P. siamensis.
On the face of it, the origins of P. jrontata seem more obscure. Comparison reveals that
the distinctive external cephalic features in P. jrontata are only an exaggerated development
of the similar features exhibited by Presbytis comata sabana. The two taxa also share a
characteristic elongation of the nasal region of the skull. Their complementary geographic
distribution and their pelage colour differentiation, show a remarkable parallel both to that
between P. comata and P. jemoralis in Sumatra, and between P. comata and Presbytis
jredericae in Java.
Brandon-Jones (1978, & in preparation) has recounted how Asian colobine geographic
distribution, supported by other biogeographic evidence, indicates that the Asian rainforest
was fragmented by the cool and dry climate which accompanied the most recent glaciation.
The approximate extent of three of these fragments is inferred from the modern geographic
distribution of P. comata. The squirrel, Callosciurus and the gibbon, Hylobates, are both
represented by a subspecies endemic to the area of north Sumatra where rainforest survived
a period of glacial drought, but Macaca and Semnopithecus are not so represented. This
implies that the latter genera were absent from Sumatra when that cold dry period began.
They must however, have occurred there at some earlier date in order for the macaque,
Macaca pagensis, to have colonized the Mentawai Islands, and the leaf monkey,
Semnopithecus auratus, to have reached Java. The absence of a macaque and Semnopithecus
endemic to north Sumatra therefore indicates the intervention of an earlier and more severe
arid period. This earlier glacial drought was presumably also responsible for the eradication
of the significant number of animals found on the Asian mainland and Java, but not on
Sumatra (Dammerman 1929). It is deduced that Callosciurus, Presbytis, Hylobates and the
orang-utan, Pongo, recolonized Sumatra after this drought and that their present distribution
is the result of a second and less severe climatic deterioration.
The simplest explanation for the distinctiveness of Presbytis potenziani is that it was the
sole member of its genus to survive the first arid period. The absence of P. potenziani on
Borneo implies that the genus had not colonized Borneo prior to the glaciation. During the
interstadial, P. potenziani gave rise to P. comata which migrated at least as far as Sumatra,
Java and Borneo. The second cool and dry climatic period contracted the geographic
distribution of P. comata probably to only the central areas of its modern distribution. With
climatic remission, its distribution expanded initially to occupy the peripheral areas of its
modern distribution. It then transmuted into P. jemoralis, P. jredericae and P. jrontata
respectively, in Sumatra, Java and Borneo. P. jemoralis gave rise to P. melalophos and P.
siamensis and, through the latter, to P. rubicunda. This analysis would allow of the
interpretation that Presbytis is a monotypic genus, were it not for the above discussed
instances of sympatric distribution. The species display a chromatic succession in pelage and
in some cases, skin pigmentation, comparable with those identified by Hershkovitz (1968,
1977).
Presbytis species diversity then, far from being the product of prolonged stability of the
humid rainforest environment, is actually the result of its recent fragmentation. Its allopatric
species distribution in Sumatra resulted from north Sumatra forming a recent centre of
population dispersal, predominantly in a west to east direction, with one branch (P.
melalophos) dispersing as far as the southern tip of the island, and another (P. jemoralis)
emigrating to the Malay peninsula and northwest Borneo. The only possible instance of
sympatry resulted from a reverse dispersal of P. siamensis from the Malay peninsula.
Presbytis sympatry in Borneo largely resulted from the greater productivity of the Sumatran
dispersal centre, probably assisted by an inferred eastward ocean current between Sumatra
Presbytis distribution in Borneo and Sumatra 75
and Borneo which favoured rafting in this direction. This colonization must have occurred
by rafting because central Sundaland would have been climatically unable to sustain the
requisite Presbytis habitat during its glacial emergence. Although Presbytis comata canicrus
has probably supplanted part of the distribution of P. Jrontata in northeast Kalimantan, in
general there has been decidedly less dispersal activity by indigenous Bornean Presbytis,
possibly because expansion was blocked by the lingering of the cool dry climate for a longer
period in central Borneo than occurred in central Sumatra.
Note added in proof: Fifteen kilometres south-west of Prapat near Lake Toba, during a visit
to Sumatra in November-December 1993,1. Vermeer (in litt.) observed a troop of Presbytis
siamensis paentulata whose identity was confirmed by the loud call. The locality is about
150 km north of the Barumun river, previously regarded as the northern limit of this sub-
species (Aimi & Bakar 1992). It is difficult to interpret this evidence other than indicating
significant sympatry between P. Jemoralis and P. siamensis.
LITERATURE CITED
AIMI, M. & BAKAR, A. 1992. Taxonomy and distribution of Presbytis melalophos group in Sumatera, Indonesia.
Primntes 33:191-206.
AIMI, M., HARDJASASMITA, H. S., SJARMIDI, A. & YURI, D. 1986. Geographical distribution of Aygula-
group of the genus Presby tis in Sumatra. Kyoto University Overseas Report of Studies on Asian Non-Human
Primntes 5:45-58.
BORNER, M. 1974. Progress report no. 4: The Sumntran rhinoceros (Dicerorhinus sumntrensis) in the provinces
of Riau and West Sumntera. World Wildlife Fund Project 884 - Sumatran Rhinoceros. 11 pp.
BRANDON-JONES, D. 1978. The evolution of recent Asian Colobinae. pp. 323-325 in Chivers, D. J. & Joysey,
K. A. (eds). Recent Advances in Primntology, Vol. 3: Primnte Evolution, Proceedings of the Sixth Congress of
the International Primntological Society, 1976. Academic Press, London.
BRANDON-JONES, D. 1984. Colobus and leaf monkeys. pp. 398-408 in Macdonald, D. (ed.). The Encyclopaedia
of Mammals Volume 1. Allen & Unwin, London. Pp. 102-112 in Macdonald, D. (ed.). All the World's Animnls:
Primntes. Torstar Books, New York.
BRANDON-JONES, D. 1993. The taxonomic affinities of the Mentawai Islands sureli, Presbytis potenziani
(Bonaparte 1856). Raffles Bulletin of Zoology 41:331-357.
CHASEN, F. N. 1940. A handlist of Malaysian mammals. Bulletin of the Raffles Museum, Singapore 15:i-xx, 1-
209.
DAMMERMAN, K. W. 1929. On the zoogeography of Java. Treubia 11:1-88.
HERSHKOVITZ, P. 1968. Metachromism or the principle of evolutionary change in mammalian tegumentary
colors. Evolution 22:556-575.
HERSHKOVITZ, P. 1977. Living New World Monkeys (Platyrrhini): with an Introduction to Primntes. Volume
1. University of Chicago Press, Chicago
HOOUER, D. A. 1962. Quaternary langurs and macaques from the Malay Archipelago. Zoologische Verhandelingen
uitgegeven door het Rijkmuseum van Natuurlijke Historie te Leiden 55: 1-64.
JENTINK, F. A. 1898. Zoological results of the Dutch scientific expedition to central Borneo. The mammals.
Notesfrom the Leyden Museum 20:113-125.
KAWAMURA, S. 1984. Distribution and vocalization of Presbytis melalophos and P. femoralis varieties in
westcentral Sumatra - a summarized report. Kyoto University Overseas Report of Studies on Asian Non-Human
Primntes 3:37-44.
KUHN, H.-J. 1967. Zur Systematik der Cercopithecidae. Pp. 25-46 in Starck, D., Schneider, R. & Kuhn, H.-J.
(eds). Progress in Primatology. First Congress of the International Primatological Society, Frankfurt am Main,
1966. Fischer, Stuttgart.
MEDWAY, LORD. 1970. The monkeys of Sundaland: Ecology and systematics of the cercopithecids of a humid
equatorial environment. Pp. 513-553 in Napier, J. R. & Napier, P. H. (eds). Old World Monkeys: Evolution,
Systematics and Behavior. Academic Press, New York.
MILLER, G. S., Jr. 1934. The langurs of the Presbytis femoralis group. Journal of Mammalogy 15:124-137.
NAPIER, P. H. 1985. Catalogue of Primates in the British Museum (Natural History) and elsewhere in the British
Isles. Part III: Family Cercopithecidae, Subfamily Colobinae. British Museum (Natural History), London.
NAPIER, P. H. & GROVES, C. P. 1983. Simiafascicularis Raffles, 1821 (Mammalia, Primates): request for the
suppression under the plenary powers of Simia aygula Linnaeus, 1758, a senior synonym. Z.N. (S.) 2399.
Bulletin of Zoological Nomenclature 40:117-118.
POCOCK, R. L 1935. The monkeys of the genera Pithecus (or Presbytis) and Pygathrix found to the east of the
Bay of BengaL Proceedings of the Zoological Society of London 1934:895-961.
TIIOMAS, O. 1895. On some mammals collected by Dr. E. Modigliani in Sipora, Mentawei Islands. Annali del
Museo Civico di Storia Naturale di Genova (2)14:660-672
TIIORINGTON, R. W., JR. & GROVES, C. P. 1970. An annotated classification of the Cercopithecoidea. Pp.
629-647 in Napier, 1. R. & Napier, P. H. (eds). Old World Monkeys: Evolution, SysteTlUltics and Behavior.
Academic Press, New York.
WEITZEL, V. & GROVES, c. P. 1985. The nomenclature and taxonomy of the colobine monkeys of Java.
International Journal of PriTlUltology 6:399-409.
WILSON, C. C. & WILSON, W. L. 1975. The influence of selective logging on primates and some other animals
in East Kalimantan. Folia PriTlUltologica 23:245-274.
WILSON, C. C. & WILSON, W. L. 1977. Behavioral and morphological variation among primate populations in
Sumatra. Yearbook of Physical Anthropology 20:207-233
WILSON, W. L. & WILSON, C. C. 1975. Species-specific vocalizations and the determination of phylogenetic
affinities of the Presbytis aygula-melalophos group in Sumatra. pp. 459-463 in Kondo, S., Kawai, M. & Ehara,
A. (eds). Contemporary PriTlUltology. Fifth International Congress of PriTlUltology, Nagoya 1974. Karger, Basel.
Diversity of sound-producing insects in a Bornean lowland rain forest
KLAUS RIEDE
Institut fur Biologie I (Zoologie), Albert-Ludwigs Universitiit Freiburg, Germany
ABSTRACT. In the canopy of a lowland mixed dipterocarp forest in Mount Kinabalu Park (Sabah,
Malaysia), animal sounds were recorded for several weeks during the dry and beginning of the rainy
season. Spectrogram analysis revealed clearly identifiable sound patterns which in most cases could be
classified at the order level (crickets, cicadas, katydids and vertebrate sounds) and attributed to certain
'ethospecies' which in several cases could be caught and identified. Song activity of cicadas and birds
was diurnal, while crickets started singing at dusk, followed by strictly nocturnal katydids. This indicates
pronounced temporal segregation at the order level by diel periodicity. During sunset, overall sound
intensity increases due to several species engaging in a 'dusk chorus' composed of chorusing cicadas,
crickets and frogs. A distinct 'acoustical guild' is observed at waterfalls, showing special adaptations of
call structure to high background noise. The value of acoustical records as a tool to assess local and
regional diversity is discussed.
KEY WORDS: bioacoustics, insects, Orthoptera, Cicadidae, sound recording, Sabah, Kinabalu N.P.
INTRODUCTION
Sound is an important medium for intermediate-range communication within rainforests,

where optical communication is hampered by restricted visibility due to vegetation and the
nocturnal life style of most rainforest creatures (Elton 1973). Consequently, as had been
noted by 19th century naturalists (Wallace 1869), among the most exquisite sensory
impressions when entering a tropical rainforest are the concertos produced by a great variety
of songsters. The general notion of increased species numbers in the tropics holds also for
those groups capable of sound production, and species-specific songs might well be an
efficient way of mate location and species recognition for species of low population density
within such complex habitats as rainforests. On the other hand, acoustic communication is
limited by abiotic and biotic constraints: habitat structure affects sound propagation (Romer
et al. 1989), and noise generated by abiotic sound sources like wind, rain and waterfalls,
impairs efficient communication. Among the biotic constraints is the sound produced by con-
and heterospecific animals. Therefore, it can be expected that signals will be shaped by
competition for acoustic transmission channels. The aim of this paper is to sketch the
composition of the 'acoustic commmunity' in a Bornean mixed dipterocarp forest and to
outline a framework for detailed ecophysiological studies of singing animals.
Study sites and periods

Investigations were made at Kinabalu National Park, Sabah, Malaysia, in a lowland mixed
Dipterocarp forest at Poring Hot Springs (500 m asl), along Langanan river trail and
Langanan waterfall (1100 m asl), and in mountain forest around the Park Headquarters
(1200m asl). A canopy walkway system with a total length of 900 m was available at
Poring. Observations were made from April to June 1991, March until May and August
1992, thereby covering different seasons and weather conditions.
77
78 KLAUS RIEDE
Recording method
Recordings in the rainforest canopy were made by four pairs of electret microphones (EM-
3) installed at four stations about 30 m apart at the walkway (Figure 1). Microphones were
fed into two 4-channel mixers and their outputs were recorded by a stereo cassette recorder
(Sony TCD5-PRO). Microphone stations could be selected by switching the respective mixer
channels. Besides simple stereo recording at one site, simultaneous recording at two sites
was possible by selecting microphones at different sites for the left and right channel,
respectively. In addition, individual songsters were recorded and localized by a highly
directional microphone consisting of an electret microphone capsule (type MCE-2000)
mounted into a parabolic perspex reflector (PBR-330, Sony, diameter 33 cm).
Besides these recordings, sound intensity, light intensity, temperature and humidity were
recorded. Overall sound intensity (dB [A]) was measured by a soundlevel meter (Noris NM-
3).
Sound analysis and classification

Sounds were analyzed with a spectrum analyzer (Spectro 2000, MEDA V) which produced
an on-line fast Fourier transform visible on a colour monitor (Figures 2-4). Sound
parameters such as carrier frequency and pulse and chirp intervals were fed into a data base
(Quattro Pro, Borland), together with information about locality and meteorological
parameters. In most cases, sound patterns could be classified and attributed to certain
'ethospecies' (Riede 1993).
Figure 1. Wiring scheme of microphones. Two channels are recorded by a stereo cassette recorder; for each
channel, microphones can be selected individually by changing the respective channel amplification by simply turning
the mixer controls.
Sound-producing insects in Bomean rain forest 79
10 kH
.. 1,1
~ 1: P ~ ",.. ~
t ll- I!
~
~ Ih !( ff .~ ~
l!
~
·'t ·
;1 i ;'
Figure 2. Sonagram of a 16 s record of the calling community in the canopy-region (Record canopy-walkway,
central platfonn, 17.4.1991, 19:40) Ordinate: frequency [kHz]; Abscissa: time [s].The broad frequency calls in the
higber frequency range between 7 and 8 kHz migbt be calls of a katydid (Tettigoniidae), the narrow frequency bands
at 6,3 kHz, 5,4 kHz, 4,7 kHz and 3,3 kHz could be cricket songs. Althougb the pulse structure of these songs
cannot be recognized at this time-resolution, the different chirp structure appears clearly.
-1 ' ~- 131313 s
- -'~ - -4
1. 2 ,0013 s
10~-
N
~
t5z
.. "" A ...........
w
:::>
ow
a:
u.
o ~- . ~- . -, - ____~
eee s I.eee s 2.eee •
Figure 3. At a higber time-resolution, the pulse structure of a cricket species singing at 6,3 kHz emerges. In the
upper half of the screen the time structure is represented by an oscillogram, the sonagram is plotted below. This
call is clearly identifiable because of the shortening of pulse intervals at the end of the chirp, and could be heard at
many sites during the whole time of the investigation. In addition, a second species with a deeper carrier-frequency
and higber pulse rate appears. Record from canopy-walkway, central platfonn, 18.4.1991, 19:50.
80 KLAUS RIEDE
~
w
~ 0
I-
...J
0
>
-1
¥4
::.
>3
()
m2
:>
faa: 1 ____ .....-.l-.
--
u.. 0
0 0.5 TIME [5] 1.5
Figure 4. Euphonic low-frequency calls of Tympanophyl/um sp. 1. The calls strongly resemble
calls of the frog species Metaphrynella sundana. For call parameters, see Table 1.
RESULTS
Song activity of the numerous members of the 'acoustic community' depends on season, time
of day and meteorological conditions. The aim of this paper is to analyze the diurnal
variation of acoustic guild composition at the group level; more detailed analysis for within-
group variation will be presented in later papers.
Characteristics of song types

Most of the songs can be attributed to vertebrates (mammals, birds and frogs) or different
insect orders such as crickets and katydids (Orthoptera: Gryllidae and Tettigoniidae) (Figure
2) and cicadas (Cicadidae). In the following account, some general properties of sound
producing mechanisms and song characteristics will be outlined.
Conspicuous, repetitive signals with a narrow-band carrier frequency between 4 and 9 kHz
are generated by the rhythmic wing-strokes of male crickets (Gryllidae). During inward
movement of the forewings, a file-and-scraper mechanism excites specialized resonance areas
which radiate sound in a narrow carrier frequency band and at a pulse rate determined by
the wing strokes (Figure 3). Besides these basic features, pulses can be arranged into
complex secondary and tertiary structures called chirps. Carrier frequency, pulse rate and
chirp structure are species-specific and are important cues for the innate releasing mechanism
of searching conspecific females (Schildberger et al. 1989).
The related orthopteran family, Tettigoniidae, basically employs the same sound producing
mechanism, but in general their songs consist of broader frequency spectra reaching up to
the ultrasonic range, due to a highly damped alar resonator. However, species producing
narrow carrier frequencies have been observed. Two species of the tettigoniid subfamily
Pseudophyllinae (Tympanophyllum sp. Krauss 1903 and Tympanophyllum aff. atroterminatum
[Brunner v. Wattenwyl 1895]) produce pure frequency songs of carrier frequencies between
1.4 and 2 kHz (Figure 4), which is the lowest frequency known for Tettigoniidae and
strongly resembles vocalisations of the frog genus Metaphrynella sundana (Peters 1867)
(Table 1). One could speculate that these Tettigoniidae acoustically mimic the well-protected
frogs which live inaccessibly in tree and bamboo cavities. By shifting to frog-like low-
frequency signals, Tympanophyllum could avoid falling into the scheme of a 'typical' high-
Sound-producing insects in Bomean rain forest 81
Table 1. Call structure of syntopic frogs and katydids (Pseudophyllinae: Tympanophyllum spp.)
Species Frequency (kHz) pulse length(ms) pulses/chirp chirp length (s) chirp interval (s)
Metaphrynella sundana 1 125 1 125 >3
Tympanophyllum sp. 1 1.4 60 2 160 0.5
Tympanophyllum aff. 2 54 4-5 430
atroterminatum
frequency tettigoniid song which might have been learned by vertebrate predators such as
bats or tarsiers. The example shows that unequivocal attribution of ethospecies to higher
systematic units is impossible in some cases, and that taxonomic identification of the
'ethospecies' can only be accomplished if the songster can be captured.
The cicadas (Cicadidae) generate their high intensity songs using tymbals situated at both
sides of the first abdominal segments. Their characteristic, species-specific songs contain
higher harmonics (Figure 5), and ethospecies can be identified reliably after some practice.
This is reflected also in traditional classification systems where an onomatopoeic terminology
describes each 'ethospecies' (e.g. "aisa-aisa" or "ngio-ngio" for certain cicada songs in the
Rungus language).
7.5~- ________________ ~
a ......
....~ - ~'j>-r, . ~ :"'¥'. -:;.'. .•,;1 ~.

~r:,·Y ..,~.·J ..1t\:';~,".4! ~;.:- ..... ~ ...·.!o~'-;"<i
~';cfvl:<wL
1.333 s ~.67 s
bz
UJ
.'.
~
aw
0:
LL
O+-~_P
e.eee • 1.333 s ~.67 s
Figure 5. Cicada songs are characterized by their great proportion of higher harmonics. Left row: oscillograms
and spectrograms of 2.7 s length. Right row: oscillograms of longer segments (20 s) of the respective songs. The
first species (type I) is Dundubia vaginata (Fabricius 1787). None of the other 'ethospecies' have been caught yet;
hence remain unidentified.
82 KLAUS RIEDE
Sound Intensity at Dusk

400
60 350
300
55
250
::5.. 50 x
200 .2
!g 45
150
40 100
35 50
30+-.,~ 0
17:16 17:45 18:14 18:43 19:12 19:40 20:09
local time
- light intensity ......... sound at ground - sound in canopy
Figure 6. Sound intensity 1m above ground and in the canopy-region, recorded in the early evening hours together
with light intensity. The 'dusk chorus' starts around 18:15, at a light intensity around 10 Ix. The sound-intensity
in the canopy-region is distinctly higher than at the ground.
Variation of overall sound intensity

Figure 6 shows the overall course of sound and light intensity between 17.00 and 20.00
at the Poring canopy walkway. Sound-intensity in the canopy is distinctly higher than at
ground level. The first peak is due to cicadas and is more pronounced in the canopy sound
track, while the second, more prolonged peak was measured at ground level and is generated
by mole crickets. Increase in sound intensity in the evening is so pronounced that one can
speak of a 'dusk chorus'. Dusk seems to be an efficient trigger for synchronisation of
chorusing, and parameters such as change of light intensity or change of spectral composition
of evening light might serve as cues for chorus members.
Analysis of the Chorus

A very pronounced temporal segregation is observed during the dusk chorus; this probably
reflects strong competition between species at that time of the day. Spectral analysis shows
that most of the noise of the early dusk chorus is generated by cicadas. Their songs are easy
to recognize because of their content of higher harmonics (Figure 5). Several cicadas have
been identified to the species level (K. Riede & A. Kroker, unpublished data).
When cicadas are fading, and exactly at 18:30, mole crickets start. Their songs contain
one principal frequency. Although these burrowing animals are confined to humid places on
the ground, their intense songs reach high up into the canopy. Two principal ethospecies of
mole crickets have been identified. Chorusing was also observed in the eneopterine cricket
[tara sp. Walker 1869 (Figure 7), but this was not heard every night, possibly due to either
migration patterns or weather-triggered activity.
Waterfall communities
Waterfalls provide a humid environment throughout the year, even during dry seasons.
A special community of frogs, crickets and katydids can be heard close to waterfalls. Their
call structure is characterized by short duration and high intensity at rather high frequencies
(Figure 8). This convergence must be interpreted as an adaptation to overcome the
background noise of the waterfall, the power spectrum of which shows a linear decrease
towards higher frequencies.
Sound-producing insects in Bornean rain forest 83
2~-
Figure 7. Two crickets forming the principal components of the late dusk chorus: a mole cricket (Gryllotalpa sp.)
generates the lower frequency band between 3.7 and 3.95.1 kHz, and the eneopterine cricket [tara sp. , sings at a
higher frequency of 5 kHz (pulse intervals 76 ms, chirp length around I s, chirp intervals around 1.6 s). After
several chirps, additional individuals of ftara sp. fall in forming a chorus, which results in a blurring of chirp
structure.
-20
iiJ
~
~
CJ)
Z +8db +6db
W
I- 2 .. 9 4.6 7 .. 41<
z
--60
I
o FREQUENCY [kHz] 10
Figure 8. Amplitude spectrum of a record at Langanan-waterfall . Three species managed to call louder than the
waterfall noise; these could be frogs or Tettigoniidae. Nwnbers below frequency maxima signify dB above waterfall
noise and carrier frequency. Sound intensity of Langanan waterfall was around 65 dB[A) at 3 m distance, with a
characteristic drop of ·3 dB per frequency increase of I kHz .
DISCUSSION
In contrast to bird songs and vocalisations from higher vertebrates, the limited vocal
repertoire of lower vertebrates and insects allows the recognition of species by analysing
stereotyped, inborn (but by no means simple!) advertisement calls. Therefore, acoustical
recording is a valuable tool for mapping and monitoring elusive, cryptic and nocturnal
rainforest organisms (Riede 1993). The classification of 'ethospecies' presented here is
preliminary and probably erroneous in some cases, but has the great advantage in as much
as it allows a non-invasive study of singing animals.
The strong convergence between orthopteran and anuran song indicates that acoustic
interference might occur across systematic units. Members from different phyla engage
within the 'acoustic guild' and must be considered together in order to understand the
constraints shaping acoustic communication among rainforest creatures. Members of the
'acoustic guild ' interact within a soundspace defined by call parameters such as frequency
and pulse intervals, and within a soundscape of abiotic sounds such as wind, rain and
acoustical habitat properties. As with radio stations, 'senders' compete for the limited
84 KLAUS RIEDE
resource 'acoustic transmission channel'. One therefore should be able to observe

competition, resource partitioning and niche segregation, and investigate ecological questions
such as the saturation of a community. Acoustic transmission channels are a common,
physically well-defined resource for a great variety of species, similar to light in a plant
community. The observed diurnal resource partitioning points to strong competition within
the acoustic guild which is especially pronounced during the dusk chorus where nocturnal and
diurnal species overlap. Waning daylight might be attractive for many species for a variety
of reasons. Dusk is well suited to trigger precisely timed, brief activity bouts between males
and females, thereby reducing overall song time. Diurnal, visually oriented predators may
stop hunting for want of light, but conspecifics would still have some residual light for
orientation. In addition, members of a chorus are difficult to locate acoustically.
An interesting aspect of bioacoustic diversity is the comparison of acoustic guilds from
different rainforests. A comparative analysis of acoustic partitioning in frogs showed strong
similarities in anuran calling assemblies from three different neotropical forest habitats
(Duellmann & Pyles 1983). The community described here indicates some basic differences
with respect to a Neotropical community from Northwest-Amazonia (Riede 1993). In the
Neotropical community, crickets (Gryllidae) sing during day and night, cicadas are much less
dominant and the dusk chorus is less pronounced. The cricket community seems to be more
densely packed, especially in the canopy, which may reflect differences in canopy
architecture and/or epiphyte load.
In Southeast Asia, the diurnal acoustic dominance of loud cicada songs, at least at certain
times of the year, may well form an important constraint displacing fainter orthopteran
songsters to nocturnal activity.
ACKNOWLEDGEMENTS
I thank Sabah Parks for generous support of my field investigations, R. Malkmus, Wiesthal,
for identification of frogs, and A. Kroker, Freiburg, for identification of cicadas. The study
was conducted with the support of the Deutsche Forschungsgemeinschaft.
LITERATURE CITED
DUELLMAN, W. E. & PYLES, R. A. 1983. Acoustic resource partitioning in anuran communities. Copeia 3:639·
649.
ELTON, C. S. 1973. The struCTUre of invertebrate populations inside neotropical rain forest. Journal of Animal
Ecology 42:55-104.
RffiDE, K. 1993. Monitoring biodiversity: Analysis of Amazonian rainforest sounds. Ambio 22:546-548.
ROMER, H., BAILEY, W. & DADOUR, I. 1989. Insect hearing in the field. Journal of Comparative Physiology
AI64:609-620.
SCHILDBERGER, K., HUBER, F. & WOHLERS, D. 1989. Central auditory pathway: Neuronal correlates of
phonotactic behaviour. pp. 423-458 in Huber, F., Moore, T. E. & Loher, W. (eds). Cricket behaviour and
neurobiology. Cornell University Press, Ithaca.
WALLACE, A. R. 1869. The Malay archipelago. Tynron Press, Scotland.
The arthropod community of bamboo internodes in Peninsular Malaysia:
microzonation and trophic structure
DAMIR KOV AC 1 and BRUNO STREIT2

IForschungsinstitut Senckenberg, Entomologie I, Senckenberganlage 25, 60325 Frankjun,
Germany
2Zoologisches Institut, 1. W. Goethe-Universitiit, Siesmayerstrasse 70, 60054 Frankjun,
Germany
ABSTRACT. The hollow culm segments (internodes) of the Southeast-Asian bamboo Gigantochloa
scortechinii Gamble were chosen as model systems for studying tropical animal communities. When the
culm wall is punctured by boring insects or otherwise injured so that a connection to the exterior is
established. the respective internode becomes filled with runoff rainwater. Thus a phytotelma - a small
water body within a plant - is created. The phytotelma-containing internodes are colonized by numerous
species of both aquatic and terrestrial animals. As a first step in the investigation of this poorly known
community a survey was taken of the animal taxa occurring in living and dead internodes. In order to
watch the interactions between the internode inhabitants more closely, artificial 'experimental internodes'
were designed in which the internodes of horizontally positioned culm sections were prepared in such a
way that their upper part formed a movable lid that could be opened for inspection. In such internodes
the processes within the internode interior such as the sequence of colonization and the trophic
interactions, could be observed. In an, several hundred animal taxa were found, many of them restricted
to bamboo internodes. and many still undescribed. This makes bamboo internodes among the most
species-rich of an phytotelmata. Several fac\ors contribute to this extraordinary diversity: the nutrients
that become available from the decomposition of the internode wan; the secluded and therefore wen
protected terrestrial zone above the water surface; and the occurrence of different types of internodes that
provide a panoply of different ecological conditions.
KEY WORDS: food webs, phytotelmata, spatial distribution, bamboo internodes, Gigantochloa
scortechinii.
INTRODUCTION
Various hypotheses have been proposed to explain the astonishing diversity and complexity
of tropical communities. Our research group is studying this and other ecological questions
in tropical ecosystems of West Malaysia. A microecosystem in the culms of the SE-Asian
bamboo Gigantochloa sconechinii Gamble was chosen as a model system. Bamboo culms
are usually subdivided into hollow segments, or internodes, whose interior is secluded from
the outside environment. However, when the culm wall is penetrated by boring insects, the
resulting holes not only allow other invertebrates to enter, but may also result in the
internodes becoming filled with runoff-water during heavy rain. In this way a habitat is
created which may be colonized by both aquatic and terrestrial organisms.
Other examples of such natural bodies of water in plants (=phytotelmata) in SE-Asia are
water-filled tree cavities (Kitching 1987, 1990, Thienemann 1934) and the insectivorous
pitcher plants (Nepenthes) (Beaver 1979a, 1979b, 1983, 1985, Thienemann 1932). These
microhabitats have received increasing attention as methodologically amenable systems for
testing hypotheses on the functioning of animal communities (Frank & Lounibos 1983,
Kitching & Beaver 1990, Mogi & Yong 1992, Pimm & Kitching 1987). They offer several
inherent advantages for the study of ecosystems: (1) they usually occur in sufficient number
to allow statistical analyses; (2) they are colonized by a complex and specialized fauna which
85
86 DAMIR KOVAC AND BRUNO STREIT
Figure I. Schematic representation of a horizontal internode that has been sawed open in such a way that its upper
half forms a movable lid that can be opened and closed ('experimental internode'). During rain the internode fiUs
with water and several distinct microhabitats are created: water (aquatic), water surface (neustonic), wet transitional
zone (semi-aquatic), and land zone (terrestrial). The cleft that remains between the top (lid) and the bottom of the
internode also becomes a distinct habitat with a specialized community. The different zones can be subdivided further
(detritus, free water, etc.), with gradual transitions between the individual zones.
can be sampled completely; (3) they are physically delimited from other habitats. Bamboo
internodes offer even more advantages: (4) the process of colonization can be started
experimentally at any time by boring artificial holes; (5) complex interactions between the
aquatic and terrestrial fauna can be investigated; (6) the wide geographical distribution of
bamboo makes possible comparative studies in different biogeographic regions and climatic
zones.
As the fauna of bamboo internodes, with the exception of the mosquito larvae, was poorly
known at the outset of this study (see Discussion), a faunal inventory of the internode
inhabitants was the first step in our investigation. It soon became apparent that different types
of internodes (living or dead, with holes or as stumps containing open water) lead to the
formation of different animal communities within them. The second important step was
taxonomic work on the internode inhabitants, many of which were still undescribed. Finally,
an investigation of the structure and dynamics of the internode community was begun, with
emphasis on trophic interrelationships and succession.
Among the different types of internodes the 'experimental internodes' were most
thoroughly investigated. These horizontally positioned internodes of felled bamboo culms,
whose upper halves could be swung open for inspection (Figure 1), allowed close monitoring
of the internode fauna over an extended period of time without causing much disturbance.
Moreover, the system could be experimentally altered by the removal of certain species. An
additional advantage was the ease of observation of the animals in these internodes which,
in contrast to open phytotelmata, do not fill up with fallen leaves. This allowed quantitative
investigations of the trophic relationships under nearly natural conditions. This paper is
intended to give a first, qualitative picture of the animal community occurring in the
experimental internodes and of their trophic interactions.
METHODS
The study was conducted in 1991 in West Malaysia, Selangor Darul Ehsan, at the Ulu
Gombak Field Studies Centre (University of Malaya), which is situated on the western slopes
of the Main Range (location: 3°19'32"N, 101°45' 16"E, altitude c. 250 m). The climate and
geology of the area have been outlined by Marshall (1971) and Bishop (1973). The
The anhropod community of bamboo internodes 87
surrounding forest was logged selectively in 1956-1958. Clumps of the large bamboo G.
sconechinii Gamble, which is indigenous to the Malay Peninsula (Holttum 1959), are
abundant in the regenerating vegetation (Figure 3A). The culms grow to 25 m. In the lower
section the length of the internodes measures c. 20-60 cm, the diameter c. 8-lO cm, and the
thickness of the waH is c. 1 cm.
The faunas of different types of internode were investigated, including living internodes,
dead internodes, bamboo stumps etc. Experimental internodes (Figure 1) were studied most
intensively for methodological reasons. These were made by felling the bamboo culm,
removing the upper section including aH side branches, and placing the remaining lower
section horizontaHy at a height of c. 1.5 m by wedging its terminal ends between other, erect
bamboo culms. Ten internodes per suspended culm were sawn through halfway near the
nodal septa at both ends of the internode, then the section between the two transverse cuts
was split on one side with parang and hammer. The upper halves of the internodes could thus
be bent upwards, with the only remaining connection between lower and upper side forming
the hinge.
One hundred experimental internodes were checked daily during a six-months period (end
of May through November 1991) to monitor the composition of the community and study its
succession and trophic interactions. After the first heavy rain aH internodes had filled with
water (c. 200-500 ml per internode). With the exception of two which became temporarily
dry, aH internodes remained water -fiHed throughout the study period. Colonization of the
internodes was possible via the two sawn gaps near the node septa, which were about 2 mm
wide. Species that spend the adult stage outside the internodes were taken to the laboratory
as larvae or pupae to be hatched. Freshly captured prey was taken from predators with a pair
of forceps and identified in the laboratory. The animals were observed using a ZeissR
prism-head lens F (8x).
RESULTS
The 'experimental internode' habitat may be subdivided into several adjacent vertical zones:
(1) aquatic zone, (2) water surface, (3) semiaquatic zone, and (4) terrestrial zone. The small
gap between the movable upper and the lower part of the internode forms an additional
microhabitat for a specialized fauna dominated by fly larvae. This latter microhabitat will not
be considered in the present paper.
Aquatic microhabitat
The aquatic food chain of the internode community is fed by both water soluble nutrients
from the inner bamboo waH and nutrient import from the outside, e.g. drowned animals.
Once water has entered the internode, organic and inorganic substances of the nutritious
internode waH become dissolved in it, stimulating the growth of bacteria and other
microorganisms. Protozoa, e.g., ciliates, soon follow, feeding, among other things, on
bacteria. Other members of the internode microfauna are small crustaceans (Copepoda and
others) and water mites. Apart from these organisms, insect larvae also dominated the aquatic
habitat. With the exception of Seines (Coleoptera), the insects involved aH belonged to the
order Diptera.
At the beginning of the succession, MegaseUa sp. (Phoridae) larvae are numerous (see
also Disney 1991, 1993). They move swiftly along the bottom, apparently scraping organic
matter off the substrate. They are also capable of ingesting floating organic particles or of
feeding on dead animals. In the course of time a substantial amount of detritus, including
dead animals from the terrestrial zone of the internode, nesting material carried in by ants
etc., accumulate at the internode floor.
Figure 3. A. Stand of the bamboo species Gigantochloa scortechinii in Ulu Gombak, Peninsular Malaysia. There
are several fallen bamboo culms in the foreground. In internodes of such fallen culms that have holes in their wall
the predators shown in B-E occur. B. A jumping spider of the subfamily Spartaeinae (length c. 7 mm) has captured
a predaceous mosquito larva, Toxorhynchites metllllicus. C. A male of the water strider Baptista (length c. 5 riun)
is sucking on a springtail (Collembola). D. The rove beetle Acylophorus sp. (length c. 6 mm) is holding a freshly
captured mosquito pupa in its mandibles while standing on the water surface. E. The predaceous mosquito larva
Toxorhynchites metallicus (length c. 7 mm) is feeding on the larva of a different species of mosquito. (With the
exception of E, all photographs taken in the field).
The arthropod community of bamboo internodes 89
The larvae of the ceratopogonids Dasyhelea grata (Johannsen) and D. assimilis

(Johannsen) stay in or on the detritus layer and feed on dead animals. Beetle larvae of the
genus Seirtes are found within the detritus layer, feeding on sizable organic particles. Larvae
of a still undescribed Foreipomyia sp. (Ceratopogonidae), which also ingest organic particles,
stay on the detritus layer or on the submerged part of the internode wall. A frequent and
conspicuous inhabitant of the detritus layer are the larvae of the chironomid Polypedilum
(Pentapedilum) convexum (Johannsen), which build vertical chimney-like structures out of
faecal pellets and other detritus. With the help of their undulating body movements they
create a water current through their tube and their fine-meshed silk web. They consume
organic particles that get trapped in the web. Furthermore, on the detritus layer larvae of the
cecidomyiid Xylodiplosis sp. are found which probably are not truly aquatic (see Discussion)
but may nevertheless be an important dietary component for aquatic predators during the
early succession stages.
In contrast to the other water inhabitants, the mosquito larvae, being able to swim or float,
also colonize the open water. They feed mainly on minute organic particles and
microorganisms which they strain from the water. During this activity some species move
along the internode wall (Uranotaenia sp. A), others lie on their backs on the floor of the
internode (Culex (Lophoceraomyia) minor (Leicester», stay in the middle layer of the water
(Uranotaenia sp. B), or hang down from the water surface (Orthopodomyia anopheloides
(Giles». Anopheles asiaticus Leicester strains food particles from the underside of the water
surface. In addition, predatory behaviour may occur in the genera Armigeres (Leicesteria)
and Anopheles. Armigeres may also be cannibalistic.
Two aquatic predators were found in the experimental internodes: mosquito larvae of the
genus Toxorhynchites (T. leicesteri Theobald and T. metallicus Leicester, Figure 3E) and an
undescribed midge larva of the subfamily Tanypodinae (Pentaneurini, near Trissopelopia).
The midge larvae move slowly and inchworm-like along the bottom hunting small fly larvae.
While searching for prey, the animals also climb onto raised objects (e.g. the tubes of
Polypedilum convexum) then extend their body into the open water waiting for prey to swim
by. In addition to live prey, dead animals are also taken. Several of the predaceous midges
have often been observed feeding simultaneously on the same object. The Toxorhynchites
spp. prey on all of the larger aquatic organisms. During the day, they tend to stay at the
bottom of the water and during the night near the surface. In addition to aquatic prey, they
also catch terrestrial animals which lie on the water surface or which have become trapped
there. The majority of prey objects were other species of mosquito larvae. The prey spectra
for both Toxorhynchites spp. were nearly identical. T. leicesteri reached its greatest
abundance before T. metallicus, at the beginning of the succession. In both species fighting
to the death was observed between the larvae until one or, rarely, two individuals per
internode remained.
The water sUrface •

Besides springtails (Collembola) and mites (Acari), water striders of the genus Baptista
(Figure 3C) and rove beetles of the genus Acylophorus (Figure 3D) were the most
conspicuous arthropods found on the water surface. Baptista sp. feeds to a large extent on
Collembola (c. 45% of the prey of the nymphs and the adults, n=278) which it hunts
actively on the water surface. The remaining prey objects were mainly terrestrial animals that
had been trapped on the water surface. A small percentage were aquatic animals caught just
below the water surface, e.g. small Seirtes or Clogmia larvae.
The adults of the rove beetle Acylophorus also stayed mostly on the water surface. They
waited for mosquito larvae and pupae (c. 58% of the prey objects, n=51) to come to the
surface for air, then seized the prey, dragged it out of the water, and ate it. Live and dead
animals on the water surface were also caught by aquatic predators, including Toxorhynchites
and the normally filter-feeding larvae of Anopheles. In two cases an Anopheles larva was seen
to seize a live Baptista nymph by the hind leg and gradually swallow it entirely.
The semiaquatic microhabitat

Between the aquatic and the terrestrial zone there is a narrow band covered by a thin
water film. This wet zone normally extends only a few millimeters vertically. The meniscus
from the aquatic zone extends into the semiaquatic zone, and the semiaquatic zone in turn
merges with the moist lower section of the terrestrial zone a little higher up.
The semiaquatic microhabitat is inhabited by various fly larvae which graze upon
microorganisms growing on the internode wall. Some of the larvae are confined to this zone,
while others occur also in the aquatic or the lower, moist part of the terrestrial zone. The
larvae of Forcipomyia sp. (Ceratopogonidae), Philosepedon sp. (Psychodidae), and Ptilona
confinis (Walker) (Tephritidae), for example, are restricted to the semiaquatic zone.
Anopheles asiaticus larvae enter the semiaquatic wne without ever leaving the meniscus
region. The larvae of the psychodid, Clogmia sp., for the most part stay in the aquatic zone
with their caudal spiracles in constant contact with the water surface, but they also enter the
semiaquatic zone. The pupae of the mosquito, Uranotaenia sp. A, usually remain in the
semiaquatic zone but flee into the water upon disturbance. Other inhabitants of the
semiaquatic wne also occur in the moist part of the terrestrial zone, e. g., larvae of the
phorid Megase/ia humida Disney, larvae of the stratiomyid Camptopteromyiajractipennis de
Meijere (Rozkosny & Kovac 1991), and larvae of several Forcipomyia species. Most larvae
of the semiaquatic (and terrestrial) zone possess some passive anti-predator defenses such as
protective secretions (Forcipomyia spp.), camouflage by covering the dorsal surface with
detritus particles (the phorid Megaselia humida, the psychodid Philosepedon sp.), or
surrounding themselves with threads of mucus (e.g., Mycetophilidae).
The inhabitants of the semiaquatic wne are preyed upon by spiders (Spartaeinae sp;
Theridiidae spp., etc.) and by Acylophorus adults and, especially, larvae. The Acylophorus
larvae remain completely in the semiaquatic zone. Unlike the adult beetles the larvae have
no hydrophobous pubescence and therefore cannot walk on the water surface. Instead they
incessantly walk along the internode wall just above the water surface hunting the fly larvae
of the semiaquatic zone. They also catch Baptista nymphs and, like the Acylophorus adults,
seize any mosquito larvae that come to the surface for air and pull them out of the water with
their mandibles.
The terrestrial microhabitat

The terrestrial fauna is supplied with nutrients from the decomposition of the inner
internode wall and from the import of organic substances from the outside, e. g., in the form
of animals seeking protection within the internode. At the high ambient relative humidity,
microorganisms and especially fungi begin to break down the internode wall. Various insect
larvae graze on microorganismic growth on the internode wall. These include the flies,
Forcipomyia species or, at later stages of succession, the larvae of Mycetophilidae and
Sciaridae; the beetles, larvae of Endomychidae, Coccinellidae, and Tenebrionidae, and the
moths, larvae of Arctiidae. Another type of decomposer are the various wood-boring beetles,
e. g. Conarthrus and Dinoderus, that stay in the internode cavity for a brief period at the
beginning of the succession, but then tunnel into the bamboo wall.
The terrestrial fauna also includes groups of detritivores which are normally abundant in
the soil. These are species requiring high humidity, e. g. wood-lice, springtails, Psocoptera,
and earwigs (see Srivastava & Kovac, in press). They have been observed to enter the
internodes at night. In addition, numerous mites, true bugs, beetles and insects from other
orders are found. In some internodes earthworms were found, mostly in the cleft between
the upper and lower part of the internode, with part of the body submerged.
The predators among the terrestrial internode inhabitants belong to the true bugs, beetles,
ants and spiders. The assassin bug Emesopsis sp. used its raptorial forelegs to catch mainly
small, soft-bodied insects like Collembola and Ceratopogonidae, but also caught larger
animals, e. g., adult mosquitoes. The staphylinid beetle Stilicopsis sp. had a similar prey
spectrum. A number of additional predaceous and non-predaceous staphylinids were found
in the internodes, but none of them was as common as Stilicopsis and Acylophorus. Ants
were found only rarely and only a few prey objects were taken from them for identification
(for example, larvae of Cecidomyiidae). As ants are the most common prey objects of the
web-building internode spiders, however, they apparently do enter the internodes regularly,
perhaps more frequently at night, when no regular checks were conducted.
The top predators in the experimental internodes were spiders. A species of salticid
(Spartaeinae) (Figure 3B) fed on terrestrial and aquatic animals (c. 50% of the prey items
each, n=84), especially mosquito larvae, that it caught at the water surface. The majority
of the remaining spiders were web-builders of different families which constructed their webs
in the upper region of the internode. The majority of species belonged to the Theridiidae,
which build irregular, three-dimensional cobwebs. Ants and beetles were their main prey.
To a lesser extent freshly emerged adults of insects with aquatic larvae were caught
(Culicidae, etc.). In a few cases an undetermined Theridiid was seen to intrude into the webs
of other spiders and prey on the occupants. The remains of the spider's prey organisms were
dropped to the water surface after some time and thus became accessible to the aquatic fauna.
Most other species of animals found in the terrestrial zone were represented by few or
even single specimens, for example various wasp and fly parasites that attack eggs, larvae
and pupae (e.g., those of Forcipomyia and Epicypta) or animals which utilize the internodes
for breeding (e.g., the crab spider Smodicinodes kovaci Ono (Thomisidae), see Ono 1993)
or for temporary hiding places. Many of these animals certainly are not regular internode
inhabitants. All of these frequently get trapped in spider webs or on the water surface and
are therefore an important component of the nutrient import into the internode.
DISCUSSION
Comparison with earlier investigations on bamboo phytotelmata

The community in bamboo internodes has, with the exception of mosquito larvae, received
little scientific attention to date, especially with respect to the terrestrial components of the
community. Leicester (1903), who studied mosquitoes on the Malay Peninsula, was the first
to mention this kind of habitat, notably the erect culms with holes. Thienemann (1932,
1934), who collected inhabitants from bamboo stumps and other phytotelmata in Sumatra,
Java and Bali, was the next to make an important contribution. MacDonald & Traub (1960)
conducted an extensive 3-year investigation on the ecology of the forest mosquitoes, mainly
in the Ulu Gombak area. This study covered c. 100 species of mosquitoes that inhabit various
internode types of Gigantochloa scortechinii and other bamboo species. In addition,
information on the origin of the holes in live bamboo culms was given (MacDonald 1960).
In the same area the bats Tylonycteris pachypus (Temminck) and T. robustula Thomas, which
roost exclusively in live internodes (Medway 1972, Medway & Marshall 1970, 1972), and
the parasitic batfly Basilia hispida Theodor (Nycterbiidae), which pupates on the internode
wall (Marshall 1970, 1971) were investigated intensively.
Kitching (1987, 1990) investigated the fauna of bamboo internodes and other phytotelmata
in Sulawesi and Papua New Guinea. He was the first to construct food webs for this habitat.
Other investigations on the inhabitants of bamboo internodes were conducted in Japan. Most
of these studies dealt with the ecology of culicid larvae in bamboo stumps and bamboo
cemetery vases (see Kurihara 1983). Mogi & Suzuki (1983), in addition, investigated the
complete fauna of live bamboo internodes with artificial holes made by the investigators. The
most dominant groups were aquatic larvae of Culicidae and Ceratopogonidae. The most
common animals in the terrestrial zone were mites, ants, collembolans and diplopods.
Probably none of these animals were bamboo specialists, because there were no naturally
occurring internodes with holes in this region and the animals occurred only in low numbers.
In the present investigation many previously undescribed internode-inhabiting species were
found, especially in the terrestrial zone, which had never before been studied in SE-Asian
bamboos. But even in the better known aquatic zone new animal taxa were discovered, for
example, the semiaquatic larvae of Tephritidae. Another family which had not previously
been found in bamboo internodes are the Cecidomyiidae. Aquatic larvae of this family
(Lestodiplosis syringopais (Hering), Beaver 1979a) were known, however, to inhabit pitcher
plants. The same genus also occurred in the experimental internodes (see Table 1), but only
as adults. In the water we found, however, larvae of the cecidomyiid Xylodiplosis. The red,
nearly fully grown larvae appeared in the water very suddenly, but then died within a few
days both under natural and laboratory conditions. The larvae probably developed in the
vascular tissue of the bamboo, as the larvae of other members of this genus do (Gagne 1985,
Rock & Jackson 1985). Xylodiplosis females were regularly seen to deposit their eggs on the
cut surfaces of the internode wall. Probably the larvae left the vessels to pupate, and the
animals in the lower part of the internode got trapped in the water and died there or were
caught by predators.
Comparison between different internode types

The kind of nutrient input and the accessibility of the bamboo internode have an important
influence on the structure of the internode community. This was realized by Leicester (1903)
Table 1. Provisional list of arthropod species encountered in experimental internodes (families, genera and species
in alphabetical order; list not exhaustive). Other macroscopic internode-inhabiting metazoans present were annelids
and molluscs. The organisms designated * have not yet been described.
CRUSTACEA
Copepoda
ARACHNIDA
Acarina
Araneae Clubionidae Clubiona sp. *
Heteropoda sp.
Gnaphosidae Aphantaulax sp.
Mysmeridae Tamasesla sp.
Ochyroceratidae Merizocera sp.*
Oonol'idae Gamasomorpha sp.
Pholcldae Spermaphora miser Bristowe
Uthina luzonica Simon
Salticidae Spartaeinae sp. *
Theridiidae Coscinida sp. A
C. sp. B
Euryopis sp. A
Eu. sp. B
Theridion sp. A
T. sp. B
INSECTA
Blattodea Blattelidae Margattea sp. *
Dermaptera Chelisochidae Euenkrates sp. *
Hamaxas sp.
Proreus sp. *
Labiduridae Allostethella guttata (Bormans)
Spongiphoridae Chaetospania feae (Bormans)
Psocoptera Lepidopsocidae Soa sp.
Pachytroctidae Tapinella sp.
Hemiptera Anthocoridae Lasiochilus sp.
Aradidae Brachyrhynchus sp.
Ceratocombidae
Miridae Cylapomorpha sp.
Cylapus sp.
J1ulvlus sp.
n. gen .• n. sp.
Pentatomidae
Table I (continued).
Hemiptera (cont.) Reduviidae Emesopsis sp. *
Veliidae Baptista sp. *
Coleoptera Biphydidae
Brenthidae
Carabidae Pentagonica sp.
Cleridae
Coccinellidae
Corylaphidae
Cucujidae
Curculionidae Conarthrus sp.
Endomychidae Eumorphus vitalisi Arrow
lndalmus sp.
Histeridae
Hydrophilidae
Lagriidae
Lathridiidae
Nitidulidae
Pselaphidae Batraxis sp.
Scirtidae Seirtes sp. A
S. sp. B
S. sp. C
Staphylinidae Acylophorus sp. *
Coproporus sp.
Eleusis sp.
Hesperus perakensis Cam.
Pinophilus sp.
SefJedophilus sp.
Sttlicopsis sp.
Sti/ode rus sp.
Thorakoehirus sp.
Tenebrionidae Leiehrodes rufojulvus Westw.
Diptera Cecidomyiidae Clinodiplosis sp. *
Feltiella sp. *
Lestodiplosis sp. *
Xylodiplosis sp. *
Ceratopogonidae Culieoides sp.
Dasyhelea assimilis (Johannsen)
D. grata (Johannsen)
Forcipomyia (Euprojoannisia) similans (Johannsen)
F. (Fore.) bikanni Chan & LeRoux
F. (Fore.) swezeyanaadjinis Chan & LeRoux
further F. (Fore.) spp. *
Chironomidae Compterosmittia sp.
Pentaneurini nr Trissopelopia*
Polypedilum (Pentapedilum) eonvexum Johannsen
Culicidae Aedes (Christophersiomyia) gOmbakensis Mattingly
Ae. (Stegomyia) sp.
Anopheles asiatieus Leicester
Armigeres (Leicesteria) doliehoeephalus (Leicester)
A. (Leicesteria) flavus (Leicester)
A. (L.) sp. A
A. (L.) sp. B
Culex sp.
Orthropodomyia anopheloides (Giles)
Topomyia (Suaymyia) spathulirostris Edwards
Toxorhynehites leieesteri Theobald
T. metallieus Leicester
Tripteroides (T.) sp.
Uranotaenia sp. A
U. sp. B
Mycetophilidae Allaetoneura sp.
Epicypta sp. A
E. sp. B
Phoridae Megaselia animaloterga Disney
M. deningi Disney
M. gombakensis Disney
M. humida Disney
M. kovaci Disney
Psychodidae Clogmia sp.
Philosepedon sp.
Sciaridae Bradysta sp.
Stratiomyidae Camptopteromyia fraetjp'ennis de Meijere
Pteetieus longlpennis (Wiedemann)
Syrphidae Graptomyza sp.
Tephritidae Pti/ona eonjinis (Walker)
Tipulidae
Lepidoptera Arctiidae
Hymenoptera Formicidae Strumigenys sp., Pheidole sp.
many other ants and parasitic Hymenoptera
and MacDonald & Traub (1960). MacDonald & Traub (1960) classified the internodes
intoseveral types: 1) artificial containers (bamboo pots placed in various situations), 2)
bamboo stumps, 3) fallen and upright split bamboos, and 4) upright bamboos with holes. The
latter group was further divided into three subgroups by the size of the holes: a) long axis
less than 5 mm, b) long axis 5-10 mm, and c) long axis more than 10 mm.
We distinguish at a basic level between live and dead internodes:
(1) Access to live, upright internodes is possible through openings resulting from
mechanical injury (e.g. fissures) or from the boring activities of animals. The most common
types of holes are made by the larvae of the leaf beetle Lasiochila goryi (Guer.), by
caterpillars of the moth family Pyralidae (3 species), and by at least one species of
woodpecker, Gecinulus viridis Blyth (MacDonald 1960, and personal observations). Less
common are openings made by adults of the long-horned beetle Abryna regispetri Paiva as
they break out of the internode (Kovac & Yong 1992). During the first stage of succession,
when the water has just entered, part of the surface tissue of the inner internode decomposes.
The bamboo is not damaged in any recognizable way by this process and the water filled
internodes, like tree cavities, may persist for many years. Our preliminary investigations
indicate that the fauna of live internodes comprises fewer species than that of dead
internodes, but there seems to be a higher proportion of specialists. The aquatic community
of the live internodes has a similar structure to that of experimental internodes, but often
there are differences at the species level. For example, Toxorhynchites magnificus (Leicester),
the most common Toxorhynchites species of live internodes, is almost completely restricted
to this internode type (MacDonald & Traub 1960, and personal observations). Some animals
occur only in live internodes - the bats Tylonycteris robustula and T. pachypus for example
(Medway & Marshall 1970).
Also in the terrestrial zone there are animals that probably live exclusively in live
internodes. One of these is an undescribed ant, Tetraponera sp. near attenuata F. Smith, that
inhabits only internodes with small holes, which may be made by pyralid or chrysomelid
larvae or by the ants themselves (Klein et al. 1993, and personal observations). These ants
get much of their nourishment from the honeydew of pseudococcids that inhabit the inner
wall of the bamboo (Klein et al. 1992). The worker ants not only keep other animals out of
the internode but they also keep the internodes dry by ingesting rainwater that has entered
and regurgitating the water to the outside entrance (Klein et al. 1993).
Additional differences in the faunal composition of live internodes derive from the fact
that some animals seem to be confined to the young shoots (e.g., the larvae of a syrphid
species of the genus Graptomyza; see Kovac 1994). There is also evidence that some animals
prefer internodes in the higher part of the culm while others prefer internodes in the lower
part (MacDonald 1960, personal observations).
(2) We found that fallen or felled bamboo culms are broken down within 1-2 years by
fungi and detritivorous animals (e.g. in the later stage by termites), with decomposition
proceeding fastest for culms near the ground. The community within the experimental
internodes corresponds closely to the fauna occurring naturally in fallen bamboo culms with
fissures or small holes. Minor differences between the experimental internodes and internodes
with small holes derive from the fact that the long, slit-like openings of the experimental
internodes allow access also to larger animals if these are sufficiently flat (e.g. some spiders).
A different faunal composition results when the water surface of the bamboo phytotelmata
is freely visible, as in longitudinally split internodes or in bamboo stumps. In that type of
internode larger aquatic animals also occur, for example, tipulid larvae, water beetles
(Dytiscidae: Lacconectus krikkeni Brancucci, Hydrophilidae: Enochrus sp. and others), the
larvae of damselflies and dragonflies (Coenagrionidae: Pericnemis stictica Selys; Libellulidae:
Lyriothemis cleis Brauer) and tadpoles of several amphibians (for example, Kalophrynus
pleurostigma Tschudi). There seem to be more species of planktonic organisms such as small
crustaceans than in closed internodes. Occasionally water measurers (Hydrometra), ripple

bugs (Microvelia) and flies (Dolichopodidae: Pelastoneurus sp.) are found on the water
surface. There are usually no animals in the freely visible terrestrial zone of these open
internodes. 'Closed' internodes with a sufficiently large opening to admit sizable animals on
the other hand, may be inhabited by both large aquatic inhabitants of open internodes and by
terrestrial animals. In such internodes, crabs (lohora johorensis intermedia Ng) and large
spiders that hunt mosquito larvae and dragonfly larvae may be found.
Differences in the community composition of closed and more or less open internodes
result from the following factors: (1) nutrient import from the outside (plant and animal
detritus) is more substantial in open internodes; (2) open internodes are colonized more easily
by microorganisms and insects. There are a number of insects, e.g. water beetles, that use
cues provided by the reflection of light at open water surfaces to find a small body of water.
(3) Large animals require large openings to be able to colonize an internode; (5) fallen plant
matter provides better hiding opportunities for the aquatic fauna.
The community composition of dead internodes depends not only on the size of the
opening but also on the position of the internode. There are, for example, animals that prefer
internodes lying on the ground (e.g. web-building spiders of the family Pholcidae) or others
that colonize only internodes near a river (e.g. the crab lohora johorensis intermedia).
Further, a 'mixed' internode fauna may form when a bamboo culm with holes topples over
and the fauna of live internodes is gradually replaced by animals preferring dead internodes.
Comparison between bamboo internodes and other phytotelmata

The main differences between bamboo internodes and other phytotelmata derive from
following special features of the internodes: (1) decomposition of part of the bamboo tissue
adds to the nutrient supply of the community; (2) the space above the water surface is
frequently secluded from the outside and thereby provides protection for a species-rich
partially specialized terrestrial community; and (3) the occurrence of different internode types
is equivalent to the availability of different habitats.
Other phytotelmata are usually open at the top and receive nutrient input from the outside
in form of, for example, fallen leaves and animals. The phytotelma plant itself usually does
not provide nutrients to the community. The plant tissue of bamboo internodes, in contrast,
may contribute significantly to the nutrient supply. In live internodes plant-born nutrients play
a significant role only in the beginning of the succession; in dead internodes the whole plant
tissue is gradually broken down and decomposers and detritivores form an important
species-rich segment of the community.
The aquatic fauna of the bamboo internodes is similar in structure to the fauna of other
Southeast Asian phytotelmata (Beaver 1979a, 1979b, Kitching 1987, 1990). In open
internodes, in particular, an aquatic community develops that shows great similarity to the
fauna of tree cavities (Kitching 1990, MacDonald & Traub 1960, and personal observations).
'Closed' internodes (access only through small holes) on the other hand, are special in that
they contain a species-rich terrestrial fauna. This terrestrial fauna interacts closely with the
aquatic fauna, as is shown in the diagram of trophic interactions (Figure 2). The
species-richness of the terrestrial fauna apparently depends on the presence of a protected
space above the water surface which provides food in addition to protection. Relatively open
and unprotected tree cavities, for example, do not have a distinct terrestrial fauna, whereas
there are several - though not many - specialized terrestrial species in the better protected
interior of pitcher plants, e.g. spiders of the genus Misumenops (Beaver 1979a).
In closed internodes spiders are the dominant predators. Spiders certainly also playa more
important role as top predators in other tropical phytotelmata than generally acknowledged
(see, for example, Schoenly et al. 1991). The senior author watched spiders at night, waiting
for aquatic prey at the edge of the water surface of open phytotelmata, both in split bamboo
l Theridiidae
co
t
web-building Araneae (e.g .• Theridiidae. Mysmerldae. Oonoplda8. Pholcidae) ;)
r--
t
Hymenoptera
..,~
Lepidoptera
Diptera "I
Spartaeinae
l,J
; "60
c
Coleoptera
Psocoptera I I
1=,.
.i?
n.. Hemiptera
4 Emesopsis I r- -
~
Dermaptera
.!l
-!-1
Blattodea Camptopteromyia
; ti>
c
eo0 CoUembola
.i?
g Myriapoda so:ops;s ,J Ptilona
,; !:l
E ~
E j+ 1=
Acari
Forcipomyill co
Isopod. f4--
~ =
eo
\k
... '--
~ -
Meg8selia
g or
C/ogmis E "
~ +
'--
E
"I I-=:t J;J l :J

TRAPPED Acylophorus
ANIMALS Baptista
t
- --
0
-
- -~ - - - - w-~ - - -- - - ·1 - -- --
l -
- - t e r sur rae e
- Anopheles
- --I
-
•
I
~
" t I I
- _I
Toxomynchitss
(;J Orthopodomyis
Aedes
I Xylodip/osi. L:
J Tanypodina. P
I Meg.selia -1 I
I
Tripteroides
--Anniget'8s
---
- --
Topomyis
I
I P
I II I
Ur8IJotasnis
Culex
co
=
.or
I DBSyhel6s
I Seirtes Forcipomyill Polypedlum
i i
.....
-.... DROWNED
ANIMALS
1+1 COARSE PARTICLE
DETRITUS
FINE PARTICLE DETRITUSI
8. SUSPENDED MATERIAL
j +
microorganisms. protozoa, sman metazoans
+
i
PLAN!' TISSUE
Figure 2. Simplified representation of the food web within the experimental internodes. The subdivision into different
microhabitats is as shown in the framed section of Figure 1. Only the trophic interactions actually observed in the
field were included. Slender arrows indicate from where individual organisms receive their nutrients. Circular arrows
signify the occurrence of cannibalism. The broad arrows represent nutrient import from the outside or from the
terrestrial to the aquatic part of the internode. Broad-framed boxes, in combination with broad arrows, indicate the
process of decomposition of dead organic matter. The trophic interactions shown are not necessarily valid for all
developmental stages. For example, Toxorhynchites preys only on the nymphs, not the adults of Baptista.
internodes lying on the ground, and in open tree cavities. Another vlsltmg predator of
phytotelma inhabitants was observed by D. H. Murphy in Singapore: The flying lizard,
Draco melanopogon Boulenger entered water-filled bamboo poles to feed on mosquito larvae.
Another interesting finding is that almost every predator will, on rare occasions, prey on
conspecifics. Only in Toxorhynchites spp. and Armigeres (Leicesteria) spp., i.e. aquatic
predators that can not escape to other internodes, cannibalism is common.
As is the case with other phytotelmata. a number of specialists that are probably restricted
to this habitat occur in bamboo internodes. For example, many species of the mosquito genus
Armigeres, subgenus Leicesteria, are closely tied in their distribution to certain species of
bamboo (MacDonald 1960). Many other internode inhabitants are probably also bamboo
specialists. The water strider Baptista was found exclusively in bamboo internodes (limnetic
ecosystems in the Gombak region were investigated intensively by the senior author for their
water strider fauna). In contrast to the specialists, there are a number of animals that inhabit
a variety of habitats and have a wide distribution; Megaselia deningi Disney (distribution:
Sri Lanka, Malaysia, Japan), for example, occurs also in Nepenthes distallatoria L. and even
in steepers of Japanese garden radishes pickled with rice-bran and salt. In addition, this
species has been reported in a case of urogenital myasis in a girl (see Disney 1991).
Preliminary comparisons indicate that for many internode inhabitants the particular species
of bamboo is not important - Baptista sp. for example, has been found in several species of
bamboo in West and East Malaysia (Sabah) (Gigantochloa scortechinii Gamble, G. ligulata
Gamble. G. levis (Blanco), Gigantochloa latifolia Rid\., Schizostachyum grande Rid\.,
Dendrocalamus pendulus Ridl.).
During the present investigation several hundred animal species were found in the different
internode types of just one bamboo species, Gigantochloa scortechinii. Up to 26 different
species (microorganisms and fungi not included) co-occurred in one single experimental
internode. This makes the bamboo phytotelmata the most species-rich phytotelmata of
Southeast Asia. There are probably several reasons why this was not recognized earlier. For
one thing, very few long-term studies on bamboo internodes have been conducted.
Furthermore, these earlier investigations were made in regions with fewer species (Japan)
or areas where bamboo culms with holes are rare or absent. Although Thienemann (1934),
who knew about the work of Leicester (1903), was looking for bamboo culms with holes on
Sumatra, Java, and Bali, he did not find any. The senior author of this article was also
unable to detect any holes in bamboo culms at several localities near settlements on Bali and
in Sabah. The reason for that failure may be that either certain species of bamboo are never
or only rarely attacked by boring insects (see Daniels 1908) or that the insect species able
to attack certain introduced bamboo species do not occur in the area.
The results of this investigation endorse Leicester's (1903) closing sentence, "The full
study of the bamboo as a breeding place for insects would, I feel sure, well repay careful and
prolonged work".
ACKNOWLEDGEMENTS
This work is part of a research project funded by the Deutsche Forschungsgemeinschaft and
supported by the Socio-Economic Research Unit of the Government of Malaysia and
Universiti Malaya. Thanks are due to R. Klein for reviewing the manuscript and improving
the English. Special thanks are due to I. Azarae for making available facilities at the Ulu
Gombak Field Studies Centre. The authors are most grateful for the taxonomic assistance
which was provided by the following taxonomists (in alphabetic order): M. Brancucci
(Basel), A. Brindle (Nelson), Chan Kai Lok (National University of Singapore), P. Chandler
(University of New Hampshire), P. S. Cranston (CSIRO, Canberra), C. Deeleman-Reinhold
(Ossendrecht), H. Disney (University of Cambridge), W. W. MacDonald (Liverpool School
of Tropical Medicine), D. Duckhouse (University of Adelaide), C. E. Dyte (University of

Reading), R.I. Gagne (USDA, Washington), M. Hamalainen (Espoo), P. Hammond
(Natural History Museum, London), D. E. Hardy (University of Hawaii), E. Heiss
(lnnsbruck), 1. K. H. Koh (Singapore), 1. Murphy (London), T. R. New (La Trobe
University, Bundoora), P. Ng (National University of Singapore), H. Ono (National Science
Museum, Tokyo), L. M. Roth (The Agassiz Museum, Cambridge, Massachusetts), R.
Rozkosny (Masaryk University, Brno), G. K. Srivastava (Zoological Survey of India,
Calcutta), G. M. Stonedahl (Natural History Museum, London), 1. R. Vockeroth (Centre for
Land and Biological Research, Ottawa), W. W. Wirth (Gainsville), Yong Hoi Sen (Universiti
Malaya, Kuala Lumpur).
LITERATURE CITED
BEAVER, R. A. 1979a. Biological studies of the fauna of pitcher plants (Nepenthes) in West Malaysia. Anna/es de
la Societe Entorrwlogique de France (Nouvelle Serie) 15:3-17.
BEAVER, R. A. 1979b. Fauna and foodwebs of pitcher plants in West Malaysia. The Malayan Nature Journal
33:1-10.
BEAVER, R. A. 1983. The conununities living in Nepenthes pitcher plants: Fauna and food webs. pp. 129-159 in
Frank, J. H. & Lounibos, L. P. (eds). Phytotelmata: Terrestrial plants as hosts for aquatic insect communities.
Plexus, Medford, New Jersey.
BEAVER, R. A. 1985. Geographical variation in food web structure in Nepenthes pitcher plants. Ecolagical
Entomology 10:241-248.
BISHOP, J. E. 1973. Limnology of a small Malayan river, Sungai Gombak. Monographiae Biologicae 22. Junk,
The Hague.
DANIELS, C. W. 1908. Breeding grounds of Culicidae. Studies from Institute for Medical Research, Federated
Malay States 3: 1-7.
DISNEY, R. H. L. 1991. The aquatic Phoridae (Diptera). Entorrwlogica Scandinavica 22:171-191.
DISNEY, R. H. L. 1993. New species of aquatic Phoridae (Diptera) from Malaysia. Aquatic Insects 15:149-158.
FRANK, J. H. & LOUNIBOS, L. P. (eds). 1983. Phytotelmata: Terrestrial plants as hosts for aquatic insect
communities. Plexus, Medford, New Jersey.
GAGNE, R. J. 1985. Descriptions of new nearctic Cecidomyiidae (Diptera) that live in xylem vessels of fresh-cut
wood, and a review of Ledomyia (s. str.). Proceedings of the Entorrwlagical Society of Washington 87:116-134.
HOLTTUM, R. E. 1958. The bamboos of the Malay Peninsula. The Gardens' Bulletin, Singapore 16:1-135.
KITCHING, R. L. 1987. A preliminary account of the metazoan foodwebs in phytotelmata from Sulawesi. The
Malayan Nature Journal 41:1-12.
KITCHING, R. L. 1990. Foodwebs from phytotelmata in Madang, Papua New Guinea. The Entomologist
109:153-164.
KITCHING, R. L. & BEAVER, R. A. 1990. Patchiness and conununity structure. In Shorroks, B. & Swingland,
I. (eds). Living in a Patchy Environment. Oxford University Press, Oxford.
KLEIN, R. W., KOVAC, D., SCHELLERICH, A. & MASCHWlTZ, U. 1992. Mealybug-carrying by swarming
queens of a Southeast Asian bamboo-inhabiting ant. Naturwissenschaften 79:422-423.
KLEIN, R., MASCHWITZ, U. & KOVAC, D. 1993. Flood control in ants: A Southeast Asian bamboo-dwelling
Tetraponera (Formicidae: Pseudomyrmecinae) bails excess water from its internode nests. Insectes Sociaux
40:115-118.
KOVAC, D. 1994, in press. Die Tierwelt des Bambus: Ein modell fUr komplexe tropische Lebensgemeinschafien.
Natur und Museum 124.
KOVAC, D. & YONG, H. S. 1992. Abryna regispetri: a bamboo long-horned beetle. Nature Malaysiana 17:92-98.
KURIHARA, Y. 1983. The succession of aquatic dipterous larvae inhabiting bamboo phytotelmata. pp. 55-77 in
Frank, J. H. & Lounibos, L. P. (eds). Phytotelmata: Terrestrial Plants as Hosts for Aquatic Insect Communities.
Plexus, Medford, New Jersey.
LEICESTER, G. F. 1903. A breeding place of certain forest mosquitoes in Malaya. Journal of Tropical Medicine
and Hygiene 6:291-292.
MARSHALL, A. G. 1970. The life cycle of Basilia hispida (Diptera: Nycterbiidae) in Malaysia. Parasitology
61:1-18.
MACDONALD, W. W. 1960. On the systematics and ecology of Armigeres subgenus Leicesteria (Diptera,
Culicidae). Malaysian Parasites XXXV-XLIX. Studies from the Institute for Medical Research, Federation of
Malaya 29:110-153.
MACDONALD, W. W. & TRAUB, R. 1960. An introduction to the ecology of the mosquitoes of the lowland
dipterocarp forest of Selangor, Malaya. Malaysian Parasites XXXV-XLIX. Studies from the Institute for Medical
Research, Federation of Malaya 29:79-109.
MEDWAY, LORD 1972. Reproductive cycles of the flat-headed bats Tylonycteris pachypus and T. robustu/a
(Chiroptera: Vespertilioninae) in a humid equatorial environment. Zoological Journal of the Linnean Society
51:33-61.
MEDWAY, LORD & MARSHALL, A. G. 1970. Roost-site selection among fiat-headed bats (Tylonycteris spp.).
Journal of Zoology, London 161:237-245.
MEDWAY, LORD & MARSHALL, A. G. 1972. Roosting associations of flat-headed bats, Tylonycteris species
(Chiroptera: Vespertilionidae) in Malaysia. Journal of Zoology, London 168:463-482.
MOGI, M. & SUZUKI, H. 1983. The biotic community in water-filled internodes of bamboos in Nagasaki, Japan,
with special reference to mosquito ecology. Japanese Journal of Ecology 33:271-279.
MOGI, M. & Yong, H. S. 1992. Aquatic arthropod communities in Nepenthes pitchers: the role of niche
differentiation, aggregation, predation and competition in community organization. Oecologia 90: 172-184.
ONO, H. 1993. An interesting new crab spider (Araneae, Thomisidae) from Malaysia. Bulletin of the National
Science Museum, Tokyo, Series A 19:87-92.
PIMM, S. L. & KITCHING, R. 1987. The determinants of food chain lengths. Oikos 50:302-307.
ROCK, E. A. & JACKSON, D. 1985. The biology of xylophilic Cecidomyiidae (Diptera). Proceedings of the
Entomological Society of Washington 87:135-141.
ROZKOSNY, R. & KOVAC, D. 1991. First description of the male and the larva of Camptopteromyia fractipennis
de Meijere from Malaysia (Diptera: Stratiomyidae). Entomologica Scandinavica 22:297-304.
SCHOENLY, K., BEAVER, R. A. & HEUMIER, T. A. 1991. On the trophic relations of insects: a food web
approach. The American Naturalist 137:597-638.
SRIV ASTAVA, G. K. & KOVAC, D., in press. Notes on some Dermaptera from Malaya with the description of
two new species. Records of the Zoological Survey of India.
THIENEMANN, A. 1932. Die Tierwelt der Nepenthes-Kannen. Archiv jUr Hydrobiologie, Supplement-Band
11:1-54.
THIENEMANN, A. 1934. Die Tierwelt der tropischen Pflanzengewiisser. Archiv jUr Hydrobiologie, Supplement-
Band 13:1-91.
The avifauna of the Belalong forest, Brunei Darussalam
CLIVEF.MANN
122 Hanswood Road, London W12 9NG, United Kingdom
ABSTRACT. The areas explored were the full length of the East (Main) Ridge to the summit of Bukit
Belalong and down to Sungai Temburong and Sungai Belalong, the length of the West Ridge from the
Field Centre back to Sungai Belalong, much of Sungai Belalong from Kuala Belalong almost to the
project boundary, and Sungai Temburong from Kuala Belalong to a little beyond Sungai Babi. 24-hour
periods of mist netting were carried out, 10 periods at Pondok Bujat, one in an alluvial forest close to
Sungai Babi, and another along Sungai Belalong at the Field Centre. Netting was carried out rwice at
the summit of Bukit Belalong. A total of 218 individuals of 58 species were captured, of which 214 of
57 species were ringed, the others being net casualties. Moult and weight were recorded, along with any
changes in retraps, as were signs of breeding activity. 198 species were recorded in the project area, with
three others identified to genus only. 25 species recorded are not forest dependent, and a further 15 are
also likely to be found outside forest. During the expedition, 20 definite cases of breeding were
discovered involving 16 species. Within the area three zones were recognised, namely riverside (Zone
1), slopes and ridges up to about 460 m (Zone 2) and slopes and ridges above 460 m (Zone 3). No true
montane species were found. Bird species demonstrate vertical stratification in tall tropical forest, with
four levels being recognised, with a tlfth category for aerial feeders: ground level (leaf litter foragers) -
20 species, nine exclusive to this stratum; understorey - 76 species, 11 exclusive; mid levels - 95 species,
10 exclusive; canopy - 75 species, 44 exclusive; aerial - three forest-dependent species are exclusively
aerial, whereas one other is also found in the canopy. Nine species are tied to waterways, of which two
(Enicurus spp.) are also litter foragers on land, and another (Ceyx erithacus) may also be found some
distance from water. 128 (116 forest dependent) species were found in the riparian zone (Zone 1); 139
(136 forest dependent) species were found on the lower slopes (up to c. 460 m, Zone 2) and 121 (114
forest dependent) on the higher slopes (above c. 460 m; ridge-top pole forest - Zone 3). 14 species of
palaearctic migrants were recorded, together with two others, viz. Ninox scutuiata and Muscicapa
daurica, which have resident and migrant populations. Moult cycles were calculated to last from 100 to
220+ days.
KEY WORDS: bird diversity, tropical rainforet, population structure, breeding, distribution, Borneo.
INTRODUCTION AND METHODS
The aims of the study were as follows:

(a) to make an inventory of the birds, and as far as possible assess population numbers.
(b) to investigate the distribution of avian species in relation to habitat, altitude and vertical
stratification.
(c) to record, through capture and ringing, the timing and rate of moult, weight changes,
population structure and movements of species of the lower strata.
(d) to obtain data on breeding.
The area was first visited by the author on 6 August 1989, accompanying Lord Cranbrook
and others on a reconnaissance trip. The next visits were during the expedition phase of the
project, when a total of 68 days were spent by the author in the project area between 21
February and 13 October 1982. The areas explored during this period were the full length
of the East Ridge to the summit of Bukit Belalong and down to Sungai Temburong and
Sungai Belalong, the length of the West Ridge from the Field Centre back to Sungai
Belalong, Sungai Belalong from Kuala Belalong almost to the project boundary, and Sungai
Temburong from Kuala Belalong to a little beyond Sungai Babi (see Wills, this volume,
Figure 1 for a map of the area). Apart from these main arteries some exploration was
101
102 CLIVE MANN
carried out in intervening areas, but time prevented much of the area from being investigated.
Apart from netting, other ornithological activities consisted chiefly of walking trails and
river banks, and observing from stationary posts and moving boats. As the main netting site,
the northern end of the East Ridge (Bujat), received a disproportionate amount of attention,
as did the area immediately surrounding the Field Centre. As well as making general
observations on birds, Lord Cranbrook carried out netting (see below), as did D.R. Wells,
who also produced a trip report on his visit to the area from 12 to 22 January 1992. These
records, together with a report by C.M. Francis, have been made use of in the present work.
A list of species recorded appears in Appendix 1. Mist netting was carried out during 10 x
24-hour periods (with nets closed during darkness) close to Pondok Bujat, in hill dipterocarp
at the northern end of the East ridge; one period in an alluvial forest close to Sungai Babi,
and another along Sungai Belalong at the Field Centre. After the author's study was
completed, Lord Cranbrook netted twice at Pondok Bujat, as did D.R. Wells. The former
also twice netted briefly at the summit of Bukit Belalong. The birds netted were ringed with
uniquely numbered metal bands provided by Universiti Malaya, and with uniquely colour-
coded plastic bands for individual identification in the field. The age and sex (where
possible) was recorded for each individual ringed, and the weight, wing, tail and gape were
measured. The state of moult was also recorded. Details of change of weight and moult
status was noted for any bird retrapped on subsequent occasions. A total of 218 individuals
of 58 species were captured, of which 214 of 57 species were ringed, the rest being net
casualties (Appendix 3).
RESULTS
The Avifauna
The number of species recorded in the project area was 198 (two only recorded before the
expedition), with three others identified to genus only (see Appendix 1). Since the end of
the expedition phase, another species (Nieuwenhuis's Bulbul Pycnonotus nieuwenhuisl) has
been recorded (R.S.R. Williams, personal communication). This species was previously
known only from one specimen from the upper Kayan collected in 1900, and one specimen
of a different subspecies from Sumatra (Smythies 1981). At the moment the record is sub
judice. The cumulative totals of the species recorded in the area by month are shown in
Figure 1, whereas Appendix 1 shows the months in which each species was recorded, and
the number and percentage of days recorded during the author's presence. Within the area
three zones were recognised, namely riverside/riparian forest (Zone 1), slopes and ridges up
to about 460 m (Zone 2) and slopes and ridges above 460 m (Zone 3). Species recorded in
each of these divisions are shown in Table 1. Appendix 1 shows the stratified and zonal
distribution of each, their diets, and indicates those which are not forest dependent, those tied
to waterways, and those which are Palaearctic migrants. The information on stratified
distribution came from observations made during the project period, reinforced and
supplemented by observations made elsewhere in Borneo and by published material such as
Smythies (1957, 1981). Dietary data, summarised in Table 2, came mostly from published
sources, e.g. Smythies (1957, 1981), supplemented by personal observations.
Although the majority of the resident species recorded undoubtedly breed in the project
area, firm evidence of breeding in equatorial forest birds is notoriously difficult to obtain.
However, during the expedition, 20 definite cases of breeding were discovered involving 16
species (Appendix 2). Breeding was detected by direct observations of active nests or
recently fledged young, the carrying of nest material or the detection of brood patches on
captured birds. Seven records are in January, three in February, five in March, one in May,
Birds of Belalong forest 103
200
175
150
125
100
75
50
Figure 1. Cumulative total of species recorded by month.
Table 1. Numbers of species in each stratum. Levels: G = ground, U = Lower, M = Mid-levels, C = Canopy,
A = Aerial; X = exclusive to that level, Non X = not exclusive to that level. Zone: 1 = riverside, 2 = slopes
to 460 m, 3 = slopes above 460 m. [Since many species occur in more than level and more than one zone, the
totals cannot be arrived at by simply totalling the numbers in a particular line or column)
Levels G U M C A TOTALS
X Non X X Non X X Non X X Non X
Zones
6 10 6 53 6 68 30 25 4 128
2 6 7 7 55 9 75 30 28 4 139
3 5 5 6 37 6 55 30 20 129
TOTALS 9 11 11 65 10 85 44 31 4
one in June, two in July and one in August. Thus, the majority of records of breeding
activity are from the early part of the year just after the Northeast monsoon rains, but the
sample is too small to be considered significant, and covers a rather systematically disparate
community.
About another 35 species could reasonably be expected to occur at times, so the present
list is perhaps within 10-15 % of the true total (see Figure 1), which will include some rarer
residents, as well as palaearctic visitors. Some species, recorded from similar habitats, and
at no great distance from the study area, in Borneo, were inexplicably absent. Whether or
not this absence is real will hopefully be revealed by future field work in the area. Of the
198 species in the project area, 25 are not forest dependent, and a further 15 are also likely
to be found outside the forest.
Bird species demonstrate vertical stratification in tall tropical forest, with four levels being
recognised, viz. ground, understorey (shrub layer), mid-level (higher parts of small
branches of tall trees) and canopy (including emergents). A fifth division accounts for aerial
feeders (G, U, M, C and A in Appendix 1).
104 CLIVE MANN
Table 2. Summary of diet preferences of birds in the study area. Numbers in brackets are non-forest dependent
species.
Exclusive Non-Exclusive Total

F : Fruit (+seeds, flowers) II 61( + I) 73
N : Nectar (+pollen) 0 10 10
I : Invertebrates 94(+7) 65( +5) 171
P: Fish 3(+2) 3(+3) II
C : Vertebrates (except fish) 5 14(+1) 20
F + N only
F + I only 57
1+ N only
1+ Conly 7
1+ P only 3
I+P+C 2
However, much of this stratification is preferential rather than absolute, with such canopy
species as hornbills and barbets occasionally found in lower strata, and ground species such
as pittas being found in the understorey, or higher, when singing. Preferences for each
species in the study area are shown in Appendix 1 and a summary of stratified distribution
in Table 1. These data came from observations in the study area and other parts of Borneo,
and from published sources, particularly Smythies (1957, 1981). Data on zonal distribution
from the study area are combined with these and discussed below.
Ground level (leaf litter foragers): Nine species are exclusively ground foragers, and 11
others use the ground for part of the time. The number of species in this guild diminishes
with increasing altitude from 16 in Zone 1 (six exclusive to this stratum), to 13 in Zone 2
(six exclusive), to 10 in Zone 3 (five exclusive).
Understorey: Only 11 species feed exclusively in the understorey, with 65 other species
which feed elsewhere as well. Greater numbers are found close to rivers and on slopes below
460 m, viz. 59 (six exclusive) and 62 (seven exclusive) as compared to 43 (six exclusive)
above 460 m. These distributions reflect the attenuation of low vegetation with increasing
altitude.
Mid-levels: Only 10 species feed exclusively in mid level vegetation, with 85 other species
that also feed elsewhere. The figures for the three zones are 74 (six exclusive to this
stratum), 84 (nine exclusive) and 61 (six exclusive).
Canopy: A total of 44 species feed exclusively in the canopy, with a further 31 that also feed
elsewhere. 55 (30 exclusive) in Zone 1, 58 (30 exclusive) in Zone 2 and 50 (30 exclusive)
above 460 m. The altitude and proximity to waterways appears to have little effect on the
distribution of canopy feeders.
Aerial: Three forest-dependent species were identified as exclusively aerial, all occurring in
the first two zones, and one in the last. Another species, occurring in the first two zones,
was identified as both aerial and canopy.
In Bornean forests 900 m is generally found to be the altitude required for montane
species. On mountains above this altitude these species may be found below 900 m, but they
will not be found on mountains that do not reach this height (Banks 1937, Chasen 1935,
Smythies 1960, C. F. Mann, personal observation). As was expected, no true montane
species were found in the project area. Lowland species, such as Rhinomyias umbratilis,
which are replaced by highland representatives at about 600-900 m on higher mountains,

were found to the top of Bukit Belalong (approximately 900 mas!).
Alluvial forest: Pitta baudi was perhaps the only species confined to this forest type, as
opposed to those in Zone 1 which are associated with rivers. However, little time was spent
investigating alluvial forest which only exists in tiny pockets within the project area.
Waterways: A total of nine species are tied to waterways, of which two (Enicurus) are also
litter foragers on land, and another (Ceyx erithacus) may also be found some distance from
water. 128 (116 forest dependent) species were recorded in this zone.
Lower slopes (up to c. 460 m): 139 (136 forest-dependent) species were recorded in this
zone.
Higher slopes (above c. 460 m; ridge-top pole forest): 121 species (114 forest-dependent)
were found in this zone.
Palaearctic migrants: 14 species were recorded, and two others, viz. Ninox scutulata and
Muscicapa daurica, which have resident and migrant populations. The former was believed
to be of the resident population, whereas all records of the latter were almost certainly of
migrant birds. Migrants, never numerous, were more frequently encountered during
November and December, but most had moved on by January (Cranbrook, in preparation).
Species Diversity
This was based on totals of each species caught in each ringing session. These are
recorded in Table 3. Figure 2 shows that a larger number of species is represented by a
smaller number of individuals (18 singletons), and vice versa (one species represented by 22
individuals). Such a community structure has been found in other areas, e.g. Kenya (Mann
1985). Table 4 gives some species diversity indices for East African forests, Kenyan
steppe/savanna, Southeast England woodland, and selectively-logged forest in Brunei
(Lamunin) from Mann (op. cit.) for comparison.
Three indices were calculated, but only McIntosh's Index (Om) could be calculated for
sites other than Belalong due to the manner in which the data was recorded, and thus only
this could be used in comparisons. Algorithms for these indices are found in Magurran
(1988). The overall totals were perhaps rather too small in some cases to give a reliable
species diversity index. There is a negative correlation between number of individuals in the
catch and Om, (r = -0.491, P = <5%). Therefore lower catches show a larger Om.
The results clearly show a greater diversity at Belalong than in East African forests or in
selectively-logged forest in Brunei. Some of the figures for savanna/steppe in Kenya are
unexpectedly close, but it should be remembered that in such habitats there is no vertical
stratification, and therefore the sample is closer to the whole avifauna of the area, rather than
a segment.
Moult
Moult was recorded on trapped birds where it was detectable. The degree of moult of the
primary wing feathers was recorded, and its progress monitored on birds retrapped. This
is shown in Figure 3. The numbers in moult are shown in Figure 4. Individuals could be
found moulting in every month, but greatest between May and August. A score of 50
indicates just completed primary moult. Extrapolation of figures on the graph indicate that
a moult cycle would last from between 100 and 220+ days, assuming that most moult would
be contained within the period required for primary moult. Two individuals (Criniger bres
and Malacopteron cinerum) demonstrate interrupted moult, as shown by the shallow angles.
Age Structure of the Community

The numbers of adults and young in the trapped samples are shown in Figure 4. The
numbers of young in all samples are very low, indicating a low rate of replacement, with
106 CLIVE MANN
Table 3 : Species diversity and diversity indices of understorey community of birds in study area. Dm = McIntosh's
index of diversity, range 0 - 1; Em = McIntosh's index of diversity (modified) (range 0 - 1), I/Ds = Simpson's
index of diversity (reciprocal), range 1 - infinity.
Date Total No. of Species Species diversity indices
Dm Em liD,
BUJat
28-29/03/91 15 11 0.86 0.92 15.10
02-03/04/91 14 10 0.89 0.95 18.2

03-04/05/91 29 14 0.82 0.92 13.1
21-22/06/91 33 20 0.92 0.95 24.8
19-20/07/91 25 14 0.87 0.95 17.7
21-22/08/91 28 17 0.89 0.96 23.6
31110-01111191 30 18 0.91 0.97 27.1
13-14/01192 14 8 0.81 0.92 10.1
20-21101192 22 13 0.85 0.91 13.0
Mean 23.3 13.9 0.87 0.94 18.1
Sg. Babi
27-28/10/91 12 9 0.88 0.94 16.5
18
16
(/)
Q)
·0
Q)
a.
(j)
10
'0
Q)
-.0
E
:J
Z
Number of birds
Figure 2. Species diversity among birds netted at Pondok Bujat, 1991-92.
greatest numbers between March and July, which shows some concurrency with observations
on breeding. This is suggestive of greater longevity as found in other tropical forests, for
example in Kenya (Mann op. cit) , where one individual of Andropadus (= Pycnonotus)
curvirostris had a minimum longevity of 19.5 years.
Abundance
No systematic study was carried out to estimate the abundance of species in the project
area (with the exception of Operation Hornbill, Lord Cranbrook (personal communication)
Birds of Beiaiong forest 107
Table 4 : Species diversity & diversity indices in other localities. Dm = McIntosh's index of diversity, range 0-1.
Locality No. of individuals No. of species Om

East African mid-Ievellhighland forests
Kakamega Forest, Kenya (2 year average) 61.3 20.3 0.68
Mau forest, Kenya 45 15 0.76
Arnani Forest, Tanzania (average of 2) 123 23.5 0.78
S.Nandi Forest, Kenya (average of 7) 56.4 22.7 0.81
Kenyan savanna/steppe
Tot 44 19 0.74
Kaimosi 62 19 0.78
Kapsabet 80 19 0.61
Barsaloi 46 10 0.35
Archer's Post 59 15 0.49
Lokichar 46 14 0.79
Lake Hannington (Bogoria) 49 21 0.83
Lake Kisima 59 12 0.68
South Horr 61 23 0.82
Southeast England woodland (average of 4) 85.5 12.3 0.64
Lamunin Forest, Brunei (average of 2) 42 18.5 0.78
o
31/3 30/4 30/5 29/6 29/7 28/8 27/9 27/10 26/11 26/12 25/1
Dates 1991192
Figure 3. Primary moult scores for birds ringed in project area. Vertical axis shows primary moult score on a scale
of 0 (no moult) to 50 (primary moult just completed). Criniger bres = .; Malacopteron magnum = .; Copsychus
pyrrhopygus = 0; Arachnothera /ongirostris = ,,; Trichastoma malaccense = .; Malacopteron cinereum = 0;
Arachnothera crassirostris = 0; ---------- = minimum slope.
and reported elsewhere) due to time constraints and the difficulty of the terrain. Comments
on numbers would be impressionistic only. Appendix 1 shows the percentage of the author's
field days between February and October 1991, when each species was recorded.
108 CLIVE MANN
35
30
en
-a
'- 25
:.a
'0 20
en
L-
a>
.D 15
E
:J
z 10
5
I~ Ul
I
0 b 10 In 10
Mar AprMayJun Jul AugSep Oct Nov Dec Jan Feb
Months
I~ N birds _ N retrap ~ N young b!dd N moult
Figure 4. Statistics of understorey birds netted at Pondok Bujat, 1991-1992. N = number.
Despite the assumption that the majority of species were resident, only three were
recorded on all days, and 22 on 50% or more days. All except five are species that are
particularly conspicuous by their vocalisations. These figures are indicative of the difficulties
encountered when trying to census birds in tropical forest. The large number of birds (79)
seen on 5 % or less days is made up of birds that are genuinely rare within the study area,
or occasionally visit the area, or are difficult to detect. A few species, e.g. Eumyias
thalassina, were only detected by netting.
Figure 4 and Table 5 show the percentage of retraps in the ringed population on each
occasion of netting. Only figures from Bujat are meaningful, as Babi was only netted once,
and Bukit Belalong summit twice, but with very small numbers involved. Over a period of
time as short as this study, the percentage of retraps would be expected to increase steadily
if a totally sedentary community was involved, and if net-shyness did not develop. The
fluctuating results here suggest that although the community was predominantly sedentary,
there is some movement into and out of the immediate area where netting was carried out.
This phenomenon was detected in tropical forest in Kenya where individuals disappeared for
up to a year or more, and then re-appeared (Mann 1985 and personal obserations). Net-
shyness is unlikely to be a significant factor when netting at such time intervals (Mann 1985).
DISCUSSION
The Avifauna
D.R. Wells (personal communication) increased the numbers of species along river banks
by 8.5%, on ridges to 460 m by 5.5%, and above 460 m by 34.4%. The overall species list
was increased by 5.8 %. The small increases for the first two areas are to be expected. It
is not uncommon to add to the avifauna of lowland dipterocarp forest even after some years
Table 5. Ringing analysis.
Locality Date Total No. of Species % retraps
Bujat 28-29/03/91 15 11 0.0
Bujat 02-03/04/91 14 10 21.4
Bujat 03-04/05/91 29 14 13.8
Bujat 21-22/06/91 33 20 13.3
Bujat 19-20/07/91 25 14 24.0
Bujat 21-22/08/91 28 17 17.9
Sg. Belalong 24-25/08/91 8 6 0.0
Sg. Babi 27-2811 0/91 12 9 0.0
Bujat 31/10-01111/91 30 18 30.0
Bt. Be1along 05111/91 6 3 0.0
Bujat 13-14/01192 14 8 42.9
Bujat 20-21101192 22 13 22.7
ofregular observation (personal observation). The larger figure for the higher groundreflects
the less time found by the author to study that area prior to Wells's visit. Also, Wells's
observations were made in January, a month not covered by other workers. Our present list
is perhaps within 10 to 15 % of the true total, which will include some rarer residents, as
well as palaearctic visitors.
Cycles
In temperate species moult and breeding cycles generally do not overlap. Both events are
a considerable drain on food resources, which at higher latitudes show marked seasonal
fluctuations. Such fluctuations are also marked in the tropics, but much less so in equatorial
forests. Seasonal breeding in humid tropical forests seems to be widespread, being recorded
from Southeast Asia (Fogden 1972, Gibson-Hill 1952, Ward 1969, Yoous 1950), East Africa
(Mann 1985, Moreau 1936, 1950), New Hebrides (Baker et al. 1940) and Central and South
America (Skutch 1950, Snow & Snow 1964). Due to both moult and breeding cycles being
very protracted in some equatorial forest species (Mann 1985 and other references cited
therein), they may overlap broadly. Generally there is a smaller number of young (and
broods) which remain dependent on their parents (and in some cases related helpers) for
extensive periods. Although annual cycles are common in equatorial bird species, cycles of
less than, or more than, one year have been detected in Sarawak (Fogden 1972) and East
Africa (Mann 1985).
Comparison with equatorial East African forests

The total number offorest-dependent species recorded in the study area (176) exceeds that
of Kakamega Forest, Kenya (147), where the author carried out an extensive study from
January 1973 to January 1975 (Mann 1985). Kakamega Forest lies on the equator at an
average elevation of 1640 m above sea level; it is both less variable in altitude, and higher
than Belalong (maximum altitude < 900 m). The catch rate in mist nets at Kakamega was
consistently higher (200-400 %) than at Belalong for a shorter total length of netting (140 m
compared to 213 m). Species Diversity Index (Dm) was lower at Kakamega than at Belalong
(means 0.68 and 0.88 respectively).
110 CLIVE MANN
ACKNOWLEDGEMENTS
Much assistance was received in the field from Jen Elkin and A. Conrad Ozog. Awang
Samhan bin Nyawa, Dr. Joe Charles, Tom Watson and other expedition members contributed
observations. Lord Cranbrook and Dr D.R. Wells contributed ringing data as well as their
observations. Dr C.M. Francis and Dr. D.R. Wells contributed trip reports. Nigel de N.
Winser, Catriona Prebble, Amanda Hardy and the staff of Kuala Belalong Field Studies
Centre were helpful in ways too numerous to mention. John Banks and Lord Cranbrook
helped with the preparation of materials. Above all, I would like to express my sincere
appreciation to the Royal Geographical Society for its support of my field work and my
attendance at the Tropical Rainforest Conference 1993 at Universiti Brunei Darussalam.
LITERATURE CITED
BAKER, J.R., MARSHALL, A.I. & HARRISON, T.H. 1940. The seasons in a tropical rain forest (New
Hebrides). Part 5 Birds (Pachycepha/a). Journal of the Linnean Society (Zoology) 41:50-70
BANKS, E. 1937. The distribution of Bornean birds. Sarawak Museum Journal IV:453-496.
CHASEN, F.N. 1935. A handlist of Malaysian birds. Bulletin Of the Raffles Museum 11:1-389.
FOGDEN, M.P.L. 1972. The seasonality and population dynamics of equatorial forest birds in Sarawak. Ibis
114:307-343.
GIBSON-HILL, C.A. 1952. The apparent breeding seasons of land birds in North Borneo and Malaya. Bulletin
of the Raffles Museum 24:270-294.
MAGURRAN, A.A. 1988. Ecological diversity. Croom Helm.
MANN, C.F. 1985. An avifaunal study in Kakamega forest, Kenya, with particular reference to species diversity,
moult and weight. Ostrich 56:236-262.
MOREAU, R.E. 1936. Breeding seasons of birds in East African evergreen forest. Proceedings of the Zoological
Society. London 1936:631-653.
MOREAU, R.E. 1950. The breeding seasons of African birds. I Land birds. Ibis 92:223-267.
SKUTCH, A.F. 1950. The nesting season of Central American birds in relation to climate and food supply. Ibis
92: 185-222.
SMYTHIES, B.E. 1960. The birds of Borneo, 1st edition. Oliver & Boyd, London & Edinburgh.
SMYTHIES, B.E. 1981. The birds of Borneo, 3rd edition. Sabah Society & Malayan Nature Society, Kota
Kinabalu.
SNOW, D.W. & SNOW, B.K. 1964. Breeding seasons and annual cycles of Trinidad land birds. Zoologica 49:1-
49.
VOOUS, K.H. 1950. The breeding seasons of birds in Indonesia. Ibis 92:279-287.
WARD, P. 1969. The annual cycle of the Yellow-vented Bulbul Pycnonotus goiavier in a humid equatorial
environment. Journal of Zoology. London 157:25-45.
WILLS, J. T. (this volume). The creation of a Geographical Information System for the Kuala Belalong Field Studies
Centre.
Birds of Belaiong forest 111
Appendix 1. Birds recorded in Belalong forest. I = alluvial forest/river banks; 2 = slopes and ridges upto c,46Om;
3 = slopes and ridges above 46Om. N = non forest; W = forested waterways; G = ground; U = understorey; M
= mid layer; C = canopy; A = aerial; # = Palaearctic migrant. Diet [P = fish; C = other vertebrate; I =
invertebrates; F = fruits, seeds, flowers; N = nectar, pollen). Obs - Frequency of recording of species in project
area. @-recorded before 21 February 1991; # = only recorded after 13 October 1991; Arabic numerals = nwnber
of days recorded by author (max. 68) between 21 February and 13 October 1991. Roman nwnerals = months
recorded.
Habitat Dist. Diet Obs Months

I 2 3
Darter Anhinga melanogaster + N P 2 IV,V
Little Green Heron Butorides striatus + N PII 9 II,IX,X
Osprey Pandion haliaetus + N# C #
Blyth's Hawk-Eagle Spizaetus alboniger + + + C C 2 VI
Wallace's Hawk-Eagle S. nanus + C C #
Rufousbellied Eagle Hieraaetus kienerii + + C C I VIII @
Lesser Fish-Eagle lchthyophaga humilis + W P 5 II,V,VIII,IX
Greyheaded Fish-Eagle I. ichthyaetus + W P V
Accipiter sp. Accipiter (?trivirgatus) + C C
Crested Serpent Eagle Spilornis cheela + + + C C/I 16 II-IV, VII-X
Blackthighed Falconet Microhierax fringillarius + C C/I I III
Crested Wood Partridge Rollulus roulroul + G F/I @
Great Argus Pheasant Argusianus argus + + + G F/I 47 II-X, @
Common Sandpiper Actitis hypoleucos + N# I 8 III ,IX ,X
Longtoed Stint Calidris subminuta + N# 1 VIII
Rednecked Phalarope Phalaropus lobatus + N# #
Thickbilled Green Pigeon Treron curvirostra + C F 4 III,IV,VII
Little Green Pigeon T. olax + C F VIII
Mountain Imperial Pigeon Ducula badia + + C F 3 III, V
Emerald Dove Chalcophaps indica + + + G F/I 3 II, III, V
Jambu Fruit Pigeon Ptilonopusjambu + + G F V
Bluerumped Parrot Psittinus cyanurus + + C F 3 III, VIII
Bluecrowned Hanging Parrot Loriculus galgulus + + + C F 5 IV, V, VII, VIII
Moustached Hawk-Cuckoo Cuculus vagans + + VIM I 32 III-V, VIII
Hodgson's Hawk-Cuckoo C. fugax + + + M/C I #
Indian Cuckoo C. micropterus + + + C 12 II, III
Banded Bay Cuckoo Cacomnntis sonneratii + + + M/C 43 II-X
fudonesian Cuckoo C. sepulcralis + + M V
Plaintive Cuckoo C. merulinus + N lIF 3 III, IV
Violet Cuckoo Chrysococcyx xanthorhynchus + + M/C I/F 8 III-V,VII,VIII,X
Little Bronze Cuckoo C. minutillus + N T #
Drongo-Cuckoo Surniculus lugubris + + + M/C I 34 II-V
Raffles's Malkoha Phaenicophaeus chlorophaeus + + + M/C I 4 II, V, VIII
Chestnutbreasted Malkoha P. curvirostris + + M I(C) #
Redbilled Malkoha P. javanicus + + + M I 14 III, IV, VI-VIII
Blackbellied Malkoha P. diardii + M I #
Shorttoed Coucal Centropus rectunguis + G/U I III
Greater Coucal C. sinensis + G/U lIF(C) VIII
Bay Owl Phodilus badius + U/M C/I VIII
Barred Eagle-Owl Bubo sumntranus + + UiM C 2 V, VIII
Buffy Fish-Owl Ketupa ketupu + N/W P/IiC VIII
Brown Wood-Owl Strix leptogrammica + + UiM lIC 3 III, VIII
Brown Hawk-Owl Ninox scutulata + U/M(#) #
Javan Frogmouth Batrachostomus javensis + UI?M 1 VIII
Grey Nightjar Caprimulgus indicus + U# #
Cave Swiftlet Aerodramus sp. + + + N/A 45 II-V, VII, VIII
Whitebellied Swiftlet Collocalia esculenta + + N/A 30 II, III, V, VIII-X
Brown Spinetail Hirundapus giganteus + A 2 IT, IV
Whitethroated Spinetail H. cautiacutus + N/A 1# 2 III
Silverrwn2ed S2inetail Rhae.hidura leuc°e.l:.8,ialis + + + A I 28 II-V, VII-X
112 CLIVE MANN
Appendix 1 (continued). Birds recorded in Belalong forest.

1 2 3
Whiskered Tree-Swift Hemiprocne comata + + CIA 3 II, III
Greyrumped Tree-Swift H. longipennis + + ?C/A I 6 III, VII, VIII
Rednaped Trogon Harpactes kasumba + + + UIM I 19 II,IV -VI, VIII,
Diard's Trogon H. diardii + + U/M lIF 3 III, VI, VIII
Scarletrumped Trogon H. duvauceli + + + u/M I 18 II-VI ,X
Banded Kingfisher Lacedo pulchella + + + M I 12 III-IV ,VIII
Storkbilled Kingfisher Pelargopsis capensis + N PII 18 V ,VII-IX
Blueeared Kingfisher Alcedo meninting + W P II
Bluebanded Kingfisher A. euryzona + W P/lIC 49 III,V-X
Rufousbacked Kingfisher Ceyx erithacus + + W/U P/I 3 VI,VIII,X
Redbearded Bee-eater Nyctiornis amicta + + + M/C I 6 III,IV
Whitecrowned Hornbill Berenicornis comatus + + + C F/C 6 II, III, VII
Black Hornbill Anthracoceros malayanus + + C F/lIC VII
Bushycrested Hornbill Anorrhinus galeritus + + + C F/I 25 II-VIII
Wreathed Hornbill Rhyticeros undulatus + + + C FII 20 II-V, VIII, X
Wrinkled Hornbill R. corrugatus + + + C F/I 16 III, IV, VIII, X
Rhinoceros Hornbill Buceros rhinoceros + + + C F/C 52 II-X
Helmeted Hornbill Rhinoplax vigil + + + C F/C 34 II-X
Brown Barbet Calorhamphus fuliginosus + + + C F/I 3 III
Goldwhiskered Barbet Megalaima chrysopogon + + + C F 27 II-V, VII-VIII
Redthroated Barbet M. mystacophanos + + + M/C FII 44 II-X
Yellowcrowned Barbet M. henricii + + + C F 33 II-VIII
Blueeared Barbet M.australis + + + C F 44 II-VIII
Malaysian Honeyguide Indicator archipelagicus + ?u/M #
Rufous Piculet Sasia abnormis + + U/M 2 IIIIIV
Checkerthroated Woodpecker Picus mentalis + + + U/M I 3 III, VIII
Crimsonwinged Woodpecker P. puniceus + M I #
Greycapped Woodpecker Picoides canicapillus + + M/C I III
Grey-&-Buff Woodpecker Hemicircus concretus + + M/C lIF III
Rufous Woodpecker Micropternus brachyurus + U/M 7 II, III
Olivebacked Woodpecker Dinopium rafflesii + + U/M 2 III
Buffnecked Woodpecker Meiglyptes tukki + + u/M 2 VI, VIII
Maroon Woodpecker Blythipicus rubiginosus + + + U/M 7 III, IV, VII, IX
Whitebellied Woodpecker Dryocopus javensis + + M/C 3 III, VII
Great Slaty Woodpecker Mulleripicus pulverulentis + + M/C I 6 III, VII, VIII
Orangebacked Woodpecker Chrysocolaptes validus + + M/C I III
Green Broadbill Calyptomena viridis + + + U/M/?C F 21 III, IV, VI-VIII
Black-&-Red Broadbill Cymbirhynchus macrorhynchus + M/C I(F/P) 8 II, V, VIII, IX
Black-&-Yellow Broadbill Eurylaimus ochromalus + + + C lIF 58 II-X
Banded Broadbill E. javanicus + + + M/C I 46 II-VIII, X
Dusky Broadbill Corydon sumatranus + C I(C) 1 III
Garnet Pitta Pitta granatina + + G lIF 8 II, V, VIII, IX
Banded Pitta P. guajana + G @
Blueheaded Pitta P. baudi + G 1 IX
Barn Swallow Hirundo rustica + N# 2 III#
Redrumped Swallow H. daurica + N# #
Pacific Swallow H. tahitica + N #
Asian House Martin Delichon dasypus + N# #
Grey Wagtail Motacilla cinerea + N# I 2 II, III
Lesser Cuckoo-Shrike Coracina jimbriata + + C lIF 7 III, V, VII
Blackwinged Flycatcher-Shrike Hemipus hirundinaceus + + C I 3 III, VII
Large Wood-Shrike Tephrodornis virgatus + + + C I III?#
Scarlet/Fiery Minivet Pericrocotus jlammeusligneus + + C IfF 5 III, V-VII
Green lora Aegithina viridissima + + M/C I 3 III, VII
Lesser Green Leafbird Chloropsis cyallopogoll + + + M/C F/I 8 III, V, VII, VIII
Greater Green Leafbird C. sonnerati + + + C F/! 4 IV,IX
Blue-winged Leafbird C. cochinchinensis + + C FII 4 III, VII, VIII
Fairy Bluebird Irena l!.uella + + + C FII 10 II, III, V!-VIII, X
Appendix 1 (continued). Birds recorded in Belalong forest.

1 2 3
Greybellied Bulbul Pyenonotus cyaniventris + + C F/I 3 II, VII
Blackheaded Bulbul P. atrieeps + C F/I 1 VIII
Black-&-White Bulbul P. melanoleueos + M/C F/I 5 IV-VI, VIII
Redeyed Bulbul P. brunneus + + + VIM F/I 10 III, VII-IX
Spectacled Bulbul P. erythrophthalmos + + + VIM F/I 31 II, III, V-VIII
Creamvented Bulbul P. simplex + + + U1M F/I 3 VI, VIII
Puffbacked Bulbul P. eutilotus + + M F/! 15 VII-X
Strawheaded Bulbul P. zeylanicus + NIV F/! I VIII
Greycheeked Bulbul Criniger bres + + + VIM F/I 29 II-X
Yellowbellied Bulbul C. phaeocephalus + + VIM F/! 41 III, V-IX
Finsch's Bulbul C. jinsehii + + + VIM F/! 7 II, III, VII-IX
Hairybacked Bulbul Hypsipetes eriniger + + + U1M F/! 26 II-VIII
Streaked Bulbul H. malaeeensis + + M/C F/! 8 III, V!-VIII
Buffvented Bulbul H. charlottae + + VIM/C F/! 6 II, III, VIII
Hookbilled Bulbul Setomis criniger + VIM F/I #
Siberian Blue Robin Erithaeus cyane + + G# F/I #
Rufoustailed Shama Copsychus pyrrhopygus + + GIV I 25 II-VIII
Whiterumped Shama C. malabaricus + + + GIV 37 II-X
Chestnutnaped Forktail Enicurus rujicapil/us + WIG IIC 10 III, VII-IX
Whitecrowned Forktail E. leschenaulti + WIG I V
Blackcapped Babbler Pellomeum capistratum + + G 13 II-V, VII, IX
Shorttailed Babbler Triehastoma malaccense + + + GIV 23 II-X
Whitechested Babbler T. rostratum + + GIV 52 III-X
Ferruginous Babbler T. bicolor + + ?GIV 8 II-V!, VIII
Horsfield s Babbler T. sepiarium + + + GIV 10 II-V, VIII
Rufouscrowned Babbler Malacopteron magnum + + + VIM lIF 68 II-X
Scalycrowned Babbler M. cinereum + + VIM IfF 22 II, III, V, VII-X
Moustached Babbler M. magnirostre + + + VIM 21 III-VIII
Sootycapped Babbler M, affine + + U1M 10 II, III, V, IX
Whitethroated Babbler M. albogulare + 'lGIV II
Chestnutbacked Scimitar-Babbler Pomatorhinus montanus + + + VIM F/! 28 II-X
Striped Wren-Babbler Kenopia striata + + + GIV 1 18 II-IV, VI-IX
Fluffybacked Tit-Babbler Maeronous ptilosus + + V 59 II-X
Blackthroated Babbler Staehyris nigricollis + + V 25 II-VI, VIII, IX
Chestnutrumped Babbler S. maeulata + + + VIM 68 II-X
Chestnutwinged Babbler S. erythroptera + + + VIM 68 II-X
Whitenecked Babbler S. leueotis + V III
Greyheaded Babbler S. poliocephala + V 4 III, IX, X
Rufousfronted Babbler S. rujifrons + + + M/C 6 IV, V, VII
Brown Fulvetta Alcippe brunneieauda + + + VIM 46 III-X
Whitebellied Yuhina Yuhina zantholeuea + + + M/C IIF 10 III, IV, VII, VIII
Flyeater Gerygone sulphurea + + C III
Arctic Warbler Phylloseopus borealis + + M/C# IV#
Yellowvented Prinia Prinia jlavi ventris + N III
Rufoustailed Tailorbird Orthotomus serieeus + + + V 57 II-X
Ashy Tailorbird O. rujieeps + V 2 III
Darknecked Tailorbird O. atrogularis + + + V 13 II ,III ,V, VII, VIII
Spotted Fantail Rhipidura perlata + + + ?VIMIC 18 II-V, VII, VIII, X
Greyheaded Flycatcher Culicicapa eeylonensis + + + VIM 31 VI-IX
Asian Brown Flycatcher Museicapa dauuriea + + + M/C(#) 19 II-IV
Verditer Flycatcher Eumyias thalassina + VIM V
Largebilled Blue Flycatcher Cyomis caerulata + VIM 31 III, V-VII, X
Bornean Blue Flycatcher C. superba + VIM 2 IV, V
Malaysian Blue Flycatcher C. tureosa + VIM 3 V, IX
Pale Blue Flycatcher C. unicolor + M/C #
Greychested Flycatcher Rhinomyias umbratilis + + + VIM II II, III, V-VIII
Olivebacked Flycatcher R. olivaeea + VIM 1 VII
Rufoustailed Fl~cather R. rut!:.cauda + + VIM 3 III
114 CLIVE MANN
Appendix I (continued). Birds recorded in Belalong forest.
Habitat Dis!. Diet Obs Months

I 2 3
Rufouswinged Monarch Philentoma pyrrhoptera + +
+ U/M 4 II, ill
Maroonbreasted Monarch P. velatum + +
+ M 10 ill, V-VIII
Blacknaped Monarch Hypothymis azurea + +
+ UIM 28 III-V, VII-IX
Asian Paradise Flycatcher Terpsiphone paratiisi + +
+ M/C 42 II-Vill
Velvetfronted Nuthatch Sitta frontalis + C I 2 ill, VI
Yellowrumped Flowerpecker Prionochilus xanthopygius + + + U/M F/IIN 15 ill, V-VIII, X
Crimsonbreasted Flowerpecker P. percussus + + UIM F/IIN VII
Yellowbreasted Flowerpecker P. maculatus + + U/M F/II?N 43 II-VI, VIII
Yellowvented Flowerpecker Dicaeum chrysorrheum + VIM F/II?N 5 III, V
Brownbacked Flowerpecker D. everetti + VIM F/I N
Orangebellied Flowerpecker D. trigonostigma + VIM F/I VIII
Plain Flowerpecker D. conc%r + VIM I #
Plain Sunbird Anthreptes simplex + + + VIM liN 13 III-VIII, X
Redthroated Sunbird A. rhodolaema + C F/I/N I ill
Rubycheeked Sunbird A. singalensis + + M/C F/I/N 6 III, VI-VIII
Scarlet Sunbird Aethopyga mystacalis + + VIM F/N(I) 3 III, VII
Crimson Sunbird A. siparaja + VIM I VIII
Purplenaped Sunbird Hypogramma hypogrammica + + + U I 15 II, III, VI-VIII, X
Little Spiderhunter Arachnothera longirostra + + + VIM liN 25 II-X
Thickbilled Spiderhunter A. crassirostris + + + ?UIM I 3 V, VIII, X
Yelloweared Spiderhunter A. chrysogenys + + + C F/I 2 III
Spectacled Spiderhunter A. jlavigaster + C ?F/I I IX
Greybreasted Spiderhunter A. ajfinis + + + UIMI?C F/I 2 VI, X
Everett s White-eye Zosterops everetti + C F/I VII
Bornean Bristlehead Pityriasis gymnocephala + + C I I III
Dusky Munia Lonchura fuscans + N F 25 I1-N, VII, Vill
Philippine Glossy Starling Aplonis panayensis + N F/I 2 VII
Hill Myna Gracula religiosa + C F/I 2 III
Greater Rackettailed Drongo Dicrurus paratiiseus + + + M/C 22 III, V-IX
Crowbilled Drongo D. annectans + + + M/C# 2 I1I,N
Bronzed Drongo D. aeneus + + + M/C I 3 III, VIII, X
Darkthroated Oriole Orio/us xanthonotus + + + C F/I 22 III, V-X
Black Magpie Platysmurus leucopterus + + M/C F/IIC 8 II, III, V, IX, X
Crested Jay Platy/ophus ga/ericulatus + + + VIM 2 III, VII
Slenderbilled Crow Corvus enca + C C/I V
Appendix 2. Breeding records from Belalong forest, Brunei Darussalam.
SEecies Date Localitx Comments

Spizaetus nanus 12-22/1/92 near KBFSC on nest; ?young
Alcedo euryzona 26/3/91 near KBFSC feeding young
Mulleripicus pulverulentus 1917/91 below E. ridge young in nest
Corydon sumatranus 25/3/91 below summit fledgeling
Hypsipetes criniger 16/1/92 Busiri nest material
Hypsipetes criniger 16/1/92 Bujat brood patch
Pomatorhinus montanus 22/2/91 near Bujat fledgeling
Pomatorhinus montanus 19/8/91 near summit nest; 3 eggs
Stachyris erythroptera 23/2/91 KBFSC fledgeling
Stachyris erythroptera 16/1/92 Bujat brood patch
Alcippe brunneicauda 13/1/91 Bujat just fledged
Orthotomus sericeus 23/2/91 KBFSC 2 fledgelings
Rhipidura perlota 16/1/92 Bujat brood patch
Hypothymis azurea 31/3/91 W. ridge on nest
Hypothymis azurea 18/7191 KBFSC on nest
Terpsiphone paradisi 12/6/91 W. ridge building
Cyomis turcosa 16/5/91 Sg. Babi nest material
Anthreptes simplex 8/3/91 E. ridge young in nest
Aracimothera longirostra 16/1/92 Bujat brood patch
Arachnothera chrysogenys 23/3/91 KBFSC young in nest
116 CLIVE MANN
Appendix 3. Birds netted in project area.
Bujat Total
Crested Serpent Eagle Spilornis cheela I I
Rednaped Trogon Harpactes Iwsumba I I
Banded Kingfisher Lacedo pulchella 2 2
Rufousbacked Kingfisher Ceyx erithacus I I
Brown Barbet Calorhamphus fuliginosus I
Goldwhiskered Barbet Megalaima chrysopogon I I
Malaysian Honeyguide Indicator archipelagicus I I
Rufous Piculet Sasia abnormis 2 2
Buffnecked Woodpecker Meiglyptes tukki 3 3
Green Broadbill Calyptomena viridis 9 9
Blueheaded Pitta Pitta baudi 0 I
Barn Swallow Hirundo rustica 0 5
Redrumped Swallow H. daurica 0 I
Asian Fairy Bluebird Irena puella I I
Black-&-White Bulbul Pycnonotus melanoleucos 7 7
Redeyed Bulbul P. brunneus I I
Spectacled Bulbul P. erythrophthalmos 10 10
Creamvented Bulbul P. simplex 3 3
Puffbacked Bulbul P. eutilotus I I
Greycheeked Bulbul Criniger bres 9 9
Yellowbellied Bulbul C. phaeocephalus 2 5
Hairybacked Bulbul Hypsipetes criniger 9 9
Streaked Bulbul H. malaccensis 3 3
Buffvented Bulbul H. charlottae I 1
Hookbilled Bulbul Setornis criniger 2 2
Siberian Blue Robin Erithacus cyane I I
Rufoustailed Shama Copsychus pyrrhopygus 5 5
Whiterumped Shama C. malabaricus I 2
Chestnutnaped Forktail Enicurus rujicapillus 0 2
Blackcapped Babbler Pellorneum capistratum 2 3
Shorttailed Babbler Trichastoma malaccense 4 5
Whitechested Babbler T. rostratum 0 1
Ferruginous Babbler T. bicolor I I
Horsfield s Babbler T. sepiarium 0 I
Rufouscrowned Babbler Malacopteron magnum 2 2
Scalycrowned Babbler M. cinereum 9 9
Moustached Babbler M. magnirostre 6 6
Sootycapped Babbler M. affine 0 2
Fluffybacked Tit-Babbler Macronous ptilosus 2 3
Blackthroated Babbler Stachyris nigricollis 0 2
Chestnutwinged Babbler S. erythroptera 12 12
Brown Fulvetta Alcippe brunneicauda 3 3
Arctic Warbler Phylloscopus borealis 1 I
Rufoustailed Tailorbird Orthotomus sericeus 0 I
Spotted Fantail Rhipidura perlata 5 5
Verditer Flycatcher Eumyias thalassina 1 I
Bornean Blue Flycatcher Cyornis superba 3 3
Greychested .Flycatcher Rhinomyias umbratilis 5 5
Rufouswinged Monarch Philentoma pyrrhoptera 3 3
Maroonbreasted Monarch P. velatum I 1
Blacknaped Monarch Hypothymis azurea 5 5
Yellowbreasted Flowerpecker Prionochilus maculatus 7 8
Crimsonbreasted Flowerpecker P. percussus 1 I
Yellowrumped Flowerpecker P. xanthopygius 9 9
Plain Sunbird Anthreptes simplex 7 7
Little Spiderhunter Arachnothera longirostra 22 24
Thickbilled Spiderhunter A. crassirostris 1 1
Greybreasted Spiderhunter A. affinis 2 2
Total 192 218
Terrestrial Ants (Hymenoptera: Formicidae) of Poring, Kinabalu Park,
Sabah
MARYATI MOHAMED, AZIZAH HUSSEIN and ARBAIN KADRI
Biology Department, Faculty of Science and Natural Resources, Locked Bag 62, 88996, Kota
Kinabalu, Sabah, Malaysia
ABSTRACT. Collections of ants were carried out for a 7 month period in Poring, Kinabalu Park,
Sabah. In total 46,999 individual ants were collected, representing 6 subfamilies. These
subfamilies were represented by 52 genera and 125 morphospecies. Percentages of species in the
various subfamilies were as follows: Myrmicinae 44.8%; Formicinae, 24.8%; Ponerinae, 14.4%;
Dolichoderinae, 13.6%; Dorylinae, 1.6% and Pseudomyrmecinae, 0.8%. The 2 genera with the
highest number of species were Pheidole and Dolichoderus with 12 each. Ninety-five species were
collected by baited pitfall traps and 71 species were collected manually. Using a one way ANOV A
no significant difference in the number of species was found between the 2 methods.
KEY WORDS: Borneo, ants, species richness, habitat quality, sampling techniques, diversity.
INTRODUCTION
The abundance and diversity of ants in the south east Asian tropical rain-forests have not
been fully investigated. To understand these ecological aspects, as well as assist in the
process of building up a reference collection of ants for Malaysia, we have been collecting
ants since 1986, both in the peninsular as well as in Sabah, the North Borneo State of
Malaysia. Previous collections have been concentrated in the lowland dipterocarp forests. The
actual roles of ants in the tropical forest ecosystem are not fully understood. Their presence
in all types of tropical forests are obvious. Burghout et al. (1992) and Wong (1984), working
on 2 different locations found that ants were consistently the most abundant arthropods in the
Malaysian forests. One of the roles played by ants thought to be highly plausible is the
recycling of nutrient into the forest soil. Being abundant and diverse, ants are a potentially
good bioindicator of the status of the forests. Greenslade & Greenslade (1984) emphasised
the potential of invertebrates and in particular ants as bioindicators of habitat quality. This
was emphasised by Majer (1987). It was our intention to determine the best method for
collection of ants and the best ant species to be used as indicators not only to assess forest
quality but also to serve as indicators of changes in the Malaysian environment.
METHODS
Our sampling plot was located at Poring Hot Spring in Kinabalu National Park, Sabah, at an
altitude of 550 m (Figure 1). Rainfall for the area from 1981 to 1990 ranged from 1220.0
to 2757.8 mm per annum. Day temperatures ranged from 25-31°C. A transect, parallel to
the canopy walkway, was chosen for the location of five 10 m x 10 m plots. Vegetation
along this transect was composed primarily of species of the families Euphorbiaceae,
Arecaceae, Olecaceae, Meliaceae and Dipterocarpaceae. Seven 3-4 day collecting periods
were spaced 3-4 weeks apart over a period of 7 months.
117
D. S. Edwards et al. (eds.), Tropical Rainforest Research - Current Issues. 117-123.
118 MARYATI MOHAMED, AZIZAH HUSSEIN AND ARBAIN KADRI
.............................
if?'tn2\·\iiiiJ• • ·.X·
.. / . ... ::::::::::::: T~.fI! .KI~itv:- :.-:.-:::: :;:
/:::::::............: ................. : ...... : .• : .......... ::: .. :: .. : ...... : .. : ../
.r:::..:..::......... .............. :. :::. :::::::::::::::::::::::::::;
Vi .•••••••••••••••••••••••••••••••••••••••••••••..••••••.
\ \.~:; ~h;·:?J
~
BRIVERS
BMAIN ROAD
1;·--'1 BOUNDARIES OF KINABALU PARKS
t::::.1 PARK AREA
Figure 1. Map showing the location of Poring Hot Springs, Kinabalu Park, Sabah, Malaysia.
Two sampling techniques were employed. First, ants were collected manually from all the
5 plots. Collection was carried out by one person for about 5-6h from 9am to 4pm, for 2
consecutive days during each collection trip. Ants on the forest floor, in the forest litter and
between crevices of rocks and bark, on tree trunks and undergrowth up to the eye level were
collected and put in vials with 75 % ethanol, and brought back to the laboratory for
identification. The second method, baited pitfall trapping, was a modified version of the
common pitfall traps used to collect forest floor arthropods. Plastic containers 7.5 cm
diameter and 12 cm high were placed in holes dug into the ground so that the lip of the
container was at ground level. Five containers were placed about I m apart at random in each
plot. Five different baits were used in each plot, one in each container. The different baits
used were canned tuna in oil (marine protein source), beefburger (Ramly brand, animal
protein source), soft white tau-foo (plant protein source), sweet (Trebor, sour plum flavour),
and pebbles as control. Each bait, about I cm square, was wrapped tightly in muslin cloth
and secured with cotton thread. The baits were positioned about 2/3 of the way down into
the containers from a wire rod. A soapy solution was put into the bottom of each container
(ratio of water:washing liquid is 50: 1). The addition of dish washing liquid was necessary
to lower the surface tension of the water so that fallen ants would sink to the bottom of the
container. The traps were left from 4pm until 7am the next morning.
Trapped ants were separated from the soap solution using a sieve. The ants were then
rinsed in clean water, picked out and preserved in 75% ethanol and placed in coded vials.
The trapping was repeated once during each visit. The two methods mentioned above were
as in Maryati (1992).
Identifications were based on keys in Bolton (1990) and Holldobler & Wilson (1990). Ants
were identified to generic level and designated as species 1, 2,3, etc. Comparison was made
in terms of species richness to see which collecting method was most successful in terms of
numbers of species of ants caught.
Ants of Poring, Kinabalu National Park, Sabah 119
Sweet 21.1%
Beef burger 21.4%
Tuna 20.4% Tau-foo 18.2%
Pebble {contr~ 18.9%
Figure 2. Percentage of species collected by various baits during the seven collecting periods.
RESULTS
Table 1 shows the overall results of the collection. Genera with 10 or more species were
confined to Pheidole and Dolichoderus. Those with between 5 and 9 species included
Pheidologeton, Tetramorium and Polyrhachis. As seen in Table 2, Myrmicinae included
almost 45 % of the species sampled at Poring, Formicinae 25 %, and Ponerinae and
Dolichoderinae each about 14%. Dorylinae and Pseudomyrmecinae made up only about 2%
of the species sampled.
Table 3 shows species composition for subfamilies for each of the 7 collections,
disregarding the collecting method. Myrmicinae had the highest representation of species
followed by Formicinae or Dolichoderinae (Table 3 I-VII). For most of the sampling times
the number of species of ponerines and dolichoderines were almost the same except for
collections I, III and VII when the number of species of Ponerinae collected were about 1.5
to 2 times greater than species of Dolichoderinae. The ratio of Myrmecinae to Formicinae
was about 1.5: 1 to 3: 1, while the ratio of Formicinae to Ponerinae or Dolichoderinae was
about 1.4: 1 to 2.7: 1.
Table 4 compares the efficiency of the various baits in pitfall trapping programme. Using
a one-way ANOV A it was found that there was no statistical difference between the
efficiency of the baits. Even the control (baited with pebbles) was seen to be trapping as
many ants as other kinds of baits, or even more (Figure 2).
DISCUSSION
Composition of terrestrial ants of Poring

Collections made in lowland forests such as Kuala Lompat, Pahang (Maryati 1991) and
Gunung Danum, Sabah (Maryati 1989) showed a similar pattern of composition to that of the
present study, with Myrmicinae and Formicinae being the 2 most highly represented
subfamilies followed by Ponerinae or Dolichoderinae. The remaining subfamilies represent
only 5-10% of the species composition or even less. Collections from Bukit-bukit Tawau,
Sabah, showed a higher representation of species in the subfamily Formicinae compared with
Table 1. List of subfamilies and genera of ants collected by pitfall trapping and manually, at Poring. Numbers
indicate the number of the morphospecies for each genus.
Subfamily Genus Pitfall traps Manual Total no. of

morphospecies
PONERINAE Diacamma 1
Discothyrea 1
Harpegnathos 1 1
Hypoponera 1 1
Leptogenys 1,2,3 1 3
Mystrium 1,2 3 3
Odontomachus 1 1,2 2
Odontoponera 1 1 1
Pachycondyla 1,2,3 3
Platythyrea 1 1
Prionopeita 1 1
Sub-total 15 8 18
DORYLINAE Aenictus 1,2 1 2
Sub-total 2 1 2
PSEUDOMYRMICINAE Tetraponera 1 0 1
Sub-total 1 0 1
MYRMICINAE Acanthomyrmex 1,2 2
Aphaenogaster 1 1 1
Cataulacus 1 1
Crematogaster 1,2,3 1,2,3,4 4
Lophomyrmex 1,2 1 2
Mayrieila 1,2 2
Meranopius 1 1
Myrmecina 1 2 2
Oligomyrmex 1,2,3 3
Paedaigus 1 1
Paratopula 1 1
Pentastruma 1,2 1 2
Pheidole 1-10 1-8, 11, 12 12
Pheidoiogeton 1-7 1 7
Proatta 1 1
Quadrustruma 1 1
Rhopaiomastix 1 1
Soienopsis 1 1
Strumigenys 1,2 2
Tetramorium 1-6 1-3 6
Vollenhovia 1 1
Wasmannia 1 1
Willowsieila 1 1
Sub-total 51 23 56
DOLICHODERINAE Dolichoderus 1-4 1-12 12
lridomyrmex 1 1 1
Tapinoma 1,2,3 1-4 4
Sub-total 8 17 17
FORMICINAE Acanthoiepsis 1 1 1
Acropyga 1 1
Anoplolepis 1 2 2
Camponotus 1,2,3 1,2,3 3
Colobopsis 1 1,2 2
Echinopla 1 1
Euprenoiepis 1,2,3 1 3
Lasius 1,2 2
Myrmoteras I 1
Paratrechina 1-5 1,2 5
Plagiolepis 1 1
Polyrhachis 1 1-6 6
Prenolepis 1 2,3 3
Sub-total 18 22 31
TOTAL NO. OF SPECIES 95 71 125
Table 2. Species percentage composition for sub-families of ants collected at Poring.
Subfamily Percentage
Ponerinae 14.4
Dorylinae 1.6
Pseudomyrmicinae 0.8
Myrmecinae 44.8
Dolichoderinae 13.6
Formicinae 24.8
Total 100
Table 3. Number of species of ants sampled during the seven collections (I-VII) in each subfamily.
Collection
II III IV V VI VII
Subfamily
Ponerinae 9 10 6 5 9 8 11
Dorylinae 0 0 0 0 0
Pseudomyrmecinae 0 0 0 0 0 0
Myrmicinae 34 30 25 29 35 32 31
Dolichoderinae 4 7 6 6 10 7 7
Formicinae 13 13 18 13 17 13 18
Total 60 60 55 54 72 60 68
Table 4. Number of species of ant sampled for the seven collections based on baits used in the pitfall traps.
Collection
II III IV V VI VII
Kinds of baits
Control (pebble) 34 45 21 40 37 23 25
Tau-foo 28 37 16 41 35 28 27
Beefburger 50 43 36 46 45 33 39
Tuna in oil 45 42 39 43 45 35 44
Sweet 30 43 33 45 47 33 41
Total 187 210 145 215 209 152 176
Myrmicinae, and a much lower representation of Dolichoderinae compared with Ponerinae

(Maryati 1991). From a highland area such as Sayap-Kinabalu Park (Maryati 1993),
Ponerinae had a higher representation of about 33 %, a feature also associated with collections
made from very disturbed forests such as Bangi forest, Selangor (Maryati 1990). These
patterns suggest firstly that collections from undisturbed primary forest would give a different
species composition from those collected from extreme habitats such as highland or very
disturbed forests. Greenslade & Greenslade (1977) also noted that a habitat which was
disturbed by human activities would have a lower diversity of ants compared to undisturbed
habitats. Majer (1983) used ants as bioindicators of habitat quality. Secondly, in the tropical
rain forests, subfamilies such as Myrmicinae, Formicinae, Ponerinae and Dolichoderinae are
big subfamilies, whereas the remaining subfamilies are small.
Analysing the species present, DoLichoderus had a relatively high number of species (12).
Our previous collections had never yielded this many species of DoLichoderus. Further work
is needed to investigate the reason for such a high number of species. There may also be
an advantage in looking into the biology and ecology of this group since one of the species,
Dolichoderus thoracicus, has been successfully used in cocoa plantations to control
infestations of the mirid Helopeltis theobromae (Khoo & Chung 1989).
Pheidole was represented by 12 species, a similar number to that found in many of the
previous collections, especially those involving baited pitfall trapping. The genus Polyrhachis
was also well represented, with 6 morphospecies collected. This genus is fairly common in
tropical rain forest and shows high diversity. During the Bukit-bukit Tawau expedition 9
species of Polyrhachis were collected (Maryati 1991).
As shown in Table 3, the species composition of the ants collected did not vary much with
time.
Comparison of collecting methods

Genera such as Pheidole could be collected efficiently by either of the two methods
described. This was possible because members of this genus inhabit forest litter and the
upper part of the forest soil. Crematogaster could also be caught efficiently by either method.
This genus is generally arboreal, making nests on living plant parts such as branches or
stems. Some even live inside plants such as within stems of Macaranga (Fiala et al. 1989).
A species of Tetraponera was also collected from the pitfall traps in spite of this genus being
largely arboreal. Oligomyrmex, Tetramorium and other small sized myrmicines such as
Myrmecina, Quadrustruma and Solenopsis were mainly caught using baited pitfall traps, and
not manually. On the other hand, as in previous studies, Polyrhachis could best be collected
manually and perhaps by fogging (Maryati 1992).
These observations suggest that for inventory work, when as many species of ants as
possible are required to be collected, all kinds of collecting techniques should be employed,
as discussed by Maryati (1992).
There was no effect of the type of bait on the success of ant collection. According to
Greenslade & Greenslade (1971), in areas with high ant diversity any advantage of a
particular bait was marginal and the use of unbaited pitfall traps was recommended.
ACKNOWLEDGEMENTS
We wish to thank Sabah Parks for allowing this research to be conducted at Poring Hot
Springs, Kinabalu National Park. This research was made possible by IRPA grant
04-07-03-007 managed by Prof. Dr. A. Latif Mohamad. The assistance of Mr. Sukup Akin
of the Botany Section of Biology Department of the Faculty of Science and Natural
Resources in the identification of plants around the sampling transect is also acknowledged.
LITERATURE CITED
BURGHOUT, T., ERNSTING, G., KORTHAL, G.W., & VRIES, T.H. 1992. Litterfall, leaf litter decomposition
and litter invertebrates in the primary and selectively logged dipterocarp forest in Sabah, Malaysia. In Marshall,
A.G. & Swaine, M.D. (eds). Tropical Rain Forest: Disturbance and Recovery. The Royal Society, London.
FIALA, B., MASCHWITZ, U., PONG, T.Y., & HELBIG, J. 1988. Studies of a South East Asian Ant-Plant
Association: Protection of Macaranga trees by Crematogaster bomeensis. Oecologia 7:463-470.
GREENSLADE, P. & GREENSLADE, P.J.M. 1971. The use of baits and preservatives in the pitfall traps. Journal
of the Australian Entorrwlogical Society 10:253-260.
GREENSLADE, P. & GREENSLADE, P.J.M. 1977. Some effects of vegetation cover and disturbance on a
tropical ant fauna. lnsectes Sociaux 2:163-182.
GREENSLADE, J. & GREENSLADE, P. 1984. Invertebrates and environmental assessment In Environmental and
Planning. Department of Environment and Planning, Australia, Journal No.3.
HOLLDOBLER, B. & WILSON, E.O. 1990. The Ants. Belknap Press of Harvard University Press, Cambridge,
Mass.
KHOO, K.C. & CHUNG, G.F. 1989. Use of black cocoa ant to control mirid damage in cocoa. The Planter
65:370-383.
MAJER, J. D. 1983. Ants: Bioindicator of Minesite Rehabilitation, Land Use and Land Conservation. Environmental
Management 7:375-383.
MAJER, J.D. 1987. Invertebrates as indicators for management. In Saunders, D.A., Arnold, G.W., Burbridge,
A.A. & Hopkins, A.J .M. (eds). Nature Conservation: The Role of Native Vegetation. Surrey Beatty & Sons, Pty.
Ltd.
MARYATI, M. 1989. The ants (Formicidae: Hymenoptera) of Gunung Danum, Lahad Datu, Sabah. Report of
Gunung Danum Expedition.
MARYATI, M. 1990. Fauna dan kelakuan semut (Forrnicidae: Hymenoptera). Kawasan Ekologi Hutan Bangi
Ekologi dan Ekologi Hutan Bangi, Kumpulan Kertas kerja 14. Universiti Kebangsaan Malaysia.
MARYATI, M. 1991. A preliminary list of ants (Formicidae: Hymenoptera) of Bukit-bukit Tawau, Sabah. Report
of Bukit-bukit Tawau Expedition.
MARYATI, M. 1992. Penggunaan Kaedah yang sesuai bagi pengumpulan semut (Formicidae: Hymenoptera).
Simposium Sumber Alam Kebangsaan Pertama.
MARYATI, M. 1993. Terrestrial ants (Formicidae: Hymenoptera) of Sayap-Kinabalu Park, Sabah. Report of
Sayap-Kinabalu Park, Sabah Expedition.
WONG, M. 1984. Understorey Foliage Arthropods in the Virgin and Regenerating habitats ofPasoh Forest, West
Malaysia. The MalaYsian Forester 47:43-69.
Temporal and spatial patterns of butterfly diversity in a lowland tropical
rainforest
A. G. ORR I and C. L. HAEUSER2
IDepartment of Biology, Universiti Brunei Darussalam, Bandar Seri Begawan 2028, Brunei
Darussalam
2Zoologisches Forschungsinstitut und Museum Alexander Koenig, Adenauerallee 150-164,
D-53113 Bonn, Germany
ABSTRACT. In nearly two years of regular sampling 324 species of butterfly were recorded from the
lowland rainforest within one kilometer of the Kuala Belalong Field Studies Centre (KBFSC), Batu Apoi
forest reserve, Temburong, Brunei Darussalam. This represents over one third of the Bornean fauna and
perhaps over half the total fauna of the area. Records are biased towards the more conspicuous
Papilionidae and Pieridae with relatively smaller proportions ofNymphalidae, Lycaenidae and Hesperiidae
being represented. Samples taken at approximately monthly intervals indicated relatively slight variation
in species richness. However there was a steady turnover in species composition, presumably the result
of changing species abundance relationships as most of the more common species experienced substantial
and irregular fluctuations in numbers. Some closely related species pairs showed phenological
dissociation in abundance, but many were phenologically unrelated or were positively associated. There
was no evidence for seasonal trends in community composition or in the phenology of any particular
species. While many species ranged freely throughout the area, some were more or less restricted to
either riverbanks and open areas, the forest understorey, or the canopy, subcanopy and tops of major
ridges. Certain Lycaenidae and Hesperiidae were extremely localized, being sometimes restricted to an
area of a few hundred m', presumably in the vicinity of their larval hostplants.
KEYWORDS: Borneo, butterflies, diversity, phenology, seasonality, species-richness, species-turnover.
INTRODUCTION
Approximately 950 butterfly species are recorded from the island of Borneo, and as the fauna
is fairly well known, the true figure is probably not more than one thousand species (Corbet
& Pendelbury 1992, Otsuka 1988, Seki et at. 1991a, 1991b). Of these, the great majority
have been recorded in North Borneo, and with the exception of a handful of high altitude
species, most could potentially occur in Brunei Darussalam, as it includes within its
boundaries most of the forest types known from the island. The extent to which butterfly
diversity is related to vegetational diversity is uncertain; previous studies suggest that most
butterfly species are to be found in lowland mixed dipterocarp forest (Cassidy 1982, 1985,
Holloway 1984), but even within this formation there is considerable floristic heterogeneity
relating to edaphic and topographic factors (Ashton 1964), which might influence the
composition of butterfly communities. The present study attempts to measure alpha diversity
of butterflies within an edaphically and floristically uniform area of lowland rainforest,
especially in an area of approximately one square kilometer surrounding the Kuala Belalong
Field Studies Centre (KBFSC).
Overall diversity, and especially measurable or perceived diversity, is expected to have
significant temporal and spatial components. We examine in particular the effects of
seasonality and species turnover, and attempt to assess the importance of microdistributional
patterns, within the habitat surveyed, on the measurement of diversity. The phenology of
selected species and the extent of temporal association or disassociation between closely
125
D. S. Edwards el al. (eds.), Tropical Rainforest Research - Current Issues, 125-138.
126 A.G. ORR AND C.L. HAEUSER
related species is also examined, with the aim of testing the hypothesis that assemblages of
closely related species will show temporal dissociation of phenologies, possibly in order to
avoid competition for adult or larval resources, as has been suggested by several other studies
(Clench 1967, Owen 1971, Owen et al. 1972) .
AREA AND METHODS
The study was conducted in the Batu Apoi Forest Reserve, Temburong, Brunei, in an area
of primary lowland mixed dipterocarp forest of approximately one square kilometer in area
and ranging from about 0-350m in elevation, in the vicinity of the KBFSC. The mean annual
rainfall for the area was about 4000 mm (Pendry & Proctor, this volume) and the dry season
was not strongly marked, with some rain falling on most days. On one occasion (14
September 1991) when helicopter transport was available a sample was also taken from
nearby Bukit Belalong (elevation 932 m) in the space of four hours.
In the KBFSC area butterflies were sampled on 15 occasions over periods of two to five
days at approximately monthly intervals between March 1991 and March 1993. There was
a long discontinuity in data collection from December 1991 to April 1992. Specimens were
netted along the east ridge trail as far as plot 2, and along the west ridge trail as far as the
second summit (Figure 1). Collecting efforts were concentrated at fallen fruit, at sunspots
and along the ridges, especially at the summits where various canopy species were hilltopping
(see Shields 1967). Baits of fermenting banana were laid along the trail at regular intervals
to attract certain fruit-feeding nymphalid species.
Every sampling period at least one estimation of species abundance patterns in the forest
was made, using the transect walk technique of Pollard (1977). The area surveyed followed
the west ridge trail to the second summit and back, and usually took place between 1O.OOh
and 13.00h. Records were based mainly on visual sightings, hence included only those
species which could be identified without capture, which excluded most Lycaenidae and
Hesperiidae. Because of unevenness in sampling different taxonomic groups and species
~-1km
Figure 1. Area around the KBFSC showing the east and west ridge trails
(dashed lines) along which sampling was done.
Butterfly diversity in lowland rainforest 127
inhabiting different microhabitats, estimates of relative abundance are only considered

comparable between equally conspicuous species of similar habits (e.g. the understorey fruit
feeding guild), which usually meant members of taxonomically related assemblages.
Every day butterflies which came to 'mud-puddle' (Arms et at. 1974) at the sewage outlet
by the river were censused. Numbers visiting the soak between 1400h and 1430h were
recorded. A majority of papilionid and pierid species came to the sewage, and as most
species could be identified without capture this yielded the most extensive and reliable data
on relative abundance. As only newly emerged males visited the soak, recruitment rates
rather than abundance are estimated, but if the rate of recruitment is reasonably steady over
periods approximating the life expectancy of the butterflies it should also reflect overall
abundance. Because the quality of the sewage source varied with the level of occupancy of
the buildings, estimations of relative abundance of a species from one sampling period to
another may be unreliable. However this is not expected to bias the measurement of the
relative proportions of different species.
RESULTS
Species richness - general trends

Three hundred and twenty five species were recorded from the lowland area described
above and an additional 24 sp were collected from Bukit Belalong. A full list of species with
summary information on abundance and distribution is provided in the Appendix. One
hundred and thirty-five species from our list are previously unrecorded in Brunei (Cassidy
1982, 1985) bringing the national tally to 492 species.
Together these samples represent about 37% of the total Bornean Fauna (Otsuka 1988,
Seki et at. 1991a, 1991b). As is to be expected the more conspicuous Papilionidae, Pieridae
and Nymphalidae are much better represented than are the Lycaenidae or Hesperiidae (Table
1). After 42 man days collecting the cumulative number of species recorded was still
increasing at a rate of about one species every two days (Figure 2), hence there is obviously
a substantial number of species in the area yet to be recorded.
Table 1. Proportional representation of families relative to the Bornean fauna (including sample from Bukit
Belalong).
Family Total species Number in Propotion of

recorded Borneo Borneo total
Papilionidae 26 44 59%
Pieridae 31 41 76%
Nymphalidae 128 236 54%
Lycaenidae 112 393 28%
Hesgeriidae 51 217 24%
Using the method of Clench (1979, see also Ragusa & Llorente-Bousquets 1990) species
richness was estimated to be 454 species in total. Overall sampling is considered too uneven
to describe species abundance relations with any degree of precision. However 54 species
can be considered common (greater than 50 records overall), 112 were moderately common,
at least sporadically (between 10 and 50 records) and the remaining 158 species were rare
(less than 10 records, often only 1).
Phenology, species turnover and seasonality

Excluding very rainy days, the number of species recorded per day ranged from 59 to 102
128 A.G. ORR AND c.L. HAEUSER
300
'"
w
u
If
'"
10 20 3D 40
DAYS OF SAMPLING
Figure 2. Cumulative species total against number of days spent sampling.
~
c
"'
w
"- 50
'"w
U
W
Q.
'"
Figure 3. Average total number of species recorded per day for the various sampling periods
from March 1991 to January 1993.
20
w
U50
Z
g
Z
:::>
ED
«
~ 30 ParantlcQQS delessertl
i=
«
...J
w
Ill:
Figure 4. Three different patterns of phenology exhibited by (a) Cupha erymanthis - even, (b) Leptosia nina -
fluctuating, and (c) Paranticops delesserti - sporadic, between March 1991 and January 1993.
with a mean of 78 (Figure 3). No regular seasonal trends in species richness were apparent,
nor did any species show regular fluctuations in abundance. Certain species were always
fairly common e.g. Cupha erymanthis, a species found mainly in open areas around the
(a) Papillonidae
1:
Q)
1.0
:Q
.....
Qj
00.5
U
t:
.Q
......
!II
~ (b) Pleridae
U
0 1.0
d -d min
!"
§
~ 0.5
&
MAMJJASON MJJASONDJ
MONTH
Figure 5. Month to month similarity (measured by Pearson's correlation coefficient) in species composition for
mudpuddling (a) Papilionidae, and (b) Pieridae. Horizontal lines mark the minimum similarity recorded on
consecutive days (asssumed to be zero), and any figure below this line probably represents a significant change in
species composition.
centre (Figure 4a), whereas other very common species fluctuated markedly and apparently
irregularly in numbers, e.g. Leptosia nina (Figure 4b). Some species were common only for
one or two sampling periods and were not seen in between, e.g. Paranticops delesserti
(Figure 4c).
Phenologies were classified as even, fluctuating or sporadic by the following scheme:
90 % observations in 10-15 sampling periods - even
90 % observations in 5-9 sampling periods - fluctuating
90 % observations in 1-4 sampling periods - sporadic
By these criteria 43 species can be considered even in their phenology, 86 species
fluctuating, and 37 species sporadic. The remaining species were represented by fewer than
ten records and do not provide adequate data.
The prevalence of species with fluctuating or sporadic phenologies is reflected in the
approximate monthly turnover of papilionid and pierid species visiting soaks (Figures 5a,
5b). Relative proportions of the various species are compared between successive sampling
periods, using a Pearson product moment correlation coefficient which ignores variation in
absolute abundance and hence variations resulting from varying attractiveness of the soak.
It was found that similarities between consecutive months were almost always lower than the
lowest correlation recorded for samples taken on consecutive days (in which turnover was
assumed to be zero). This is true of both families, analysed independently, and indicates a
continually changing community composition.
There was little evidence to suggest that species abundance patterns are repeated on a
seasonal or yearly basis. High correlations were recorded for mudpuddling Papilionidae
between May 1990 and September 1991 (r = 0.85) and for mudpuddling Pieridae between
July 1990 and October 1990 (r = 0.78). There are no environmental correlates to suggest
an explanation for these convergences. On the other hand there was no evidence of species
abundance patterns repeating themselves on a yearly seasonal basis (Figure 6), indicating that
the greatest similarity was between consecutive months. Samples from the same months in
successive years tended to be dissimilar, suggesting a lack of any annual cycle.
APRl
MARl
JUL1-----,
SEPl
AUGl
MAYl
OCT2
OCTl
SE
JUL2
AUG2
.
':-~=o
.
10 0.5
Figure 6. Relative similarity in species composition for different months throughout the sampling period March 1991
to January 1993. Nwnerals afer months indicate year (l = 1991,2= 1992,3 = 1993). Average linkage cluster analysis
of Pearson correlation coefficients.
r-lW!~
L-~ycles
~-D
Figure 7. Similarity of phenologies of Graphium species: The sister species eurypylus and evemon group into
different clusters. Average linkage cluster analysis of Pearson correlation coefficients.
100
~
C
0::
w
Q.
!3::::; 50
u.
ffi...
...
1i!
MAR APR MAY JUL AUG SEP OCT p.¥JV I MAY JUL AUG SliP OCT p.¥JV JAN
Figure 8. Positive association of phenologies in two species of Eurema, E. andersoni and E. simuiatrix, mudpuddling
at sewage (tau = 0.43, p<0.05), between March 1991 and January 1993.
MAR APR MAY JU L AUG SliP OCT NOV I MAY JUL AUG ·SliP OCT NOoi JAN
Figure 9. Dissociation of phenologies of two species of Faunis, F. stomphax and F. gracilis, feeding at fermenting
banana baits (tau= -0.44, p< 0.05), between March 1991 and January 1993.
Table 2. Association/dissociation of phenologies of close species pairs. Only figures for Eurema and Faunis are
significant at 0.05 level.
SPECIES PAIR KENDAL'S TAU
Graphium evemon/G. eurypylus -0.15
Eurema andersonilE. simulatrix 0.43*
Cepora iudith/C. pactolicus -0.02
Prioneris cornelia/Po philonome -0.05
Erites elegans/E. argentata -0.27
Lexias dirtea/L. pardalis -0.20
Faunis gracilis/F. stomphax -0.44*
Cirrochroa tycheiC. emalea -0.17
Paralaxita orghne/P. telesia 0.23
Phenological relationships between closely related species were variable. A significant

positive association was found between Eurema simulatrix and E. andersoni (Pieridae)
(Figure 7), and a significant negative association was found between Faunis stomphax and
F. gracilis (Nymphalidae) (Figure 8). However, most commonly, when the phenologies of
closely related species pairs were compared there was no significant association or
dissociation (Table 2). When the degree of similarity of phenologies was calculated for five
Graphium species (Papilionidae) (Figure 9), the sister species G. evemon and G. eurypylus
fell into different clusters, indicating at least a lack of similarity in their phenologies.
However there was no significant negative correlation between them (Table 2).
Spatial patterns of diversity

Many species appear restricted to particular microhabitats, here divided into riverbanks
and open areas, forest understorey and ridges (Table 3). Many of the 'ridge' species were
probably found mainly in the canopy and subcanopy, but were either visible from the ridge
or descended lower than usual along the crest. Information is available for about half the
species recorded, the remainder being represented by too few records for conclusions to be
drawn. About half of these are recorded as being in some way restricted in their
distribution. Major trends observed were: 1. Many nymphal ids were restricted to open
habitats - these were mostly danaines and nymphaline species which also occur in secondary
forest; 2. Many nymphalids were restricted to the understorey - these included most
satyrines, amathusiines and the euthaliine nymphalines; 3. Many lycaenids were restricted
to the ridge - these were mostly canopy dwelling theclines. Spatial separation of species in
different microhabitats is a major source of unevenness in sampling, certain situations being
much easier to sample than others.
About 12 % of species were classified as local. Many of these species were restricted to
very small areas, sometimes of only a few hundred square metres (Figure 10), and were
found in only one place along the 1.5 km transects on either side of the KBFSC. Presumably
these species, mostly Lycaenidae, were restricted to the immediate vicinity of their larval
Table 3. Microdistribution and habitat preference by family (including 176 species, representing 54 % of all species).
Family General Open/Riverine Ridge Understorey Local

Papilionidae 15 2
Pieridae 25 5
Nymphalidae 62 21 5 31 14
Lycaenidae 23 17 4 27
Hesgeriidae 7 3 3
WEST RIDGE EAST RIDGE
Figure 10. Sites along the east and west ridge transects where highly localized species occur (the transects have been
transformed onto a vertical plane at right angles to the river). 1. Trix scopula, 2. Paralaxita orphne, 3. Paralaxita
telesia, 4. Dacalana sp. 5. Laxita tenata, 6. Poritia erycinoides, 7. Horaga syrinx, 8. Flos morphina.
hostplant. Several of these species, such as Flos morphina, Laxita tenata, Paralaxita telesia,
Ihrix scopula and Boraga syrinx, were common in a very limited area and almost always
present. Typically one male was visible, maintaining a territory in a sunspot. When it was
removed it was very shortly replaced by another and this was in turn replaced when
captured, and so on until at least three or four individuals had been recorded. Some species
were slightly less localized but still restricted in their distribution. For example, the
powerfully flying Charaxes durrifordi was found only on the east ridge between Pondok
Bujak and Plot 2. Of 23 species with a highly restricted distribution, 19 were found in only
one place. This suggests that the area covered was too small to provide an adequate sample
of localized species.
DISCUSSION
The number of species recorded from Kuala Belalong in this study substantially exceeds the
totals recorded for Mulu National Park, Sarawak, (276 species, Holloway 1984), and Endau
Rompin, in Peninsular Malaysia (232 species, Tan et al. 1990) and is similar to the figure
recorded for Kinabalu National Park and other localities in Sabah (340 species, Barlow et al.
1971). To put this result in perspective, the difficulty of comparing sampling efforts in
different localities with different personnel must be taken into account, as well as the time
frame within which the collections were made. The longest period of sampling for any of
the other studies above was five months (continuous) at Mulu by Holloway and colleagues
who collected 276 species from a range of habitat types. After five months periodic
sampling at Kuala Belalong, 212 species had been recorded from a single forest type but
overall sampling effort was probably considerably less.
It is well recognized, at least in seasonal environments, that measurement of species
richness must take into account variation in species' phenologies, the so called 'temporal
component of diversity' (Hill 1988, Shapiro 1975). This is especially true of temperate
environments, or of tropical environments with well defined wet and dry periods, where
species occur in a predictable sequence throughout the year (Owen 1971, Shapiro 1975).
Within climatic constraints, species phenology is thought to be a function mainly of larval
food availability (Slansky 1974), adult food availability (Clench 1967, Hill 1992) and mutual
avoidance by species competing for either resource (Clench 1967, Owen 1971). In tropical
areas, even where there is pronounced seasonality, there may be considerable variation in
the phenologies of different butterfly families (Owen 1971, Owen et al. 1972), which tends
to smooth out overall variation throughout the year. Some studies have found slightly greater
diversity and/or abundance during the wet season (Owen 1971), whereas others report greater
diversity during the dry season (Emmel & Leck 1970, Fox et al. 1965), but given the
different family level trends, all such claims should be treated with caution because of the
high probability of sampling bias in favour of particular groups.
At Kuala Belalong, which is without pronounced seasonality in rainfall and temperature,
Butteifly diversity in lowland rairiforest 133
species phenology is mostly unpredictable. Apart from a small proportion of very abundant
or very predictable species which are always reasonably common, the majority of species
appear to undergo substantial and unpredictable fluctuations in abundance, resulting in a
continual and fairly steady turnover in species abundance relationships and, hence in species
actually sampled. Probably the main reason why this study has recorded substantially more
species than comparable surveys such as Holloway (1984), is because it has continued over
a longer period of time.
When considering the phenologies of closely related species three types of relationship are
possible:
1. No interaction - phenologies of different species are independent.
2. Temporal association - similar species have similar phenologies because they are
tracking the same set of environmental factors.
3. Temporal dissociation - similar species have dissimilar phenologies either because
they are avoiding competition for larval or adult resources, or because they are
adapted to different conditions.
Because of the difficulty in obtaining adequate sample sizes, phenological relationships
among species assemblages are often unclear. Therefore a lack of correlation, positive or
negative, between the phenologies of species pairs may just be a result of inadequate sample
sizes of one or both species. This is particularly true when sister species are compared,
since as a rule one species is always much rarer than the other in a given area. Nevertheless
cases of dissociation in Faunis species and of positive association in Eurema species were
evident, and it seems probable that numerous other cases of temporal dissociation occur.
Previous studies emphasize the role of competition in promoting temporal dissociation
(Clench 1967, Owen 1971, Owen et al. 1972), but in none of these studies can other
explanations be ruled out. In general, species abundance and phenology seems likely to be
influenced by a complex set of biotic interactions, including the phenology of resource
availability, levels of predation and parasitoid attack, and within this set of potential factors
interspecific competition might sometimes be the most significant factor.
Beta diversity, or turnover in space, is probably a less important component of total
diversity in butterflies than in poorly dispersing groups such as Coleoptera (Mawdsley, this
volume) or Microlepidoptera (Robinson & Tuck, this volume). However Pinheiro & Ortiz
(1992) report high beta diversity in fruit feeding butterfly species within the understorey in
gallery forest in central Brazil; at Kuala Belalong, greater turnover than was evident may be
revealed by a more rigorous sampling technique. However, assessing microdistributional
patterns is fraught with difficulties; for example many mudpuddling species are essentially
canopy dwellers and only young males come to drink salt. In many species the
microdistribution of males and females is apparently quite different. Many species are too
rare to provide reliable data. Moreover within the rainforest there is always a characteristic
vertical stratification (DeVries 1988, Papageorgis 1975) which inevitably introduces bias into
any sampling proceedure. At Kuala Belalong the localization of so many species to one site
only, suggests that the forest was inadequately sampled.
The most important single finding to emerge from this study is that the lowland forest in
the Batu Apoi Forest Reserve supports an exceptionally rich diversity of butterflies, with in
all probability nearly half the entire Bornean fauna occurring within a single square
kilometre. Presumably many additional species would have been recorded had more
sampling been conducted on Bukit Belalong. Therefore the area certainly deserves special
conservation efforts on the part of the global community.
ACKNOWLEDGEMENTS
We would like to thank the organizers of the 1991192 UBD/RGS Brunei Rainforest
expedition, especially Catriona Prebble and Amanda Simpson, and the staff of the KBFSC,
including Samhan bin Nyawa and Kamariah Abu Salim for their practical assistance en route
to and while at the KBFSC. AGO was supported by a Universiti Brunei Darusalam research
grant. CLH was supported by a grant from the Deutsche Forschungsgemeinschaft.
LITERATURE CITED
ARMS, K., FEENY, P. & LEDERHOUSE, R.C. 1974. Sodiwn: stimulus for puddling behaviour by tiger
swallowtail butterfly butterflies, Papilio glaucus. Science, N. Y. 185:372-374.
ASHTON, P.S. 1964. Ecological Studies in the Mixed Dipterocarp Forests of Brunei State. Oxford Forestry
Memoirs 25.
BARLOW, H.S., BANKS, H.J. & HOLLOWAY, J.D. 1971. A Collection of Rhopalocera (Lepidoptera) from Mt.
Kinabalu, Sabah, Malaysia. Oriental Insects 5:269-296.
CASSIDY, A.C. 1982. An annotated checklist of Brunei buttertlies, including a new species of the genus
Catapaecilma (Lycaenidae). The Brunei Museum Journal 5:202-272.
CASSIDY, A.C. 1985. An enlarged checklist of Brunei buttertlies (Lepidoptera: Rhopalocera) including
descriptions of one new species and two new subspecies. Brunei Museum Journal 6:135-168.
CLENCH, H.K. 1967. Temporal dissociation and population regulation in certain hesperiine butterflies. Ecology
48:1000-1006.
CLENCH, H.K. 1979. How to make regional lists of buttertlies: some thoughts. Journal of the Lepidopterists'
Society 33:215-231.
CORBET, A.S. & PENDELBURY, H.M. 1992. The butterflies of the Malay Peninsula. Fourth edition revised by
J.N. Eliot. Malayan Nature Society, Kuala Lwnpur. xiv + 577 pp.
DEVRIES, P.J. 1988. Stratification of fruit feeding nymphalid butterflies in a Costa Rican rainforest. Journal of
Research on Lepidoptera 26:98-108.
EMMEL, T.F. & LECK, C.F. 1970. Seasonal changes in organization of tropical rainforest butterfly populations
in Panama. Journal of Research on Lepidoptera 8:133-152.
FOX, R.M., LINDSEY, A.W. Jr., CLENCH, H.K. & MILLER, L.D. 1965. The buttertlies of Liberia. Memoirs
of the American Entomological Society 19: 1-438.
HILL, c.J. 1988. The temporal distribution of butterfly species richness in a subtropical region of Australia.
Proceedings of the Ecological Society of Australia 15:115-126.
HILL, C.J. 1992. Temporal changes in abundance of two Iycaenid butterflies (Lycaenidae) in relation to adult food
resources. Journal of the Lepidopterists' Society 46:173 -181.
OTSUKA, K. 1988. Butterflies of Borneo, Volwne 1. Tobishima Corp., Tokyo.
OWEN, D.F. 1971. Tropical Butterflies. Clarendon Press, Oxford. 214 pp.
OWEN, D.F., OWEN, 1. & CHANTER, D.O. 1972. Seasonal changes in the relative abundance and estimates of
species diversity in a family of tropical buttertlies. Oikos 23:200-205.
PAPAGEORGIS, C. 1975. Mimicry in neotropical butterflies. American Scientist 63:522-532.
PINHEIRO, C.E.G. & ORTIZ, J.V.C. 1992. Communities of fruit-feeding buttertlies along a vegetation gradient
in central Brazil. Journal of Biogeography 19:505-511.
POLLARD, E. 1977. A method for assessing changes in abundance of butterflies. Biological Conservation
12: 115-134.
RAGUSO, R. A. & LLORENTE-BOUSQUETS, 1. 1990. The buttertlies of the Tuxlas Mountains, Veracruz,
Mexico, Revisited: Species-Richness and Habitat Disturbance. Journal of Research on Lepidoptera 29: 105-133.
ROBINSON, G.S. & TUCK, K.R. 1993. Diversity and faunistics of small moths (Microlepidoptera)in Bornean
rainforest. Ecologiclli Entomology 18:385-393.
SEKJ, Y., TAKANAMI, Y. & OTSUKA, K. 1991a. Butterflies of Borneo, Volwne 2, No.1: Lycaenidae. (In
Japanese and English). Tobishirna Corp. 252 pp., 72 pI.
SEKJ, Y., TAKANAMI, Y. & OTSUKA, K. 1991b. Butterflies of Borneo, Volwne 2, No.2: Hesperiidae. (In
Japanese and English). Tobishima Corp. 172 pp., 48 pI.
SHAPIRO, A.M. 1975. The temporal component of butterfly species diversity. pp. 181-195 in Cody, M.L. &
Diamond, J .M. (eds). Ecology and evolution of communities. Harvard University Press, Cambridge.
SHIELDS, O. 1967. Hilltopping. Journal of Research on Lepidoptera 6:69-178.
SLANSKY, F. 1974. Relationship of larval fOOd-plants and voltinism patterns in temperate butterflies. Psyche
81:243-253.
TAN, M.W., KIRTON, L.G. & KIRTON, C.G. 1990. Composition and distribution of butterflies in Rompin-Endau,
especially along Sungei Kinchin and its vicinity. pp. 193-212 in Proceedings of the International C01!/erence
on Tropical Biodiversity 'In Harmony with Nature', 12-16 June 1990, Kuala Lumpur, Malaysia. Malaysian
Nature Society, Kuala Lwnpur.
Appendix. Checklist of the butterflies of Belalong. Abundance classes (ABUND.): 1 - common (+ moderately
common, 10-50; + + very common >50),2 - sporadic (> 10 records limited to one or two recording periods),
3 - rare « 10, frequently one record only). Distribution classes (DIST.): 1 - general (within 1.5 Ian radius of
KBFSC); 2 - localised (+ localised, + + very localised); 3 - Bkt Belalong (submontane). Habitat preference
(H): rv - open areas or riverine, u - understorey, rg - ridge.
SPECIES ABUND. DIST. H. I SPECIES ABUND. DIST. H.

2 3 2 2 2
PapiHooidae I Eurema alilha +
Trogonoplera brookiana ++ + Eurema lilaha ++ +
TroiiUs miranda + + Eurema tominia + +
TroiiUs amphrysus + N~pWdae-D
Pachliopta aristolochiae + + rv Tiramula seplenlrianis + + +
Atrophaneura nox + + Paranlica aspasia + + rv
Atrophaneura neptunus + Paranlica agleaides + + rv
Papilio kama + + Jdeapsis gaura ++ + +
Papilio iUmolion ++ + Jdeapsis vulgaris + + + rv
Papilio nephelus + + liUa Iynceus + +
Papilio helenus + liUa slolli ++ +
Papilio fuscus ++ + Euploea moiUsla + rv
Papilio iswara + + Euplaea camaralzeman + + rv
Papilio memnon + rv Euplaea sylvester + + rv
Chilasa slaleri + + Euplaea mulciber ++ + + rv
Chilasa paradoxa + Euploea midamus +
Meandrusa poyeni + Euploea phaenarele +
Graphium sarpedon ++ + Euploea radnmanthus ++ +
Graphium evemon ++ + Euploea algea + + rv
Graphium eurypylus ++ Euplaea eyndhavii + +
Graphium bmhyc/es ++ Nymphalidae - Satyrinae
Graphium agamemnon ++ + Melanin's leda + u
Paranticops ramaceous + + Coeliles epimintha + + u
Paranticops delessenii + + Caeliles euptychaides + + u
Pmhysa anliphales ++ + Neon'na lowii + + rv
Lamproptera curius + En'les elegans ++ + u
Lnmproplera meges + Eriles argentina + + u
Pieridae Ragadia makula ++ + u
uplasia nina ++ + rv Mycalesis maianeus + +
Delias ninus + + rg Mycalesis orseis + + rv/u
Delias eumolpe + Mycalesis fuscum + +
Delias hyparele ++ + rg Mycalesis horsfieldi +
Delias heningia + + + rg Mycalesis anapila ++ +
Delias singhapura + Mycalesis kina + + rg/u
Prioneris philonome + + Mycalesis amoena +
Pn°oneris cornelia + + Yplhima Jasciala + rv
Cepora paclolicus ++ + Yplhima hanburyi + + rv
Cepora iudilh ++ + Elymnias hypermneslra +
Appias Iyncida ++ + Elymnias nesaea + + rg
Appias paulina + + Elymnias kuenslleri +
Appias indra ++ + rg Elymnias penanga +
Appias nero + + I N~phalide - Amatbus~e
Appias cariUna ++ + Faunis kirata : + + u
Saletara liberia ++ + I Faunis gracilis ; + + u
Ixias pyrene + + I Faunis slomphax ++ + u
Hebomoia glaucippe + + I XanlhOlaenia busin's ++ + u
Pareronia valeria + + rg I Amalhusia phidippus + + u
Dercas verhueli + + I Amalhusia masina +
Calopsilia pomona + Amalhuxidia amylhaon + + u
Gandaca harina ++ + Zeuxidia amylhyslus + + u
Eurema hecabe + Zeuxidia doubledayi + + u
Eurema ada ++ + Zeuxidia aurelia + u
Eurema /aCleo/a + + lhaumanlis klugius + + u
Eurema blandn ++ + 1haumanlis adana + +
Eurema simulatrix ++ + Thauna alin's ++ +
Eurema andersonii ++ + Discophara sondaica +
Eurema san· + + Discophara necha +
136 A.G.ORR AND C.L.HAEUSER
Appendix. Checklist of the butterflies of Belalong. Abundance classes (ABUND.): I - common (+ moderately
common, 10-50; + + very common >50), 2 - sporadic (> 10 records limited to one or two recording periods),
3 - rare « 10, frequently one record only). Distribution classes (DIST.): I - general (within 1.5 Ian radius of
SPECIES ABUND. DIST. H. I SPECIES ABUND. mST. H.

2 3 2 2 2
Nymphalidae - Nymphalinae I Bassarona leula + + U
CuphiJ erymanlhis : ++ + + rv EUlhalia adonia + +
Phalanla alcipe : + + + Eu/halia whileheadi + +
Vagrans egisla + + + Euthalia aconlhea + +
Vindula erola : ++ + rv EUlhalia monina + +
Padaca /asciala + EUlhalia kando +
Ci"ochroa Iyche + + Lexias dirlea ++ + u
Ci"ochroa emalea + Lexias parcialis + + u
Ci"ochroa malaya + + Lexias canescens + +
Ci"ochroa Salellila + Eulaceura osteria +
Cirrochroa orissa + HeTana sumartrana +
Ten"nos lerpanti.er ++ + u Agalasa calydonia + + rg
Terinos clarissa + Pro/hoe franck + + rg
Celhosia hypsea + + rv Polyura schreiber +
Rhinopalpa polynice + Polyura alhama + + rv
Doleschallia bisallide : ++ + + rv Polyura moon· +
Dicho"agia nesimachus + + Polyura dRlphis + +
Chersonesia rahria l+ + rv Charaxes dam/ordi + + rg
Chersonesia peraka + + + Charaxes dislant; +
Chersonesia n·sa + Charaxes bomeensis + +
Cyrestis nivea + + Charaxes bemardus + +
Cyrestis theresae + Lycaenidae - Riodininae
Panloporia hordonia + Zemeros emesoides + + rv
Pan/oporia dindinga + Dodona elvira + +
Panloporia paraka + + Abisara gem + + rg
Panlaporia aurelia ++ + Abisara kausambi +
Lasippa liga + + Paraiaxila lelesia + ++ + u
Lasippa viraja + Paraiaxila orphna + ++ u
Neptis hylas + Laxila leneta + ++ u
Neptis daryodona , + + Lycaeniclae - Lycaeninae
Neptis nala + Porilia erycinoides + ++ rg
Neptis leucoporos ++ + + rv Porilia sumalrae +
Neptis miah + Porilia plateni +
Neptis clinia + Simiskina pharyge + ++
Neplis harila + Simiskina phalena + ++
Neptis amerada + + rv Deramas neLvis + ++
Neplis anjana + Deramas yasodo + ++
Alhyma pravara + Miletus gopara ++ rg
Alhyma asura + MiLelus ceLlarius + + rg
Alhyma kanwa , + + + Allolinus subviolaceous + +
Alhyma rela + Allolinus bomeensis + +
Alhyma adanora + Allolinus horsefieldi +
Alhyma selenophora + I Allolinus sa"asles +
Alhyma nefte + + I AIIOIinus nicholsi +
Sumalia ciaraxa + I AIIOIinus subslrigmus +
Modam procris ~ + + I Logonia regina + + rg
Pandila sinope + + : Logonia malayica ++ + rg
Lebadea marlha + Logonia marmorala +
Parlhenos sylvia + + rv : Discolampa ethion ++ +
Tanaecia munda + Calela elna ++ +
Tanaecia aruna + : Neopilhecops mlmora + + u
Tanaecia orphne ~ ++7 + u Acylolepis puspa ++ +
Tanaecia peLea : ++1 + u I Acylolepis riple + +
Tanaecia godartU + u I Piau/ella cossaea + +
Tanaecia iapis ~ ++ + u : Celaslrina lavendalaris + +
Dophla evelina ~ + + u I Callenya lenya +
Bassarona danya : ++ + u I Cebrella pel/ecebra +
I
Appendix. Checklist of the butterflies of Belalong. Abundance classes (ABUND.): I - common (+ moderately
common, 10-50; ++ very common >50),2 - sporadic (>10 records limited to one or two recording periods),
3 - rare « 10, frequently one record only). Distribution classes (DIST.): I - general (within 1.5 Ian radius of

2 2 2 2
famides virgularus + : Dacalaru! sinhara + +
famides bochus + Dacalana lowii + +
lamides caeruleus + + Thrix scopula + ++ rg
famides pura ++ + Mantoides gamn + +
lamides aralus + + Suasa Usides +
famides elpis + + Remelana jangala + + +
Jamides lugine + + Hypolyeaena amasa + +
famides zebra + Hypolyeaena Iheeloides +
lamides celeno + Hypolycaena amnbilis + +
Naeaduba subperusia + rg Hypolycaena erylus + +
Nacaduba hermus + rg Deutkrix epijarbas +
Nacaduba calaun'a + Deudorix slaudingeri +
Nacaduba beroe + Rapala pherelimn + ++ rg
Nacaduba kurava + Rapala varuna +
Prosolas alUla + Rapala tkmilia +
Prosolas nora ++ Viraehola subgullala + + +
Prosolas pia + Curetis santana +
Anthene emolus ++ + Curelis felderi +
Anthene lycaenina ++ + Curetis spenhis +
Arhp, pseudocentaurus + Curetis tagalica + +
Arhopala kinabala + + Hesperiidae
Arhopala hypomura + Bibasis sena +
Arhopala aedias + Hasara proxissima +
Arhopala epimura + + rg Hasara schoenherr + + rv
Arhopala lurida (?) + Hasara chromus +
Arhopala della (?) + Hasora quadnpunclala +
Arhopala Iropaea + Hasara villa + +
ArhopaJa moore; + + Choaspes plateni +
ArhopaJa anlimula + Capila phaneus +
Arhopala inornara (?) + Charnu'on fieulnea + +
ArhopaJa democrilus + + rg Charmion species +
Arhopala denla (?) + Celaenorrhinus bazi/anus +
Arhopala elopura (?) + Tapena lhwailesi +
Arhopala alilaeus (?) + Darpa striata + + rg
Arhopala zambra + + rg Mooreana trichoneura +
Arhopala bomeen.<;is + + rg Tagiades japerus + +
Arhopala vihara + Tagiades gana + +
Arhopala arvina (?) + Tagiades lavalus +
Arhopala labuana + Tagiades parra +
Flos julgida + + Tagiades loba + +
Flos morphina + ++ rg Tagiades waters/radii +
[raola distant; + Otkntoplilum pygela +
lrama rochana + Amella verones
Calapaecilrna elegans + Halpe sikkima +
Loxura cassiopeia + Iambrix sre/lifer/salsala +
Exooy/ides Iharis + + Koruthaialos rubicula + +
Horaga syrinx + ++ rg Korurhaialos sindu +
Chen'lra jreja + + Psolos ju/igo +
Rilra aurea + Ancistroides n;grita +
Ticherra acre + Ancistroides armalus +
Drupadia ravindra + I NOlOcrypra paralysos + +
Drupadia rufolaenia + I NOlOcrypra clavaza +
Drupadia rheda ++ + I Cupirha pu"eea +
Drupadia niasica + I Zographerus kuru +
Drupadia cinesia + I OeTane microlhyrus + +
Drupadia cineus + + rg : Hyarorus adrasrus +
Dacalana vidura + + I Ismn kuala +
I
138 A.G.ORR AND C.L.HAEUSER
Appendix. Checklist of the butterflies of Belalong. Abundance classes (ABUND.): 1 - common (+ moderately
common, 10-50; + + very common >50), 2 - sporadic (> 10 records limited to one or two recording periods),
3 - rare « 10, frequently one record only). Distribution classes (DIST.): 1 - general (within 1.5 Ian radius of

2 2 2 2
Pyroneura jlavia + rg II Gangara lebadea +
Pyroneura lalOia + En"nota thrax +
Pyroneura niasana + I Erinota sybirita +
Pyroneura aurantiaca + + rg I Hidari dasoena +
LotonglL. calalhus + : Eetion elia +
'U/a exceLlans + Acerbus anlhea +
Gangara thyrsis + : Pirdana hye/a +
Gangara sanguinocculus +
A quantitative inventory of trees in one hectare of mixed dipterocarp
forest in Temburong, Brunei Darussalam
AXEL DALBERG POULSEN 1 , IVAN C. NIELSEN 1 , SYLVESTER TAN2 and HENRIK

BALSLEV I
IDepartment of Systematic Botany, Institute for Biological Sciences, Aarhus University,
Nordlandsvej 68, DK-8240 Risskov, Denmark.
2Forest Department Headquarters, Kuching, Sarawak, Malaysia.
ABSTRACT. One hectare of mixed dipterocarp forest in Temburong, Brunei, was studied in a square
one hectare plot. A 5 x 5 metres grid (horizontal projection) was laid out and marked with stakes and
strings. All trees ~ 10 cm dbh were marked with numbered aluminium plates and, for each tree, its
coordinates in the plot were marked, its dbh and height were measured, and a voucher was collected for
botanical identification. For each species the total and relative densities, dominance, and frequencies, and
Importance Value Index (WI) were calculated. For each family the total and relative values for density,
diversity, and dominance, and Family Importance Value (FIV) were calculated. The plot contained 550
individual trees with dbh ~ 10 cm and these were identified as belonging to 231 species divided among
43 families. The total basal area of trees with dbh ~ 10 cm dbh was 40.79 square metres. The average
height of the trees was 22.5 m, the tallest was 62 m, and the closed canopy was found at 30-40 metres.
Dipterocarpaceae ranked highest in FlY with a score of 91.0 (out of 300 possible) followed by
Euphorbiaceae with an FlY score of 34.5. The remaining 41 families all had FlY scores of less than
15, the lowest being Dilleniaceae with FlY score of 0.6. Shorea parvifolia (Dipterocarpaceae),
Elateriospermum tapos (Euphorbiaceae), Shorea laevis (Dipterocarpaceae) and Gluta laxiflora
(Anacardiaceae) were the most important species with lVI values between 18.2 and 12.9 (of 300
possible).
KEY WORDS: diversity, density, dominance, frequency, importance value, Borneo.
INTRODUCTION
The aim of the present study was to establish a permanent plot that could serve as an area
of reference for future studies in the floristic diversity in both Brunei Darussalam and
adjacent areas of NW Borneo. An equally important purpose was to create a framework for
studies in other biological disciplines. The mixed dipterocarp forest at Belalong was subject
to the classical study by Ashton (1964), where the study area chosen was "an ill-defined area
at the meeting place of the Temburong and Belalong rivers". Ashton included trees above
12 inches girth (approximately 9.7 cm dbh). The present plot was placed near acre plots no.
31-34 of Ashton (op. cit. Figure l). In the nearby Gunung Mulu National Park (Sarawak)
Proctor et al. (1983) surveyed a one hectare plot in a dipterocarp forest site at 200 to 250
m altitude on "heavy-textured red-yellow podzolics with a surface layer (up to about 15 cm
thick) of reddish-brown fibrous organic matter" (op. cit., p. 242). In that study, all trees
above 10 cm dbh were enumerated.
THE STUDY AREA
The study was conducted in Batu Apoi Forest Reserve, Temburong District, Brunei (115°10'
E; 4°32' N) in Northern Borneo. The soil is derived from the Setap Shale Formation.
Precipitation is above 4000 mm per year with less than 200 mm in January-February and
139
140 AXEL POULSEN, IVAN NIELSEN, SYLVESTER TAN & HENRIK BALSLEV
..
. ' .. :;
.. e
.' . . ~ -:'P.
e· ...
'.
e0'-"
. ,'
• ..
....
."
.. . :.
.'
• 41
t
[10,30[
[30,SO[
.
. . '.
, .
", ',- .~
:~. "
..
.~
. .. ...
.
:
..
. ,
,
• [SO,70[
it
.. ,. ... .
:..
'
:,
..
~
. e. . .•. ..' e [70,90[
~
:.
. .
'.0. .
e.
:' .:: . • : .: .. e .. '.
'" ..
'.
e [90,1S0[
•.....
.' '.'
-.
.' . ..
'"
e· • '"
•
.. .. .. :...... •• '._ • '. • • eO ': •
. ,
...... .' ." .. ~ '
. '.' .!. ... :0 .•.••.. .• 0'. "
Figure 1. Map of the 550 trees ~ 10 em dbh on a I-ha plot (100 x 100 m). The dot size represent trees of
different dbh classes (em).
July-August. A one hectare plot was laid out on the ridge west of Kuala Belalong Field
Studies Centre on a steep slope in primary, mixed dipterocarp forest at about 250 m
elevation. The difference between highest and lowest point of the sample plot was about 60
m.
A grid of 400 squares, each one measuring 5 x 5 metres in horizontal projection was laid
out, using measuring tape and compass. The corners of each grid were marked with 40 cm
aluminium tubes which were then connected with nylon string. The total study plot of one
hectare was then surveyed for trees with dbh ~ 10 cm which were marked consecutively
with numbered aluminium tags. Dbh was measured using a tree caliper. For stems with
irregular circumference (Le., not circular) the average of the largest and smallest diameter
was used. For buttressed trees the diameter was measured just above the buttresses (if they
reached over breast height). Tree height was measured using a clinometer, based on the
angles to the top and base of the tree and the distance from observer to tree. Voucher
specimens of all individual trees were collected whenever possible. Duplicates are deposited
at BRUN and AAU (acronyms according to Holmgren et at. 1990) where identifications were
carried out. Unvouchered trees were identified in the field by comparison with vouchered
individuals.
For calculations the following definitions were used:
Density: Number of individuals per hectare.
Basal Area (Dominance): The area occupied by a cross-section of the stem at breast
height.
Frequency: The number of sample units (5 x 5 m) in which a species is found.
Diversity: The number of species of a family encountered in the sample.
Relative density, dominance, frequency, and diversity are calculated as the percentages
of the sum of values of all species or families.
Importance Value Index (IVI) for a species is the sum of its relative density, dominance
and frequency (Curtis & Macintosh 1951).
Inventory of trees in mixed dipterocarp forest 141
Family Importance Value (FlV) for a family is the sum of its relative density, diversity
and dominance (Mori et al. 1983).
Density
The total density in the sample plot was 550 trees. This was unequally divided among the
231 species encountered (Appendix). The highest densities found for a species were those
of Elateriospermum tapos (Euphorbiaceae) with 35 individuals per hectare, Gluta laxiflora
(A nacardiaceae) with 26 individuals, and Vatica odorata (Dipterocarpaceae) with 21
individuals per hectare. Three species had densities between 10 and 20 individuals/hectare
(Shorea parvifoiia, Moulthonianthus leembrugianus, Hydnocarpus sumatrana). The
remaining 225 species all had densities below 10 individuals per hectare, 143 species (62%)
being represented by only a single individual in the plot.
Densities for families were also unequally divided among the 43 families represented
(Table 1). Dipterocarpaceae were represented by 133 individuals and Euphorbiaceae by 93
individuals. Anacardiaceae, Ebenaceae, Flacourtiaceae, and Myristicaceae all had between
25 and 34 individuals, whereas the remaining families all had less than 25 individuals. Eight
families (Dilleniaceae, Ixonanthaceae, Linaceae, Loganiaceae, Myrsinaceae, Olacaceae,
Sabiaceae, and Thymelaeaceae) were very rare, each being represented by a single individual
in the study plot. The distribution pattern of trees reveals no obvious clustering.
Dominance
The total basal area of trees;;:': 10 cm dbh was 40.79 square metres. The most dominant
species were Shorea parvifolia and Shorea laevis, representing 12.3% and 14.7%,
respectively, of the basal area. Eighteen species had more than 1 % relative dominance; 13
of them belong to Dipterocarpaceae (see Appendix). The distributions of trees in specific
size classes appear to be random (Figure 1.)
At the family level Dipterocarpaceae account for over half of the basal area at the study
site, followed by Euphorbiaceae with 8.5 % of the basal area. The remaining 41 families
each have less than 5 % of the basal area, 30 of them less than 1 % (Table 1).
Frequency
Frequency conveys information about how evenly a group of individuals is distributed
over the study plot. It therefore only makes sense to discuss this parameter for taxa with
several individuals in the sample. If a taxon has many of its individuals located in a few
sample units, i.e. it has a clumped distribution, then its relative frequency would be markedly
lower than its relative density. Six species (Gluta laxiflora, Shorea parviflora, Vatica
odorata, Elateriospermum tapos, Moultonianthus leembruggianus, Hydnocarpus sumatrana)
all have over 10 individuals in the sample. For each them there is no dramatic difference
between their relative density and their relative frequency (see Appendix), indicating that
none of them have a clumped distribution at the study site.
Diversity
The most diverse family is Dipterocarpaceae which is represented by 28 species,
Euphorbiaceae coming a close second with 21 species. Myristicaceae (17 sp.), Burseraceae
(11 sp.), Ebenaceae (11 sp.), and Myrtaceae (11 sp.) are also species rich (Table 1). The
total number of species with dbh ;;:.: 10 cm in the plot of 231 is comparable to the estimate
of about 215 using the species-area curve of a dipterocarp forest from Mulu (Proctor et al.
1983). Ashton's (1964) species-area curve for the Belalong area gives a figure of only 90
species per hectare for trees exceeding 9.7 cm diameter even though this number is obtained
lumping rectangular 1-acre subplots over a larger area. Although Ashton's plots were more
strictly confined to a single habitat (ridge or hillside) thereby minimising beta-diversity, the
extreme difference between his and our results can only be interpreted as an underestimation
by him (the more likely) or overestimation by us (less likely to be the major cause) of the
number of rare species, often based on inadequate fragmentary samples (Ashton, personal
communication) .
Table 1. Families of trees represented at Belalong in the plot west of Kuala Bela10ng Field Studies Centre showing
the number of species, individuals, basal area, relative diversities (reI. div.), densities (reI. dens.), dominance (reI.
dom.), and Family hnportance Value (FlV).
Family sp. indiv. basal area reI. div. reI. dens. reI. dom. FlY
Alangiaceae 1 3 0.05 0.43 0.55 0.12 1.10
Anacardiaceae 6 34 1.90 2.60 6.18 4.65 13.43
Annonaceae 8 14 0.26 3.46 2.55 0.65 6.66
Apocynaceae 4 7 1.77 1.73 1.27 4.33 7.34
Arecaceae 1 2 0.05 0.43 0.36 0.12 0.91
Bombacaceae 3 4 0.08 1.30 0.73 0.18 2.21
Burseraceae 11 13 0.52 4.76 2.36 1.27 8.40
Celastraceae 3 0.08 0.43 0.55 0.19 1.17
Chrysobalanaceae 5 5 0.10 2.16 0.91 0.25 3.32
Clusiaceae 9 13 0.29 3.90 2.36 0.70 6.96
Crypteroniaceae 2 0.Q2 0.43 0.36 0.05 0.84
Dilleniaceae 1 0.01 0.43 0.18 0.03 0.64
Dipterocarpaceae 28 133 22.31 12.12 24.18 54.70 91.00
Ebenaceae 11 25 0.92 4.76 4.55 2.25 11.56
Euphorbiaceae 21 93 3.48 9.09 16.91 8.54 34.54
Fabaceae 3 3 0.21 1.30 0.55 0.52 2.37
Fagaceae 6 7 0.80 2.60 1.27 1.97 5.84
F1acourtiaceae 9 37 1.54 3.90 6.73 3.77 14.39
Icacinaceae 2 6 0.13 0.87 1.09 0.31 2.27
Ixonanthaceae 1 0.32 0.43 0.18 0.79 1.40
Lauraceae 7 7 0.44 3.03 1.27 1.09 5.39
Lecythidaceae 2 2 0.05 0.87 0.36 0.13 1.36
Linaceae 1 0.09 0.43 0.18 0.23 0.84
Loganiaceae 1 I 0.06 0.43 0.18 0.14 0.76
Melastomataceae 5 6 0.29 2.16 1.09 0.72 3.97
Meliaceae 9 9 0.26 3.90 1.64 0.63 6.16
Moraceae 7 9 0.63 3.03 1.64 1.54 6.21
Myristicaceae 17 26 1.04 7.36 4.73 2.55 14.64
Myrsinaceae 1 1 0.03 0.43 0.18 0.08 0.69
Myrtaceae 11 16 0.67 4.76 2.91 1.63 9.30
OIacaceae 1 1 0.06 0.43 0.18 0.14 0.75
Polygalaceae 6 18 0.63 2.60 3.27 1.54 7.41
Proteaceae 2 0.06 0.43 0.36 0.15 0.95
Rosaceae 3 3 0.29 1.30 0.55 0.72 2.56
Rubiaceae 8 10 0.24 3.46 1.82 0.59 5.88
Sabiaceae 0.01 0.43 0.18 0.03 0.64
Sapindaceae 1 2 0.05 0.43 0.36 0.13 0.92
Sapotaceae 7 10 0.33 3.03 1.82 0.82 5.67
Sterculiaceae 5 10 0.59 2.16 1.82 1.45 5.43
Thymelaeaceae 0.Q2 0.43 0.18 0.04 0.66
Tiliaceae 2 2 0.04 0.87 0.36 0.09 1.32
Ulmaceae 2 0.02 0.43 0.36 0.05 0.85
Verbenaceae 4 0.05 0.43 0.73 0.13 1.29
Importance Value Index (IVI)

The Importance Value Index (lVI) lumps three ecological parameters (density, dominance,
frequency) that measure quite different features of a forest. In itself it therefore hardly has
an ecological meaning, and ranking species by it must be taken only as a way of seeing
which species come out as important elements of the forest even if they are looked upon from
quite different viewpoints in ecological terms. Dipterocarpaceae is conspicuous in the IVI
ranking, accounting for five (Shorea parvifolia, S. laevis, S. pauciflora, Vatica odorata, V.
endertit) of the top 10 species, but Euphorbiaceae is also conspicuous with three
(Elateriospermum tapos, Hydnocarpus sumatrana, Moultonianthus leembruggianus) species
among the top 10, and Anacardiaceae (Gluta laxiflora) and Apocynaceae (Dyera costulata)
are also represented (see Appendix).
Family Importance Value (FlV)

The Family Importance Value (FIV) is another index of questionable ecological meaning,
mixing the different parameters of diversity, density and dominance. Nevertheless it gives
some indication of importance regardless of one particular parameter. Dipterocarpaceae is
by far the most important family accounting for 91.0 (of 300 possible) points in FlV (Table
1). If individuals less than 30 cm dbh are omitted 59 species remain, 22 of these belonging
to the Dipterocarpaceae; this result emphasizes the importance of the family among the
canopy trees. Euphorbiaceae again is important with a FIV of 34.5. A group of five
families (Myristicaceae, Flacourtiaceae, Anacardiaceae, Myrtaceae, Ebeneaceae) have FIV's
between 10 and 15, and the remaining 36 families all have a FIV less than 10. Ashton
(1964) also found Dipterocarpaceae followed by Euphorbiaceae to be the two most important
families in his plots in the same area. In the nearby Gunung Mulu National Park (Sarawak)
Proctor et at. (1983) found that the five most important families, based on percent basal area
and actual numbers of individuals, were: Dipterocarpaceae (43.2 %/114), Leguminosae
(5.1 %/15), Sapotaceae (4.8%/52), Myrtaceae (4.05%/65), and Euphorbiaceae (2.8%/70) out
of a total for all families of 57 m2 and 778 individuals at or above 10 cm DBH.
Height
In the sample 87 (16%) trees were over 30 m high, 25 (5 %) were over 40 metres, and
five (1 %) were over 50 metres. The tallest tree, a Shorea laevis, measured close to 62
metres and had a dbh of 110 cm. This species was also the tallest emergent in Ashton's
profile diagrams (1964) from the same area. Of the 87 trees ;::: 30 m dbh, 47 were
Dipterocarpaceae. The closed canopy of the forest was at 30-40 metres above the ground
and appears irregular (Figure 2).
CONCLUSIONS
With 231 species, the plot is among the most species-rich recorded in Northern Borneo for
trees ;::: 10 cm dbh in a one hectare plot of mixed dipterocarp forest. This high number may
partially be attributed to the inclusion of both ridge top and hill side habitats. The number
of species may also be slightly overestimated because a rather narrow concept of morpho-
species was applied compared with, for example, that used by Ashton (1964). The number
of species recorded increases markedly when 10 cm dbh is used as a lower limit, rather than
30 cm.
70
-E
.E 60
Shorea iaevis ...,.. .
Cl
'il)
I 50
' ..
40
30
20
10
0
0 20 40 60 80 100 120 140 160 DBH (em)
Figure 2. Relation between dbh and height of trees <!: 10 cm dbh. Note the emergent Shorea laevis.
ACKNOWLEDGEMENTS
We are indebted to P. S. Ashton, M. M. 1. van Balgooy, C. Kalkman, and W. 1. 1. O. de

Wilde for help with identification. The study was financed by the Danish Natural Science
Research Council (Grant no. 11-9018).
LITERATURE CITED
ASHTON, P. S. 1964. Ecological studies in the mixed Dipterocarp forests of Brunei State. Oxford Forestry Memoirs
25:1-75, 17 tables, 70 figures, 2 appendices.
CURTIS, J. T. & MACINTOSH, R.P. 1951. An upland forest continuum in the praire-forest border region of
Wisconsin. Ecology 32:476-496.
HOLMGREN, P. K., HOLMGREN, N.H. & BARNETT, L.C. 1990. Index Herbariorum. Part I: The Herbaria
of the World. New York Botanical Garden, Bronx.
MORl, S. A., BOOM, B.M., CARVALINO, A.M. & DOS SANTOS, T.S. 1983. Ecological importance of
Myrtaceae in an eastern Brazilian wet forest. Biotropica 15:68-70.
PROCTOR, J, ANDERSON, J.M., CHA!, P. & VALLACK, H.W. 1983. Ecological studies in four contrasting
lowland rain forests in Gunung Mulu National Park, Sarawak. Journal of Ecology 71:237-260.
Appendix l. Tree species sampled at Belalong in the plot west of Kuala Belalong Field Studies Centre arranged
alphabetically in families. One voucher number is given after the name. Number of individuals, basal area,
frequency, relative density (reI. dens.), dominance (reI. dom.), frequency (reI. fre.) and Importance Value Index
(IVI) are given for each species.
SEeciesNollcher Ind. bas.a fre. reI. den. rel.dom. reI. fre. NI
Alangiaceae
Alangiumjavanicum (BI.) Wang 50 0.05 0.55 0.12 0.58 1.25
Anacardiaceae
Gluta laxiJIora Ridl. 21 26 1.58 22 4.73 3.88 4.27 12.88
Gluta wallichii (Hk. f.) Ding Hou 146 I 0.02 0.18 0.05 0.19 0.42
Gluta sp. I 367 2 0.05 2 0.36 0.11 0.39 0.86
Semecarpus helerophylla BI. 72 0.04 0.55 0.09 0.58 1.22
Swinlonia joxwonhyi Elmer 8 0.Q3 0.18 0.08 0.19 0.46
Swintonia sp. I 248 0.18 0.18 0.44 0.19 0.82
Annonaceae
Alphonsea elliplica Hk. f. & Thoms. 288 2 0.06 2 0.36 0.15 0.39 0.90
Monocarpia marginalis (Scheff.) Sinclair 330 2 0.Q3 2 0.36 0.08 0.39 0.83
Polyallhia hypoleuca Hk. f. & Thoms. 46 3 0.06 0.55 0.14 0.58 1.27
Polyallhia sp. 104 0.03 I 0.18 0.08 0.19 0.46
Xylopia malayana Hk. f. & Thoms. 39 0.02 2 0.36 0.06 0.39 0.81
Xylopia sp. 251 0.02 0.18 0.04 0.19 0.42
genus indet sp. 1 282 0.01 I 0.18 0.03 0.19 0.41
genus indet sp. 2 131 0.02 2 0.36 0.06 0.39 0.81
Apocynaceae
Aislonia .• cholaris (L. ) R. Br. 153 0.16 0.18 0.39 0.19 0.77
Dyera coslulala (Miq.) Hk. f. 87 4 1.58 4 0.73 3.88 0.78 5.38
Kibalalia sp. 510 I 0.01 0.18 0.03 0.19 0.40
Tabaernaemonlana corymbosa Roxb. 75 0.02 0.18 0.04 0.19 0.41
Arecaceae
Oncosperma horridum BI. 53 0.05 2 0.36 0.12 0.39 0.87
Bombacaceae
Durio excelsus (Korth.) Bakh. 349 0.01 0.18 0.02 0.19 0.40
Durio kUlejensis (Hassk.) Becc. 427 0.02 I 0.18 0.05 0.19 0.43
Durio sp. I. sect. bastia 269 0.04 2 0.36 0.11 0.39 0.86
Burseraceae
Canarium sp. 538 0.10 0.18 0.24 0.19 0.62
Dacryodes coslala (Benn.) Lam 83 0.07 0.55 0.18 0.58 1.31
Dacryodes roslrala (BI.) Lam 448 0.05 0.18 0.12 0.19 0.49
Dacryodes rugosa (BI.) Lam 267 0.01 0.18 0.02 0.19 0.40
Dacryodes sp. I 10 0.01 0.18 0.03 0.19 0.40
Dacryodes sp. 2 71 0.02 0.18 0.04 0.19 0.41
Dacryodes sp. 3 302 0.14 0.18 0.34 0.19 0.72
Dacryodes sp. 4 417 0.05 0.18 0.12 0.19 0.50
Dacryodes sp. 5 501 0.01 0.18 0.03 0.19 0.40
Santina megaphylla Kalkman 37 0.01 0.18 0.03 0.19 0.40
Triomma malaccensis Hk. f. 171 0.05 0.18 0.13 0.19 0.50
Celastraceae
Lophopetalum sp. I 480 0.08 0.55 0.19 0.58 1.32
Chrysobalanaceae
Atuna racemosa ssp. racemosa Raf. 531 0.Q3 0.18 0.08 0.19 0.46
Licania splendens (Korth.) Prance 32 0.02 0.18 0.04 0.19 0.42
Paraslemon grandifrucms Prance 155 0.01 0.18 0.03 0.19 0.41
Parinan' canarioides Kostenn. 259 0.03 0.18 0.06 0.19 0.44
Parinari oblongijolia Hk. f. 324 0.01 0.18 0.03 0.19 0.41
Clusiaceae
Calophyl/um nOMsum Yes que 69 0.01 0.18 0.02 0.19 0.40
Garcinia penangiana Pierre 399 0.01 0.18 0.03 0.19 0.41
Garcinia sp. 1 481 0.01 0.18 0.02 0.19 0.40
Appendix (continued). Tree species sampled at Belalong in the plot west of Kuala Belalong Field Studies Centre
arranged alphabetically in families. One voucher number is given after the name. Number of individuals, basal
area, frequency, relative density (reI. dens.), dominance (reI. dom.), frequency (reI. fre.) and Importance Value
Index (lV1) are given for each species.
S~ecisNouhr Ind. bas.a fre. rel.den. rel.dom. rel.fre. IV!
Clusiaceae (continued)
Garcinia sp. 2 49 2 0.02 2 0.36 0.05 0.39 0.80
Gareinia sp. 3 98 2 0.07 2 0.36 0.16 0.39 0.92
Garcinia sp. 4 517 0.01 1 0.18 0.03 0.19 0.40
Mesua beeeariana (BaHI.) Kostenn. 15 2 0.08 2 0.36 0.20 0.39 0.95
Mesua grandis (King) Kostenn. 405 0.06 0.18 0.14 0.19 0.52
Mesua oblongifolia (R.SI.) Kostenn. 390 0.02 2 0.36 0.05 0.39 0.80
Crypteroniaceae
CrypleronifJ panieulala BI. 303 2 0.02 2 0.36 0.05 0.39 0.80
Dilleniaceae
Dillenia beecariana Martelli 412 0.01 0.18 0.03 0.19 0.40
Dipterocarpaceae
Anisoplera laevis Ridl. 224 0.21 0.55 0.52 0.58 1.65
Dipleroearpus eaudatus ssp. penangianus
(Foxw.) Ashton 13 6 1.28 6 1.09 3.13 1.17 5.39
Dryobalanops beeearii Dyer 4 0.23 2 0.55 0.56 0.39 1.50
Dryobalanops laneeolala Burck 135 0.15 0.18 0.36 0.19 0.73
Hopea dryobalanoides Miq. 143 0.38 0.55 0.94 0.58 2.06
Hopea sangal Korth. 512 0.01 0.18 0.02 0.19 0.40
Parashorea smylhiesii Wyatt-Smith ex Ashton 148 2 0.24 2 0.36 0.59 0.39 1.35
Shorea amplexieaulis Ashton 60 7 0.64 6 1.27 1.58 1.17 4.01
Shorea corifusa Ashton 178 0.14 0.18 0.33 0.19 0.71
Shorea Mmaliosa Ashton 211 6 1.12 4 1.09 2.76 0.78 4.62
Shorea fagueliana Heim 85 8 0.67 8 1.45 1.65 1.55 4.66
Shorea fallax Meijer 245 3 0.42 2 0.55 1.02 0.39 1.96
Shorea ferruginea Dyer ex Brandis 5 6 0.77 6 1.09 1.89 1.17 4.14
Shorea foxworthyi Sym. 149 0.01 0.18 0.04 0.19 0.41
Shorea hypoleuea Meijer 34 2 0.15 0.36 0.37 0.19 0.93
Shorea laevis Rid!. II 7 5.98 7 1.27 14.66 1.36 17.29
Shorea leprosula Miq. 23 1 0.07 1 0.18 0.18 0.19 0.55
Shorea maeroplera ssp. maeroplerifolia Dyer 95 6 0.42 5 1.09 1.03 0.97 3.09
Shorea maxwelliana King 467 3 0.30 3 0.55 0.74 0.58 1.87
Shorea obseura Meijer 204 0.44 0.18 1.07 0.19 1.44
Shorea parvifolia Dyer 6 17 5.03 14 3.09 12.34 2.72 18.15
Shorea paueiJIora King 45 4 1.83 4 0.73 4.48 0.78 5.99
Shorea seaberrima Burck 126 0.01 1 0.18 0.03 0.19 0.41
Shorea scrobiculata Burck 444 0.05 0.18 0.12 0.19 0.49
Shorea sp 400 0.03 0.18 0.08 0.19 0.45
Valiea endertii Siool. 62 16 0.78 15 2.91 1.91 2.91 7.74
Valiea oMrala (Griff.) Sym. 28 21 0.90 16 3.82 2.20 3.11 9.12
Valiea sp. 1 397 0.04 0.18 0.11 0.19 0.48
Ebenaceae
Diospyros bomeensis Hiern 59 6 0.09 1.09 0.22 1.17 2.48
Diospyros sp. 1 48' 6 0.23 6 1.09 0.57 1.17 2.83
Diospyros sp. 2 25 0.01 1 0.18 0.02 0.19 0.40
Diospyros sp. 3 511 0.Q3 0.18 0.08 0.19 0.46
Diospyros sp. 4 525 1 0.02 1 0.18 0.05 0.19 0.42
Diospyros sp. 5 31 I 0.17 1 0.18 0.41 0.19 0.79
Diospyros sp. 6 152 2 0.09 2 0.36 0.21 0.39 0.96
Diospyros sp. 7 375 0.02 1 0.18 0.05 0.19 0.43
Diospyros sp. 8 317 I 0.02 0.18 0.05 0.19 0.42
Diospyros sp. 9 362 1 0.17 I 0.18 0.43 0.19 0.80
Diospyros sp. 10 278 4 0.07 4 0.73 0.17 0.78 1.67
Euphorbiaceae
Aporusa sp. 1 81 4 0.05 4 0.73 0.13 0.78 1.63
Aporusa sp. 2 292 1 0.02 0.18 0.05 0.19 0.43
Aporusa sp. 3 319 0.02 0.18 0.05 0.19 0.43
area, frequency, relative density (reI. dens.), dominance (reI. dom.), frequency (reI. fre.) and hnportance Value
Index (!VI) are given for each species.
S~iesNouchr Ind. bas. a fre. reLden. reLdom. reLfre. NI
Euphorbiaceae (continued)
Baccaurea /atifolia King ex Hk. f. 252 0.01 0.18 0.02 0.19 0.40
Baccaurea sp. 1 80 0.01 0.18 0.03 0.19 0.40
Baccaurea sp. 2 212 0.04 0.18 0.09 0.19 0.47
Baccaurea sp. 3 455 0.03 0.18 0.08 0.19 0.45
Baccaurea sp. 4 477 0.01 0.18 0.02 0.19 0.40
Blumeodendron calophyllum Airy Shaw 120 0.01 0.18 0.03 0.19 0.40
Drypetes kikir Airy Shaw 272 2 0.15 0.36 0.38 0.39 1.13
Drypetes sp. 1 99 1 0.01 1 0.18 0.02 0.19 0.40
Elateriospermum tapos B!. 55 35 2.36 28 6.36 5,78 5.44 17.58
Koilodepas longifolium Hk. f. 356 0.04 0.55 0.10 0.19 0.84
Macaranga kingii var plaryphy/Ia Hk. f.381 0.01 0.18 0.02 0.19 0.40
Macaranga repando-den/ata Airy Shaw 280 0.13 1.45 0.32 1.55 3.33
Macaranga trachophylla Airy Shaw 326 0.01 0.18 0.02 0.19 0.40
Mallotus sp. 1 74 0.07 7 1.27 0.16 1.36 2.79
Moultonianthus leembruggianus (Boer!. & Koord.)
v. Stoonis 105 16 0.40 13 2.91 0.99 2.52 6.42
Neosconechinia kingii (Hook. f.)
Pax & Hoffin. 167 0.02 2 0.36 0.04 0.39 0.79
Pimelodendron grilJithianum (Muell. Arg.)
Benth. 30 4 0.08 4 0.73 0.19 0.78 1.69
Ptychopyxis javanica (J.1. Smith) Croiz. 114 0.01 0.18 0.02 0.19 0.40
Fagaceae
Castanopsis megacarpa Gamble 29 0.05 0.18 0.13 0.19 0.51
Castanopsis sp. 1 325 0.24 0.18 0.59 0.19 0.97
£ithocarpus sp. 1 16 2 0.21 2 0.36 0.52 0.39 1.27
£ithocarpu sp. 2 117 1 0.28 1 0.18 0.67 0.19 1.05
Quercus chrysotricha A. Camus 493 0.01 0.18 0.03 0.19 0.40
Quercus sp. 1 193 0.01 0.18 0.03 0.19 0.40
Flacourtiaceae
Hydnocarpus bomeensis Sleum. 271 3 0.08 0.55 0.19 0.58 1.32
Hydnocarpus pinguis Sleum. 184 2 0.29 2 0.36 0.70 0.39 1.45
Hydnocarpus sumatrana (Miq.) Koord. 17 15 0.61 13 2.73 0.11 0.39 0.86
Hydnocarpus woodii Merr. 82 3 0.19 0.55 0.08 0.58 1.21
Hydnocarpus sp. 1 36 2 0.04 2 0.36 0.08 0.19 0.46
Hydnocarpus sp. 2 432 3 0.03 3 0.55 1,50 2.52 6.75
Hydnocarpus sp. 3 513 0.03 0.18 0.47 0.58 1.59
Ryparosa baccaureoides Sleum. 350 0.02 0.18 0.06 0.19 0.44
Ryparosa kosternrnnnsii Sleum. 77 7 0.24 6 1.27 0.58 1.17 3.02
lcacinaceae
Gonocaryum minus Sleum. 12 0.09 0.55 0.23 0.58 1.36
Stemonurus umbellatus Becc. 187 0.03 0.55 0.08 0.58 1.21
Ixonanthaceae
lrvingia malayana Oliv. e! Benn. 130 0.32 0.18 0.79 0.19 1.16
Lauraceae
Beilschmedia sp. \07 0.04 0.18 0.11 0.19 0.48
Endiandra clavigera Kosterm. 257 0.04 0.18 0.10 0.19 0.48
Endiandra sp. 1 44 0.16 0.18 0.39 0.19 0.77
Endiandra sp. 2 547 0.01 0.18 0.03 0.19 0.41
Phoebe sp. 86 0.05 0.18 0.12 0.19 0.50
genus inde! sp. 1 181 0.12 0.18 0.30 0.19 0.68
genus indet sp. 2 306 0.01 0.18 0.02 0.19 0.40
Lecythidaceae
Barringtonia sp. 1 449 0.01 0.18 0.03 0.19 0.40
Barringtonia sp. 2 127 0.04 0.18 0.\0 0.19 0.48
Index (IV!) are given for each species.
SeeciesNollcher Ind. bas.a fre. rel.den. rel.dOffi. rel.fre. IVI
Leguminosae
Archidendron microcarpum (Benth.)
Ivan Nielsen 494 0.04 0.18 0.\0 0.19 0.47
Crudia sp. 431 0.11 0.18 0.26 0.19 0.64
Koompassia excelsa (Becc.) Taubert 458 0.07 0.18 0.16 0.19 0.54
Linaceae
CtenoLophon parvifolius Oliv. 96 0.09 0.18 0.23 0.19 0.60
Loganiaceae
Norrisia maior Soler. 469 0.06 0.18 0.14 0.19 0.52
Melastomataceae
MemecyLon sp. 1 35 0.02 0.18 0.05 0.19 0.43
Memecylon sp. 2 343 0.01 0.18 0.02 0.19 0.40
Memecylon sp. 3 177 1 0.04 0.18 0.11 0.19 0.49
Prernandra sp. 1 215 2 0.21 2 0.36 0.51 0.39 1.27
Prernandra sp. 2 347 0.01 0.18 0.02 0.19 0.40
Meliaceae
Aglaia sp. I 304 0.02 0.18 0.04 0.19 0.41
Aglaia sp. 2 333 0.02 0.18 0.04 0.19 0.41
Aglaia sp. 3 456 0.05 0.18 0.13 0.19 0.51
Aglaia sp. 4 457 0.01 0.18 0.02 0.19 0.40
Aglaia sp. 5 487 0.06 0.18 0.15 0.19 0.52
Azadirachta excelsa (Jack) Jacobs 162 0.06 0.18 0.14 0.19 0.52
Disoxylum cauliflorum Hiero 264 0.01 0.18 0.02 0.19 0.40
Sandoricum borneense Miq 479 0.01 0.18 0.03 0.19 0.40
genus indet (alive no leaves) 465 0.02 0.18 0.06 0.19 0.43
Moraceae
Artocarpus lanceifolius Roxb. 373 I 0.02 1 0.18 0.05 0.19 0.43
Artocarpus nitidus Tree. 2 2 0.11 2 0.36 0.26 0.39 1.02
Artocarpus odoratissima Blanco 232 2 0.10 2 0.36 0.26 0.39 1.01
Artocarpus lamaran Becc. 520 I 0.17 0.18 0.02 0.19 0.40
Artocarpus sp. I 537 0.01 0.18 0.43 0.19 0.80
Ficus sp. I 175 0.20 0.18 0.49 0.19 0.86
Ficus sp. 2 433 0.02 0.18 0.04 0.19 0.41
Myristicaceae
Gymnacranthera conlracla Warb. 144 0.08 0.18 0.21 0.19 0.58
Gymnacranthera farquhariana (HOOk. f. & Th.)
Warb.374 0.05 0.18 0.13 0.19 0.50
HorsfteLdia bomeensis de Wilde 70 0.10 0.18 0.26 0.19 0.63
Horsfteldia polyspherula (Hook. f. emend. King)
Sinclair 378 0.04 0.55 0.10 0.58 1.22
Horsfteldia reticu/ata Warb. 509 0.01 0.18 0.04 0.19 0.41
Horsfteldia wa/lichii'(Hook. f. & Th.) Warb. 328 0.06 0.18 0.15 0.19 0.52
Knema latericia Elmer ssp. ridleyi (Gandoger)
de Wilde 18 2 0.02 2 0.36 0.06 0.39 0.81
Knema lalifolia Warb. 19 2 0.03 2 0.36 0.08 0.39 0.83
Knema lingUiformis (Sinclair) de Wilde 100 2 0.08 2 0.36 0.19 0.39 0.94
Knema membranifolia H. Winlder in Engl. 546 0.01 0.18 0.02 0.19 0.40
Knema oblongata Merr. ssp. oblongata 290 0.01 0.18 0.03 0.19 0.41
Knema percoriacea Sinclair foona sarawakensis
de Wilde 411 0.01 0.18 0.02 0.19 0.40
Knema sp. I woodii/korthalsii/pulcbra 408 0.06 0.18 0.15 0.19 0.52
Knema sp. 2 glauca group 535 0.02 0.18 0.04 0.19 0.41
Myristica bornunsis Warb. 3 0.07 0.55 0.18 0.58 1.31
Myristica iners BI. 201 2 0.35 2 0.36 0.85 0.39 1.60
Myristica maxima Warb. 102 2 0.03 2 0.36 0.07 0.39 0.82
arranged alphabetically in families. One voucher nwnber is given after the name. Nwnber of individuals, basal
Index (lVI) are given for each species.
SEeciesNoucher Ind. bas. a fre. reLden. reLdom. reLfre. IVI
Myrsinaceae
Ardisia sp. 464 0.03 0.18 0.08 0.19 0.45
Myrtaceae
Eugenia sp. 1 52 1 0.01 1 0.18 0.02 0.19 0.40
Eugenia sp. 2 353 2 0.03 2 0.36 0.07 0.39 0.82
Eugenia sp. 3 205 0.03 2 0.36 0.06 0.39 0.81
Eugenia sp. 4 179 4 0.31 4 0.73 0.77 0.78 2.27
Eugenia sp. 5 156 0.04 0.18 0.09 0.19 0.46
Eugenia sp. 6 462 0.Q3 0.18 0.07 0.19 0.45
Eugenia sp. 7 112 0.15 0.18 0.36 0.19 0.74
Eugenia sp. 8 287 0.01 0.18 0.02 0.19 0.40
Eugenia sp. 9 305 0.01 0.18 0.03 0.19 0.41
Eugenia sp. 10 233 0.03 0.18 0.07 0.19 0.44
Eugenia sp. 11 274 0.03 0.18 0.07 0.19 0.45
Olacaceae
Strombosia sp. 0.06 0.18 0.14 0.19 0.51
Po1yg.1.ceae
XanIhophylum .denotus Miq. 484 1 0.01 0.18 0.03 0.19 0.41
XanzhophyUum sp. 1 40 6 0.38 1.09 0.94 1.17 3.20
Xanlhophyllum sp. 2 65 3 0.04 0.55 0.10 0.58 1.23
Xanlhophyllum sp. 3 110 2 0.06 2 0.36 0.14 0.39 0.90
Xanlhophyllum sp. 4 54 0.05 0.55 0.11 0.58 1.24
XanIhophyllum sp. 5 64 0.08 0.55 0.21 0.58 1.33
Proteaceae
Helicia sp. 41 2 0.06 0.36 0.15 0.39 0.91
Rosaceae
Maranlhes corymbosa 81. 406 0.20 0.18 0.50 0.19 0.87
Maranlhes corymbosis 81. 168 0.03 0.18 0.06 0.19 0.44
Prunus beccarii Ridl. 370 0.07 0.18 0.16 0.19 0.54
Rubiaceae
Adina sp. 413 0.01 0.18 0.03 0.19 0.40
Canlhium sp. 1 0.01 0.18 0.03 0.19 0.40
Neonauclea sp. 1 489 0.01 0.18 0.03 0.19 0.40
Neonauclea sp. 2 387 0.01 0.18 0.02 0.19 0.40
Psydrax sp. 1 388 0.01 0.18 0.03 0.19 0.40
Psydrax sp. 1 507 0.04 0.18 0.11 0.19 0.48
Urophyllum sp. 73 0.13 0.55 0.32 0.58 1.45
genus inde! 172 0.02 0.18 0.04 0.19 0.41
Sabiaceae
Meliosma sp. 460 0.01 0.18 0.04 0.19 0.41
S.pind.ceae
Nephelium sp. 357 2 0.05 2 0.36 0.13 0.39 0.88
S.polaceae
Madhuca sp. 1 318 0.05 0.18 0.12 0.19 0.49
Madhuca sp. 2 227 1 0.02 0.18 0.05 0.19 0.43
Pa/aquium sp. 1 111 2 0.02 2 0.36 0.06 0.39 0.81
Pa/aquium sp. 2 527 1 0.04 1 0.18 0.09 0.19 0.47
Payena sp. 1 43 2 0.09 2 0.36 0.23 0.39 0.98
Pouleria sp. 2 159 0.05 1 0.18 0.13 0.19 0.50
genus inde!. 331 2 0.06 2 0.36 0.14 0.39 0.89
Index (WI) are given for each species.
SpeciesNouchcr Ind. bas.a fre. reLden. reLdom. reLfre. IVI
Sterculiaceae
Heritiera sumalrana (Miq.) Ko,term. 124 0.01 0.18 0.02 0.19 0.40
Scaphium linearicarpum (Mast.) Pierre 285 0.28 0.91 0.68 0.97 2.56
Scaphium macropodum (Miq.) Blume ex Heyne 47 0.25 0.18 0.61 0.19 0.99
Slerculia sconechinii King 277 2 0.03 2 0.36 0.08 0.39 0.83
Slerculia 'p. 226 0.02 0.18 0.05 0.19 0.43
Thymelaeaceae
Gonoslylus 'p. 27 0.02 0.18 0.04 0.19 0.42
Tiliaceae
Grewia cinnamomifolia (Burrett) Stapf ex
Ashton 473 0.01 0.18 0.03 0.19 0.40
Penrace laxiflora Merr. 355 0.03 0.18 0.07 0.19 0.44
Ulmaceae
Gironniera nervosa Planch. 254 2 0.02 2 0.36 0.05 0.39 0.80
Verbenaceae
Teijsmanniodendron sarawakanum
(H.H.W. Pearson) Kosterm. 323 4 0.05 4 0.73 0.13 0.78 1.64
Forest bird diversity in Oceania
JIRO KIKKAW A
Centre for Conservation Biology, the University of Queensland, Brisbane, Qld 4072,
Australia
Forest bird diversity in Oceania reflects biogeographical patterns and historical factors of
distribution rather more than the degree of niche differentiation that the forest birds are
expected to have undergone. There are five main gradients of diversity found to the east and
south of Wallace's Line, and these are summarised below in terms of species richness.
1) On the Australian continent climatic cycles of the past produced more dry-adapted
forms in wetter periods than wet-adapted forms in drier periods (Brereton & Kikkawa 1963).
This is because the continent has primarily been a dry continent dominated by sclerophyllous
vegetation since the late Tertiary. Thus a prominent, biogeographical, gradient is found
from the greatest diversity in open eucalypt forests (45 species per 8 ha and a high
beta-diversity) to coastal closed forests (38 species per 8 ha and a low beta-diversity) and to
inland savannahs and woodlands (25 species per 8 ha and a high beta-diversity) in the
mid-latitude range (Kikkawa 1968).
2) The second gradient on the continent is developed in rainforest vegetation in which the
latitudinal gradient associated with the resources of the forest is seen in forest birds (Kikkawa
1991). This gradient parallels the size of the area and in the wet tropical region, where the
rainforest area is the largest on the continent, the diversity is greatest with 100 species,
decreasing to 80 species in subtropical, and 42 species in temperate rainforests.
3) The third gradient on the continent is seen along the moisture gradient associated with
seasonality of rainfall. From the wet tropics of the Cardwell-Cooktown region of north
Queensland the forest bird fauna is attenuated towards the wet-dry region of Cape York
Peninsula (Kikkawa et al. 1981a) and the Northern Territory (Kikkawa et al. 1981b). On
the drier side rainforest vegetation occurs in smaller patches of monsoon forests containing
many dry-defoliating tree species. Here the number of forest bird species declines to 85 with
many replacements with dry-adapted forms.
4) In contrast to Australia, New Guinea is topographically very complex for its area, with
great opportunities for speciation among the wet-adapted forms. Many of them exhibit
conspicuous altitudinal zonation (Diamond 1972). An altitudinal gradient of diversity is
found from the lowland rainforests (184 species at the Brown River, 147 species at the
Trans-Gogol) to highlands (144 species in the Central Highlands) and mountain tops (8
species) (Driscoll & Kikkawa 1989, Kikkawa & Williams 1971b). This pattern is
resource-related but, when the entire island is considered, the altitudinal gradient is even
more pronounced, reflecting allopatry due to geographical isolation of rainforest areas in
different altitudinal zones (Kikkawa & Williams 1971a).
5) The oceanic islands of the South Pacific have been colonised largely by birds of the
Western Pacific region (Mayr 1945). The underlying diversity is therefore related to the
distance from this source of colonisation. This gradient is shown for many plant and animal
groups (Stoddart 1992). The number of forest bird species decreases from New Guinea (445)
to Solomon Islands (127), New Caledonia (52), Central Polynesia (34) and south-eastern
Polynesia (40). The only exception is found in the Hawaiian Islands where most of the forest
birds are derived from the New World (Diamond 1984). These include the endemic family
151
152 JIRO KIKKAWA
of honeyceepers which have undergone spectacular radiation on the islands. On a much

smaller scale speciation is evidenced among the white-eyes (Zosterops) in Micronesia and
pigeons and parrots in Polynesia and Fiji. The fauna is impoverished on smaller and remote
islands of Central Pacific (e.g. 12 species on Marshall, Nauru, Kiribati, Tokelau Islands
(Pratt et at. 1987» (Whiffin & Kikkawa 1992).
In historical times many exotic species have been introduced into islands either
accidentally or on purpose. Some of them have established themselves in indigenous forests
(e.g. New Zealand). Indigenous species were vulnerable to introduced influences on the
islands and many forest birds became extinct in the past hundred years. These influences
include habitat destruction caused by forest clearing, heavy grazing and disturbance by
introduced mammals, competition from newly established species, predation by introduced
predators (e.g. black rats, stoats, cats) and diseases (e.g. bird malaria).
These trends are significant modifiers of diversity gradients seen in the forest birds of the
South Pacific region.
LITERATURE CITED
BRERETON, I. Ie G. & KlKKAWA, I. 1963. Diversity of avian species. Australian Joumal of Science 26:12-14.
DIAMOND, I.M. 1972. Avifauna of the Eastern Highlands of New Guinea. The Nuttall Ornithological Club,
Cambridge, Massachusetts. 438 pp.
DIAMOND, I.M. 1984. Biogeographic mosaics in the Pacific. pp. 1-14 in Radovsky, F.I., Raven, P.H. & Sohmer,
S.H. (eds). Biogeography of the tropical Pacific. Bishop Museum, Honolulu.
DRISCOLL, P.V. & KlKKAWA, I. 1989. Bird species diversity of lowland tropical rainforests of New Guinea
and Australia. pp. 123-152 in Harmelin-Vivien, M. & Bourliere, F. (eds). Vertebrates in complex tropical
systems. Springer-Verlag, New York.
KlKKAWA, I. 1968. Ecological association of bird species and habitats in eastern Australia: similarity analysis.
Journal of Animal Ecology 37:143-165.
KlKKAWA, I. 1991. Avifauna of Australian rainforests. Pp. 187-196 in Werren, G. & Kershaw, P. (eds). The
rainforest legacy: Australian national rainforests study. Volume 2. Australian Heritage Commission, Canberra.
KlKKAWA, I., MONTEITH, G.B. & INGRAM, G.I. 1981a. Cape York Peninsula: a major region of faunal
interchange. pp. 1695-1742 in Keast, A. (ed.). Ecological biogeographY of Australia. W. Iunk, The Hague.
KlKKAWA, I., WEBB, L.I., DALE, M.B., MONTEITH, G.B., TRACEY, I.G. & WILLIAMS, W.T. 1981b.
Gradients and boundaries of monsoon forests in Australia. Proceedings of the Ecological Society of Australia
11:39-52.
KlKKAWA I. & WILLIAMS, W.T. 1971a. Altitudinal distribution of land birds in New Guinea. Search 2:64-65.
KlKKAWA, I. & WILLIAMS, W.T. 1971b. Ecological grouping of species for conservation of land birds in New
Guinea. Search 2:66-69.
MAYR, E. 1945. Birds of the Southwest Pacific. Macmillan, New York. 316 pp.
PRATT, H.D., BRUNER, P.L. & BERRETT, D.G. 1987. The birds of Hawaii and the tropical Pacific. Princeton
University Press, Princeton. 409 pp. + 45 plates.
STODDART, D.R. 1992. Biogeography of the tropical Pacific. Pacific Science 46:276-293.
WHIFFIN, T. & KlKKAWA, I. 1992. The status of forest biodiversity in Oceania. Joumal of Tropical Forest
Science 5: 155-172.
The fauna and flora of a semi-deciduous forest in Perlis, Peninsular
Malaysia
DIONYSIUS SHARMAI, HYDEIR KAMARUDIN2, HALIJAH IBRAHIM 3 and ONG

HEAN CH00I3
IWWF Malaysia, Locked bag No. 911, lIn Sultan P.O., 46990 Petaling laya, Selangor,
Malaysia
2Malayan Nature Society, No. 485, lIn 5/53, 46000 Petaling laya, Selangor, Malaysia
3Department of Botany, University of Malaya, 59100 Kuala Lumpur, Malaysia
ABSTRACT. A study was conducted to survey the fauna and flora at Mata Ayer Forest Reserve (3176
hectares) and the proposed Bukit Wang Mu Forest Reserve (2768 hectares), Periis, Peninsular Malaysia.
Thirty seven mammal species were recorded, including five primates and five ungulates. The sighting
of the Stump-tailed macaque (Macaca arctoides) at Mata Ayer confirms its existence in Peninsular
Malaysia. Thirty five species of reptiles were recorded. Eighty eight species of birds, including six
species of hornbills were observed. One hundred and forty three non-timber plant species (herb, climber,
shrub, sedge, grass, rotan) were identified. The study sites are important as water catchment areas. The
establishment of a transfrontier reserve is recommended.
KEY WORDS: Mata Ayer Forest Reserve. Bukit Wang Mu Forest Reserve, transfrontier reserve.
INTRODUCTION
Perl is (6°15' - 6°43'N, 1()(f6' -1()(f23 'E) is the smallest and northernmost State of Peninsular
Malaysia, covering an area of some 81,063 hectares. It experiences a strongly monsoonal
climate with an average annual rainfall of 1,735 mm (WWFM 1984) and a 3-4 month dry
season. The original forest over most of Perlis (semi-deciduous with Burmese affinities) has
been cleared for cultivation, or converted by cutting and fire into secondary Schima-bamboo
forest.
The main forested areas (unalienated forested State land and forest reserves) left in Perl is
lie in a continuous belt along the western border (Figure 1). The forests lie mainly on hilly
terrain of limestone with numerous caves. The limestone forms part of the Setul Formation.
Within this forested belt is the Mata Ayer Forest Reserve (3176 hectares) and the proposed
Bukit Wang Mu Forest Reserve (2768 hectares). Both these areas lie adjacent to Thaleban
National Park (NP) in Thailand and are important water catchment areas for the Timah-Tasoh
Dam, situated in the south. Forest clearance projects, mainly for agriculture, have occurred
in most of the valley-floor compartments of Mata Ayer. Due to this, the IUCN Commission
on National Parks and Protected Areas felt compelled to place Thaleban NP on their list of
'Threatened Protected Areas of the World' (IUCN 1988). Since the forest at Mata Ayer and
Bukit Wang Mu is unique to Peninsular Malaysia, and in view of the importance of the area
for water catchment, the potential clearance of the forests and consequent threat to Thaleban
NP, surveys were conducted to investigate the fauna and non-timber flora. The objective of
the study was to collate the scientific information with the intention of developing a viable
management plan for the area.
153
154 D. SHARMA, H. KAMARUDlN, H. IBRAHIM AND H.C. ONG
/ .I . . . .·,.
-
...
Stertl boundorr
RoIIwoy Ira
( For.., ' ...ne NuncIor,.

Wet. cotdWNnI boundary
~I •• I' 30S,I04
Figure 1. Existing and proposed nature reserves in Periis, Penninsular Malaysia.
METHODS
Surveys were carried out at Mata Ayer between May and August 1992 and at Bukit Wang
Mu in October and November 1992. Three line transects, each a kilometre long were cut
in Mata Ayer and Bukit Wang Mu respectively. These were walked daily between 0700-
0930 hours. Any mammal, bird or reptile observed along the transect was recorded. Calls
of known animals (e.g. gibbons, dusky langurs, barbets, argus pheasants) were recorded as
well. Animal tracks (especially of ungulates) and remains (e.g. skull, bones, quills, scales)
were recorded when encountered. Night transect walks were conducted between 2100-2330
hours to investigate the nocturnal fauna. One hundred small mammal cage traps were set
along the transects to investigate the squirrel, treeshrew and rat diversity. Traps were placed
on the ground and up in the trees (two metres off the ground) at 25 metre intervals. At each
locality surveyed, plants were collected for herbarium specimens and wherever possible,
samples were selected for planting in the Botanic Garden, Department of Botany, University
of Malaya. Specimens were identified with the aid of reference books (e.g. Henderson 1974,
Holttum 1966, Keng 1969, Whitmore 1977) and by reference to herbarium specimens lodged
in the Department of Botany, University of Malaya.
Fauna and flora of a semi-deciduous forest 155
RESULTS
Thirty eight mammal species from twenty one different families were confirmed in the area
(Appendix 1). In addition to this, several flying squirrels were seen but could not be
identified to species level. Five species of primates and five ungulates were recorded. The
most commonly seen primate was the White-handed gibbon (Hylobates lar). A troop of wild
Stump-tailed macaques (Macaca arctoides) was observed in compartment 6, Mata Ayer FR.
The group was estimated to include at least eight individuals. Two individuals were observed
feeding on the ripe fruits of Baccaurea kunstleri (Sharma, in preparation). The troop was
approximately 500 metres from the Malaysia/Thailand border.
Eighty eight species of birds from 34 families were recorded during the surveys (Appendix
1). The six species of hornbill recorded were the White-crowned hornbill (Berenicornis
comatus), Bushy-crested hornbill (Anorrhinus galeritus), Wreathed hornbill (Rhyticeros
undulatus), Rhinoceros hornbill (Buceros rhinoceros), Great hornbill (Buceros bicornis) and
Helmeted hornbill (Rhinoplax vigil). The most common was the Great hornbill, frequently
seen in pairs (n=12), although small flocks of three were observed as well (n=3). Two
species of threatened Malaysian birds, the Chestnut-necklaced partridge (Arborophila
charltom) and Helmeted hornbill (Rhinoplax vigil), were recorded (see Collar & Andrew
1988).
Thirty five species of reptiles, including 21 species of snakes, five species of batagurid
turtles, one trionychid and a tortoise, were recorded (Appendix 1). The striped racer (Elaphe
taeniura) was often seen in caves, and together with the Oriental whip snake (Ahaetulla
prasina), were the two most commonly encountered snakes. The Asian leaf turtle (Cyclemys
dentata) was frequently encountered near slow-flowing streams, and three specimens were
collected from inside caves. 1uvenile specimens of the Malayan softshell (Dogania subplana)
(carapace length 50mm) and Malayan flat shell (Notochelys platynota) (carapace length
88mm) were collected from stagnant pools.
One hundred and forty three non-timber plant species from 39 families were identified
(Appendix 2). These include 17 species of orchids and seven species of palms. The Bukit
Rongkit limestone hill flora in the Mata Ayer FR was interesting in that it includes seven
species of orchids, two species of Begonia (B. guttata and B. phoeniogramma) and
Maxburetia rupicola, a palm which is endemic to Peninsula Malaysia and of limited
distribution. Of the 23 zingiberaceous species recorded, 36 % were in flower, 33 % sterile,
9% in bud stage, 9% were over flower and 5 % were in fruit. Three zingiberaceous species,
Zingiber zerumbet, Globba sp. and Camptandra sp., found on the slope of Bukit Rongkit
Utara, Mata Ayer FR, have not previously been recorded in limestone habitats.
DISCUSSION
The Asian elephant (Elephas maximus) and the Asian two-horned rhinoceros (Dicerorhinus
sumatraensis) no longer occur in the Mata Ayer FR as a result of habitat loss (large scale
conversion of lowland forests to agricultural land). The remaining forested areas are on hilly
terrain (mainly limestone) and are not suitable habitats for these species. A confirmed
sighting of the Stump-tailed macaque in northern Peninsular Malaysia was made by Syed
Abdullah Idid, then Deputy Game Warden for Perlis/Kedah (Medway 1978). During their
primate survey of Peninsular Malaysia, Marsh and Wilson (1981) spent several days at Mata
Ayer but did not encounter these macaques. They were, however, informed by local people
that the macaques may sometimes still be seen. In addition to this, a group of these
macaques were reportedly seen in the State of Kelantan (Charles, 1.K., personal
communication). The Stump-tailed macaque was described as being 'possibly extinct' in
156 D. SHARMA, H. KAMARUDIN, H. IBRAHIM AND H.C. ONG
Peninsular Malaysia (Bennett 1991) as there were no confirmed sightings for the last fifteen
years. The sighting of M. arctoides at Mata Ayer therefore further substantiates previous
observations and confirms that this macaque still occurs in Peninsular Malaysia. Populations
may wander to and fro across the international border. In Thailand, M. arctoides is now
uncommon throughout its range and occurs only in isolated populations. It is nearing
extinction in Peninsular Thailand, where it was once abundant (Roonwal & Mohnot 1977,
Southwick & Siddiqi 1970).
Several other mammals recorded at Mata Ayer and Bukit Wang Mu are important in
conservation terms. Mammals likely to be threatened include the Moonrat (Echinosorex
gymnurus) and the Pencil-tailed tree mouse (Chiropodomys gliroides). Mammals which are
already threatened include Low's squirrel (Sundasciurus lowii), Panther (Panthera pardus)
and Serow (Capricomis sumatraensis).
The valley-floor compartments of Mata Ayer FR which have been cleared for agriculture,
were regarded as being important for several bird species including the Streaked-breasted
woodpecker (Picus viridanus), Dusky eagle owl (Bubo coromandus), Brown fish owl (Ketupa
zeylonensis), Hooded pitta (Pitta sordida) and the extremely rare Green-backed Narcissus
flycatcher (Ficedula narcissina elizae) (Wells 1987). Only the Hooded pitta was observed
during the present surveys restricted to the remaining valley compartments.
The Wildlife at Mata Ayer FR and Bukit Wang Mu is threatened by hunting and trapping.
Snare traps for Wild pig and Serow were seen, as well as bamboo traps for catching small
rodents. Remains of wildlife found (with signs that it had been eaten) include Pangolin
(Manisjavanica), Malayan porcupine (Hystrix brachyura), Dusky langur (Presby tis obscura),
Lesser mousedeer (Tragulus javanicus), Greater mousedeer (Tragulus napu), Spiny turtle
(Heosemys spinosa), Giant pond turtle (Heosemys grandis), Malayan flat-shelled turtle
(Notochelys platynota) and Elongated tortoise (Indotestudo elongata).
The limestone flora of Peninsular Malaysia has been well studied (e.g. Chin 1977, 1979,
1983a, 1983b, Henderson 1939, Kiew 1991). Several herbaceous plants found on limestone
have great potential as ornamentals, especially those from the Balsaminaceae, Begoniaceae
and Gesneriaceae. The species collected and identified at Mata Ayer FR to date are only a
small fraction of the total number that may exist in the area. A systematic resurvey is
required to fully document the flora.
The retention of natural forest cover at Mata Ayer and Bukit Wang Mu is crucial as these
areas play a vital role in the hydrological cycle for the State of Perl is and are important
catchment areas for the Timah-Tasoh Dam (Kumari & Ariffin 1988, WWFM 1984). During
the drier months, demand for water for domestic consumption, industry and agriculture will
have to be met. The loss of surface water, stagnant or flowing, will have immediate
consequences on both the terrestrial and aquatic fauna.
On an international level, the establishment of a transfrontier reserve will serve to protect
the fauna and flora along the common border. Any degradation or reduction of the forests
at Mata Ayer and Bukit Wang Mu will not only extirpate Malaysian fauna and flora but also
exert pressure on the biodiversity existing in Thaleban NP, Thailand.
ACKNOWLEDGEMENTS
We would like to thank the then Director of the Perlis Forestry Department, En. Borhanudin
bin Haji Arshad, and the Assistant Director, En. Zainal Abidin, for permission to carry out
field work. Thanks also to the staff of the Police Field Force, staff of the Balai Polis Kaki
Bukit, Raleigh International, Caving Group of the Malayan Nature Society, Dr. Mikaail
Kavanagh and Dr. Isabelle Louis for their assistance and support. This work was funded by
the World Wide Fund for Nature, Malaysia.
LITERATURE CITED
BENNETT, E.L. 1991. Diurnal Primates. pp. 150-172 in Kiew, R. (ed.). The State of Nature Conservation in
Malaysia. Malayan Nature Society, Kuala Lumpur.
CHIN, S.C. 1977. The limestone flora of Malaya 1. Gardens' Bulletin Singapore 30:165-219.
CHIN, S.C. 1979. The limestone flora of Malaya 2. Gardens' Bulletin Singapore 32:64-203.
CHIN, S.C. 1983a. The limestone flora of Malaya 3. Gardens' Bulletin Singapore 35:137-190.
CHIN, S.C. 1983b. The limestone flora of Malaya 4. Gardens' Bulletin Singapore 36:31-91.
COLLAR, N.J. & ANDREW, P. 1988. Birds to watch. The ICBP world check list of threatened birds. ICBP
Technical Publications No.8. ICBP, Cambridge.
HENDERSON, M.R. 1939. The flora oflimestone hills of the Malay Peninsula. Journal of the Malayan Branch of
the Royal Asiatic Society 17: 13-87.
HENDERSON, M.R. 1974. Malayan Wild Flowers. 2 Volumes. The Malayan Nature Society, Kuala Lumpur.
HOLTTUM, R.E. 1966. A Revised Flora of Malaya. Volume II. Ferns of Malaya. Govt. Printing Office, Singapore.
IUCN 1988. IUCN Commission on National Parks and Protected Areas. Newsletter No.45. Gland, Switzerland.
KENG, H. 1969. Orders and Families of Malayan Seed Plants. University of Malaya Press, Kuala Lumpur.
KIEW, R. 1991. Botanical survey of limestone hills in Perak. Unpublished Report.
KUMARI, K. & ARRIFIN, I. 1988. Conservation Issues Critical to the Timah-Tasoh Dam and its Catchment at
Mala Ayer Forest Reserve in Perlis. WWF Malaysia, Kuala Lumpur.
MARSH, C.W. & WILSON, W.L. 1981. A Survey of Primates in Peninsular Malaysian Forests. Universiti
Kebangsaan Malaysia, Kuala Lumpur.
MEDWAY, LORD 1978. The Wild Mammals of Malaya (Peninsular Malaysia) and Singapore. Oxford University
Press, Kuala Lumpur. 131 pp.
ROONWAL, M.L. & MOHNOT, S.M. 1977. Primates of South Asia. Ecology, Sociobiology and Behaviour.
Harvard University Press, London. 421 pp.
SOUTHWICK, C.H. & SIDDIQI, M.P. 1970. Primate population trends in Asia, with specific reference to the
rhesus monkeys ofIndia. Papers and Proceedings of the 11th Technical Meeting of the IUCN (New Delhi, Nov.
1969) 1:135-147.
WELLS,D.R. 1987. A letter to the President, Chairman, Trustees and Scientific Advisory Committee, Wordl Wide
Fund for Nature, Malaysia, entitled 'Water and Wildlife Conservation in Perlis - A Crisis'.
WHITMORE, T.C. (ed.) 1977. Tree Flora of Malaya. 4 Volumes. Longman, Malaysia.
WWFM 1984. Proposals for a Conservation Strategy for Perlis. A paper submitted to the State Government of
Perlis. World Wide Pund for Nature, Malaysia, Kuala Lumpur.
APPENDIX 1. Checklist of the fauna occurring at Mata Ayer Forest Reserve and the proposed Bukit Wang Mu
Forest Reserve, Perlis, Peninsular Malaysia.
Farnil:!: Sl!ecies Common name

MAMMALS
Erinaceidae Echinosorex gmnurus Moonrat/Gymnure
Soricidae Crocidura fu 19inosa South-east Asian white-toothed shrew
Cynocephalidae f{,nocephalUS variegatus Colugo
Pteropodidae nopterus sphinx Short-nosed fruit bat
Rhinolophidae Rhinol0p'hus ~nis Intermediate horseshoe bat
Hipposlderos iadema Diadem roundleaf horseshoe bat
Vespertilionidae Greater flat -headed bat
Tupaiidae Tupaia glis Common treeshrew
Tupaia minor Lesser treeshrew
Cercopithecidae Presbytis obscura Dusky langur
Presbytis melol%hos Banded langur
Macaca fascicu ris Long-tailed macaque
Macaca arctoides Stump-tailed macaque
~Iobatide Ifllobates lar White-handed gibbOn
anidae anis javanica Malayan pangolin
Sciuridae Ratufa bicolor Black giant squirrel
Callosciurus notatus Plaintam squirrel
Callosciurus caniceps Grey-bellied squirrel
Callosciurus prevostii Prevost's squirrel
Sundasciurus lowii Low's squirrel
Lariscus insignis Three-striped ground squirrel
Petauristinae Unidentified
Muridae Ch:;;;gOdOmys gliroides Pencil-tailed tree mouse
Ra a tiomanicus Malaysian wood rat
Rattus bowersii Bower's rat
Rattus rajah Brown sYciny rat
Leopoldamys sabanus Long-tai ed giant rat
Hystricidae ~strix brachyura Malayan porcupine
Ursidae elarctos malayanus Malayan sun bear
Viverridae Paradoxurus hermaphroditus Common palm civet
Paguma larvata Masked palm civet
Herpestes urva Crab-eatmg mongoose
Felidae Panthera pardus Leopard
Rhinocerotidae Dicerorhinus sumatrensis Asiatic two-horned rhinoceros
Suidae Sus scrofa Wild pig
Tragulidae Tragulus javanicus Lesser mousedeer
Tragulus napu Greater mousedeer
Cervidae Muntiacus muntjak Barking deer
Bovidae Capricomis sumatrensis Serow
BIRDS
Accipitridae Elanus caeruleus Black-shouldered kite
~ilOmS cheela Crested serpent-eagle
?vizaetus cirrhatus Chan~ebl
hawk-eagle
zizaetus aiboniger Blyth s hawk-eagle
Phasianidae borophila charltoni Scaly-breasted partridge
Gallus gallus Red junglefowl
Argusianus argus Great argus
Turnicidae Tumix suscitator Barred buttonquail
Gruidae Rallina fasciata Red-legged crake
Columbidae Treron vemans Pink-necked tgeon
Chalc0fuhaPS indica Gren-wi~ pigeon
Psittacidae Loricu us galgulus Malay lori eet
Cuculidae Cuculus micropterus Indian cuckoo
Cacomantis merulinus Plaintive cuckoo
Chrysococcyx xanthorhynchus Violet cuckoo
Phaenicophaeus diardi Black-bellied malkoha
Phaenicophaeus chlorophaeus Raffles's malkoha
Centropus sinensis Greater couca1
Caprimulgidae Eurostopodus temminckii Malaysian eared nightjar
Caprimulgus macrourus Lafj~e-tild nigh~ar
Appodidae Collocalia esculenta White-bellied swi let
Rhophidura leucopygialis Silver-rumKed swift
~us affinis House swi
'Psiurus batasiensis Asian palm swift
Trogonidae arpectes duvaucelii Scarlet-rumped trogon
Alcedinidae Lacedo pulchella Banded kingfisher
Meropidae Merops leschenaulti Chestnut-headed bee-eater
Nyctyomis amictus Red-bearded bee-eater
APPENDIX I (Continued). Checklist of the fauna occurring at Mata Ayer Forest Reserve and the proposed Bukit
Wang Mu Forest Reserve, Perlis, Peninsular Malaysia.
Family Species Common name

BUcerotidae Berenicomis comatus White-crowned hornblll
Anorrhinus galeritus Bushy-crested hornbill
Rhyticeros undulatus Wreathed hornbill
Buceros rhinoceros Rhinoceros hornbill
Buceros bicomis Great hornbill
Rhino~lax vigil Helmeted hornbill
Capitonidae Mega ima chiffisopogon Gold-whiskered barbet
Megalaima ra eSli Many-colored barbet
Megalaima mystacophanos Red-throated barbet
Megalaima australis Blue-eared bar bet
Picidae Picoides canicapilius GrerycaPPed woodpecker
Eurylaimidae Corydon sumatranus Dus broadbill
Eurylaimus javanicus Banded broadbill
Eu,&laimus ochromalus Black and yellow broadbill
Ca ptomena viridis Green broadbill
Pittidae Pitta sordida Hooded pitta
Pitta guajana Banded pitta
Hirundinidae Hirundo daurica Red-rwnped swallow
Campephagidae Hemipus picatus Bar-winged flycatcher shrike
Chloropseidae Chioropsis sonnerati Greater green leafbird
Pycnonotidae Pycnonotus atriceps Black-headed bulbul
Pycnonotus melanicterus Black-crested bulbul
Pycnonotus eutilotus Puff-backed bulbul
Pycnonotus brunneus Red-eyed bulbul
~cn.otus e'JJ!,roPthalmos ~ectald bulbul
nmger jinsc II msch's bulbul
Criniger bres Grey-cheeked bulbul
HYPslpetes malaccensis Streaked bulbul
Dicruridae Dicrurus aeneus Bronzed drongo
Oriolidae [rena puella Asian fairy-bluebird
Paridae Melanochlora sultaneae Sultan tit
Tamiliidae Trichastoma malaccense Short-tailed babbler
Trichastoma rostratum White-chested babbler
Malacopteron magnirostre Moustached babbler
Alcil!fve brunneicauda Brown fulvetta
Stac yris erythroptera Chestnut-winged babbler
Stachyris rujiceps Rufous-capped babbler
Macronous gularis StriKYd tit-babbler
Macronous ~tiIOSU Flu -backed tit-babbler
Yuhina zant oleuca White-bellied ¥uhina
Turdidae Cal!sychus malabaricus White-rwnped shama
Emcurus rujicapullus Crested-naRed forktail
'i!f:,0G.honus caeruleus Blue whist ing trush
Sylviidae P y loscopus coronatus Eastern crowned warbler
Orthotomus sutorius Common tailorbird
Orthotomus atrogularis Dark-necked tailorbird
Orthotomus rujiceps Ashy tailorbird
Muscicapidae Culicicapa ceylonensis Grey-headed flycatcher
!Pr.pothYmis azurea Black-naped monarch
repsiphone paradisi Asian paradise flycatcher
Motacillidae Dendronanthus indicus Forest wagtail
Laniidae Lanius cristatus Brown shrike
Sturnidae Acridotheres tristis Common myna
Nectariniidae Hypogramma hYPofJrammicum Pu~le-nad sunbird
Arachnothera longlrostra Litt e spi erhunter
Arachnothera affinis Grefbreasted ~iderhunt
Dicaeidae Prionochillus maculatus Yel ow-breaste flowerpecker
Dicaeum trigonostigma Orange-bellied flowerpecker
Dicaeum cruentatum Scarlet-backed flowerpecker
Ploceidae Lonchura striata White-rwnped munia
REPTILES
TY~hOIidae J(~hlops diardi muelleri Diard's blind snake
Co ubri ae phe taemura Striped racer
Zaocys caranatus Keeled rat snake
Oligodon purpurascens Brown kukri snake
Xenelaphls hexagonotus Malayan brown snake (?)
Calamaria lumbricoidea Variable reed snake
Chrysopelea ornata Golden tree snake
Boiga jaspidea Jasper cat snake
Boiga drapiezii White-spotted cat snake
Psamnwtiynastes pulverulentus Mock viper (?)
APPENDIX I (continued). Checklist of the fauna occurring at Mata Ayer Forest Reserve and the proposed Bukit
Wang Mu Forest Reserve, Perlis, Peninsular Malaysia.
Family Species Common name

Colubridae Ahaetulla prasina Onental whip snake
Dryophios rubescens Keel-bellied whip snake
Xenochrophis triangu/igerus Triangle keel back
Xenochrophis piscator Chequered keel back
Macropisthodon rhodomelas Blue-necked keel back
Elapidae Bun8arus candidus MalaLan krait
Mattcora bivirgata Blue alayan coral snake
Naja nat: sputatrix Black spittmg cobra
Ophiop gus hannah King cobra
Viperidae Agkistrodon rhodostorna Malayan pit viper
Trimeresurus surnatranus Sumatran~it viper (?)
Bataguridae i1eclemys dentata Asian lea turtle
eosemys grandis Giant pond turtle
Heosemys spinosa ~iny turtle
Notochelys platynota alayan flat shell
Testudinidae Indotestudo elongata Elongated tortoise
Trionychidae Dogania subp/ana Malayan softshell
Varanidae Varanus salvator Mal~n water monitor
Varanus nebulosus Clou ed monitor
Scincidae Mabuya multifasciata Common sun skink
Agamidae Bronchocela cristatellus Green crested lizard
Calotes versicolor Changeable lizard
Draco melanopogon Black-bearded flying dragon
Draco vo/ans Common flying dragon
Gekkonidae Gekko smithi Large forest gecko
AMPHmIANS
Pelobatidae Leptobrachium hendriksoni Red-eyed Sround toad
Leptobrachium nigrops Black-eye ground toad
Mffo0PhryS aceras Short-horned toad
Bufonidae Bu 0 asper Giant toad
Ranidae Rana hosei Poisonous rock frog
Rana limnocharis Grass frog
Rana bZ;hi Malayan giant frog
Rana c leonota Copper-cheeked frog
Rana n~rovita
RhaCOKhoridae Polype tes leucomystax Malayan house frog
Micro ylidae Microhyla heymonsl Answering froglet
Microhyla bomeensis Rufous fr0ltlet
Kaloula pulchra Banded bu Ifrog
Kalophrynus p/eurostigrna Red-sided sticky frog
APPENDIX 2. Checklist of the non-timber flora recorded at Mata Ayer Forest Reserve, Perlis, Peninsular Malaysia.
Family Species Family Species

Annonaceae Anaxagorea javanica Orchidaceae Oberonia sf..
Aquifoliaceae flex cymosa Paphiopedz/um barbatum
Araceae Alocasia lowii Phalaenopsis cornu-cervi
Arisaema roxburghii Renanthera histrinica
Arisaema wrayi Sarcochilus stenoglottis
Homalomena pendula Palmae Calamus sp.
Raphidophora sp.a Caryota mitis
Raphidophora sp.b Daemonorops longiper
Schismatoglottis wallichii Licuala sp.
Scindapsus hederaceus Maxburretia rupicola
Scindapsus scortechini Pinanga malaina
Scindapsus sp.a Plectocomia grifJithii
Scindapsus sp.b Pandanaceae Pandanus sp.
Scindapsus sp.c Pentaphragmataceae Pentaphra8,ma albiflora
Typhonium sp. Piperaceae Piper kuroi
Araliaceae Ii;assaiopsis mitis Piper ramipilum
Aristolochiaceae Apama tomentosa Polypodiaceae Drymoglossum piloselloides
Thottea sp. Rubiaceae Argostemma pictum
Asclepiadaceae Hoya diversifolia Hedyotis congesta
Begoniaceae Begonia gutta Heydotis philippensis
Begonia phoeniogramma fxora javanica
Chloranthaceae Chloranthus ofJicinalis Ixora nigricans
Commelinaceae Aneilema sp. fxora pendula
Forrestia grifJithii Lasianthus cyanocarpus
Pollia sorzol{onensis Lasianthus pterospermus
Pollia thyrsiflora Morinda sp.
Compositea Erechthites valeriannifolia Mussaenda sp.a
Costaceae Costus globosus var.a Mussaenda sp.b
Costus globosus var.b Mussaenda glabra
Costus speciosus Mycetia lanceolata
Cucurbitaceae Trichosanthes wallichiana Mycetia malayana
Cyperaceae
~
perus diffusus Oldenlandia sp.
polytrum nemorum Ophiorrhiza discolor
apama sp. Pavetta indica
Seleria lithosperma Uncariapedicellata
Dennstaedtiaceae Microlepia speluncae Unidentified species
Dilleniaceae Acrotrema costatum Rutaceae Zanthoxylum nitidum
Dioscoreaceae Dioscorea sp. Solanaceae Solanum ferox
Euphorbiaceae Antidesma velutinosum Taccaceae Tacca integrifolia
Antidesma sp. Tacca leontopetaloides
Chaetocarpus castonocarpus Verbenaceae Clerodendrum laevifolium
Croton cascanillioides Vitaceae Leea indica
Macaranga trichocarpa Vitis sp.a
Mallotus porterianus Vitis sp.b
Flagellariaceae Hanguana malayana Vitis sp.c
Gesneriaceae Didymocarpus corchorifolia Zingiberaceae Alpinia sp.
Paraboea sp. Ammomum biflorum
Gramineae Leptaspis urceolata Ammomum sp.a
Hxpoxidaceae Curculigo sp. Ammomum sp.b
Lihaceae Dracaena umbritica Ammomum sp.c
Peliosanthes violacea Ammomum sp.d
Marantaceae Donax grandis Boesenbergia curtisii
Donax parviflora Boesenbergia plicata
Phrynium terminale Camptandra sp.
Moraceae Ficus pisocarpa Elettariopsis sp.a
Musaceae Musa sp. Elettariopsis sp.b
Myrtaceae Eugenia sp. Elettariopsis sp.c
Orchidaceae Bulbophyllum membranaceum Elettariopsis cf smithiae
Claderia viridiflora Etlingera sp.a
Cymbidium pubescens Etlingera if littoralis
Dendrobium bifarium Etlingera maingayi
Dendrobium flabellum Globba sp.
Dendrobium germinatum Hedychium sp.
Dendrobium leonis Kaempferia pulchra
Dendrobium pandaneti Zingiber zerumbet
Dendrobium secundum Zingiber sp.a
Habenaria sp. Zingiber sp.b
Hetaeria sp. Zingiber sp.c
Luisia SB.
Brunei and biodiversity - the Kew-Brunei checklist project
M.J.E. COODE, J. DRANSFIELD and D.W. KIRKUP
The Herbarium, Royal Botanic Gardens, Kew, Richmond, Surrey TW9 3AE, United Kingdom
The small state of Brunei Darussalam, lying on the north-west coast of Borneo, is of
botanical importance and diversity out of all proportion to its size. The biological richness
of Brunei's rainforests is very well appreciated by the Brunei authorities. But, before the
potential of the resource can be realised, it is essential to know what actually occurs within
the forest.
The flora of Borneo is covered to some extent (c. 20%) by Flora Malesiana and other
specialist accounts (e.g. Dipterocarpaceae - Ashton 1964, Rhododendron - Argent et al.
1988). The trees of Brunei itself are listed by Hassan bin Pukul & Ashton (1966, reprinted
recently) which was based on Ashton's years in Brunei in the fifties. The trees of Sarawak
are listed by Anderson (1980). Some tree groups are also covered by the two volumes of the
Trees of Sabah (Cockburn 1976, 1980). Whitmore et al. (1990) have recently published
several parts of a Flora of Indonesia (trees of timber size), including a checklist for
Kalimantan.
So why is Kew involved? The reasons, not surprisingly, are partly historical and partly
pragmatic.
In 1986 Wong Khoon Meng arrived in Brunei with a brief to rejuvenate and expand the
Forest Herbarium. He realised that what was needed above all was a checklist of all species
of vascular plants in Brunei which would not only be essential and indeed fundamental for
proper development planning in Brunei itself but also would be valuable for the surrounding
areas of Sarawak, Sabah and Kalimantan. He could not hope to accomplish such a task in the
relative isolation of Brunei, without ready access to material from those surrounding areas,
without types and without literature. Kew is blessed with these things.
Meanwhile, John Dransfield was seeking a focus for rejuvenating South-East Asian
interests at Kew. A joint Kew-Brunei Checklist would serve both purposes; they began
planning and, in March 1990, an accord between the then Director of Forestry Haji Yasin
and the Keeper of the Herbarium at Kew, Gren. Lucas, was signed. In brief, this accord
envisaged a programme of joint exploration of Brunei for 5 years during which the Forestry
Department would provide all logistic support in Brunei while Kew would send specialists
at regular intervals; Kew would also mastermind the production of the checklist of vascular
plants itself and a database at the project's end as complete as possible as regards taxa. The
checklist would be at least one step 'ahead' of the existing lists - it would contain specimen
citations and, where practicable, keys and short descriptions; it was agreed that it would be
unrealistic to expect a comparable level of detail throughout. Taxa will be separated out to
the best of the specialist's abilities and named where possible, but in certain of the intractable
groups - Rubiaceae and Ebenaceae, for instance, which need basic revision - it will not be
possible to provide determinations other than species a, species b etc.
The expertise of Robert Johns, with experience in New Guinea before he came to Kew,
and David Edwards, with experience in Brunei, will permit the ferns and fern-allies to be
treated separately and in more detail.
Part of the agreement was the understanding that the top set of any collection would
remain in the Forestry Department Herbarium at Sungai Liang (BRUN) and the second set
163
D. S. Edwards et al. (eds.), Tropical Rain/orest Research - Current Issues, 163-168.
164 M.J.E. COODE, J. DRANSFIELD AND D.W. KIRKUP
would come to Kew. Six other institutes (in alphabetical order: A, KEP, L, SAN, SAR,
SING) were earmarked to receive the next 6 duplicates, while it was agreed that collaborating
specialists would receive the third set, whether in that list or not.
Obviously, both sides have already benefitted from the collaboration. Both 'lead' herbaria
have been enormously enriched. Kew botanists have been able to collect in tropical rainforest
under near-ideal conditions. Two staff from the Brunei Herbarium have received training at
Kew and a third is expected to begin training in October. Friendship and trust has been built
up on both sides.
At Kew, much of the initial work was done by Leonard Forman who retired in 1991.
Mark Coode was put in charge of the 'Kew end' towards the end of 1991 while John
Dransfield remains responsible for overall policy and for liaison at all levels in Brunei. the
team has been joined for some of his time by Don Kirkup, who manages the data-base.
Happily, Leonard Forman has carried on working, in his retirement, on the identification of
the collections. In 1992, Wong Khoon Meng, having set the system up in Brunei, moved to
Sandakan and his place has been taken recently by Dr Idris Said.
THE DATABASE
Recent development of botanical databases has seen a division into two different but
overlapping areas; specimen-based databases which contain information relating to individual
specimens and taxon-based databases which deal with plant names and associated references
to the literature. Once up and running, databases of the two types may be related through
classes of information which both share (such as that of plant name) and represent a powerful
taxonomic tool.
The Brunei database is a specimen-based database. The range of information held in such
databases depends on its end use and so designs can vary greatly. A highly specialised design
might involve exhaustive and detailed descriptors of specimens for the purpose of
classification, on-line identification, generating 'automatic' descriptions or 'automatic' key
writing. The Brunei database is of a more 'general purpose' design than this (Table 1),
incorporating a more limited range of plant descriptors together with locality and habitat data,
date of collection and ancillary information such as local names and uses. The individual
specimen records are retrievable through the unique combination of collector name and
number (or collector and date where s.n.).
It is not envisaged that the finished checklist should be generated automatically from the
database; rather, a skeleton listing family by family be produced which summarises all the
available information for each species. This will provide a framework into which other
information (from the literature or re-examination of the specimens) can be incorporated.
The utility of a database like this increases with the number of records added. Interim
listings include a gazetteer and a list of taxa, ad hoc queries include a listing of rheophytes.
Of course, at the end of the project Kew would expect to hand over control of the
database to Brunei for Bruneian botanists to expand and use.
The database has been built up so far in two main phases.
Phase 1
As part of Kew's response to the agreement, three 'sandwich course students' worked
consecutively over a period of nearly three years, combing the Kew Herbarium systematically
for collections of flowering plants from Brunei. Label details were captured and in the end
about 2900 collections were recorded, almost all of these Ashton collections in the BRUN
series (undoubtedly, there will be more at BRUN).
The Kew-Brunei checklist project 165
Table 1. Structure of the Brunei Checklist database. All numeric fields are integers, apart from numbers 23 & 25
which have two decimal places.
Field Field Name Type Width Field Field name Type Width
SPNUM Numeric 6 20 LOCALTEXT Character 40
2 FAMILY Character 30 21 GEOLOGY Character 40
3 GENUS Character 25 22 HABITAT Character 40
4 SPI Character 25 23 LAT Numeric 5
5 SPAUTHI Character 30 24 NS Character
6 SSPVAR Character 3 25 LONG Numeric 6
7 SP2 Character 25 26 EW Character
8 SPAUTH2 Character 30 27 ALT Numeric 6
9 COLLECTOR Character 30 28 DETBY Character 30
10 NUMBER Character 15 29 DETYEAR Numeric 2
11 ADDCOLL Character 40 30 HABIT Character 40
12 COLLDD Numeric 2 31 FLOWEROBS Character 40
13 COLLMM Numeric 2 32 FRUITOBS Character 40
14 COLLYY Numeric 2 33 SIZE Character 15
15 CENTURY Numeric 2 34 NOTES Memo 10
16 COUNTRY Character 20 35 USES Character 40
17 MAJORAREA Character 30 36 VERNACULAR Character 40
18 MINORAREA Character 30 37 WILDCULT Numeric 6
19 LOCALITY Character 30
The data were captured into a single DBASE3 + table on a laptop computer; the
portability of the machine allowed the operator to move from site to site around the
herbarium. The students were 'biologically oriented' and were competent to make decisions
regarding the extraction of information from the specimens into the database without
involving large amounts of specialists' time. Printouts were checked for errors by senior
staff.
Phase 2
This is a continuing process involving the addition of collections new to the database.
Collections made under the Collaborative Agreement have also been recorded - Wong's 2000
collections and Kew's 4850 (with more to be processed). Together with collections made by
other members of the Forest Department staff and by colleagues visiting from overseas the
total now stands at about 10,000.
At Kew, the project collectors all input their own data. Originally, a DBASE3+
programme, LABELS 3, was used for data input and herbarium specimen label production.
Although somewhat limited in its original form, it was uncompiled and therefore easily
166 M.l.E. COODE, l. DRANSFIELD AND D.W. KIRKUP
customised by the project to produce acceptable labels. LABELS 3 has now been superseded
by a revised database structure compatible with the BRAHMS programme being developed
at the Oxford Forestry Institute, a structure gaining wider currency. Data entry is direct into
a single DBASE4 table and label production is via an in-house programme (essentially a mail
merge process).
We also need to compare the Kew listings with the BRUN card index and add those
records only at BRUN and we hope also to include at least the fern component of the
Herbarium here in the University of Brunei Darussalam. There are also van Niel's collection
of c. 1500 numbers from the early sixties, lodged in Leiden but never systematically written
up.
Database Management: Following the completion of phase 1 came a shift in requirements,
from data capture to both data capture and manipulation, the additional workload involving
updating determinations as the newly collected material was named; producing labels and
listings; editing, cleaning up the data; routine housekeeping tasks.
The replacement of the laptop computer with a faster machine with a larger storage
capacity (486/33mhz 120mb HDD) was essential at this juncture. At the same time the less
easily solved problem of maintaining the database integrity had to be addressed. At present
this is achieved by restricting access to computer-literate users only. New records and name
changes are entered into temporary files which are used to update the main database. All
queries are channelled through trained users.
A longer-term solution is for some sort of database management program, e.g. BRAHMS,
which offers verification of data at entry and has built-in security procedures, safely allowing
unskilled users access to the database for basic queries and edits. The BRAHMS program is
still evolving and the Checklist Project has been closely involved in its recent development
and evaluation. In addition, BRAHMS has recently been adopted by the Tree Flora of Sabah
and Sarawak. Future developments of BRAHMS will include its own GIS.
Kew will also deliver the checklist in book form, for normal publication.
ASSESSMENT
The question has often been asked: Why does Kew put so much effort into Brunei? The
flora's well-known already; Ashton was there for years. Perhaps the following figures will
show that the effort has not been misplaced, that the flora was not well-known, and that, if
we thought Brunei was rich before, now we can say it has twice the genera (Table 2) and at
least twice the number of species (Table 3) we knew it had.
As an example: for Palms, fifteen species were known in the collections at the start of the
project; there are now 125. For Elaeocarpus, Ashton listed 15 of which ten can be accepted
with the names he applied to them; one was added in the seventies and the tally now is 12
of his original. 15 and 16 more at least (including four from our recent trip to Gunung Pagon)
out of a total for Borneo of about 66. So the increase in known taxa, while due mostly to
recent collections of non-timber species, also includes many trees which Ashton missed. This
Table 2. Figures from the database. Figures for genera collected since 1988 will be slightly inflated by uncorrected
errors being read as 'new' genera - but we expect to 'replace' as many 'genera' as are lost when errors are
eliminated by genuinely new generic records amongst specimens currently recorded to family only.
Genera, pre-1988, in Kew Genera collected since Total known genera
478 529 1007

The Kew-Brunei checklist project 167
Table 3. Nwnbers ofdetennined specific epithets for Brunei. The same proviso applies to this table as to Table 2.
Note _ no account taken of specimens named simply to genus, which outnwnber those so far identified to species.
Species pre-1988 in Kew Species coUected since Total named species

1425 1626 3051
Table 4. Approximate nwnbers of species in a range of groups in which the taxonomy is reasonably well-known.
For Bamboos, we are indebted to S. Dransfield; Ferns - R. Johns; Aroids - P. Boyce; Palms - J.D.; Elaeocarpus
_M.C.; Begonia - M. Sands; Loranthaceae - O.K.; Dipterocarps - Ashton (1982) and Forestry Department figures;
Mapania - D. Simpson (1992).
Borneo total Brunei total Rous!! Percenta&e

Bamboos: 50-60 c. 10 20
Ferns: 1000-1200 25-30
Aroids: c.220 c. 110 50
Palms: 302 125 45
Elaeocarpus 65-70 c.28 40
Begonia 120 17 14
Loranthaceae 77 27 35
and from published accounts:
Dipterocarps 267 171 55
Mapania 26 13 50
is not to lessen the importance of Ashton's achievement - but to emphasise rather the
richness of the flora. For instance, let us consider the easily accessible locality known as
'Bukit Teraja', visited by Ashton and subsequently by at least 5 teams under the collaborative
venture. Ashton found Sloanea javanica at Bukit Teraja, but later visits by others did not.
D. Kirkup and M. Coode found at least one Elaeocarpus and a very odd member of the
Euphorbiaceae there, both new to Brunei, which none of the others found. J. Dransfield
found Pterisanthes grandis, a new species of Calamus and a curious Monocarpia
(Annonaceae). In fact, each trip to Bukit Teraja has resulted in finding species that all other
teams have missed.
For those groups we at Kew have looked at more closely, a few figures will emphasise
the increase in knowledge (Table 4). From these we can hazard a guess that either about one
third of all Borneo's species are present in this one small area of less than 1 % of the land
surface (probably nearer one half of Borneo's flora below 2000 m) or that hundreds, even
thousands, of new species still await discovery in the surrounding areas, in numbers
sufficient to bring the figures to more reasonable percentages area for area. While
undoubtedly new species will be found elsewhere in Borneo, we do not believe that their
number will materially alter these figures and, for us, the first conclusion is the more likely.
Finally, and most importantly, each new record increases the known value of Brunei's forests
as a 'gene-pool' island for the future.
LITERATURE CITED
ANDERSON, J.A.R. 1980. A Checklist of the trees of Sarawak. Dewan Bahasa dan Pustaka Lawangan Sarawak,
for Forest Department, Sarawak.
ARGENT, G., LAMB, A., PHILLIPS A. & COLLENETTE, S. 1988. Rhododendrons of Sabah. Sabah Parks
Publication number 8. Kota KinabaJu.
168 M.J.E. COODE, J. DRANSFIELD AND D.W. KIRKUP
ASHTON, P.S. 1964. A Manual of the dipterocarp trees of Brunei State. Oxford University Press, London.
ASHTON, P .S. 1982. Dipterocarpaceae. In Flora Malesiana Ser. I, Vol. 9, Pt. 11:237-552. M. Nijhoff, The Hague.
COCKBURN, P.F. 1976. Trees of Sabah. Vol. 1. Sabah Forest Record 10. Borneo Literature Bureau for Forest
Department, Sabah. Kuching, Sarawak.
COCKBURN, P.F. 1980. Trees of Sabah. Vol. 2. Sabah Forest Record 10. Borneo Literature Bureau for Forest
Department, Sabah. Kuching, Sarawak.
HASAN BIN PUKUL & ASHTON, P.S. 1965, reprinted 1988-9. A Check-List of Brunei Trees. Forest Department,
Brunei Darussalam.
SIMPSON, D.A. 1992. A revision of Mapania (Cyperaceae). Royal Botanic Gardens, Kew.
WHITMORE, T.C., TANTRA, I.G.M. & SUTISNA, U. 1990. Tree Flora of Indonesia. Checklistfor Kalimantan.
Parts I, 11.1, 11.2. Forest Research and Development Centre. Bogor, Indonesia.
Stalk-Eyed Flies: Caprices of Evolution?
DIETRICH BURKHARDT and INGRID DE LA MOTTE
lnstitut fUr Zoologie der Universitiit, D 93040 Regensburg, Germany
ABSTRACT. A report is given on the species of stalk-eyed flies found around the Kuala Belalong Field
Studies Centre (Brunei Darussalam, Borneo) during the Brunei Rain Forest Project 1991-92 of the
University of Brunei and the Royal Geographical Society. Twelve of thirteen species of stalk- eyed flies
found belong to the family Diopsidae. Ten of these are new species which are yet to be described. Local
diversity of that family is much higher than hitherto assumed. Some of the species closely resemble their
sister species present on the Malay Peninsula but do not interbreed with them. Allometric data of body
length and eyespan are given for two species; some features of their behaviour are described. The only
non diopsid stalk-eyed fly found, Themara hirtipes (Rondani), belongs to the family Tephritidae. Here
only males have eyestalks. Allometric data of males are presented, together with field observations on
territoriality, mating behaviour and reproductive behaviour.
KEY WORDS: Diopsidae, eyestalks, male competition, Tephritidae, Borneo, Brunei Darussalam.
INTRODUCTION
Stalk-eyed flies have their compound eyes on the tips of long, rigid lateral extensions of the
head. Eyestalks are a typical male feature in a few species of several dipteran families,
including the Tephritidae, Platystomatidae, and Drosophilidae. In the family Diopsidae,
however, both sexes of all species have eyestalks (Burkhardt & de la Motte 1987a, Feijen
1989, Grimaldi & Fenster 1989, Shillito 1971). In large diopsid species the approximately
2500 facets of each compound eye have small receptive angles. These provide a high visual
acuity, often well below one degree in the frontal direction (Burkhardt & de la Motte 1983).
Comparable high visual acuity is mostly found in much larger insects such as honey bees and
dragonflies. The eyestalks of diopsids provide several selective advantages by improving
visual capacities: the visual field of each compound eye exceeds the perimeter of its
respective hemisphere. Thus both eyes together provide panoramic vision and additionally
binocular vision for two thirds of the entire sphere, leaving monocular vision to only the
lateral parts. Eyestalks also allow them to look out from hiding places and, most
importantly, a wide eye separation allows a long range estimate of an object's size. This
last ability is crucial for recognizing and sizing up conspecifics. In species with sexual
dimorphism in eyestalk length, the male's eyespan is strongly and steeply correlated with its
body size: 25 % increase of body length corresponds to 75 % increase of eyespan (Burkhardt
& de la Motte 1987b). Thus eyespan signals a male's strength in a much exaggerated way.
This facilitates male competition and female choice. With the sensory advantages of eyes
on long stalks a male is better equipped to size up a competitor's strength, yet simultaneously
the eyespan is an enhanced signal of the male's own strength.
RESULTS
Diopsidae
In South-East Asia diopsids are found close to creeks and rivers. They form a substantial
portion of the dipteran fauna along these waterways in areas where the vegetation is
169
D. S. Edwards et at. (eds.). Tropical Rainforest Research - Current Issues, 169-174.
170 DIETRICH BURKHARDT AND INGRID DE LA MOTTE
undisturbed. Diopsid larvae feed on decaying plant material and, in rainforests, aid in the
decomposition of organic matter, as do other dipteran larvae. For Borneo seven diopsid
species have been listed (Steyskal 1975). During two short stays with the Brunei Rainforest
Project in the Belalong area we found twelve species most of which are yet to be described
(H. Feijen, personal communication). This indicates that local diversity of this family is
much higher than has been assumed hitherto. Some of the Bruneian species show a close
resemblance with sister species from the Malay Peninsula, but speciation seems to have
progressed since, in our cultures only one of these species, Megelabops sexguttata (Brunetti),
crossbreeds with its counterpart from the Malay Peninsula.
In some diopsid species, males and females have eyestalks of about the same length and
are monomorphic in that character. The two largest species from our Brunei samples,
Eurydiopsis argentifera (Bigot) and a Cyrtodiopsis species (sp. nov., close to C.
quinqueguttata (Walker», belong to this group. In the monomorphic species eyespan mea-
surements have a standard deviation of 3 % to 5 % of the mean, which corresponds to the
typical variability of such insect body measurements. In other species, especially in the
genus Teleopsis, males have much longer eyestaiks than females. In these species variability
of eyespan is extremely high and the standard deviation may reach 12 % of the mean in
females and 30 % of the mean in the males with their longer eyes talks.
Of all the diopsids we have found in the Belalong area the smallest (two species of the
Sphyracephala group) showed the closest affinity to water. We found them on emergent
boulders and logs in the river or close to the river banks. The two largest species
Eurydiopsis argentifera and a Cyrtodiopsis species (sp. nov., close to C. quinqueguttata),
were sometimes found distant from the rivers, several hundred metres up the slopes. The
predominantly dimorphic Teleopsis species kept to the river banks or fairly close to smaller
creeks uphill.
Among the new Bruneian diopsids a Teleopsis (sp. nov., close to T. belzebuth (Bigot» sets
a new record: the eyespan of large males may reach 20 mm, with a body length of only 8
mm. This large eyespan makes eye cleaning a difficult task. We propose the name Teleopsis
belalongensis in the description to come.
In some of the sexual dimorphic species females outnumber males. Typically the flies
gather at sunset in sleeping societies, usually on an aerial rootlet, a dried tendril or a dry vein
of a rotten leaf. In most cases a large male is the first to settle on a sleeping site, followed
by smaller males and females. Given a choice, females favour a sleeping site already
occupied by a large male (Burkhardt & de la Motte 1989). Lesser males are often driven
away by a large male. Males of equal size frequently engage in highly ritualized fights to
win a harem (de la Motte & Burkhardt 1983). This behaviour was thoroughly studied in the
Malay species Cyrtodiopsis whitei (Curran), and first observations on two of the Bruneian
species, Teleopsis pallifacies (Feijen) and Teleopsis belalongensis, reveal a similar behaviour:
the males display fighting activity on the sleeping sites. Figures 1 to 3 show the allometric
data of the Bruneian species in comparison with those of Cyrtodiopsis whitei. Here we also
find a close concordance.
Tephritidae
On the West ridge of the Belalong study area we encountered a stalk-eyed member of the
tephritid family, Themara hirtipes (Rondani). In tephritids the female has a normal fly's
head configuration; eyestalks are purely a male attribute. As in the diopsids, male eyespan
is strongly correlated with the body length (Figure 4). These data are derived from
measurements of the population we found on a rotting log on the West ridge slope at an
altitude of about 180 m. We measured the visual field of the eyes of males using the
pseudopupil method (cf. Burkhardt & de la Motte 1983). The width of the visual field in
Stalk-Eyed Flies 171
20 T.p. T. b.
20
E mm
E mm
15
15
10
10
5
5
mm mm
o 5 L 10 o 5 L 10
Figure I. Correlation between eyespan (E) and body Figure 2. Correlation between eyespan (E) and body
lengtb (L) in Teleopsis pallijacies, sp.nov. Abscissa: lengtb (L) in Teleopsis belalongensis, sp. nov.
body lengtb in mm, ordinate: eyespan in mm. Abscissa: body lengtb in mm, ordinate: eyespan in
Triangles: males, squares: females. Large symbols mm. Values for tbe males are small triangles. The
indicate tbe means. Dotted line: one to one ratio. large triangle indicates tbe mean value of tbe 11
males, tbe female symbol marks tbe mean value of
tbe densely clustered measurements from 8 females.
tephritid males is about the same as in diopsid males. We observed the population of more
than 50 flies for ten days. Usually they arrived at the log in the morning, stayed there during
daytime and left for an unknown destination in the late afternoon. About 50 of the flies were
captured in small tubes. We took them to the camp to take body measures and to apply
marking numbers and released them at the log on the following morning. Some of the
marked males returned on ten successive days. The males fought for territories on the log
or sometimes on the ground close to the log.
Sizing up the opponent is the initial behaviour in male/male encounters. The smaller of
the two usually turns away. Males of about equal size make contact with their heads and
eyestalks and try to push each other off the premises. Such contests may last for many
minutes and are sometimes repeatedly resumed by the same pair of males. The winner stays
in the territory, the looser runs or flies away. Females that pass by are courted. Successful
courtship leads to copUlation, which lasts 20 minutes on average. Shortly thereafter, the
female begins to deposit eggs, searching with her mouthparts and the abdomen for insect-
made holes in the log. Normally, after copulation and during egg deposition the male guards
the female; this may last for one hour. While mate guarding, even a small male is very
pertinacious and successfully fights off much larger intruders. We even observed a small
male alternately fighting with two intruders simultaneously. Another female passing the
territory will cause most males to turn their interest from the egg laying female to the
newcomer.
20 T.h.
E mm
12
E
[mm]
10
8
¥ .••...
15
10
6
4
-l
5
2
.....
......
.....
mm
0 2 4 6 8 10 o 5 L 10
Figure 3. Correlation between eyespan (E) and body Figure 4. Correlation between eyespan (E) and body
length (L) in Cyrtodiopsis whitei (specimens from length (L) in Themara hirtipes. Abscissa: body
Malay Peninsula). Abscissa: body length in mm, length in mm. ordinate: eyespan in mm. Males:
ordinate: eyespan in mm. Crosses: values for males, triangles, females: squares; the larger symbols
circles: values for females. The large male and indicate the mean values.
female symbols are positioned at the respective mean
values.
DISCUSSION
The majority of the newly found diopsid species in the study area belong to the genus
Teleopsis (see Table 1). As a revision of this genus is in preparation (H. Feijen, personal
communication), a full description of these species will be given later.
Eyestalks of flies were often regarded as a hypertely, yet several earlier authors speculated
on special visual and various other functions (see Feijen 1989). Recently eyestalks have been
mostly discussed in the context of male/male encounters, since head to head encounters
between males can be observed in many fly species (see Feijen 1989, Grimaldi & Fenster
1989). Indeed, such male/male encounters have been found in nearly every stalk-eyed
species that has been thoroughly investigated.
Thus, at first glance, stalk-eyed flies may strike one as another showcase for a 'caprice
of evolution', like horned beetles and the like. But a closer look reveals there is more to it:
there are the enhanced visual capacities including a large binocular field which allows
accurate size estimation over large distances, there is the unambiguous signal advertising a
male's strength, there is the possibility of using a stalked eye to peer around corners, and,
finally, the flies can use their eyestalks to push an opponent off a wanted territory
(Burkhardt & de la Motte 1987a, Burkhardt & de la Motte 1987b, Burkhardt & de la Motte
1989, de la Motte & Burkhardt 1989).
Stalk-Eyed Flies 173
Table 1. List of stalk eyed flies from the Belalong area, Brunei Darussalam.
Diopsidae
1. Teleopsis belalongensis, sp. nov.
2. Teleopsis pallijacies, sp. nov. (Feijen, description unpublished)
3. Teleopsis sp. nov.
4. Teleopsis sp. nov.
5. Teleopsis sp. nov., close to T. discrepans (Walker)
6. Megaiabops (= Teleopsis) sexguttata (Brunetti)
7. Cyrtodiopsis sp. nov., close to C. dalmanni (Wiedemann)
8. Cyrtodiopsis sp. nov., close to C. quinqueguttata (Walker)
9. Eurydiopsis spec., probably E. argentifera (Bigot)
10. Pseudodiopsis (= Sphyracephala) sp. nov.
11. Pseudodiopsis (= Sphyracephala) sp. nov., close to P. detrahens (Walker)
Tepbritidae
12. Themara hirtipes (Rondani)
One more undescribed species of the genus Teleopsis was found. There are only two specimens (one male and one
female) which look very much like Teleopsis burkhardti (Feijen sp. nov., unpublished) from the Malay Pensinsula.
Resin-embedded samples (males plus females) of each of the listed species 1 to 12 are in the custody of the Brunei
Musewn.
The question naturally arises whether in a competitive situation larger males do indeed
have higher reproductive success than smaller ones. This assumption has been proved in one
species using genetically marked pure strains of Cyrtodiopsis whitei (Burkhardt et al. 1991,
Burkhardt et ai. 1993).
ACKNOWLEDGEMENTS
We are grateful for all the help we received from colleagues at the University of Brunei
Darussalam and for their gracious hospitality. We thank. the RGS (London) for inviting us
to participate in the UBD/RGS Brunei Rain Forest Expedition 1991-92 and for supporting
our visit to Brunei in April 1993. We thank Mrs. Eckert and Mr. Schrepfer for taking care
of our Diopsids.
LITERATURE CITED
BURKHARDT, D. &. MOTTE, I. DE LA 1983. How Stalk-Eyed Flies Eye Stalk-Eyed Flies: Observations and
Measurements of the Eyes of Cyrtodiopsis whitei (Diopsidae, Diptera). Journal of Comparative Physiology
AI51:407-421.
BURKHARDT, D. & MOTTE, I. DE LA 1987a. Selective Pressures, Variability, and Sexual Dimorphism in Stalk-
Eyed Flies (Diopsidae, Diptera). Naturwissenschaften 72:204-206.
BURKHARDT, D. & MOTTE, I. DE LA 1987b. Physiological, Behavioural, and Morphometric Data Elucidate
the Evolutive Significance of Stalked Eyes in Diopsidae (Diptera). Entomclogica Generalis 12:221-233.
BURKHARDT, D. & MOTTE, I. DE LA 1989. Big 'Antlers' are Favoured: Female Choice in Stalk-Eyed Flies
(Diptera, Insecta), Field Collected Harems and Laboratory Experiments. Journal of Comparative Physiology
AI52:649- 652.
BURKHARDT, D., MOTTE, I. DE LA & YONG, H. S. 1991. PGM: A Suitable Genetic Marker in Studying
Mating Success of Cyrtodiopsis whitei (Diopsidae, Diptera). Naturwissenschaften 78:76-78.
BURKHARDT, D., MOTTE, I. DE LA & LUNAU, K., in press 1993. Signalling Fitness: Larger Males Sire More
Offspring. Studies of the Stalk-Eyed Fly Cyrtodiopsis whitei (Diopsidae, Diptera). Journal of Comparative
Physiology A.
FEDEN, H. 1989. Diopsidae. pp. 1-122 in Griffiths, G. C. D. (ed.). Flies of the Nearctic Region, Vol. lX, Part
12. E. Schweizerbart'sche Veriagsbuchhandlung, Stuttgart.
GRlMALDI, D. & FENSTER, G. 1989. Evolution of Extreme Sexual Dimorphisms: Structural and Behavioral
Convergence Anwng Broad-Headed Male Drosophilidae (Diptera). American Museum Novitates No. 2939, 25pp,
New York.
MOTTE, I. DE LA & BURKHARDT, D. 1983. Portrait of an Asian Stalk-Eyed Fly. Naturwissenschaften
70:451-461
SHILLITO, J. F. 1971. Dimorphism in Flies with Stalked Eyes. Zoological Journal of the Linnean Society 50:297-
305.
STEYSKAL, G. C. 1975. Family Diopsidae. pp. 32-36 in Delfinado, M. & Hardy, D. E. (eds). A Catalog of the
Diptera of the Oriental Region, Volume Ill. Hawaii University Press, Honolulu.
Small mammal diversity in riparian and dipterocarp habitats in Belalong
forest, Brunei Darussalam.
JOSEPH K. CHARLES
Universiti Brunei Darussalam, BaruJar Seri Begawan 2028, Brunei Darussalam
ABSTRACT. A small mammal study carried out in the forest of Belalong in the Temburong District of
Brunei Darussalam, showed that 18 species of small mammals (insectivores, tupaiids and rodents)
occurred in the dipterocarp forests as compared to 14 species in the riparian habitats. While' some small
mammals were exclusive to either the dipterocarp or riparian habitats, others were recorded in both
habitats. It is inferred that competition between Maxomys surifer and M. rajah has displaced the former
to the riparian areas. Habitat disturbance of a riparian area appears to have led to a popUlation increase
in two murid species, Maxomys muelleri and Rattus e.xulans.
KEY WORDS: Oriental, insectivores, tupaiids, rodents, competition, habitat disturbance, Borneo.
INTRODUCTION
Considerable tectonic uplift in geologically recent times in the southern Temburong District
of Brunei has given rise to steep slopes and sharp ridges (Dykes, this volume). This has
affected the extent of riparian areas in the Belalong valley of Ulu Temburong. The structure
and composition of the vegetation and the invertebrate fauna in these areas perhaps determine
the presence or absence of certain vertebrate fauna there. Some of the vertebrates may not
be found in the adjacent hill dipterocarp forests. A small mammal study was carried out
from February 1991 to February 1992 during the Brunei Rainforest Project that was centred
in the Kuala Belalong Field Studies Centre (KBFSC) in Belalong (Temburong District,
Brunei) to survey the diversity in both riparian and dipterocarp habitats and the movements
of some small mammal species in different forest habitats. A wildlife survey carried out in
Bukit Tudal, Ulu Temburong, about 25 km south-east of the Kuala Belalong Studies Centre
included small mammals but their diversity in relation to different habitats was not
investigated (Bennett et al. 1987).
METHODS
The study was carried out in two forest types; a variety of survey techniques were employed.
Forest types
Dipterocarp forest: two plots were chosen, as follows:
(a) A 0.9 ha plot at 85 m altitude on the West Ridge slope, close to Sungai (Sg.)
Belalong (SM Plot 1).
(b) A 1.0 ha plot at 330 m altitude on the East Ridge (SM Plot 2).
Riparian forest: two plots were chosen, as follows:
(a) An undisturbed 0.22 ha plot (Riparian Plot 1) at 50 m altitude along the Sg. Belalong.
(b) A disturbed 0.22 ha plot at 50 m altitude where the Kuala Belalong Field Studies
Centre is situated (Riparian Plot 2).
The distance between Riparian Plots 1 and 2 was about 150 m. Riparian Plot 2 and SM
175
176 JOSEPH K. CHARLES
Plot 1 were contiguous. It was not possible to use larger study plots in the riparian forest
due to to the limited forest fringe available. The Kuala Belalong Field Studies Centre is
situated in the valley between the East and West Ridges.
Survey techniques
There are a variety of trapping and marking techniques and their relative merits are
discussed by Barnett (1992). The following methods were chosen and modified where
necessary, based on previous experience in small mammal research.
Trapping by cage traps

At all sites, standard wire-mesh cage traps for small mammals were set at 10m grid
intersection points. The traps were modified to be collapsible and hence permit portability
in the forest. The bait used was banana. Previous experience in trapping small mammals
in forests in Malaysia and a series of bait acceptance trials in various forests in Brunei
(Lamunin, Sg. Akar and Kampung Sg. Dadap) have shown that bananas ('pisang rasthali')
were the most acceptable bait under the prevailing conditions (unpublished data).
Unbaited traps were fixed in the open position and left in the respective habitats for one
to two weeks. They were then pre-baited for 3 months twice weekly, followed by a 2 to 3
week period of no baiting before trapping began. This method was to permit acceptability
of the traps as part of the habitat and to prevent trap shyness in the small mammal species
resident in the study plots. To prevent irrecoverable damage to plot vegetation, no trapping
session exceeded seven days.
Since the overall study also included investigation of home ranges and population size,
marking was necessary. Trapped mammals were first etherised, measured and marked
before being released at the point of capture. Previously marked animals were released
immediately. Permanent marker pens (Staedtler) and dyes (e.g. Malachite Green, Carbol
fuschin) were used to mark the fur; these marks remained discernible for 2 - 3 months.
Animals trapped with faint marks were re-marked.
Trapping by pit-fall traps

Buckets buried in the ground with drift fences leading to them were used as pitfall traps
to catch very small mammals such as shrews that could not be trapped in the small mammal
traps. These traps were located outside the plots to avoid disturbance to the study areas.
Nocturnal surveys
When the weather was favourable (when there was no rain or heavy mist), nocturnal
surveys using powerful torches (12 volt Coleman Halogen Spotlights producing
approximately 100 watts or 750,000 candle power beam) were conducted for 2 to 3 hours
per survey. A distance of 500 m of regular trails was walked at the rate of 150 m per hour
from 1900-2200 hrs. The area was carefully scanned for nocturnal mammals. Reflected
eyeshine showed the presence of a mammal. Identifications were confirmed using a
searchlight and binoculars. This technique was employed in all habitats but much of the time
surveys were affected by inclement weather.
RESULTS & DISCUSSION
Though the study covered diversity and home ranges, the results reported here only deal with
small mammal diversity, competition between two murids and murid densities in the
disturbed and undisturbed riparian areas.
Small mammal diversity 177
Species richness in the lowland dipterocarp and riparian forests

The term 'small mammal' usually refers to non-flying mammals less than 1 kg in adult
weight and in practice includes insectivores, rodents and marsupials. The results presented
in this section covers insectivores, tupaiids and rodents and excludes mustelids and viverrids,
some of which also occur in both forest types.
Eighteen species of small mammals were recorded in the dipterocarp forest plots in the
East and West ridges as compared to 14 in the riparian habitat (Table 1).
Borneo's smallest shrew, Savi's Pygmy Shrew, Suncus etruscus, was caught in the pitfall
traps in the dipterocarp forest. This is the first record for Brunei. The Bukit Tudal wildlife
survey (Bennett et al. 1987) showed the presence of the South-east Asian White-toothed
Shrew, Crocidura fuliginosa rather than Suncus etruscus in that area.
Of the 8 species of tupaiids (tree shrews) found in Borneo, one is confined to mountains
(Mountain Treeshrew, Tupaia montana) while 7 are recorded in the lowland forests. In the
Table I. Small mammals of the Lowland Dipterocarp Forest (DF) and Riparian (R) areas in Belalong.
DF R
Order Insectivora
Soricidae
Suneus etruseus, Savi's Pygmy Shrew *
Order Scandentia
Tupaiidae
Tupaia tana, Large Tree Shrew *
Tupaia pieta, Painted Tree Shrew * *
Tupaia minor, Lesser Tree Shrew *
Tupaia gracilis, Slender Tree Shrew *
Tupaia dorsalis, Striped Tree Shrew *
Order Rodentia
Muridae
Rattus exulans, Polynesian Rat *
Rattus tiomanieus, Malaysian Field Rat *
Sundamys muelleri, Muller's Rat *
Maxomys surifer, Red Spiny Rat *
Maxomys rajah, Brown Spiny Rat * *
Maxomys whiteheadi, Whitehead's Rat * *
Leopoldnmys sabanus, Long-tailed Giant Rat * *
Niviventer cremoriventer, Dark-tailed Tree Rat * *
Sciuridae
Callosciurus notatus, Plantain Squirrel * *
Callosciurus prevostii, Prevost's Squirrel * *
Surulaseiurus hippurus, Horse-tailed Squirrel *
Surulasciurus lowii, Low's Squirrel * *
Exiliseiurus exilis, Plain Pygmy Squirrel * *
Ratufa ajjinis, Giant Squirrel * *
Rheithrosciurus macrotis, Tufted-ground Squirrel *
Hystricidae
Theeurus erassipinis, Thick-spined Porcupine *
TOTAL 18 14
East and West ridges of Belalong, 5 were recorded in the dipterocarp areas (Tupaia tana, T.
picta, T. minor, T. dorsalis and T. gracilis), while only T. picta was present in the riparian
habitats. T. glis, found on Bukit Tudal (Bennett et al. 1987), is absent at Belalong.
Eight species of forest rats were trapped, of which only 4, Maxomys rajah, M. whiteheadi,
Leopoldamys sabanus and Niviventer cremoriventer, were found in the dipterocarp forest.
These and 4 others, namely Rattus exulans, R. tiomanicus, Sundamys muelleri and Maxomys
surifer, were recorded in the riparian areas. The latter 4 species were resident and exclusive
to the riparian zone, and did not occur in the dipterocarp habitat 35 m higher up the slope
from the riverine areas. The Red Spiny Rat, M. surifer, is a rat of the dipterocarp forests
in Malaysia (Harrison 1969, Charles, unpublished data) and Borneo (Payne et at. 1985), but
at Belalong it was restricted to the riparian areas at very low numbers. The hypothesis that
the unusual distribution of this species is the result of competition will be discussed later.
Of the 7 species of squirrels recorded in the forest, 4, namely Callosciurus notatus, C.
prevostii, Sundasciurus hippurus, and S. lowii, were trapped while the other 3, Exilisciurus
exilis, Ratufa affinis and Rheithrosciurus macrotis, were seen. Apart from C. prevost;; and
R. macrotis, the other 5 squirrels ranged into the riparian areas during the course of their
daily activities. R. macrotis is a very rare Bornean endemic.
One porcupine species, Thecurus crassipinis, was seen in the forest study plot.
Some tupaiids, murids and sciurids occur in both the dipterocarp and riparian habitats; this
could be because the forest small mammals living close to the riparian fringe included parts
of it in their activity ranges. Observations of marked animals and preliminary analyses of
the home ranges of some species, e.g. Tupaia picta, Maxomys rajah, M. whiteheadi,
Leopoldamys sabanus, Niviventer cremoriventer, Callosciurus prevostii and Ratufa affinis
(which are not presented in this paper) have borne out this suggestion.
Some small mammals may be able to live and utilise the resources in both habitats -
Sundasciurus lowii and Exilisciurus exilis seen in the dipterocarp forests were also resident
and territorial in Riparian Plot 2.
Some of the other species are probably confined to the riparian habitats owing to
competition for some resource. Only a detailed population and resource partition study could
throw light on this aspect. Rattus exulans, Rattus tiomanicus, Sundamys muelleri and
Maxomys surifer were resident in the riparian zone though both habitats (Riparian Plot 2 and
SM Plot 1) were contiguous but had a 35 m altitudinal difference. The composition and
structure of the vegetational communities, food preferences, the availability of food and
microhabitats and the need for other environmental resources could determine the occurrence
of these species in the riparian habitat alone.
Is there competition between Maxomys surifer and Maxomys rajah?

The Red Spiny Rat, Maxomys sUrifer, is normally a dipterocarp forest rat; at Belalong it
is confined to Riparian Plots 1 and 2.
M. rajah,.a dipterocarp forest rat, ranges into the riparian areas at Belalong and is
probably M. surijer's competitor. The hypothesis is that M. rajah has probably displaced M.
surifer in the dipterocarp forest and the latter has been restricted to the riparian habitats, a
sub-optimal habitat for its survival.
That some kind of competition has occurred or is occurring between the two species can
be inferred from the following observations:
- Though trapping has been going on for one year in the study plots, M. surifer has so far
not been trapped in the dipterocarp forest but occurs in very low numbers in the contiguous
riparian habitat (9 per hectare) where M. rajah is still dominant (23 per hectare).
- The lower limit to similarity between competing sympatric species according to
Hutchinson (1959) is a 1.3 size ratio and a figure of 2 for the corresponding weight ratio
(Hutchinsonian ratios). This 'rule' holds true for many taxonomic groups (Das 1992, Giller
1984).
The Hutchinsonian ratios for both M. surifer and M. rajah were calculated for mean
Head-Body length measurement and Body Weight (Table 2; sample sizes were small for
those caught in the riparian areas M. surifer = 2; M. rajah = 5). The ratios for Head-Body
(1.13 ) and Body-Weight (1.36) were lower than that expected for sympatric congeners to
coexist. This implies that the species should not co-exist in the same habitat unless they
exploit different resources or adopt different foraging strategies. It has been shown in some
anurans that sympatric species overlap broadly in diet but much less in microhabitat usage
though the most closely related species overlapped the least (Das 1992).
When the Hutchinsonian ratios for the assemblage of four resident murids in Riparian
Plots 1 and 2 (Rattus exuians, R. tiomanicus, Maxomys surifer, Sundamys muelleri) are ex-
amined (Table 3), the Head-Body ratios generally fit into lower limit of similarity while the
Body-Weight ratios are not very clear though they approach the value of 2.0 (values used are
not means but the maximum and minimum values for Head-Body and Body-Weight
measurements to give the range of the ratios). It implies that these species show some degree
of coexistence. To what extent the other 4 dipterocarp forest murids (Table 1) compete with
the resident riparian murids during their ranging activities in Riparian Plots 1 and 2 is
difficult to know.
Hutchinsonian ratios for the same measurements for the forest murids in SM Plot 1 (West
Ridge) (Maxomys whiteheadi, Niviventer cremoriventer, M. rajah and Leopoldamys sabanus)
indicate that there is competition between the first two species but not between the last three
(Table 4). Preliminary observations have shown that while N. cremoriventer is more
arboreal, M. whiteheadi is more terrestrial and this spatial separation perhaps reduces
competition for food.
Detailed analyses of food (and other resources) and environmental requirement analyses
including foraging tactics are required to elucidate the difficult question of whether there is
coexistence or competition between M. surifer and M. rajah. In a resource-abundant
environment, two ecologically similar species can coexist indefinitely with complete dietary
overlap. Until there is data to show that food or some other resource is limiting in the study
area, the question about the role of competition between these two species will remain
Table 2. Hutchinsonian ratios for Head-Body (fiB) and Body weight measurements (mean value for one species
divided by the corresponding value for the adjacent species).
HB Ratio Body Ratio

(mm) Weight
Maxomys rajah 165.5 133.9
Maxomys suriler 187.0 1.13 181.5 1.36
Table 3. Hutchinsonian ratios for Head-Body (fiB) and Body weight measurements for the resident assemblage in
Riparian Plots 1 & 2. (Ratios are based on maximum and minimum values for each measurement.)
HB ratios Body Weight ratios
Rattus exuklns
Rattus tiomanicus 1.36 - 1.38 1.73 - 1.92
Maxomys suriler 1.0 . 1.32 1.54-2.18
Sundamys muelleri 1.15 - 1.34 1.3 - 2.07

Table 4. Hutchinsonian ratios for Head-Body (lIB) and Body weight measurements for the resident assemblage in
SM Plot 1, West Ridge. (Ratios are based on maximum and minimum values for each measurement).
HB ratios Body weight ratios
Maxomys whiteheadi
Niviventer cremoriventer 1.03 - 1.07 1.2 - 1.8
Maxomys rajah 1.3 - 1.4 1.8 - 2.2
Leopoldamys sabanus 1.3 - 1.6 2.4 - 2.6
unanswered. The occurrence of M. surifer in the Riparian Plots 1 and 2 only might be a local
phenomenon since this species occurs in the dipterocarp forest of Bukit Tudal (Bennett et at.
1987) and in the Peradayan Forest Reserve about 25 km north of the Belalong forest
(personal observation).
Murid diversity and population densities in Riparian Plot 1 and Riparian Plot 2
Riparian Plot 1 was an undisturbed area while Riparian Plot 2 was selectively cleared to
build the Field Studies Centre. Some of the riparian vegetation around the buildings has been
allowed to regenerate but the habitat has been altered. Many cut tree stumps and logs have
been left lying around. This situation allowed investigation of the effect of disturbance on
a riparian habitat as seen in the diversity and density of resident murids.
Eight murid species occurred in both the areas (Table 1) - Rattus exulans, R. tiomanicus,
Sundamys muelleri, Maxomys rajah, M. whiteheadi, M. surifer, Leopoldamys sabanus and
Niviventer cremoriventer. Disturbance to Riparian Plot 2 has not affected murid diversity
between 1991 and 1993.
The 3 murids resident in both riparian areas and not recorded for the dipterocarp forest
habitats were S. muelleri, R. exulans and R. tiomanicus. S. muelleri is a common inhabitant
of secondary and beach vegetation habitats and heath forest (kerangas). It is also an urban
pest in some parts of Bandar Seri Begawan where it is being replaced by Rattus rattus
(Charles, unpublished data). Rattus exulans and Rattus tiomanicus are known to be
agricultural pests in Peninsular Malaysia and, though they are not yet regarded as crop pests
in Brunei Darussalam, they have the potential to become pest species.
Total removal of the 3 murids (the trap-out technique) was carried out for seven days per
period, inundating both riparian areas with cage traps at approximately yearly intervals
during March/May 1991, April 1992 and February 1993. Ideally, trapping should have been
done during the same months for both riparian areas, but this was not always possible. In
1991, Riparian Plot 1 could only be sampled in May and not in March, as was Riparian Plot
2 (Table 5). No trapping was done in the undisturbed Riparian Plot 1 in Feb 1993. From
the number of new animals captured daily and the cumulative number of captures, population
size was estimated. In 1991, the densities of R. exulans and R. tiomanicus in the undisturbed
and disturbed areas were similar, R. exulans being slightly less common in the disturbed plot.
There were more S. muelleri in the disturbed area than the undisturbed. A year later in
1992, S. muelleri and R. exulans showed a clear increase in the disturbed area while their
numbers fluctuated slightly in the undisturbed plot. Densities of R. tiomanicus did not
change. In February 1993, there was a very sharp increase in the R.exulans population while
S. muelleri remained at the level of the previous years. No R. tiomanicus were trapped
during the 1993 removal. Perhaps the density had fallen below a certain threshold when
trapping the animals would be difficult.
The results indicate that habitat disturbance to Riparian Plot 2 has led to a population
increase of S. muelleri and R. exulans when compared to the control situation (Riparian Plot
1). R. tiomanicus population was unaffected over a single year but its absence in the 1993
Table 5. Densities of three murid species in riparian areas (numbers per hectare).
Species Riparian Plot 1 Riparian Plot 2

(undisturbed) (KBFSC)
May 91 Apr 92 Mar 91 Apr 92 Feb 93

Sundamys muelleri 37 30 50 62 55
Rattus exulans 31 37 25 81 250
Rattus tio1TlfJnicus 12 15 12 12
sampling might be symptomatic of changes in the welfare factors essential to its survival. The
high density of R. exulans after two years can be attributed to the creation and abundance of
new microhabitats by the piling up of logs in various sections of Riparian Plot 2 and the
appearance of buildings and new cover. Evidently the area of human disturbance has
produced a conducive environment for the survival and reproduction of S. muelleri and R.
exulans. The increase in population size has now produced a problem in the Field Studies
Centre - both species have invaded the buildings for foraging and nesting.
CONCLUSIONS
1. Eighteen species of small mammals were recorded in the dipterocarp forests and 14 from
the riparian areas of Belalong.
2. The record of Suncus etruscus, an insectivore, is the first for Brunei Darussalam.
3. Of 7 tupaiids found in Borneo, 5 have been recorded in the dipterocarp forests of
Belalong. Only T. picta ranged into the riparian habitat.
4. Four dipterocarp forest murids ranged into the riparian areas where 4 other murids were
exclusive to that habitat.
5. Of 7 dipterocarp forest sciurids, 5 ranged into the riparian areas during the the course of
their daily activities.
6. Only one hystricid was recorded in the dipterocarp forest.
7. It is inferred that M. surifer, a forest rat, competes with M. rajah and the former has been
displaced to the riparian fringe. This needs to be investigated further.
8. A comparison of the densities of 3 murids (exclusive to the riparian zone) in disturbed and
undisturbed riparian areas showed that S. muelleri and R. exulans have increased their
numbers as a result of habitat disturbance while the density of R. tiomanicus has remained
unchanged.
ACKNOWLEDGEMENTS
I have great pleasure in thanking Samhan bin Nyawa, Ashri bin Hj Ahmad, Dk Mahani bte
Pg Momin, Munah bte Hj Yahya and Sarimah bte Abu Bakar for their enthusiastic assistance
in fieldwork, and the Standard Chartered Bank (Brunei) for financial support to cover the
costs of their visits to the Kuala Belalong Field Studies Centre. This study was supported
by a research grant from Universiti Brunei Darussalam.
LITERATURE CITED
BARNETT, A. 1992. Expedition Field Techniques Small rruunmals (excluding bats). Expedition Advisory Centre,
London. 76 pp.
BENNETT. E.L., CALDECOTT, 1.0., & DAVISON, G.W.H. 1987. A wildlife survey of Ulu Temburong,
Brunei. Brunei Museum 10umaI6:121-166.
DAS, I. 1992. Trophic ecology of a community of South Indian anuran amphibians. Unpublished D. Phil. thesis,
University of Oxford, Oxford.
DAVIES, A.G. & PAYNE, J.B. 1982. A Faunal survey ofSabah. World Wildlife Fund, Malaysia, Kuala Lwnpur.
DYKES, A. P., this volwne. Analysis of factors contributing to the stability of steep hillslopes in the tropical
rainforest of Temburong, Brunei Darussalam.
GILLER, P.S. 1984. Community structure and the niche. Chapman and Hall, London. 175 pp.
HARRISON, J.L. 1969. The abundance and population density of mammals in Malayan lowland forests. Malayan
Nature 10umaI22:174-178.
HUTCHINSON, G. E. 1959. Homage to Santa Rosalia, or why are there so many kinds of animals? American
Naturalist 93:145-159.
PAYNE, I., & FRANCIS, C. M. 1985. A Field Guide to the Mammals of Borneo. The Sabah Society, Kota
Kinabalu, Sabah.
Species radiation of predatory empidoids (Insecta, Diptera) in lowland
rainforest in northern Papua New Guinea
PATRICK GROOTAERT
Department of Entomology, Royal Belgian Institute of Natural Sciences, Vautierstraat 29, B-

1040 Brussels, Belgium
ABSTRACT. Empidoid flies are among the most abundant flies in the lowland rainforests of northern
Papua New Guinea. The Empidoidea, as considered here, comprise the families Hybotidae, Empididae,
Atelestidae, Microphoridae and Dolichopodidae. Only one representative of the Empididae, a very
abundant family in temperate regions, is known from Papua New Guinea. No Atelestidae nor
Microphoridae are reported from the area. On the other hand, the Hybotidae and Dolichopodidae have
radiated enormously. The stouter hybotids which catch prey with their raptorial hind legs in flight are
present but not very common. The small leaf- and especially the soil-dwelling microhybotids are very
abundant and very rich in species and they may well be the most important predators of micro-arthropods.
The genus Elaphropeza (2-3 mm), the vicariant of the genus Platypalpus in temperate regions, with its
numerous species, can be found running on leaves. The genus Nanodromia, known currently only from
Papua New Guinea, also has numerous species that are found on leaves or on soil. They possess a
piercing proboscis and raptorial front legs. Their small size (about 1 mm) suggests that they feed only
on very small arthropods such as mites or collembolans. The dolichopodids occupy two major habitats.
The larger, brilliantly green shining Sciapodinae are very commonly seen on leaves of bushes and trees.
The other dolichopodids are generally found in humid areas such as the borders of pools, streams or in
swamps. The genus Sympycnus, with its relatives, is probably the largest genus with more than 200
species still to be described. The males exhibit complicated secondary sexual characters which may
reflect a fast radiation in a new environment (cf. species radiation in Hawaiian flies). Their biology and
behaviour is unknown. Since some dolichopodid species are known to be predators of mosquito larvae,
they warrant special attention.
KEY WORDS: Diptera, Empididoidea, species radiation, Sympycnus.
INTRODUCTION
Empidoid flies are among the most abundant flies in the rain forests of lowland Papua New
Guinea. An enormous radiation of species is seen in the families Dolichopodidae and
Hybotidae. The larger, brilliantly shining Sciapodinae (Dolichopodidae) are very common
and they are readily seen sitting on leaves of scrubs and trees. The smaller Hybotidae are
also very abundant but less conspicuous and therefore poorly known. Unfortunately we have
no idea about their habitat preferences or community structure. At most 10% of the existing
fauna has been recorded and named. Even these species are generally only recorded
incidentally by non-dipterist explorers. On the whole we estimate that about 90% of the
empidoid flies in Papua New Guinea remain undescribed.
The present study is based on material collected during three short expeditions to the
northern coast of Papua New Guinea (Madang province, Bogia district). Samples were taken
in mangroves, swamp forests and in secondary and primary rainforest at sea level (Bunapas,
183
184 PATRICK GROOTAERT
""km
.... -
Figure 1. Map of Papua New Guinea with a detail of Hansa Bay and Laing Island
Awar, Boroi bush) and in the hills (alt. 200 m) near Bogia (Apingan, Mangigim) (Figure 1).
A large collection of Sympycninae from all over Papua New Guinea was offered to us for
study by Dr. Neil Evenhuis of the Bernice P. Bishop Museum in Honolulu.
RESULTS
Current taxonomic knowledge

In recent years many new species of empidoid flies have been reported from Papua New
Guinea (De Meyer & Grootaert 1990, Grootaert 1984, 1987, 1992, Grootaert & Meuffels
1984, 1987, 1988, 1989, in press, Grootaert & Van de Velde 1988, McAlpine 1982,
Meuffels & Grootaert 1984, 1985, 1986, 1987, 1988). Table I gives a list of genera and the
number of species known from Papua New Guinea. The second column contains the number
of species mentioned in the catalogue of the Diptera of the Australasian and Oceanic regions
(Bickel & Dyte in Evenhuis 1989); the third column contains the number of species published
after 1989 and those seen by the author but not yet published. These estimates are cautious
and are in fact based only on observations on the northern coast of Papua New Guinea.
The superfamily Empididoidea comprises the families Empididae, Hybotidae,
Microphoridae, Atelestidae and Dolichopodidae (Chvala 1983). Apart from a single Empis
species no other representatives of the family Empididae, and more particularly the
Empidinae, are known from Papua New Guinea. The Empidinae are mainly nectar feeders,
Species radiation of empidoids in Papua New Guinea 185
Table 1. Empidoidea of Papua New Guinea. The second column gives the nwnber of species· cited in the catalogue
of the Australasian and Oceanian regions (Evenhuis 1989); the third column gives the nwnber of species published
after 1989, together with those seen by Meuffels & Grootaert but unpublished, and the fourth column gives the swn
of columns two and three. M: genera with marine or mangrove species.
No. of Species No. of Species published/ Total No. of

Taxon Ecology
{1989} recognised since 1989 sl1ecies
Empidoidea
Empididae
Empidinae
Empis
Hybotidae
Hybotinae
Hybos 1
Syndyas 1 1
Syneches 6 6
Tachydromiinae
Crossopalpus 1 2 3
Drapetis 2
Elaphropeza 6 20 26
Micrempis 1 I
Platypalpus 4 4
Tachydromia 1
Chersodromia 5 5 M
Nanodromia 18 18
DoJichopodidae
Sciapodinae
Amblypsilopus 9 9
Chrysosoma 36 37
Krakatauia 7 7
Mesorhaga 1 2 3
Plagiozopelma 4 4
Neurigoninae
Neurigona
Dolichopodinae
Lichtwardtia I 1
Metaparaciius 1 1
Paraclius 4 3 7
Medeterinae
Medetera 16 16
Hydrophorinae
Cymatopus 2 2 4 M
Thinophilus 7 7
Thinolestris 1 M
Diaphorinae
Asyndetus 7 4 11
Chrysotus 2 1 3
Crytophleps 1 1
Diaphorus 12 5 17
Mischopyga 2 2 M
Sympycninae
Chaetogonopteron 1 1
Phrudoneura 1 1
Sympycnus 29 100 129
Teuchophorus 17 2 19 M
Acropsilus 9 9
Rhaphiinae
Physopyga M
Incertae sedis
Phacaspis 2 2 M
Total 159 135 294
a feeding habit that is derived from predatory feeding. However, in almost all species the
male hunts prey and offers it to the female before mating. The subfamily is very abundant
and rich in species in the temperate zone of the holarctic region and they are also present in
Australia and New Zealand. Apparently they failed to colonize successfully even the
temperate mountainous areas in Papua New Guinea. The small family Atelestidae is not
known from the Australasian region. The Microphoridae are represented by only two species
in Australia, none are known from New Guinea.
The families Hybotidae and Dolichopodidae have radiated extensively in Papua New
Guinea. The genera Hybos, Syndyas and Syneches, belonging to the ancestral Hybotinae, are
rather stout flies with raptorial hind legs. They catch their prey in flight. This subfamily
is not very rich in species and is not very common. In contrast, the small leaf- and soil-
dwelling microhybotids belonging to the subfamily Tachydromiinae have radiated greatly.
These hybotids have the front legs and sometimes the mid legs raptorial instead of the hind
legs. They sit or run, preferentially, on horizontal surfaces such as leaves where they search
for prey. Prey generally consists of small flies or hymenopterans upon which they jump,
hold them firmly with the fore and mid legs, pressing them against the substrate while trying
to pierce them with their short but strong proboscis. It is expected that more than 90% of
the tachydromiine species still have to be described. They represent probably the most
important predators of micro-arthropods.
The genus Elaphropeza, which is very rich in species in all tropical regions in the world,
is also very diverse in Papua New Guinea. As can be seen in Table I, only six species were
known previously from Papua New Guinea, but in the small area under study at least 20
species have now been recognized. Elaphropeza is the tropical vicariant of the genus
Platypalpus in temperate climates. Apart from some bristles on the legs few other male
secondary sexual characters (MSSC) have developed. The abdominal pheromone glands
show a high morphological variability among the species. The second, third, fourth and/or
fifth tergite may be modified and may bear leaf-like bristles and ridges. Underneath the
terga are pheromone glands. So instead of visual recognition, chemical communication is
highly diversified in Elaphropeza.
The new genus Nanodromia (Grootaert 1993) is known only from Papua New Guinea and
it is also very rich in species. The genus consists of minute flies (1-2 mm). They are very
abundant and densities of several tens of individuals per m2 are commonly observed. They
can be seen running on leaves of shrubs and trees or on fallen leaves and on the soil. Their
small size suggests that they feed only on small arthropods such as mites, collembolans and
small flies. The male secondary sexual characters are also strongly developed in the genus
Nanodromia. In most species the tarsal segments of the front legs are coloured or annulated.
By analogy with the dolichopodids they are probably waved in front of the female before and
during mating. The mid legs in males are enlarged and the femora and tibiae bear bristles
and spines which are distinct at the species level. These raptorial looking legs are not
present in the female. They could be used to handle prey, perhaps to offer prey to the
female, but such a habit has never been observed in Tachydromiinae. They may function
as a clasping organ to hold the female during mating. One species has the tips of the wings
curled up in the male. Apart from the tibial glands no other glands with possible pheromone
secreting functions have been observed to date. It is clear that in Nanodromia visual
communication prevails over chemical communication to permit discrimination among species
and sexes. Phylogenetically, Nanodromia is closely related to the marine genus
Chersodromia. A number of species also live on the coast but many are associated with
mangrove, swamp forests and drier rainforests.
A number of smaller genera such as Drapetis, Chersodromia, Platypalpus and related new
genera occur in Papua New Guinea, but are not as diverse as Elaphropeza and Nanodromia.
Habitat and niche

The larvae of empidoids and their biology are poorly known. Except for the genus
Medetera, all others live in the soil and are thought to be predatory like the adults.
Dolichopodids are generally found in humid areas near swamps or on the borders of puddles
or streams. Although there is a preference for these humid soils, only the larvae of the
Hydrophorinae have adaptations for an aquatic existence. The hybotids on the other hand
are generally found in drier habitats.
As can be seen in Figure 2, several genera show habitat preferences. On the leaves of
herbs, shrubs and trees , the large, brillantly green shining Chrysosoma, the somewhat smaller
Sympycnus, the often yellow to brown coloured Elaphropeza or the minute Nanodromia are
frequently observed. The latter is also active on the soil or on fallen leaves. On tree trunks,
in a vertical position, sit Sciapus , Medetera and Tachydromia . The larvae of Medetera live
in the burrows of scolytid beetles under bark and feed on the beetle larvae, while the adult
Medetera have been observed feeding on aphids. Nothing is known about the fauna of the
canopy but this is likely to be the habitat of the Sciapodinae.
canopy Chrysosoma
"P0
( 'sky hunters' Syneches )

~
~
~
Sciapus
tree trunks Medetera
Tachydromia
Chrysosoma
Sympycnus
Elaphropeza
Nanodromia
Diaphorus
Figure 2. Habitat preference of the major empidoid genera in Papuan rainforest

Species radiation in Sympycnus: a case study

The genus Sympycnus is a very large taxon with a worldwide distribution and it is very
prolific in the tropics. As can be seen from Table I, the genus is very rich in species. Only
a very small portion of the Papuan fauna limited to the northern coast has been described by
Meuffels & Grootaert (1987) but at least a hundred species await description (Meuffels &
Grootaert, in preparation). The enormous development of male secondary sexual characters
(MSSC) is very striking. The antennal arista can bear a terminal thickening comparable to
some kind of flag (cf. the related genus Yumbera (Bickel 1992». The eye facets may be
enlarged. All pairs of legs may be modified. The fore legs bear peculiar bristles on the
femora, tibiae and tarsi (Figure 3). The tarsi may be somewhat compressed (Figure 4) and
may be carry various types of bristles. The terminal tarsal segments may be coloured white
and bear white or black leaf-like bristles. By analogy with other dolichopodids, the tarsi are
probably waved in front of the female during display. The mid legs bear also various hairs
and bristles (see Meuffels & Grootaert 1987; e.g. the cork-screw bristle) of which the
function is unknown. The hind legs have in many species the two basal tarsal segments
shortened (Figure 5). The second segment bears an extension called the c1idium (Figure 6).
Its morphology is specific but its function is completely unknown. Even the wings may be
modified: the tip of the wing may bear a black spot (S. sticticus) or the base may be
shrunken and bear long bristles (genus Chaetogonopteron) or the median vein (vein M) may
be abbreviated (Phrudoneura abbreviatus).
The male genitalia in dolichopodids may be very complex suggesting the presence of many
Lock-Key-Mechanisms (LKM) or even associated with sperm competition. In contrast to the
MSSC, the male genitalia in Sympycnus are rather simple in structure (Figure 7) but
nevertheless specific. The LKM's are apparently poorly developed. They consist of clasping
organs, the telomeres, the intromittent organ (aedeagus) and a guiding piece (hypandrium).
The cerci are simple in structure and have is no apparent function. Apart from some
jumping up and down on leaves, display and mating have not yet been observed in
Sympycnus species.
Two very closely related genera occur together with Sympycnus - Phrudoneura, originally
described by Meuffels & Grootaert (1987) as a subgenus of Sympycnus, and the genus
Teuchophorus. Teuchophorus is also very rich in species in northern Papua New Guinea
with no fewer than 16 species described, the same number as are known from the rest of the
world (Meuffels & Grootaert 1986). However the distinction between Teuchophorus and
Sympycnus is sometimes unclear in Papua New Guinea. In the Holarctic region, male
Teuchophorus can be easily recognized by the presence of a lanceolate stigma. This is the
swelling of the tip of the R1 (radial vein) and the costa which forms a distinct visual mark.
In northern Papua New Guinea, however, different degrees of the swelling of the stigma
occur and it is even absent in many species. Species lacking the stigma could well be placed
in Sympycnus but have to be considered as intermediate between the two genera. A character
analysis and, eventually, a redefinition of the genus Sympycnus and related groups is
necessary (Bickel 1992).
Research on habitat preferences and niches is preliminary. Table 2 (refer also to Figure
1 for geographical location) shows the Sympycnus-Teuchophorus-Phrudoneura assemblages
occurring in a transect from high saline mangrove over brackish water mangroves, swamp
forests to primary and secondary rainforests on drier soils. Species diversity is rather low
in the saline mangroves but increases, with up to 12-14 species occurring together, when the
salinity decreases.
Figures 3 - 7. Sympycnus compressipes Meuffels & Grootaert, male. 3. fore leg with white apical tarsal segments.
4. compressed first tarsal segment of fore leg. 5. hind leg. 6. clidium, appendage on second tarsal segment of hind
leg. 7. genitalia (a: aedeagus; c: cerci; h: hypandrium; t: telomeres). Scale bars 0.1 mm.
DISCUSSION
The fauna of Papua New Guinea is interesting for various reasons: (1) its very high
biodiversity and the origin of this diversity i. e. the speciation mechanisms responsible for
that diversity; (2) the presence of ancestral groups and; (3) the high degree of development
of MSSC in many groups and its relationship with the modes of speciation.
Gressitt (1982a) and several other authors (in Gressitt 1982b) state that the insect fauna
of Papua New Guinea has a closer relationship with the S.E. Asian fauna than with the
Australian fauna and that its origin is probably to be found in the wet tropical rainforests of
S.E. Asia rather than in the drier savanna fauna of Australia. This hypothesis seems not
unlikely, but since only a very small part of the Papuan empidoid fauna is known and even
Table 2. Habitat preference of species of the Sympycnus - complex (Sympycnus, Phrudoneura and Teuchophorus)
in Hansa Bay, Bogia district, northern Papua New Guinea (refer to Figure 1 for locations). 1. Laing Island: brackish
mangrove; 2. Awar: mangrove with brackish water; 3. Sisimangum: mangrove with low salinity; 4. Awar: puddle
with fresh water; 5. Warawaranga: river with fresh water; 6. Bunapas: swamp forest; 7. Awar bush: secondary
rainforest; 8. Boroi bush: secondary rain forest; 9. Apingan: primary rainforest in hills (all. 200 m); 10. Mangigim:
secondary rainforest (alt. 50 m).
Species Locality
2 3 4 5 6 7 8 9 10
S. teuchophoroides x x
S. campsicnemoides x x x
S. aechmophorus x
S. gigas
S. sticticus x
S. anachrostichus x x
S. tricomiger x
S. ictericus x x x
S. araneipes x x x
S. palJipes x x x x
S. priapus x
S. torrenticola x
S. ripicola x x
S. ctenophorus x x
S. nigrisqunmis x
S. aJipes x
S. miritarsis x x x x
S. candidimanus x
S. leucotarsus
S. albise rratus X X
S. compressipes x x
Ph. abbreviatus x x
T. femoratus x x
T. conspicuus x x
T. pusio x x x
T. denticulatus x x x
T. costalis x x
T. spatulifer x x
T. quadratus x x x X x
T. digitatus x x x
T. fuscicornis x x
T. pectinatus x x
T. vanaartseni x x x
T. parcearmatus x x
T. uncinatus x x x
T. grandior x x x
T. laingensis x
No. of taxa 2 5 9 8 9 2 12 14 13 3
less is known of that of adjacent areas, we will have to wait before anything definitive can
be said on this case.
It is remarkable that the boundaries between certain taxa which are very distinct in other
zoogeographical regions fade away in Papua New Guinea. This is seen in Teuchophorus-
Sympycnus and in the genus Platypalpus (Grootaert & Van de Velde 1988). This
phenomenon has already been observed by Gressitt (1982) in beetles and by McAlpine (1982)
in platystomatid flies. Gressitt suggests that the intermediate forms persisted here because
they did not suffer too much during the glacial periods. The presence of these 'missing
links' and other ancestral groups (De Meyer & Grootaert 1990) is of course interesting
phylogenetically.
Rainforests are well known for their high diversity of plants and animals. However, the
numbers of each species per unit of surface are generally rather low. This poses a mate-
location problem for the species involved. Two strategies seem to have evolved in the
empidoids: a strategy to find the mate directly or the strategy of finding the mate's habitat.
In the first strategy there must be long distance communication through pheromones which
aggregate the species. In the second type of strategy a species recognizes its correct habitat
through structure (vegetation), relative humidity, light intensity or other physical parameters.
Adult dolichopodids are very faithful to their habitats and, although they are excelllent flyers,
they remain in their habitat even if it is only a few square meters in size. Aggregations of
many individuals, even belonging to various species, may thus occur in the same small
habitat. Correct partner recognition is important to avoid false matings and over these short
distances visual recognition becomes very important.
In both dolichopodids and hybotids there are many species in which the MSSC are well
developed. Most of the MSSC play an important role in display and in mate recognition.
The high development of MSSC suggests that there exists a sexual selection by the females.
This selection could be enhanced by the co-occurrence of several related species in large
numbers in the same area and the high development may be due to a reinforcement
phenomenon (cf. Butlin 1989). In Sympycnus and Nanodromia the strategy of early species-
and partner-recognition by MSSC is adopted, whereas in Elaphropeza chemical
communication helps to aggregate members of the same species, and visual MSSC act only
over short distances. Many species of dolichopodids which have no chemical, acoustical or
other type of long distance communication, probably find their partners because they stay
close to their habitats (borders of ponds, rivers) and never stray from them. Visual MSSC
are then sufficient in species recognition. Elaphropeza has a wider habitat range and
members of this genus need long distance communication to find their partners.
ACKNOWLEDGEMENTS
The author acknowledges a grant from the National Fund for Scientific Research (F.K.F.O.
project 2.9008.90) to study biodiversity in Papua New Guinea. The author also thanks his
colleague Henk Meuffels for many interesting discussions on the subject.
LITERATURE CITED
BICKEL, D.J. 1992. The Australian Sympycninae (Diptera: Dolichopodidae): Introduction and Description of a New
Genus, Yumbera. Invertebrate Taxononomy 6:\005-\017.
BUTLIN, R. 1989. Reinforcement of premating isolation. Pp. 158-179 in Otte & Endler (eds). Speciation and its
consequences. Sinauer Associates Inc. Massachusetts.
CHV ALA, M. 1983. The Empidoidea (Diptera) of Fennoscandia and Denmark. II General Part. The Families
Hybotidae, Atelestidae and Microphoridae. Fauna Entomoiogica Scandinavica 12:1-279.
DE MEYER, M. & GROOTAERT, P. 1990 (1991). Pipuncu1idae (Diptera) from Papua New Guinea. The genus
Cephalosphaera, Cephalops and Beckerias. Zoologica scripta 19:403-412.
EVENHUIS, N. 1989. Catalog of the Diptera of the Australasian and Oceanian Regions. Bishop Museum Press.
1155 pp.
GRESSITT, J.L. 1982a. Ecology and biogeography of New Guinea Coleoptera (Beetles). Pp. 709-734 in Gressitt,
J.L. (ed.). Biogeography and Ecology of New Guinea. Monographiae biologicae 42.
GRESSITT, J.L. 1982b. Zoogeographical summary. pp. 897-918 in Gressitt, J.L. (ed.). Biogeography and
Ecology of New Guinea. Monographiae biologicae 42.
GROOTAERT, P. 1984. Two new species of Platypalpus Macquart (Diptera, Empidoidea, Hybotidae) from Papua
New Guinea. Bulletin et Annales de la Societe royale beige d'Entomologie 120:283-290.
GROOTAERT, P. 1987. A new Tachydromia from Papua New Guinea (Diptera, Empididoidea, Hybotidae).
Bulletin et Annales de la Societe royale beige d'Entomologie 123:141-144.
GROOTAERT, P. 1992. Description of new species in the genus Chersodromia (Diptera, Empidoidea, Hybotidae)
from the northern coast of Papua New Guinea. Bulletin et Annales de la Societe royale beige d 'Entomologie
128:235-244.
GROOTAERT, P. & MEUFFELS, H.J.G. 1984. Dolichopodidae (Diptera) from Papua New Guinea. n. Some
new species of the genus Thinophilus Wahlberg, 1844, from the lowland. Indo-Malayan Zoology 2:209-223.
GROOTAERT, P. & MEUFFELS, H.J.G. 1987. Dolichopididae (Diptera) from Papua New Guinea. V. A new
species of Cryptophleps Lichtwardt, 1898, with notes on other species. Indo-Malayan Zoology 4:145-152.
GROOTAERT, P. & MEUFFELS H.J.G. 1988 (1989). Dolichopodidae (Diptera) from Papua New Guinea. VII.
Description of Thinolestris gen. nov., a new beach dwelling Hydrophorinae. Indo-Malayan Zoology 5:31-39.
GROOTAERT, P. & MEUFFELS H.J.G. 1989. Dolichopodidae (Diptera) from Papua New Guinea. IX.
Mischopyga gen. nov. and Physopyga gen. nov., two new genera from the mangrove in Papua New Guinea.
Indo-Malayan Zoolology 6:195-206.
GROOTAERT, P. & MEUFFELS, H.J.G. in press. Dolichopodidae (Diptera) from Papua New Guinea. X.
Description of new species of the marine genus Cymatopus Kertesz, and observations on their biology and
territorial behaviour. Invertebrate Taxonomy.
GROOTAERT, P. & VAN DE VELDE, I. 1988. A new Platypalpus species from Papua New Guinea (Diptera
Empidoidea Hybotidae). Bulletin et Annales de la Societe royale beige d'Entomologie 124:223-227.
McALPINE, D.K. 1982. The acalyptrate Diptera with special reference to Platystomatidae. Pp. 659-673 in
Gressitt, J.L. (ed.). Biogeography and ecology of New Guinea. Monographiae biologicae 42.
MEUFFELS, H.J.G. & GROOTAERT, P. 1984. Dolichopodidae (Diptera) from Papua New Guinea. I: The genus
Cymatopus Kertesz with a discussion on Abateti Miller and Cemocarus gen. nov. Indo-Malayan Zoology 1: 141-
158.
MEUFFELS, H.J.G. & GROOTAERT, P. 1985. Dolichopodidae (Diptera) from Papua New Guinea. ill: New
species of the genus Diaphorus Meigen, 1824. Indo-Malayan Zoology 2:9-28.
MEUFFELS, H.J.G. & GROOTAERT, P. 1986. Dolichopodidae from Papua New Guinea. IV. New species of
the genus Teuchophorus Loew, 1857. Indo-Malayan Zoology 3:217-262.
MEUFFELS, H.J.G. & GROOTAERT, P. 1987. Dolichopididae (Diptera) from Papua New Guinea. VI. Some
new species of the genus Sympycnus. Indo-Malayan Zoology 4:317-397.
MEUFFELS, H.J.G. & GROOTAERT, P. 1988 (1989). Dolichopodidae (Diptera) from Papua New Guinea. Vill.
Phacaspis, a new genus incertae sedis from the mangrove. Indo-Malayan Zoology 5:311-319.
The theory and practice of estimating regional species richness from local
samples
NICK MAWDSLEY
NERC Centre for Population Biology, Imperial College at Silwood Park, Ascot, Berkshire,
& Department of Entomology, The Natural History Museum, Cromwell Road, London, United
Kingdom
ABSTRACT. Current knowledge of the species richness and spatial distribution of many taxa is
insufficient for the basic description of species assemblages, for research into their ecology and
evolutionary relationships and perhaps most importantly as the basis of management decisions for their
conservation and evaluation. Here I consider the basic problems of sampling diverse insect faunas and
how it affects our ability to estimate the species richness of larger areas. This problem is conceptually
separated into that of estimating local species richness and species turnover (beta diversity). Using
published data for the 64 species of umbellifer (Apiaceae) in Britain it is seen that the error associated
with regional species richness estimates is that of estimating species turnover. Moreover widely separated
localities tend to greatly overestimate species turnover. But the situation is even worse in the tropics
where assemblages contain many more rare species making sampling even more difficult. A number of
species richness estimators are compared using data from the canopy fogging of the ant faunas of trees
in Brunei. These give a wide range of values and each estimator appears to have its own inherent biases.
Essentially the description, recording and estimation of tropical biological diversity requires good quality
data which can only be generated by an increase in the amount of standardised field collections being
made.
KEY WORDS: species turnover, species spatial distribution, Apiaceae, Dipterocarpaceae, Dryobalanops.
INTRODUCTION
The basic question of "How many species are there?" is, and will be for some time, a major
question for biologists to answer (May 1988, 1990, 1992a, Stork 1993). A decade ago,
Erwin (1982) redefined the problem and suggested that the global total for insects alone may
be around 30 million species, highlighting our ignorance of the insect faunas of tropical
forests, arguably the most diverse communities on earth. In recent years numerous studies
have revealed much of the richness of insects in species and functional significance at the
local scale (e.g. Basset & Arthington 1992, Casson & Hodkinson 1991, Erwin 1983,
Hammond 1990, Janzen 1988, Stork 1991, Wilson 1987 and many others), but despite this
wealth of knowledge we are still in no position to use these local samples to estimate the
species richness of larger geographic regions except for just a few relatively well-known taxa
such as the larger moths and butterflies (e.g. Holloway 1987).
A variety of approaches have been used to estimate global and regional species totals
(see Hammond 1992, May 1992a, Stork 1993). Methods range from collating the views of
taxonomists (Gaston 1991), to extrapolating up from local samples using assumptions
concerning the host-specificity of tropical insects (Erwin 1982, Thomas 1990), and using the
ratio of insect:plant species numbers (Gaston 1992, Hodkinson & Casson 1991). Although
the use of several independent methods is useful for overcoming the methodological biases
of various estimates, most do not directly increase our knowledge of the distribution of
tropical floras and faunas within regions.
193
194 NICK MAWDSLEY
Local samples taken in the tropical system under investigation have the important
property that they do provide us with such information. The insights they yield into the
diversity of tropical regions are independent of the patterns and assumptions drawn from
temperate systems or from disparate taxa. The use of principles gleaned from a small and
biased sample of the world's species to predict species numbers of diverse tropical groups
is potentially unreliable (Hammond 1992, 1994a). For example, the general trend for species
ranges to be narrower at lower latitudes (France 1992, Stevens 1989) suggests that there will
be a low species overlap of local faunas in the tropics relative to temperate regions
(MacArthur 1969, 1972), and any general theory or estimate of tropical species richness must
incorporate these regional-scale patterns. Overall there is a paucity of data on the
distribution of species and the degree of faunal overlap between sites in the tropics, despite
there being a major need for such data, both for gaining an understanding of biological
diversity and for its preservation (Gentry 1992, Pressey et al. 1993, Vane-Wright et at.
1991).
The challenge to conservationists trying to preserve biological diversity is to make fast,
reliable and economic assessments of the relative species richness of different localities and
regions world-wide and to assess the distribution of species between such elements (Beattie
1993, Coddington et at. 1991, Cranston & Hillman 1992). For the hyperdiverse groups such
as arthropods, fungi and microorganisms, statistical methods and empirical data must be used
to assess species richness which will need to extrapolate outside the extent of the data,
thereby potentially multiplying biases and errors (Colwell & Coddington 1994). Clearly we
must have a better understanding of these methods and their application to estimate the
species richness of any given area, be it a hectare plot or a continental landmass. Moreover
it is in the tropical regions of the world where the bulk of species are found that these
methods will be of most use.
Attempts to estimate tropical insect species richness from local samples taken in tropical
forests have produced figures spanning the full range of global species richness estimates
from 2 million species (Hodkinson & Casson 1991) to 80 million and more (Hammond 1992,
Stork 1988). These huge differences arise because we do not have sufficient data, nor do
we know how to scale up in this way. It is therefore the intention of this paper to consider
some of the basic problems involved. I will show that in theory the estimation of regional
richness from local samples is a relatively straightforward process, but that in practice we
are faced with some major conceptual and methodological problems. I then go on to analyse
the process of scaling up from local to regional scales using data on the distribution of British
umbellifers and I compare a variety of methods that have been produced for this purpose
using samples of canopy ants fogged from a number of trees at Kuala Belalong in Brunei
Darussalam, Borneo. Finally I consider the resources required to estimate the species
richness and distribution of a hyperdiverse taxon such as the Coleoptera (beetles) in a region
such as the island of Borneo.
SPATIAL SCALE AND SPECIES RICHNESS
Regional species richness describes the species richness of large, subcontinental areas where
the large-scale biogeographic processes of speciation and geographic dispersal dominate
(Cornell & Lawton 1992, Ricklefs 1987). Within regions there exists spatial heterogeneity
in the environment at all scales, ranging from large-scale differences in landscapes, geology
and climate, in habitat structure and type, through to the finer-scale patterns of microclimate
and resource distribution. Tropical forests, especially, are noted for their many different
forest types as well as their high floristic richness (Whitmore 1984). In Amazonia for
example, Erwin (1988) considers there to be many more types of floristic assemblage than
Estimating species richness 195
botanists have yet described (see also Gentry 1988), a feature that he argues is partly behind
the great diversity of tropical insect faunas.
It is against this background of floristic and environmental variation that species are
adapted and distributed. Within regions, species distributions are highly variable; some will
be widespread, but most will have restricted geographical distributions (gamma-rarity) and/or
be restricted to certain habitats (beta-rarity) (Cody 1986, Gaston 1990, Rabinowitz et at.
1986). The result is that as we move from one locality to another there will be a turnover
in the local species composition, and we will accumulate species in our total sample (the
familiar species-area relationship; Connor & McCoy (1979), Gleason (1922), Williamson
(1988». Therefore regional richness is related to local richness by the species turnover
between localities within a region, so that (Schluter & Ricklefs 1993, Whittaker 1977):
= mea,n local mean species (1)
specIes richness X tuJnov~r per
localIty
where
Regional species turnover mean species turnover per locality x number of localities (2)
regional species richness (3)
mean local specIes rlchlless
(Whittaker's (1977) beta diversity measure)
Theoretically, we simply need to know the mean local richness, the mean species turnover
per locality and the number of localities in a region to estimate the regional species richness.
In the context of extrapolating from local samples, species turnover per locality can be
thought of as the rate at which we accumulate species per additional locality in our total
sample (and hence a non-linear function of the number of localities) or as a measure of the
average pair-wise similarity (beta diversity) of all the localities within a region. The slope
of the nested species-area curve is therefore a measure of species turnover. Colwell &
Coddington (1994) use the term complementarity to describe the biotic distinctness of two
sites - here I use the term species turnover to describe this same concept simply because
turnover involves the notion of change independent of scale. Furthermore the concept of
species turnover is also well established in the ecological literature in reference to temporal
patterns of turnover (MacArthur & Wilson 1967).
Critical to this approach of the measurement of local richness and species turnover is an
appreciation of spatial scale, and especially the preservation of the same scale of
measurement for studies to be comparative. Local richness is typically defined as the species
richness of an homogenous community (Whittaker 1977). But homogeneity is not a feature
of the natural world at any scale and the continuum concept of community organisation in
space presents a more realistic picture of how species are distributed in response to
environmental change (Austin 1985, Auerbach & Shmida 1993), emphasising the
individualistic nature of species distributions (Gleason 1939, Hengeveld 1990). The patterns
of species turnover will therefore be variable and continuous at alI spatial scales, with
different patterns and processes dominant at different scales within the spatial heirarchy of
ecological systems (Ricklefs 1987). Hence definitions and measures of local richness should
ideally consider the scale at which biotic interactions, abiotic fluctuations and disturbances
occur (Cornell & Lawton 1992, Lewinsohn 1991). This implies that the scale of
measurement should be different for different taxa. For example, small staphylinid beetles
perceive and interact with the environment and each other at a much finer scale than do
birds, and so the "grain" of the sampling regime used for staphylinids should be smaller than
that for birds (Wiens 1989). Selection of appropriate scales must rely on the biological
intuition of the investigator or upon initial investigations of species' distribution patterns
(Carlile et at. 1989, Greig-Smith 1964, Wiens 1989). This is the essential problem in
196 NICK MA WDSLEY
comparing the patterns of diversity between different taxa.
THE EFFECT OF SAMPLING ON LOCAL INVENTORY
It is a huge task to get a full species list for the insect fauna of a locality. The only
reasonable option is sampling the flora and fauna in a quantifiable way which allows
comparisons to be made between sites and among taxa (Greig-Smith 1964, Krebs 1989,
Southwood 1978). However there are two factors that affect the results we gain by
sampling, these being that the number of species collected in a sample is related to the area
sampled and to the sampling effort used (Gleason 1922, Greig-Smith 1964, Preston 1948,
Sanders 1968, Williams 1943).
Sampling and Area

The sampling strategy used to enumerate local sites is crucial because measurement of
local richness and species turnover is scale dependant. The relationship between species
richness and area illustrates that as we increase our local scale then our estimate of local
richness increases; and by definition that for species turnover decreases by an equivalent
amount. Thus when estimating regional diversity, our measures of local richness and species
turnover must be made at the same scale. Usually this will involve using equivalent sampling
methodologies in enumerating local richness and species composition, for instance using the
same sized plots for plants or the same trapping methods for animals. It would clearly be
invalid to directly compare the species richness and floral turnover between two forest plots
of 1 hectare and 5 hectares.
Sampling and Sampling Effort.

Sampling effort has several effects upon enumerating local insect faunas that are crucial
to consider. First, there is no single sampling method that will result in a full species list
for a locality - different sampling techniques collect different portions of the total fauna at
anyone site (Southwood 1978). For example, canopy fogging is biased towards flightless
arboreal insect taxa, whereas traps depending upon flight activity such as light and malaise
traps will be biased towards highly mobile flying groups (Hammond 1990, Mawdsley, in
prep.). Generally sampling methodology has more effect on the similarity between samples
than site-specific factors (Casson & Hodkinson 1991). In practice it is difficult to combine
data from different methods to yield total local richness estimates, requiring careful
calibration of methods at sites of known total faunas (Hammond 1994b). Standardisation of
a variety of methods to even out the biases of single methods into a standard 'sampling
package' used for sampling insect faunas in tropical forests worldwide would maximise the
potential of local samples for use in comparative studies and in estimating regional and global
species richness (Gadagkar et at. 1990, Stork 1994).
A second problem, especially when quantitative data is required, is that many methods for
sampling insects such as light and malaise traps are relative methods in that they do not give
absolute 'standing crop' estimates of the abundance and species richness in a given area.
Patterns of abundance seen in Malaise or light trap samples, for example, are do not directly
mirror those of the insect assemblage sampled. Instead, some species are over-represented
while others are under-represented, and this is determined by a complex interplay between
the nature of the method and species-specific behaviour (Muirhead-Thompson 1987).
Knowledge of the bias inherent in the method is crucial here and requires careful
interpretation. Moreover the bias associated with any method is likely to vary between
habitats and faunas thereby complicating matters further.
Perhaps the most evident point about sampling is that by definition only a certain
proportion of the local fauna that is susceptible to being sampled by any given method will
be collected (Sanders 1968). A given sample will only collect a certain number of
individuals and species alike, and incomplete sampling in this way may be solely responsible
for the differences in local richness and the composition between samples collected at
different sites (Colwell & Coddington 1994). Methods such as rarefaction have been
developed to overcome the problem of incomplete sampling in estimating the local richness
of sites for comparative purposes (Sanders 1968, Gage & May 1993, Heck, van Belle &
Simberloff 1975). However undersampling proves an even greater problem for estimates
of species turnover. A null hypothesis for between-site comparisons would be that between-
site differences are entirely due to random sampling from the same local community. This
difference, which we can define as "pseudoturnover", will always be part of the measured
species turnover and so must be estimated (Nilsson & Nilsson 1985). Only when we are
dealing with complete inventories will the contribution of pseudoturnover to estimates of
species turnover be equal to zero, however complete inventories of local insect faunas are
extremely difficult to get (e.g. Hammond 1990). Pseudoturnover may be estimated either
by replication of field samples at one locality (which in practice may be difficult because of
fine-scale heterogeneity, particularly in tropical forests, and 'pseudoreplication' due to the
presence of uncontrolled spatial autocorrelation) or by bootstrapping techniques (see Krebs
1989, Robinson & Tuck 1993).
Finally, the temporal extent of collecting, in a similar way to area, will affect measured
local richness. Seasonal patterns of species occurrence vary between taxa and guilds, and
species accumulation curves rarely plateau even after many months sampling (Basset &
Arthington 1992, Hammond 1990). Moreover, the pattern of species accumulation with time
will vary between taxa and guilds. For example, herbivore assemblages, which often have
peak abundances correlated with periods of leaf production (Basset & Arthington 1992), can
be more effectively sampled in a shorter time frame than taxa and guilds that do not show
such temporal concordance. Thus the timing and temporal extent of any sampling
programme will affect estimates of local richness and species turnover through temporal
turnover.
Tourist species and 'mass effects '.

Local dispersal of individuals from areas where their populations are at high densities
('source populations') to unfavourable areas which could not support a viable population of
that species alone ('sink populations') (see Pulliam 1988) will enhance levels of local richness
and lower the amount of species turnover between sites. These 'mass effects' of immigration
have been documented for desert and tropical forest plant communities (Hubbell & Foster
1986, Shmida & Wilson 1985) and entomologists have used the term 'tourists' to describe
immigrant, non-resident species (Moran & Southwood 1982, Stork 1987a). Several studies
have classified up to 20-30% of arboreal insect species as tourists (Basset & Arthington
1992, Gaston et al. 1993, Stork 1987a), and although this phenomenon distorts our
measurements of local richness and species turnover, it is debatable to what extent tourists
should be discounted as local community members (Moran & Southwood 1982). For
example they may still provide services as pollinators and become food for predators.
However, in a qualitative analysis of species diversity and species turnover where species are
recorded only in terms of presence and absence, a singleton tourist species would be
accorded the same weight as the dominant species in the assemblage. The effect of tourists
in qualitative analyses is therefore to cause an overestimation of local species richness but
an underestimation of species turnover. Large samples may suffer from this especially
because the proportion of species in a sequence of samples that are tourists increases as
further samples are added (P.M. Hammond 1994b, and personal communication).
198 NICK MAWDSLEY
METHODS OF ESTIMATING REGIONAL RICHNESS FROM LOCAL SAMPLES
Estimates of tropical insect richness derived from local samples have used assumptions about
the degree of host-specificity of tropical insects (Erwin 1982, Thomas 1990) (i.e. insect plant
species richness ratios (Gaston 1992» rather than to the spatial patterns of local richness and
species turnover, principally because these latter patterns are so poorly known. Generally we
can identify four types of methods commonly in use for estimating the species richness of an
area from local samples (Table 1), the first three of which represent the data-analytic approach
and the latter the sampling-theoretic approach (Bunge & Fitzpatrick 1993, Colwell &
Coddington 1994, Hammond 1994b). For a review of the theoretical aspects of various
sampling models not listed here see Bunge & Fitzpatrick (1993).
Despite detailing a variety of statistical methods that are directed at the estimation of species
richness, Bunge & Fitzpatrick (1993) conclude that the problem is "quite resistant to statistical
solution, essentially because no matter how many classes have been observed, there may still
be a large number of very small classes" i.e. rare species. They conclude that estimators of
coverage, that is the proportion of the total species richness that are represented in the samples,
are more reliable and are better understood, and they recommend Chao's method which utilises
the concept of coverage. Evidence from studies of single temperate forest plant communities
(Palmer 1990, 1991) and computer-simulated communities (Baltanas 1992) using the species-
area relationship, the lognormal species-abundance distribution and the jackknife suggests that
we should use these methods with caution. Regional species richness was usually
underestimated, except for species-area extrapolation applied to plant communities, which
provided an overestimate. All the measures considered by these authors were shown to be
biased, but the jackknife was the most reliable and the species-area extrapolation the least
(Baitanas 1992, Palmer 1990). However, in terms of estimating regional richness the
jackknife is likely to be of little value for hyperdiverse taxa that have not been sampled
intensively because it cannot give values greater than twice the number of species
encountered in samples (Smith & van Belle 1984). The species-area relationship, despite its
TABLE 1: Methods used for estimating the species richness of an area from local samples.
A. Data-Analytic Approach
1. Extrapolation of species accumulation curves (palmer 1990, Baltanas 1992, Soberon & Llorente 1993).
2. Fitting parametric models
(a) Relative abundance e.g. log-series, log-normal or negative binomial etc. (Fisher et al. 1943, Krebs 1989,
Palmer 1990, Coddington et al. 1991);
(b) Fitting parametric models of body size distributions (May 1978)
3. Ratios of known values of species richness (Colwell & Coddington 1994, Hammond 1994):
(a) Hierarchical
i. For example, Hemiptera:insect species richness ratio - e.g. Hodkinson & Casson 1991
(b) Non-hierarchical
i. For example, insect:plant species richness ratio - e.g. Gaston 1992
ii. Described:undescribed species richness ratio - e.g. Hodkinson & Casson 1991, Hodkinson
1992, Hodkinson & Hodkinson 1993)
B. Sampling-theoretic Methods
4. Non-parametric statistical methods (Colwell & Coddington 1994)
(a) The first-order and second-order jackknife (Heltsche & Forrester 1983, Smith & van Belle 1984,
Krebs 1989)
(b) The Bootstrap (Smith & van Belle 1984, Krebs 1989)
(c) Chao's methods (Chao 1984, 1987, see Bunge & Fitzpatrick 1993, Colwell & Coddington 1994)
shortcomings, is perhaps the most useful of method for estimating the species richness of large
areas because area is the most easily measured correlate of species richness and its nature and
behaviour is relatively well understood (e.g. Boecklen & Simberloff 1986, Connor & McCoy
1979, Gould 1979, Kelly et al. 1989, Martin 1981, Miller & Wiegert 1989, S imberloff 1992,
Stevens 1989, Williamson 1988). Why it behaves badly as an estimator of species richness
and how to overcome this will be discussed later.
Although the number of species is the quantity of interest in species richness estimation,
without any measure of the variance associated with a mean estimate it is difficult to interpret
in any common sense way. Variances are as (if not more) important than the actual mean
estimate itself. Variances associated with data-analytic measures must be gained by
replication, whereas for the sampling-theoretic measures there are a formulae available for the
calculation of variance (see Colwell & Coddington 1994).
A final point that relates to the process and dangers of extrapolation in general, and notably
to that of the species-area or accumulation curves and integration of distributions of relative
abundance for these methods have no intrinsic maximum as do the non-parametirc methods
(Colwell & Coddington 1994), is the assumption that the patterns of species accumulation
outside the range of the data collected are the same as those within the data. This point has
been the focus of much debate concerning high estimates of deep-sea species richness of 10
million species (Grassle & Maciolek 1992, May 1992b, Poore & Wilson 1993). The patterns
of local richness and species turnover among a set of local samples need to be representative
of the patterns within the whole region (Baltanas 1992). Representativeness is the crucial
factor and is extremely difficult to objectively obtain.
ESTIMATING THE REGIONAL RICHNESS OF A KNOWN BlOTA: UMBELLIFERS

IN BRITAIN
At present there have been few evaluations made of the efficiency of any of these methods in
estimating the species richness of a known regional biota. Miller & Wiegert (1989) fit a
negative binomial model to species accumulation curves for the regional flora of the southern
Appalachians in the United States, but they do this by randomly sampling the regional pool
of species and individuals in a computer-simulation. This acts to erase the spatial
heterogeneity present in the region, making it essentially a problem of estimating local species
richness without species turnover. Other studies based on single communities (Baltanas 1992,
Palmer 1990, 1991) would suggest that these methods would be even less reliable for larger
areas where there is greater heterogeneity in species distributions and where a smaller
percentage of the regional biota is contained in anyone local sample.
The distribution of the British flora has been well documented through the efforts of a great
many professional and amateur naturalists (Perring & Walters 1976). In Britain there are 64
species of umbellifer (Apiaceae) whose presence and absence was recorded for the 145 50km
x 50km grid squares in Britain. [The choice of this group of plants is arbitrary. It was chosen
because it contains a manageable number of species with well-studied biology.] These data
were then used to answer the following questions: (1) How well can we estimate this regional
richness of 64 species from local samples of grid squares? (2) What sampling regime should
we use to maximise the representativeness of our local samples with respect to the regional
patterns of local richness and species turnover? The main focus of this analysis was to use the
species-area relationship and to partition such estimates into estimated local richness and
estimated species turnover. Species-area curves were extrapolated by a power function which
showed the best fit to each sample (see Connor & McCoy 1979).
200 NICK MAWDSLEY
How Many Localities Do We Need To Sample?

Grid squares sampled randomly from the national grid clearly showed a positive correlation
between the number of squares sampled and the accuracy of the estimate of regional richness
(Greig-Smith 1964, Krebs 1989, Southwood 1978). Relatively good estimates of regional
richness were gained when 40 grid squares (28 % of the total) were sampled, yielding a mean
regional estimate of 76 species (Figure 1 and Table 1). For most tropical regions 28 % of the
land area inventoried is unrealistically high, and in most countries, National Parks and forest
reserves rarely amount to more than 5% of the land area (Hartshorn 1983). Floristic and
faunal lists of even these protected areas are scarce. In terms of being able to estimate regional
(or national) species richness from such data, it seems likely that large errors would creep in.
For the umbellifers in Britain when only 4 % (N =5) of grid squares were used, regional
richness was overestimated by 250 %, giving a mean estimate of 155 species (Figure 2 and
Table 1). The bias of these estimates is always positive (overestimation). Given that 4 % of
the area of a region for which data on floral and faunistic composition is known, how can we
improve these estimates of regional richness?
Are Random Local Samples Representative?

The mean local richness and species turnover of umbellifers for all 145 grid squares in
Britain is 27.6 and 2.32 respectively. In effect our samples need to be sufficiently
representative to yield good estimates of these two quantities. How well do different levels
of sampling achieve this? Estimates of mean local richness were close to the population value
after 14 % of grid squares (N =20) had been sampled (Figure 3), but where only 7 % (N = 10)
or less of grid squares were sampled, the variance of the estimates was high and mean local
richness was poorly estimated.
200 200
180 180
160 160
til Ul
w 140 w 140
U
w U
w
a. 120 a. 120
til Ul
w w
> 100 > 100
~
..J 80 ~
..J 80
::J ::J
:e
::J 60
:e::J 60
() ()
40 40
20 20
0 0
0 29 58 87 116 145 0 29 58 87 116 14!
NUMBER OF GRID SQUARES (N) NUMBER OF GRID SQUARES (N)
Figure 1. Extrapolated species accumulation curves Figure 2. Extrapolated species accumulation curves
for umbellifers (Apiaceae) in Britain when 40 grid for umbellifers (Apiaceae) in Britain when only 5 grid
squares are randomly sampled (equivalent to 28 % of squares are randomly sampled, equivalent to 4% of
the area of Britain). Each curve represents a single the area of Britain. The same method of
sample of 40 grid squares. This is randomly randomisation was used as for Figure I. Note that
subsampled to generate randomised data of sample error and the variance of these estimates are much
sizes of 5, 10, 20 and 40 squares to which the power higher than those in Figure 1.
function if fitted.
40 2.5 r---,----,----T""-----,r----,
• TURNOVER PER GRID SQUARE

35 o TOTAL TURNOVER ~
'"'"w
~
III:
Z ~ 2.0
:I:
0 30
o
+ r+
Z
a: III:
.J
« ... :::>
0
0 25
;--r- '"
III I',
.J U
Z ~ 1.5
«
III J VI
::IE
20
I~ Il I
11
15 1.0 '-----'--'--'-"--
- --'-___--'-___- - . J
5. 10. 20. 40. 5. 10. 20. 40.
NUMBER OF GRID SQUARES (N) NUMBER OF GRID SQUARES (N)
Figure 3. The effect of the number of grid squares Figure 4. The effect of the number of grid squares
sampled randomly on the mean and standard error of sampled randomly on the mean and standard error of
mean local richness estimates. These are calculated mean species turnover estimates. These are calculated
from five replicates (n=5) for each sample size. The from five replicates (n=5) for each sample size. The
known mean local richness across the UK for known mean species turnover per neighbouring grid
umbellifers is 27.6. square is 1.27 i.e. 27 % .
Mean species turnover is more difficult to estimate than mean local richness because the
total turnover within any sample is highly dependant upon N (Figure 4). This is because of
the species-area relationship where larger samples contain more species in total, and this
indicates the dependence of extrapolating species-area curves on estimating species turnover.
We can compare species turnover from different sample sizes by comparing the amount of
species turnover per grid square. For the whole dataset, species richness and species turnover
scales with area (N) by NO. 12 , and so by calculating (total sample turnover/NO. 12) we have a
measure of species turnover per unit area. The population value of this 'species turnover
density' that we want to estimate is 2.32/145°.12 = 1.27. Figure 4 shows that our estimates of
this quantity are consistently over this by about 40% (note that species turnover has a zero
value when equal to 1, i.e. when local richness = regional richness).
Therefore if we sample sufficiently to estimate mean local richness, we have still not
sampled enough to accurately estimate mean species turnover. Overestimates of regional
richness using species-area curves seem to be because we overestimate species turnover. How
then can we improve our estimate of species turnover and therefore regional richness?
The Representativeness of Regional Species Turnover Estimates

The observation that distant sites are usually more different in their environmental attributes
than near sites (Bell et al. 1993, Williamson 1988) leads to the prediction that species turnover
will be correlated with the distance between sites. This relationship has been observed for a
variety of taxa along large-scale transects in Britain (Harrison, Ross & Lawton 1992), for
continental mammal faunas (Flessa 1980) and for canopy insects in Brunei (Stork 1987b). Its
implications for extrapolating species-area curves are that if our sample localities are widely
separated then the measured species turnover between them will overestimate the true turnover
between localities.
This problem can be clearly observed with umbellifers in Britain. If replicate samples of
5 grid squares (4 % of the total) are taken from a contiguous block of adjacent squares (i.e.
202 NICK MAWDSLEY
localised sampling) rather than randomly from throughout the country (Figure 5), then
estimates of regional richness are significantly better (t-test; t=4.19, 13 d.f., p<O.ool)
(Figure 2 cf. Figure 6; Figure 7; Table 2). Partitioning the regional estimates into local
richness and species turnover (Figures 8 & 9; Table 2) it can be seen that it is species turnover
only that is significantly different between random and local sampling regimes. In addition,
for localised and random sampling extrapolations combined, the error of regional estimates is
proportional to measured species turnover per grid square among all these samples (Figure 10:
error of estimate (y) = 0.003(species turnover per grid square (x» + 1.062; R2 = 0.636; F-
ratio = 22.693; p < 0.001).
A .....,/1 It.. ""1'1
lI:!~ l ' -!---J.Jl.--:::< ~-=rL.,:

,Ut U
..j---hO!!I ...,--
U UoHOTH
el',.-.II ""- •.'M_
RANDOH LOCALISED
Figure 5. Distribution maps of Seseli libanotis (Apiaceae) in Britain showing

random and localised sampling distributions.
200 200 r-~,
180 180
.,w
180 ....
::IE
180
140 ;:: 140

g <II
.,
Do. 120
w
...
<II
120
w U 100
:0- 100 w
;:: Do.
...':::>" 80
III
oJ
80
::IE
:::>
u
80 '"0z 80
40
aw 40
a:
20 20
0 0
0 2Q 58 87 1111 '45 lOCAL RANDOM
NUMBER OF GRID SQUARES CN) SAMPLE DISTRIBUTION
Figure 6. Extrapolated species accumulation curves Figure 7. The mean and standard error of regional
for umbellifers (Apiaceae) in Britain when 5 grid species estimates of umbellifers (Apiaceae) in
squares (4 %) are sampled from a localised area, as Britainwhen samples are taken either locally or
opposed to being sampled at random (see Figure 2 for randomly from the national grid. These are calculated
comparison). from seven replications (n=7) for each sample
distribution.
TABLE 2. The effuct of sample size on the estimation of regional species richness, local species richness and species
turnover. Figures are mean estimate ± standard error of the mean for five replicates.
NUMBER OF PROPORTION REGIONAL LOCAL REGIONAL SPECIES

SQUARES OFTOTALN ESTIMATE RICHNESS TURNOVER
SAMPLED (N) (%) (n=5) ESTIMATE ESTIMATE + 1 (n=5)
(n=5)
5 4 155 ± 13 26.2 ± 2.5 1.40 ± 0.03

10 7 29.0 ± 2.3 1.38 ± 0.08
20 14 89 ± 18 27.7 ± 0.7 1.42 ± 0.03
40 28 76 ± 6 27.7 ± 0.5 1.39 ± 0.02
145 100 64 27.6 1.27
2.0 50
'"«a: loB
45
:>
0 40
<II
<II
C 1.B <II
;;: w 35
z - -
+
~
%:
a: U 30
w 1.4 ;;: ~
~
"- oJ
II: 25
...0'"
W
r+
U
> 1.2
0 20
z I'
..
II:
en
:>
1.0
Z
C
w 15
::Ii .
I'
I~
w 10
(j O.B
w
"- 5
<II
0 .8 0
LOCAL RANDOM LOCAL RANDOM
SAMPLE DISTRIBUTION SAMPLE DISTRIBUTION
Figure 8. TIle mean and standard error of "species Figure 9. The mean and standard error of mean local
turnover per grid square" estimates of urnbellifers richness estimates of urnbel1ifers in Britain when
(Apiaceae) in Britain when samples are taken either samples are taken either 10cal1y or randomly from the
locally or randomly from the national grid. These are national grid. These are calculated from seven
calculated from seven replications (n=7) for each replications (n=7) for each sample distribution.
sample distribution.
Correcting For Overestimation By Species-Area Extrapolations

A solution to the problem of overestimation by species-area extrapolation from random
samples may be found by considering the source and magnitude of the error in the estimates
of regional richness when samples are taken randomly. The overestimation of species turnover
is dependant upon sample size, mainly in response to the factor of site separation as outlined
above. The more sites that are sampled, the closer they are, and the percentage error in the
estimate of species turnover shows a strong negative power relationship with the proportion
of localities in the sample (Figure 11):
ERROR,pecies tllroover = 4.20(N-O·91 )
This relationship gives us the potential to correct species-area extrapolated estimates of

regional richness so long as we are confident that we have sampled sufficiently to get
204 NICK MAWDSLEY
..."'
150 "' 26
!(
::E
::E
;::
(I)
;::
(I) "'
a: 21
(I)
(I)
"' "'
>
0
100
.,"'Z Z
..'"
a:
16
~
.....
a:
Z
0
(I)
"'
i3
..."' 11
.."'
(I)
S
w 50 z
a:
u. ::E
0 6
a: u.
0 0
a: a:
a: 0
a:
"' 0
0.9 1.0 1.1
a: 1
1.2 1.3 1.4 1.5
"' 0 20 40 60 80 100
SPECIES TURNOVER PER GRID SQUARE PROPORTION OF GRID SQUARES SAMPLED (Yo)
Figure 10. The relationship between the error of the Figure II. The relationship between the error of the
regional richness estimates and species turnover per total species turnover estimate and the proportion of
grid square. Each point represents the error and grid squares randomly sampled. The points describe
species turnover of a single extrapolation. Error is the extent to which random sampling overestimates
measured as the difference between regional estimate species turnover for sample sizes of 5, 10, 20 and 40
and the known regional richness of 64 species. All grid squares.
extrapolations using five grid squares (both random
and localised; n= 14) have been used.
representative samples of local richness and species turnover. Presumably it has some form
of inverse relationship with the species-area relationship.
To summarise, the species turnover estimated among local samples of adjacent localities
represents a truer measure of the mean species turnover per locality than widely separated
sites. This is important because species turnover is the quantity that we need to measure
with most accuracy for accurate extrapolation to a regional total. However, random samples
do give a better estimate of mean local richness with less variance, but because the effect of
species turnover upon the regional estimate is greater, localised sampling performs better
(Table 3). An important result is that a given increase in the number of localities sampled
does more to reduce the error in species turnover estimation in the early stages of a regional
inventory when only a few localities have been sampled (Figure 11). For umbellifers
(Apiaceae) in Britain, this analysis suggests that we need species lists of localities
representing about 10% of the area of Britain to give sufficient data to extrapolate up to the
regional total. Miller & Wiegert (1989) suggest a value of 2-3 % for their study of the
regional flora of the southern Appalachians. This is likely to be a lower limit because of
their random computer-simulated sampling methodology (see above).
TABLE 3: The effect of localised versus random sampling on the estimation of regional species richness, local
species richness and species turnover. Figures are mean estimate ± standard error of the mean for five replicates.
LOCALISED SAMPLING RANDOM SAMPLING

REGIONAL RICHNESS ESTIMATE 84 ± 10 155 ± 13
LOCAL RICHNESS ESTIMATE 33.2 ± 3.9 26.2 ± 2.0

REGIONAL SPECIES TURNOVER ESTIMATE 1.08 ± 0.04 lAO ± 0.03
THE TROPICAL PERSPECTIVE OF ESTIMATING REGIONAL RICHNESS.
The distribution and abundance patterns of tropical species are markedly different to those
of temperate systems. Tropical systems are more species-rich, have populations at lower
densities (Basset & Arthington 1992, Currie & Fritz 1993, Elton 1973, 1975, Watanabe &
Ruaysoongnern 1989), and samples of insects and plants from tropical forests usually have
a highly inequitable distribution of abundance with many species being relatively rare (e.g.
Barlow & Woiwod 1989, Stork 1991, and Figure 12). These features are correlated with
inaccuracy of species richness estimates (Baltanas 1992, Mignoti & Meeden 1992), making
the goal of producing accurate estimates even more difficult. For tropical regions it is
probable that to estimate species richness with any certainty we will have to sample a greater
proportion of localities than less diverse regions (Mignoti & Meeden 1992).
A tropical species accumulation curve tends to rise faster than a comparable one from
temperate regions and the final asymptote will also be higher. The species accumulation
curve for ants fogged from seven trees of the same species (Dryobalanops beccarii -
Dipterocarpaceae) in Brunei illustrates the problems that tropical biologists face in estimating
species richness in that the curve shows no sign of levelling off (Figure 13; Mawdsley
unpublished data). For this dataset, the problem of pseudoturnover would be very great if
it was used as one of many local faunal lists to estimate regional richness. However we can
use these samples to compare the different methods by attempting to estimate the total species
richness of canopy ants in the locality at Kuala Belalong. Ideally a variety of real and
simulated datasets should be used to explore the performance of these methods. The results
presented here are illustrative and limited in their use of a single dataset drawn from a
universe of unknown total species richness. Initially it is worth considering what we are
extrapolating to for this is far from obvious. The samples are taken from seven trees
aligned along a transect 6km long and the tree is taken to be the unit of sampling effort.
Each tree is at the large end of the size scale (80-120cm dbh) and trees like this are found
at a density of only a few per hectare (e.g. Newberry et al. 1992).
30
25
II)
CII
'u
CII
20
Co
'"
..
i:
'0
15
...
CII
.c
E
:l
10
Z
0
Oro 1 21Q 4 &to 16 321Q 64 12& 10 256 512101024
Abundance class (number of individuals)
Figure 12. Species-abundance distribution for Brunei canopy ants sampled from seven trees by canopy fogging. Note
that these samples have a relatively low proportion of singleton species (abundance class = I individual) compared
to many other more diverse taxa.
206 NICK MAWDSLEY
So are the various methods estimating the total species richness of these seven trees only?
Or of the area of rain forest in the 6km between them? Or of just the part of this area that
is characterised by the seven trees sampled? For example, they are all the same species and
thus do not incorporate any species with specific plant-associations. These are important
points and applicable to all sampling methods. Their effect will depend both on the taxon
considered, the method used and the scale and design of the sampling programme.
Estimates for ant species richness from the samples taken at Kuala Belalong in Brunei are
given in Table 4. For the species accumulation curve I have fitted two models (Figure 14,
Colwell & Coddington 1994, Miller & Wiegert 1989, Soberon & Llorente 1993):
1. Logarithmic model:
S = 93.0x log 10 (N+l)-15.6 R2 = 0.992
where S = the cumulative number of species and N = the sampling effort or number of sites
and
2. Negative binomial model:

5 = Smaxx (l-exp (-0.15 x N» R2 = 0.997
Log models do not have an asymptote and are considered to be useful for speciose taxa
(Soberon & Llorente 1993). In contrast, the negative binomial has an asymptote (S~
which is the estimate of interest. The first-order jackknife is dependant upon the number of
singletons in the data while the second-order jackknife and both Chao's methods are functions
of the number of singletons and doubletons. Colwell & Coddington (1994) note that Chao
#1 provides a lower bound estimate but that it performs well with data containing many rare
species as do most tropical insect data. For hyperdiverse tropical insect faunas the non-
parametric estimators will certainly underestimate because their estimates are constrained
70 1-- Negative binomial model _ .. - Log mo

III
·u" 60 250
"Co
III
C 50 III 200
co
'0... 40
III
""
.--.-
J:
"E
.0
30 ..
u
.1: 150
::l
"..
.2: 20
.
c:;
Co
III
100
OJ C
"S
E 10 < 50
::l
0
0 0
0 2 3 4 5 6 7 0 50 100 150
Sampling Effort (number of trees fogged) Sampling Effort (number of trees)
Figure 13. Species-accumulation curve for ants Figure 14. Extrapolated species accumulation curves
(Formicidae:Hymenoptera) fogged from the canopies for the ant data shown in Figure 13. Two models
of seven Dryobalanops beccarii trees at Kuala have been used, a log-model and a negative binomial
Bela10ng FSC, Brunei. model (see text for details).
TABLE 4: Estimates of species richness using local samples of ants fogged from the canopies of seven Dryobalanops
beccarri trees in Brunei (see text for discussion).
Estimation Method Species richness estimate
Observed species richness (7 rrees) 69

Species accumulation - log model (extrapolation to 30 trees) 123 ± 30
Species accumulation - negative binomial model 110
First-order jackknife 105 ±9
Second-order jackknife 98 ± N/A
Bootstrap 84 ± N/A
Chao's method #1 108 ± 12
Chao's method #2 136 ± 21
2 x 2 contingency tables 102 ± 1
Integration of log-series (30 trees; 6000 individuals) 138
Integration of log-series (5000 trees; 106 individuals) 267
Integration of log-series (18,000 trees; 3.5 x 106 individuals) 300
to less than twice the observed number of species for the jackknife and bootstrap methods,
and to (S2obs + 1)/2 (Colwell & Coddington 1994, Smith & van Belle 1984). Additionally
these ignore the percentage of the population that has been sampled (Mignoti & Meeden
1992) unlike the negative binomial in which the probability of gaining another new species
in a series of samples is inversely proportional to the number already accumulated (Colwell
& Coddington 1994). Two other methods used are 2 x 2 contingency tables which estimate
the joint absences between samples (i.e. the number of species not recorded in two samples)
given the assumption of random sampling (see Pimm et al. 1994) and integration of the area
under a log-series abundance distribution fitted to the observed data (Magurran 1988). The
assumption of the former that individuals are encountered independently is also an assumption
of the rarefaction method and one that clearly is sometimes not tenable (Gage & May 1993).
The estimates are shown in Table 4 and may be divided into: (1) the non-asymptotic
estimates (log model of the species accumulation curve; integration of the log-series
abundance distribution) where it is necessary to select a value of sampling effort (i.e. trees
or individuals sampled) along the x-axis for estimation of species richness, and, (2) the
asymptotic or non-parametric measures which give a single estimate independent of any
sampling effort parameter. The outstanding pattern in Table 4 is that the asymptotic
measures mostly have produced broadly similar estimates of about 100-110 species, whereas
the non-asymptotic measures are still rising at this level of species richness (Figure 14). The
conclusions drawn from the British umbellifer analysis that random sampling overestimates
species turnover per locality means that these non-asymptotic measures are likely to be
overestimates. This coupled with the fact that the non-parametric estimates are probably
underestimates (see above) suggests that the true answer lies between 100 and 300 species!
Is this a satisfactory estimate? Almost certainly not. The range of estimates in Table 4
is great and although we can generalise about the biases of these methods, the adage that
"any method is only as good as the data" must be borne in mind. Generally the estimation
of species richness from local samples demands the data to be both representative and of
sufficient quality and coverage to be able to estimate with any practical certainty the species
richness of a larger region. In the illustrative example of the Bruneian ants it is unlikely that
208 NICK MAWDSLEY
either of these requirements have been met with only seven trees of the same species being
sampled.
There is also the problem of what is this estimating? For example, if these samples
represents a 6km x 6km area, and there are about 5 large trees per hectare, then the 18,000
trees contained in this area are predicted to be home to roughly 300 to 380 species of ant
(integration of log-series and log-model of species accumulation respectively, Table 4),
values which seem too large and which have no empirical support (Wilson 1987). Certainly
these methods are estimating only the pool of species that are susceptible to fogging (for ants
however this is probably a high proportion) and will therefore consider only a proportion of
the total fauna. These uncertainties demand increased sampling to gain better estimates.
For taxa other than ants the problem is greater. Although the dominant insect taxa in
terms of numbers of individuals and biomass (Erwin 1983, Stork 1991, Wilson 1987), ants
are a relatively species-poor group. There is good evidence, both empirical '(Wilson 1987)
and from consideration of the basic biology of different insect taxa, that the species turnover
rates of other taxa such as beetles will be appreciably higher. Species accumulation curves
for these groups are likely to be steeper and will continue rising to reach a much higher
plateau, lowering the accuracy of our predictions of species richness from a given set of local
samples.
SPECIES RICHNESS ESTIMATION: THE SCALE OF THE PROBLEM.
The way forward to estimate regional insect species richness directly as previously discussed
requires: (1) better estimates of total local richness because we still do not know how many
species in total are found in, for example, one hectare of rain forest and (2) greater
understanding of the patterns of species turnover and distribution among different taxa within
a variety of regions. More complete inventories necessitate an increase in the amount of
standardised and therefore comparable field collections and records of all poorly known taxa
being made, by no means a small undertaking. The value of such information cannot be
underestimated and is itself paramount to the goal of sustainable use of the world's natural
resources. Only through careful observation of the real world will understanding of its
organisation and the magnitude of its diversity become apparent (Hammond 1994a,b).
As a model, how long would it take to make representative collections of beetles in
Borneo with the aim of discovering the magnitude of beetle species richness and the
distribution of species within Borneo? If we consider, for arguments sake, that 5% is a
minimum total area required to be surveyed to estimate regional richness relatively accurately
(Figure 6), and a "typical" locality is about 50,000 hectares in size (i.e. 20km x 20 km) (e.g.
an ATBI site: Janzen & Hallwachs 1993, Yoon 1993), then we would need to survey about
100 such localities to cover 5 % of the area of Borneo.
How long would this take? My PhD has involved putting together such a collection at a
single locality in Brunei (11000 specimens representing 3164 species), and the total person-
hours needed for this has been about 10,000 hours. To survey 100 localities would therefore
take 106 hours, roughly 500 person-years. In just 5 years, 100 people (a mix of specialist
taxonomists, mostly trained parataxonomists plus a few data analysts) could provide an
answer to the regional diversity of Bornean beetles and detailed information on their
distribution which would have both great significance both in scientific and management
terms. More reliable data might involve sampling perhaps 40,000 specimens per locality,
increasing the resources and time required to complete such a programme but not
proportionally so. Comparable estimates made for samples of spiders collected in Bolivia
(Coddington et al. 1991) suggest that to inventory a single 'average' tropical site for a fauna
of roughly 800 species might take between 115 and 300 person-days and require a sample
of 3000-8000 specimens. 100 sites might therefore take 100 person-years. As anyone who
has ever completed faunal surveys knows, the time-consuming element is sorting, especially
for difficult and poorly known taxa. Integration of basic research with training of
taxonomists and parataxonomists can facilitate much of the preparation and initial sorting of
samples taken for biological inventory (see Janzen et al. 1993).
The problem of estimating the regional richness in a species-rich group such as Bornean
beetles may be simplified by having as a goal the estimation of mean local richness and mean
species turnover within the region. The process of scaling up will be best accomplished by
the use of ratios and of relative species turnover rates (including comparison of the slopes
of nested species area curves) at a variety of scales based on systematically collected samples
(Hammond 1994b). Certainly all methods that scale up from local to regional to global
faunas must include a comparison of species turnover rates for different taxa at different
scales (Hammond 1994a). Explicit consideration of species turnover, long overlooked by
animal ecologists, will certainly illuminate the problem of scaling up. Knowledge of the
variation of the landscape and environment, and distribution of forest types within a region,
is required to determine the localities for sampling to achieve these estimates. For local
richness, a wide variety of localities is the requirement, but for species turnover estimates
these should be localised. For both, localities need to be representative of the environmental
variation within the region.
There is still much to be learnt about the methodology of inventorying and estimating
the species richness of regions, however recent initiatives such as the IUBS-SCOPE-
UNESCO 'DIVERSITAS' programme (di Castri et at. 1992) and 'All Taxa Biodiversity
Inventories' ATBI (Janzen & Hallwachs 1993, Yoon 1993) lend weight to the need and
importance of such ventures and the need for standard means in which to analyse and
interpret such data (Solbrig 1991). Furthermore, since the signing of the Convention on
Biological Diversity in June 1992 in Rio de Janeiro (Johnson 1991; note Article 7 of the
Convention on Biological Diversity) governments are obliged to make efforts to increase the
knowledge of their biodiversity and natural resources as has been done by organisations such
as InBio in Costa Rica (GAMEZ 1991). Clearly such a major co-operative enterprise is not
beyond the realms of reality - it only needs the will to uncover the true dimensions of the
world's biological diversity.
ACKNOWLEDGEMENTS
I am extremely grateful to Peter Hammond for lengthy discussions on the subject of sampling
and methodology and for bringing to my attention the paper by Colwell & Coddington
(1994). John Lawton and Nigel Stork commented on an earlier version of the manuscript.
Special thanks are due to all the members of the RGS/UBD Brunei Rainforest Project who
supported and helped with the fieldwork, notably Impin Lasa, Rupert Hanbury-Tennison,
Toby Howes, Catriona Prebble, Amanda Simpson, and Mark Welsh. This work forms part
of a Natural History Museum funded studentship.
LITERATURE CITED
AUERBACH, M. & SHMIDA, A. 1993. Vegetation change along an altitudinal gradient on Mt. Hermon, Israel-
no evidence for discrete communities. Journal of Ecology 81:25-33.
AUSTIN, M.P. 1985. Continuum concept, ordination methods and niche theory. Annual Review of Ecology and
Systematics 16:39-61.
BALTANAS, A. 1992. On the use of some methods for the estimation of species richness. Oikos 65:484-492.
BARLOW, H.S. & WOIWOD, I.P. 1989. Moth diversity of a tropical forest in Peninsular Malaysia. Journal of
210 NICK MA WDSLEY
BASSET, Y. & ARTHINGTON, A.H. 1992. The arthropod community associated with an Australian rainforest
tree: abundance of component taxa, species richness and guild structure. Australian Journal of Ecology 17:89-98.
BEATTIE, A.J. 1993. Rapid biodiversity assessment. Proceedings of the biodiversity assessment workshop,
Macquire University - 1993. 92 pp.
BELL, G., LECHOWICZ, M. J., APPENZELLER, A., CHANDLER, M., DEBOIS, E., JACKSON, L.,
MACKENZIE, B., PREZIOSI, R., SCALLENBERG, M., & TINKER, N. 1993. The spatial structure of the
physical environment. Oecologia 96:114-121.
BOECKLEN, W.J. & SIMBERLOFF, D. 1986. Area-based extinction models in conservation. Pp .247-276 in Elliot,
D.K. (ed.). Dynamics of Extinction. Wiley, New York.
BUNGE, 1., & FITZPATRICK, M. 1993. Estimating the number of species: a review. Journal of the American
Statistical Association 88:364-373.
CARLILE, D.W., SKALSKI, J.R., BATKER, I.E., THOMAS, 1.M. & CULLINAN, V.I. 1989. Determination
of ecological scale. Landscape Ecology 2:203-213.
CASSON, D. & HODKINSON, I.D. 1991. The Hemiptera (Insecta) communities of tropical rain forest in Sulawesi.
Zoological Journal of the Linnean Society 102:253-275.
CHAO, A. 1984. Nonparametric estimation of the number of classes in a population. Scandinavian Journal of
Statististics 11 :265-270.
CHAO, A. 1987. Estimating the population size for capture-recapture data with unequal catchability. Biometrics 43:
783-791.
CODDINGTON, J.A., GRISWOLD, C.E., DAVILA, D.S., PENARANDA, E. & LAR CHER, S.F. 1991.
Designing and testing sampling protocols to estimate biodiversity in tropical ecosystems. pp. 44-60 in Dudley,
E.C. (ed.). The Unity of EVolutionary Biology: Proceedings of the Fourth International Congress of Systematic
and Evolutionary Biology.Dioscorides Press, Portland, Oregon.
CODY, M.L. 1986. Diversity, rarity, and conservation in Mediterranean-climate regions. Pp.122-152 in Soule,
M.E. (ed.). Conservation Biology: the science of scarcity and diversity. Sinauer Associates, Sunderland, MA.
COLWELL, R. K., & CODDINGTON, J. A., in press. Estimating the extent of terrestrial biodiversity through
extrapolation. Philosophical Transactions of the Royal Society London B 345.
CONNOR, E.F. & McCOY, D. (1979). The statistics and biology of the species-area relationship. American
CORNELL, H.V. & LAWTON, J.H. 1992. Species interactions, local and regional processes, and limits to the
richness of ecological communities. Journal of Animal Ecology 61:1-12.
CRANSTON, P., & HILLMAN, T. 1992. Rapid assessment of biodiversity using 'biological diversity technicians' .
Australian Biologist 5:144-154.
CURRIE, D. 1., & FRITZ, J. T. 1993. Global patterns of animal abundance and species energy use. Oikos 67:56-
68.
DI CASTRI, F., ROBERTSON VERNHES, J. & YOUNES, T. 1992. Inventorying and monitoring biodiversity:
a proposal for an international network. Biology International Special Issue 27: 1-28
ELTON, C. S. 1973. The structure of invertebrate populations inside neotropical rain forest. Journal Of Animal
Ecology.
ELTON, C. S. 1975. Conservation and the low population density of invertebrates inside neotropical rain forest.
Biological Conservation 7:3-15.
ERWIN, T.L. 1982. Tropical forests: their richness in Coleoptera and other arthropod species. Coleopterist's
Bulletin 36:74-75.
ERWIN, T.L. 1983. Beetles and other arthropods of the tropical forest canopies at Manaus, Brazil, sampled with
insecticide fogging techniques. pp. 59-75 in Sutton, S.L., Whitmore, T.C. & Chadwick, A.C. (eds). Tropical
Rain Forest: ecology and management. Oxford, Blackwell.
ERWIN, T.L. 1988. The tropical forest canopy: the heart of biotic diversity. pp. 123-129 in Wilson, E.O. &
Peters, F.M. (eds). Biodiversity. National Academy Press, Washington DC.
FRANCE, R. 1992. The North American latitudinal gradient in species richness and geographical range of
freshwater crayfish and amphipods. American Naturalist 139:342-354.
FISHER, R.A., CORBET, A.S. & WILLIAMS, C.B. 1943. The relation between the number of individuals in a
random sample of an animal population. Journal of Animal Ecology 12:42-58.
FLESSA, K. W. 1980. Biological effects of plate tectonics and continental drift. Bioscience 30:518-523.
GADAGKAR, R., CHANDRASHEKARA, K., & NAIR, P. 1990. Insect species diversity in the tropics: sampling
methods and a case study. Journal of the Bombay Natural History Society 87:337-353.
GAGE, 1. D. & MAY, R. M. 1993. A dip into the seas. Nature 365:609-610.
GASTON, K.J. 1990. Patterns in the geographical ranges of species. Biological Reviews 65:105-129.
GASTON, K.J. 1991. The magnitude of global insect species richness. Conservation Biology 5:283-296.
GASTON, K.J. 1992. Regional numbers of insect and plant species. Functional Ecology 6:243-247.
GASTON, K.J., BLACKBURN, T.M., HAMMOND, P.M. & STORK, N.E. 1993. Relationships between
abundance and body size: where do tourists fit? Ecological Entomology 18:310-314.
GAMEZ, R. 1991. Biodiversity conservation through facilitation of its sustainable: Costa Rica's national
biodiversity institute. TREE 6:377-378.
GENTRY, A. H. 1988. Changes in plant community diversity and florsitic composition on environmental and
geographical gradients. Annals of the Missourri Botanical Garden 75: 1-34.
GENTRY, A.H. 1992. Tropical forest biodiversity: distributional patterns and their conservation significance. Oikos
63:19-28
GLEASON, H.A. 1922. On the relation between species and area. Ecology 3:158-162.
GLEASON, H.A. 1939. The individualistic concept of the plant association. American Midland Naturalist 21:92-
110.
GOULD, S. J. 1979. An allometric interpretation of species-area curves: the meaning of the coefficient. American
GRASSLE, J.F. & MACIOLEK, N.J. 1992. Deep-sea species richness: regional and local diversity estimates from
quantitaive bottom samples. American Naturalist 139:313-341.
GREIG-SMITH, P. 1964. Quantitative Plant Ecology. Second edition. Butterworths, London.
HAMMOND, P.M. 1990. Insect abundance and diversity in the Dumoga-Bone National Park, N. Sulawesi, with
special reference to the beetle fauna of lowland rain forest in the Toraut region. Pp. 197-254 in W.J. Knight
& J.D. Holloway (eds.). Insects and the Rain Forests of South East Asia (Wallacea). Royal Entomological
Society of London.
HAMMOND, P.M. 1992. Species inventory. Pp.17-39 in Groomhridge B. (ed.). Global Biodiversity, Status of the
Earth's Living Resources. Chapman & Hall, London.
HAMMOND, P. M. I 994a. Described and estimated species numbers: an objective assessement of current
knowledge. In Allsopp, D., Hawksworth, D.L. & Colwell, R.R. (eds.). Microbial diversity and ecosystem
function. CABI, Wallingford.
HAMMOND, P. M. 1994b. Practical approaches to the estimation of the extent of biodiversity in speciose groups.
Philosophical Transactions of the Royal Society London B 345:000-000.
HARRISON, S., ROSS, S. & LAWTON, J.H. 1992. Beta diversity on geographical gradients in Britain. Journal
of Animal Ecology 61:151-158.
HARTSHORN, G.S. 1983. Wildlands conservation in Central America. In Sutton, S.L., T.C.Whitmore & A.C.
Chadwick (eds.). Tropical Rain Forest: ecology and management. Blackwell, Oxford.
HECK, K. L. J., VAN BELLE, G., & SIMBERLOFF, D. 1975. Explicit calculation of the rarefaction diversity
measurement and the determination of sufficient sample size. Ecology 56:1459-1461.
HELTSHE, J., & FORRESTER, N. E. 1983. Estimating species richness using the jackknife proceedure. Biometrics
39:1-11.
HENGEVELD, R. 1990. Dynamic Biogeography. Cambridge University Press. 249 pp.
HODKINSON, I.D. & CASSON, D. 1991. A lesser prediliction for bugs: Hemiptera (Insecta) diversity in tropical
HODKINSON, I. D. 1992. Global insect diversity revisited. Journal of Tropical Ecology 8:505-508.
HODKINSON, I. D., & HODKINSON, E. 1993. Pondering the imponderable: a probability-based approach to
estimating insect diversity from repeat faunal samples. Ecological Entomology 18:91-91.
HOLLOWAY, J.D. 1987. Macrolepidoptera diversity in the Indo-Australian tropics: geographic, biotopic and
taxonomic variation. Biological Journal of the Linnean Society 30:325-341.
HUBBELL, S.P. & FOSTER, R.B. 1986. Commonness and rarity in a neotropical forest: implications for tropical
tree conservation. Pp. 205-231 in Soule, M.E. (ed.). Conservation Biology: the science of scarcity and
diversity. Sinauer Associates, Sunderland, MA.
JANZEN, D.H. 1988. Ecological characterisation of a Costa Rican dry forest caterpillar fauna. Biotropica 20: 120-
135.
JANZEN, D. H., HALLWACHS, W., JIMENEZ, J., & GAMEZ, R. 1993. The role of parataxonomists, inventory
managers and taxonomists in Costa Rica's national biological inventory. Pp 223-254 in Reid, V.W. (ed.).
Biodiversity P~ospecting. World Resources Institute, Washington DC.
JANZEN, D. H., & HALLWACHS, W. 1994. All Taxa Biodiversity Inventory ATBI of Terrestrial Ecosystems:
a generic protocol for preparing wildland biodiversity for non-damaging use. Draft report of an NSF Workshop
16-18 April 1993, Philadelphia, Pennsylvania.
JOHNSON, S.P. 1991. The Earth Summit: the United Nations Conference on Environment and Development
(UNCED). International Environmental Law and Policy Series. Graham & TrotmanlMartinus Nijhoff. 532 pp.
KELLY, B. J., WILSON, J. B., & MARK, A. F. 1989. Causes of the species-area relation: a study of islands in
Lake Manapouri, New Zealand. Journal of Ecology 77:1021-1028.
KREBS, C.J. 1989. Ecological Methodology. Harper & Row, New York. 654 pp.
LEWINSOHN, T.M. 1991. Insects in flower heads of Asteraceae in southeast Brazil: a tropical case study in
species richness. Pp.525-559 in Price, P.W., Lewinsohn, T.M., Fernandes, G.W. & Benson, W.W. (eds.).
Plant-animal interactions: evolutionary ecology in tropical and temperate regions. Wiley/Interscience, New York.
212 NICK MAWDSLEY
MACARlHUR, R. H. 1969. Patterns of communities in the tropics. Biological Journal of the Linnean Society 1: 19-
30.
MACARlHUR, R. H. 1972. Patterns in the Distribution of Species. Harper & Row, New York.
MACARlHUR, R. H., & WILSON, E. O. 1967. The theory of island biogeography. Princeton University Press,
Princeton, N.J. 198pp.
MAGURRAN, A. E. 1988. Ecological Diversity and Its Measurement pp. Croom Helm, London.
MARTIN, T. E. 1981. Species-area slopes and coefficients: a caution on their interpretation. American Naturalist
118: 823-837.
MAY, R. M. 1978. The dynamics and diversity of insect faunas. Pp. 105-125 in Mound, L. A. & Waloff, N.
(eds), The diversity of insect faunas. Royal Entomological Society, London.
MAY, R.M. 1988. How many species are there on earth? Science 241:1441-1449.
MAY, R.M. 1990. How many species? Philosophical Transactions of the Royal Society London B 330:293-304.
MAY, R.M. 1992a. Past efforts and future prospects towards understanding how many species there are. pp.71-81
in Solbrig, O.T., van Emden, H.M. & van Oordt, P.G.W.I. (eds). Biodiversity and Global Change. IUBS
Monograph 8.
MAY, R.M. 1992b. Bottoms up for the oceans. Nature 357:278-279.
MIGNOTI, S. A., & MEEDEN, G. 1992. Estimating the total number of distinct species using presence and
absence data. Biometrics 48:863-875.
MILLER, R. I., & WIEGERT, R. G. 1989. Documenting completenes, species-area relations, and the species-
abundance distribution of a regional flora. Ecology 70: 16-22.
MORAN, V.C. & SOUTHWOOD, T.R.E. 1982. The guild composition of arthropod communities in trees.
Journal of Animal Ecology 51:289-306.
NEWBERRY, D. M., CAMPBELL, E. 1. F., LEE, Y. F., RIDSDALE, C. E., & STILL, M. I. 1992. Primary
lowland dipterocarp forest at Danum Valley, Sabah, Malaysia: structure, relative abundance and family
composition. Philosophical Transactions of the Royal Society London B 335,341-356.
NILSSON, I.N. & NILSSON, S.G. 1985. Experimental estimates of census efficiency and pseudoturnover on
islands: error trend and between-observer variation when recording vascular plants. Journal of Ecology 73:65-
70.
PALMER, M.V. 1990. The estimation of species richness by extrapolation. Ecology 71:1195-1198.
PRESSEY, R. L., HUMPHRIES, C. I., MARGULES, C. R., VANE-WRIGHT, R. I., & WILLIAMS, P. H. 1993
Beyond opportunism: key principles for systematic reserve selection. TREE 8: 124-128.
PRESTON, F. W. 1948. The commonness, and rarity, of species. Ecology 29:254-283.
PULLIAM, H.R. 1988. Sources, sinks, and population regulation. American Naturalist 132:652-661.
RABINOWITZ, D., CAIRNS, S. & DILLON, T. 1986. Seven forms of rarity and their frequency in the flora of
the British Isles. pp. 182-205 in Soule, M.E. (ed.). Conservation Biology: the science of scarcity and rarity.
Sinauer Associates, Sunderland, MA.
RICKLEFS, R.E. 1987. Community diversity: the relative roles oflocal and regional processes. Science 235:161-
171.
ROBINSON, G. S., & TUCK, K. R. 1993. Diversity and faunistics of small moths Microlepidoptera in Bornean
rainforest. Ecological Entomology 18:385-393.
SANDERS, H. L. 1968. Marine benthic diversity: A comparative study. American Naturalist 102:243-282.
SCHLUTER, D., & RICKLEFS, R. E. 1993. Species diversity. An introduction to the problem. Pp 1-10 in
Ricklefs, R.E. & Schluter, D. (eds). Species diversity in ecological communities: historical and geographical
perspectives. The University of Chicago Press, Chicago and London.
SHMIDA, A. & WILSON, M.V. 1985. Biological determinants of species diversity. Journal of Biogeography
12:1-20.
SIMBERLOFF, D. 1992. Do species-area curves predict extinction in fragmented forest? pp. 75-89 in Whitmore,
T.C. & Sayer, I.A. (eds). Tropical Deforestation and Species Extinction. Chapman & Hall.
SMIlH, E.P. & VAN BELLE, G. 1984. Nonparametric estimation of species richness. Biometrics 40: 119-lE29.
SOBERON, I. & LLORENTE, 1. 1993. The use of species accumulation functions for the prediction of species
richness. Conservation Biology 7:480-487.
SOLBRIG, O.T. 1991. From genes to ecosystems: a research agenda for biodiversity. Report of a IUBS-SCOPE
UNESCO workshop, Harvard Forest, Petersham, Ma. USA. IUBS.
SOUTHWOOD, T.R.E. 1978. Ecological methods: with particular reference to the study of insect populations.
Second edition. Chapman & Hall.
STEVENS, G.C. 1989. The latitudinal gradient in geographical range: how so many species coexist in the tropics.
American Naturalist 133:240-256.
STORK, N.E. 1987a. Guild structure of arthropods from Bornean rain forest trees. Ecological Entomology 12:69-
80.
STORK, N.E. 1987b. Arthropod faunal similarity of Bornean rain forest trees. Ecological Entomology 12:219-
226.
STORK, N.E. 1988. Insect diversity: facts, fiction and speculation. Biological JOUT7Ul1 of the Linnean Society
35:321-337.
STORK, N.E. 1991. The composition of the arthropod fauna of Bornean lowland rain forest trees. JouT7Ullof
STORK, N.E. 1993. How many species are there? Biodiversity and Conservation 2:215-232.
STORK, N.E. 1994. Inventories of biodiversity: more than a question of mnnbers. pp. 81-100 in Forey, P.,
Humphries, C.I. & Vane-Wright, R.I. (eds). Systematics and Conservation Evaluation. Systematics Association
Special Volume 50. Clarendon Press, Oxford.
THOMAS, C.D. 1990. Fewer species. Nature 347:247.
VANE-WRIGHT, R.I., HUMPHRIES, C.I. & wn..LIAMS, P.H. 1991. What to protect? Systematics and the
agony of choice. Biological Conservation 55:235-254.
WATANABE, H. & RUAYSOONGNERN, S. 1989. Estimation of arboreal arthropod density in a dry evergreen
forest in Northeastern Thailand. JOUT7Ul1 of Tropical Ecology 5: 151-158.
WIENS, I.A. 1989. Spatial scaling in ecology. Functional Ecology 3:385-397.
WILSON, E.O. 1987. The arboreal ant fauna of Peruvian Amazon forests: a first assessment. Biotropica 19:245-
251.
WHITMORE, T.C. 1984. Tropical Rain Forests of the Far East. Oxford University Press.
WHITTAKER, R.H. 1977. Evolution of species diversity in land communities. Evolutionary Biology 10:1-67.
wn..LIAMS, C. B. 1943. Area and number of species. Nature 152:264-267.
wn..LIAMSON, M. 1988. Relationship of species number to area, distance and other variables. pp. 91-155 in
Pimm, S.L. (ed.). Analytical Biogeography.
YOON, C. K. 1993. Counting creatures great and small. Science 260:620-622.
Rainforest dynamics: the need for new paradigms
PETER J.GRUBB
Depanment of Plant Sciences, University of Cambridge, United Kingdom
ABSTRACT. A desire to understand the maintenance of the very high species-richness of tropical
lowland rain forest provides the setting for this paper. It is widely accepted that both niche-differentiation
and chance effects among ecologically very similar species play an important part in maintaining richness.
The purpose of the paper is to provide an improved background of ideas on rainforest dynamics, against
which to consider differentiation in the regeneration niche, and against which to plan management and
conservation. Paradigms deemed to need review are the following: the all-importance of canopy gaps,
the existence of a spectrum of tolerance in which the relative positions of species are not dependent on
size of individual, the association of large seed size with shade tolerance, the idea that soil seed banks
are composed of species demanding sizeable canopy gaps. and the all-importance of shade as opposed to
competition for water or mineral nutrients in limiting regeneration. Recent evidence bearing on these
ideas is reviewed briefly, and an alternative set of paradigms is offered.
KEY WORDS: forest dynamics, regeneration niche, gaps, litter, tolerance, seed mass, dispersal, seed
banks, root competition.
INTRODUCTION
The setting for this paper is a desire to understand the maintenance of the very high
species-richness of tropical lowland rainforest, a topic recently reviewed in the context of
conservation by Primack (1992). It is generally accepted that niche-differentiation,
particularly differentiation in the regeneration niche sensu Grubb (1977), plays a significant
part, but it is also accepted that chance effects among ecologically similar species, as argued
by Hubbell (1979), are also important. The relative importance of niche-differentiation and
chance effects among consistently sparse species is no nearer resolution than it was when
discussed by Grubb (1986) and Hubbell & Foster (1986). A third approach to the problem
has been the hypothesis that intraspecific density-dependent effects will keep all species
sparse and so allow room for many to co-exist (Connell 1971, Janzen 1970). However, in
a critical long-term study in Panama it has been established by Hubbell, Condit & Foster
(1990) that the density-dependent effects in most species at their site are too weak to make
a significant contribution to maintenance of species-richness.
My chief concern in this paper is to try to provide an improved background of ideas on
rainforest dynamics, against which to think about differentiation in the regeneration niche and
against which to plan management and conservation, including re-establishment of forest on
cleared land. The paradigms which have proved valuable in helping us to assemble a mental
picture of the rain forest community as a 'working mechanism' (cf. Watt 1947), but which
now seem inadequate, have been built up by the influential text-books of Richards (1952) and
Whitmore (1975, 1984) and numerous reviews, most notably perhaps those in the volumes
edited by Tomlinson & Zimmermann (1978) and Medina, Mooney & Vazquez-Janes (1984),
and those by Bazzaz & Pickett (1980), Brokaw (1985), Budowski (1970), Denslow (1987)
and van Steenis (1958a, b).
In Table 1 I set out the paradigms to be reviewed. After discussing each of them, I put
forward some tentative replacements (Table 2). Other paradigms are in need of replacement,
215
D. S. Edwards et al. (eds.). Tropical Rain/orest Research - Current Issues, 215-233.
216 PETER 1. GRUBB
Table 1. Paradigms concerrullg rainforest dynamics tbat are in need of revision.
I. Canopy-gaps are readily defined and all-important for regeneration.

2. Species can be arranged along a spectrum of shade tolerance; it is implicit tbat tbe position of anyone
species does not change relative to tbe positions of otber species witb increased size of individual .
3. Light-demanding plants have small seeds while shade-tolerant plants have large seeds; large seeds are
primarily adaptive in tolerating shade.
4. The species represented in tbe soil seed bank are light-demanders; germination is triggered primarily by a
high ratio of red to far-red radiation and/or strong diurnal fluctuation in soil temperature.
5. The light regime is all-important for regeneration; root competition for water and/or mineral nutrients is not
important.
Table 2. Paradigms concerning rainforest dynamics proposed in place of tbose set out in Table I.
I. Gaps in tbe litter layer as well as gaps in tbe canopy are important for regeneration; tbere is a continuum in
conditions from large canopy gaps to microsites in tbe understorey where tbe daily total radiation available
for plant growtb is significantly increased by minute gaps in tbe canopy.
2. The amount of light needed to sustain a given relative growtb rate increases as tbe size of tbe individual
increases, but tbe degree of increase in light demand varies widely; tberefore tbere are many cross-Qvers in
shade-tolerance and it is not sufficient to tbink in terms of a constant spectrum of relative tolerance. The
ecological shade tolerance, unlike tbe physiOlogical shade tolerance, may increase witb size of individual, if
tbe importance of absolute growtb rate for survival declines after tbe earliest stages of establishment; such
transitIOns can generate furtber cross-overs in relative tolerance.
3. There are wide and strongly overlapping ranges of seed size in shade-tolerant and light-demanding plants;
minute seeds are typical of species establishing at litter-free sites in shade or in tbe open on nutrient-poor
substrata; common species requiring canopy gaps for establishment have small ratber tban minute seeds, and
tbeir greater size is of significant value in competition during establishment; very large seeds are not
primarily adaptive in establishment under shade but in attracting large animals as dispersers.
4. The soil seed bank contains species which display a very wide range of shade-tolerance, and in rainforests
not subject regularly to a dry season includes species witb minute seeds suited to establish at litter-free sites
in deep shade.
5. Periodic competition for water from large trees is a major factor for tbe growtb and survival of seedlings in
shade; competition for mineral nutrients is most important for plants in canopy gaps where tbey are growing
faster than III shade; tbere is commonly an interaction between nutrient supply and degree of shade on growtb
rate, and tbe extent of limitation by nutrients varies widely between species.
but I do not have room to consider them in any detail. For example, it is often thought that
there is a strong negative relationship between growth rate and longevity oftrees. However,
Lieberman et al. (1985, 1990) have shown that this is not true in the lowland rain forest
whichthey studied in Costa Rica; many understorey trees display the combination of slow
maximum growth-rate and relatively short life, while many canopy and sub-canopy trees
which are shade-tolerant when young but which respond opportunistically to increased light
display the opposite combination of fast maximum growth-rate and relatively long life.
Received views on the interpretation of 'defence' in leaves of trees with different degrees of
shade-tolerance are also in need of revision, but have been discussed recently by Edwards
(1989) and Grubb (1992), and will not be treated further here. In at least some rainforests
regeneration by sprouting, rather than from seedlings, is much more important than
generalized accounts would suggest.
PARADIGMS NEEDING REVISION
All-importance of canopy gaps

Most of the literature on rain forest dynamics is written as though canopy gaps can be
Rainforest dynamics 217
defined relatively precisely and are all-important for the regeneration of rain forest trees.
In the major reviews it is accepted that the exact size of gap needed varies a great deal with
species, but still the emphasis is on 'gap' versus 'non-gap' conditions (Brokaw 1985;
Denslow 1987). I think this perspective needs to be revised for two reasons. First, there
is a growing appreciation that there is a continuum of conditions in the forest from 'gap' to
'non-gap', and this has important implications for our ideas on the requirements and
responses of juvenile plants. Secondly, evidence is accumulating that gaps in the litter,
acting on a different scale from gaps in the canopy, are all-important for the establishment
of many species.
Popma, Bongers, Martinez-Ramoz & Veneklaas (1988), who used data from the Las
Tuxtlas forest in Mexico, were among the first to demonstrate that canopy gaps do not have
sharp edges in terms of micrometeorological characteristics. Brown (1993) has published
parallel results for the Danum Valley forest in Sabah. The idea of a continuum from fully
open to very deeply shaded micro-sites was emphasized by Lieberman, Lieberman & Peralta
(1989) with their memorable expression 'forests are not just Swiss cheese'. The latter
authors used data from La Selva in Costa Rica for the subcanopy tree Casearia arborea
(listed as a 'pioneer' by Putz & Appanah 1987) to illustrate the way in which growth can
respond continuously to the degree of canopy openness. Others, e.g. Pearcy (1983) who
worked in Hawaii, have also shown how slight differences in the light regime in the deeply
shaded understorey can have a major impact on the growth rates of saplings.
The importance of litter gaps, as opposed to canopy gaps, was shown by Molofsky
&Augspurger (1992) who reported experiments done at Barro Colorado Island in Panama on
the impact of litter cover on germination and establishment of species with seeds of widely
different sizes. For species with small seeds (less than about 5 mg dry mass), absence of
litter was shown to be vital for establishment. Litter-free sites occur not only in canopy gaps
but also in deep shade, for example where standing dead trees or small living trees tip up
without making an appreciable canopy gap, where trunks sliding down slopes in the shade
scrape off the topsoil, where micro-landslips occur, where rocks or large roots protrude,
where animals dig or scrape, and where humans create paths along slopes leaving bare banks
on the uphill sides (Figure 1). In montane forests and very high-rainfall lowland forests the
trunks of tree ferns also provide litter-free sites for establishment.
Kohyama & Grubb (1994) have shown that in a warm temperate rain forest in southern
Japan the mean dry mass values of the seeds of the fourteen most abundant shade-tolerant
trees span a range of four orders of magnitude (from 0.074 mg for Rhododendron tashiroi
to 670 mg for Litsea acuminata) and have argued that variation in litter depth at micro-sites
on the scale of seeds and young seedlings is critically important. Larger-seeded species can
cope with accumulated litter, and make use of deeper softer soil, while very small-seeded
species cannot become established through litter but can become affixed to a hard substratum
and can come to rest on slight irregularities on a steep micro-slope (Figure 2). Although the
need for a gap in the litter layer is the prime requirement for the strongly shade-tolerant
species with minute seeds, it is possible that most individuals become established at
micro-sites which also benefit from tiny canopy gaps - say 1-2 % daylight under cloudy
conditions as opposed to < 1 %; critical observations are needed. The responses of these
plants to sun flecks also require investigation. I consider the incidence of this type of plant
in tropical lowland rain forest later on when discussing the received paradigm for seed size.
Existence of a constant spectrum of tolerance

Trees are conventionally ordered along a linear scale of tolerance from very strongly
light-demanding to very strongly shade-tolerant. In temperate deciduous forests of Europe
and North America tree species are commonly assigned to five categories of tolerance (Baker
1949, Ellenberg 1963). In primaeval forest all but the most light-demanding trees become
218 PETERJ.GRUBB
Figure I. A cross-section through a rainforest showing sites (arrowed) where a litter-free substratum is combined
with deep shade. From left to right: trunk of tree-fern, pit and mound of small tipped-Up standing-dead tree, heap
and pit made by an animal, outcropping rock, major roots above soil surface, micro-landslide, bank on uphill side
of path made by humans, topsoil scraped away by sliding fallen trunk.
l
Figure 2. A comparison of a large-seeded species able to grown down through, or up through, deep litter and
penetrate relatively soft soil with its tap root, and a small-seeded species needing a litter-free micro-site for
establishment and having a diffuse root system suited to anchor the plant and grow over a hard substratum if
necessary.
established in shade so that 'light-demand' generally refers to the phase after germination and
early establishment. The mental picture most widely diffused is that the most
light-demanding species persist only a short time in deep shade (probably much less than one
year on average), but respond to the formation of a sizeable gap above them with rapid
growth, while the most shade-tolerant persist for very long periods in deep shade (probably
for decades) and achieve only moderate growth rates when a sizeable gap is formed above
them. For temperate trees Canham (1989) considered a mechanism which would make this
antithesis between persistence in deep shade and growth rate in bright light inevitable.
However, in a critical study of six species of canopy-forming tree over six years at La Selva
in Costa Rica, Clark & Clark (1992) found no trade-off between the two variables. I suggest
that although there is likely to be an antithesis between extreme types one should not expect
a precise trade-off as the most important physiological property determining shade-tolerance
is likely to be ability to reduce the respiration rate, while the most important in determining
ability to respond to increased light are (a) flexibility in the pattern of allocation between
organs and (b) flexibility in allocation between biochemical fractions of the cells.
In most of the literature on both temperate and tropical forests no reference is made to
tree species changing relative positions along the spectrum of tolerance as their individuals
increase in size. However, there is increasing evidence that the relative positions do change.
One approach to this problem is to emphasize changes in physiological tolerance that
accompany increase in plant size. Givnish (1988) emphasized that the light requirement for
survival or a given relative growth rate (increase in dry mass per unit dry mass per unit time)
of a tree increases as that tree becomes larger, as a result of the increase in the ratio between
the mass of respiring tissues and the mass of photosynthesizing tissues. Dr T. L. Poulson
(personal communication, August 1989) has suggested that we may envisage different species
having different slopes for the allometric relationship between light requirement for
persistence (compensation point) or a given relative growth rate and tree size. In this way
species will cross over in their relative light-requirement when successively larger
size-classes are considered. It seems that there is not a single case of such a cross-over
rigorously proven in the literature. However, in the forest described by Poulson & Platt
(1989) Poulson's unpublished records show that as a seedling Ostrya virginiana is more
light-demanding (for persistence) than Acer saccharum, but that it increases in relative shade
tolerance to the point where - as a 'subcanopy species' - it ends its life in the shade of the
Acer (a 'canopy species'). Dr C. D. Pigott (personal communication, April 1989) has found
the same for Carpinus betulus (subcanopy) in relation to TWa cordata (canopy) in Europe,
and Fowells (1965) reports the same for Lithocarpus densiflora (subcanopy) in relation to
Sequoia sempervirens in California. The Ostrya, Carpinus and Lithocarpus appear to be
examples of the 'late-tolerant' type of Oldeman (1990), or 'cryptic pioneer' of Hawthorne
(1993). It seems very likely that many examples of this kind of cross-over will be found
in tropical rain forests, especially among subcanopy trees and treelets of the understorey.
Oldeman & van Dijk (1991) give examples from three tropical areas: Microdesmis puberula
and Myrianthus liberica in Cote d'Ivoire, certain Inga spp. in French Guyana, and Piper
amalgo in Central America. Ellison, Denslow, Loiselle & Brenes M (1993) found that for
several understorey Melastomataceae in Costa Rica there were significantly more established
juveniles in gaps than in the understorey. Possibly related functionally are such Australian
species as Acronychia vestita, Euodia xanthoxyloides and Melicope broadbentiana reported
by Hyland & Whiffin (1993) to be 'understorey species' but also 'favoured by disturbance'.
Herbs of the understorey may be involved as well as woody species, e.g. some of the clonal
Zingiberaceae such as Afromomum spp. in Africa (W.R. Hawthorne, personal
communication, November 1993).
The key to understanding many of these species may lie in assessing ecological tolerance
rather than physiological tolerance at different stages of development. While we expect
physiological tolerance of shade always to decrease with plant size within a species,
ecological tolerance of shade may show an absolute increase. What may matter most for
survival through the establishment phase is absolute growth rate; in practice only plants
above a certain size may have an appreciable chance of withstanding the hazards of
short-lived drought, herbivory or damage by trampling animals or falling debris. Once the
220 PETER J. GRUBB
plants are established, absolute physiological tolerance of shade may become relatively more
important and enable the larger juveniles and the adults to persist in the understorey.
Another approach which has revealed cross-overs in tolerance has been attention to the
gap size needed at different stages in growth among species that all mature in the canopy.
For example, in temperate deciduous forest in eastern North America Betula aliegheniensis
was found to need gaps of about the size made by single tree-falls to establish but could grow
to maturity in such modest-sized gaps, while Fraxinus americana could establish in the
absence of a canopy gap but needed very large gaps, such as are caused by cyclones, to
reach its full potential during onward growth (Merrens & Peart 1992). Tropical parallels
certainly seem to exist. In northern Queensland an example of a species needing formation
of a canopy gap at least the size of a single treefall for establishment but able to mature in
a such a gap would be Polyscias murrayi, which has a narrow crown and rapid height
growth, while an example of a species able to establish in the absence of a sizeable canopy
gap but needing a large gap to reach full potential would be Flindersia brayeleyana which
has a wide crown at maturity (A.K. Irvine, personal communication, April 1993).
Large seed size as an adaptation to shade

The idea that shade-tolerant species of plant have large seeds and light-demanders small
seeds goes back at least as far as the classic text of Salisbury (1942). In fact, in respect of
plants in northern Europe, he proved the idea only for herbs, and he did not present a
comparable analysis for shrubs or trees. Indeed, in the temperate deciduous forests of
Europe the largest-seeded trees are not the most shade-tolerant (Grubb 1977). Nevertheless
it is widely accepted that in tropical rain forests shade-tolerators have large seeds and
strongly light-demanding species have small seeds (Clark & Clark 1992, Foster 1986, Foster
& Janson 1985, Ng 1978, Swaine & Whitmore 1988). Dissent from this view has been
expressed by Ellison et al. (1993), Kelly & Purvis (1993) and Putz & Appanah (1987).
In this account the mass values quoted refer to the true seed plus fibrous endocarp when
one is present. For species with an aril, wing or pappus on the seed this was removed before
the mass was determined.
The line between 'shade-tolerators' and 'light-demanders' is inevitably arbitrary because
at anyone stage of development there is a continuum of tolerance. In the following account
I treat as 'light demanders' species with the characterstics of 'pioneers' sensu Swaine &
Whitmore (1988), plus the most light-demanding of the non-pioneers, i.e. species which
routinely become established in shade but cannot persist there for long (probably less than
twelve months) and depend on the early creation of a sizeable gap for further growth.
Species considered here as 'shade-tolerators' persist longer (but for very varying times) as
seedlings in deep shade, say < 2 % daylight under cloudy conditions, and achieve a wide
range of growth rates when released in a canopy-gap. Species which we suspect to be
'Iate-tolerants' sensu Oldeman are omitted from the analysis at this stage. In theory, the
shade-tolerators could be subdivided on their degree of shade-requirement. This point leads
to a further problem. It is virtually certain that some species can regenerate under a wider
range of shade than others; for example, in the Malay Peninsula the treelet or shrub Ficus
grossularoides germinates freely in the shade (Raich & Gong 1990), and can be numerous
as shaded juveniles in high forest (D.1. Metcalfe & author, unpublished), but it appears to
establish at open micro-sites as well and can certainly be common in secondary forest (Sim,
Tan & Turner 1992). Apart from the problems of definition, we suffer in most parts of the
tropics from a lack of critical quantitative data on degrees of tolerance. In practice, we are
often dependent on the accumulated experience of those with an extensive knowledge of local
natural history.
It was mentioned above that in a Japanese warm temperate rain forest Kohyama & Grubb
(1994) found that the commonest shade-tolerant species ranged in mean seed dry mass over
four orders of magnitude. In this case shade-tolerance has been proven not only by extensive
quantitative stand analysis but also by obervations on mapped plants over about 15 years. In
the same forest the commonest strongly light-demanding species cover a parallel but less
extensive range (from c. 1 mg for Aralia elata to c. 730 mg for Aleurites cordata; P.J.
Grubb & T. Kohyama, unpublished).
Recent work in Singapore (Metcalfe & Grubb, submitted) has revealed a very wide range
in mean seed dry mass in shade-tolerant trees and shrubs in coastal hill dipterocarp forest.
Species which appear to establish in deep shade on litter-free sites, notably on the steep banks
uphill of the paths through the Bukit Timah Nature Reserve, and which certainly persist and
grow at micro-sites with about 1 % of daylight under cloudy conditions, have values for seed
dry mass as low as 50 mg in the case of the 20-m-tall tree Ptemandra echinata
(Melastomataceae) and 30 mg for the 3-m-tall treelet or shrub Urophyllum hirsutum
(Rubiaceae). Kiew (1988) drew attention to the minute seeds of several genera ·of autotrophic
shade-tolerant herbs in Malaysian rain forests, and Metcalfe & Grubb (submitted) have
obtained values of 5-15 mg for species of Gesneriaceae, Melastomataceae,
Pentaphragmataceae and Rubiaceae of the type to which she referred. Although tiny-seeded
shade-tolerant herbs are rare in temperate deciduous forest, a few such species are known
(Struik 1965).
Are the very wide and overlapping spectra of seed mass values of shade-tolerators and
light-demanders to be found in other areas too? That is certainly what was found by Foster
& Janos (1985) for 202 species of trees, shrubs and vines at Manu in Peru; 'pioneers' had
wet seed mass values ranging from 7 mg to 58.6 g, and 'mature forest' species 1.2 mg to
54.7 g, and although the mean values for the two groups were highly significantly different
they were similar in absolute terms (1.39 and 2.19 g respectively). In the last year my wife
and I have determined mean seed dry mass for numerous species of tropical lowland rain
forest plants in northern Queensland (Figure 3). There is a general trend for the maximum,
median and mean values of mean seed mass to be lower in species which form smaller
individuals at maturity, especially if the strange herbaceous cycad Bowenia spectabilis is
excluded (tall and medium-sized trees > small trees and shrubs > herbs). Within each of
these three growth-forms there is wide overlap between the ranges of seed size for
shade-tolerators and light-demanders. However, there is no shade-tolerant tree with seeds
as small as those of the Ptemandra spp. of the Malay Peninsula, and there are very few
shade-tolerant small trees and shrubs with seeds of < 100 mg, and no herbs in this range.
The most likely explanation is the fact that the Australian forests have a much more strongly
seasonal pattern of rainfall so that plants growing slowly from minute seeds in deep shade
are at risk of desiccation virtually every year. Quite likely the even drier periods during the
Pleistocene (cf. Hopkins et al. 1993, Kershaw 1988) should be seen as equally responsible
for the complexion of the present flora. The thesis that dry spells, on whatever time-scale,
are responsible is supported by a consideration of the habitats of the shade-tolerant species
with minute seeds « 100 mg) known in northern Australia. The shrub Cyrtandra baileyana
of the Gesneriaceae (mean seed dry mass 10 mg) is virtually confined to the edges of streams
and to deeply incised ravines. The one Ptemandra species in Australia, P. coerulescens, is
confined to swampy forest (Hyland & Whiffin 1993); in Singapore and Western Malaysia
(where mean seed dry mass values of 25-30 mg have been recorded, D. J. Metcalfe
unpublished) this species occurs in swampy forest but is not confined to it. One other
species should be mentioned: Saurauia andreana, a small tree in the Actinidiaceae; it
combines minute seeds (57 mg) and ability to flower and fruit in deep shade, but is not
shown in Figure 3 because it is suspected of being a 'late-tolerant'. It is most often found
in the same situations as the Cyrtandra.
Thus the picture of seed size that we now have is quite different from that commonly
canvassed. Where prolonged effectively moist periods occur, shade-tolerant species may
222 PETER 1. GRUBB
Dry mass Trees Small trees Herbs

(mg) ST LD ST LD ST LD.
100,000 4 I ~ I ~ I
..
:'
...
10,000 -: .
II.
:
r l'
f··
·..
-... - .
·.......
1,000
100 -" ·
·..
... ·
S.
10 -.· , -u
·....
· ·· .
•
0.1
0.01
Figure 3. The distributions of mean seed dry mass values for species of trees (tall and medium height at maturity),
small trees and shrubs (commonly maturing at < 9 m tall), and herbs in rain forests of northern Queensland
(unpublished data of P.I. & E.A.A. Grubb): ST, shade-tolerant; LD, light-demanding (see text for definitions of
these two groups).
display approximately the same range of mean seed mass as the light-demanders. The
smaller-seeded species need litter-free sites for establishment, as emphasized above, and such
sites are produced in the shade in a variety of ways summarized in Figure 1. We consider
this issue still further below in connection with the 'seed bank' in the soil.
Turning from shade-tolerant species to light-demanders, it is a striking fact that the
common light-demanding species of secondary forests in the Malay Peninusula and the island
of Borneo, species of such genera as Commersonia, Glochidion, Macaranga, Mallotus and

Trema, undoubtedly small-seeded relative to, say, dipterocarps or laurels of primary forest,
actually have much larger seeds than the tiny-seeded species we have just discussed as being
markedly shade-tolerant during establishment such as species of Pternandra and Urophyllum
(mean dry mass values in the range 2-20 mg vs 20-50 mg). So the question of interest
becomes not 'Why are the common weedy trees small-seeded?' but 'Why are the common
weedy trees as large-seeded as they are?'. In fact, there are indeed tiny-seeded weedy trees
in the Malay Peninsula and Borneo. Several are characteristic of strongly infertile soils:
species of Adinandra, Fagraea, Melastoma and Ploiarium (cf. Sim et at. 1992). They have
mean seed dry mass values in the range 30-1100 mg (Metcalfe & Grubb, submitted). The
only Macaranga species on such soils (M. heynei) has one of the smallest Macaranga seeds
we have studied (3.2 mg vs a range of 3-30 mg). The same trend is found in northern
Queensland where the dominant weedy trees of roadsides on relatively fertile basalt-derived
soils have seeds with dry mass values in the range 5-20 mg (species of Alphitonia, Mallotus,
Omalanthus and Polyscias) while four of the six weedy small trees and shrubs with tiny seeds
shown in Figure 2 (range 49-800 mg) are all characteristic of much less fertile soils on
granite or metamorphic rocks (species of Abrophyllum, Geniostoma, Melastoma,
Rhodomyrtus). In the Blue Mountains of Jamaica the characteristic colonizers of
nutrient-poor stony micro-sites on landslips studied by Dalling (1992) all have very small
seeds (notably species of Clethra, Cyrilla, Lyonia and Vaccinium) , while the colonizers of
micro-sites with more fine earth, and dominants of secondary forest on non-truncated profiles
have larger seeds. I suggest that the seeds of weedy species on more fertile soils are as large
as they because that gives them a necessary advantage in competition during establishment.
We are lacking comparative studies on relative growth rate for these species but for other
plant groups it has been found to be strongly negatively correlated with seed mass;
nevertheless seedlings of larger-seeded species may be larger than those from smaller-seeded
species for prolonged periods after germination (Maranon & Grubb 1993). By hypothesis
competition among colonizers is less critical on soils which supply smaller amounts of
nutrients. I have recently argued that where plants invade soil free of roots the impact of
competition among invaders will be greater on soil that is richer in nutrients, but that where
plants invade soil already fully occupied by roots the impact of root competition on
colonizers will be greater on soil that is inherently nutrient-poor (Grubb 1994). Certainly
the same trend to smaller seeds in plants of open sites on more nutrient-poor substrata is seen
in communities quite different from those of the wet tropical lowlands - notably the annual
grasslands of the Mediterrranean Basin and the commuities of annuals on sand dunes of
northern Europe (Maranon & Grubb 1993). Qualitatively the same argument may apply to
the tiny-seeded species suited to establish under damp conditions in deep shade on litter-free
soil; they have to tolerate strong competition for light from established trees but the impact
of competition for resources from fellow colonizers may be appreciably less than for species
which are suited to establish in high irradiance.
Not all light-demanding tropical rainforest trees with very small seeds are associated with
nutrient-poor soil. Nauclea ojficinalis (dry mass 30 mg) and Anthocephalus chinensis =
Neolamarkia cadamba (wet mass 50 mg, Foster & Janos 1985) are associated in the least
disturbed areas with rivers, and tiny-seededness may perhaps be related to establishment at
very wet micro-sites.
So far my discussion has centred on small and minute seeds. Two points have to be made
now about especially large seeds. First, we have to consider what is the factor to which
particularly large seeds in general are adapted. Secondly, we have to treat the large-seeded
gap-demanding species.
Foster (1986) attempted a comprehensive discussion of the adaptive value of large seeds
for tropical rainforest trees, and considered large size in relation to temperature, moisture,
224 PETER 1. GRUBB
'light intensity', soil factors, competition, predation, pathogens and mycorrhiza. There is no
convincing evidence that very large seeds (say, 10-100 g dry mass) allow trees to establish
at micro-sites which are more deeply shaded or drier than large seeds (say, 1-10 g).
Physiological properties of the shoot and root are more important in that context than the
difference between large and very large seeds. I suggest that the primary adaptive value of
very large seeds is not to be found in connection with any of the physical factors traditionally
considered by plant ecologists but in attracting relatively large animals as dispersers:
mammals generally, but in the special case of northern Queensland and Papua New Guinea
cassowaries - the huge flightless rainforest birds (Casuarius spp.; see Stocker & Irvine 1983).
I suggest that, other things being equal (e.g. nutritive value and taste), larger seeds, or seeds
with a larger mass of flesh around them, will be more attractive to larger dispersers whose
livelihood arguably depends on intake per unit time (Martin 1985) - a larger fruit or seed
may be taken in as quickly as a smaller one. There is a related issue: the provision of an
adequately nutritious food to support the disperser. In the context of dispersal by squirrels
in northern temperate deciduous forest attention has been paid to the relative amounts of
energy supplied by various large seeds among species of Carya, Juglans and Quercus
(Stapanian 1986). Perhaps equally important is the supply of protein.
A distinction should be made between species in which the flesh around the seed is the
attractant to dispersers, and those in which the seed itself is the attractant. In the Lauraceae
of northern Queensland my wife and I have found a negative relationship between the
nitrogen concentration of the embryo (probably proportional to protein concentration) and the
mean dry mass of the embryo (Figure 4, circles). The relatively large seeds (say, mean dry
mass 5-10 g) believed to be dispersed mainly by cassowaries and the musky rat kangaroo
(Hypsiprymnodon moschatus) rather than by flying birds have embyos which are relatively
unnutritious in protein terms, but this is irrelevant to the animals which consume the flesh.
However, in the largest-seeded species, which are eagerly sought by white-tailed rats
:§ 30 • Pc
Cl
E
c
0 • FP
n;... 20 • Al>
...c 0 0 8 0
G)
0
c
0 GIll PT
•
0
0 10 0
0
• •
Z
[]
TEO 0
OCD 0
[]
•
C NP
co
G)
:: 0
.1 1 10 100
Mean dry mass of embryo (9)
Figure 4. The mean concentration of nitrogen in the embryo as a function of mean embryo dry mass in a range of
Lauraceae (circles), and in a subjective selection of other tree species in northern Queensland (squares). Open
symbols for species dispersed only by birds, and filled symbols for species dispersed by mammals as well as birds
or by mammals alone. The birds use only the flesh, while the mammals mostly use the embryo. AD, Athertonia
diversifolia; FP, Fontainea picrosperma; NP, Niemeyera prunifera; PC, Pouteria castanosperma; PT, Prunus
tumeriana; TE, Triunia erythrocarpa. The Lauraceae (in order of increasing mean embryo dry mass) are Neolitsea
dealbata, Cryptocarya grandis, Beilschmiedia brunnea, Litsea leefeana, Cinnamnmum laubatii, Beilschmiedia
recurva, Endiandra bessaphila, Cryptocarya angulata, Endiandra monothyra, Cryptocarya pleurosperma, Endiandra
montana, Beilschmiedia tooram, Endiandra sankeyana, E. xanthocarpa, Cryptocarya oblata, Endiandra insignis,
E. palmerstonii and Beilschmiedia bancrojtii. Unpublished data of P.J. & E.A.A. Grubb and G.D. Jones.
(Uromys caudimac-ulatus) and scattered by them (Hyland & Whiffin 1993, Wellesley-
Whitehouse 1983), the nitrogen concentration in the embryos is higher than would be
expected by extrapolation from the smaller-seeded species (Figure 4, filled circles); clearly
this feature could be regarded as adaptive. Perhaps it is not by chance that nuts in general
are found by humans to provide not only plenty of energy but an effective source of protein!
Three of the most widespread very large-seeded Lauraceae in northern Queensland (mean
seed dry mass 18-43 g) have relatively thick and hard endocarps (c. 28-44% of seed dry
mass) probably broken into only by the white-tailed rat (A.K. Irvine, personal
communication, April 1993) - rather as some of the largest seeds of the Leguminosae in the
neotropics can be broken into only by certain ground-dwelling rodents (Forget 1990,
Hallwachs 1986) and some northern temperate seeds can be broken into only by certain
squirrels, Sciurus spp. (cf. Stapanian 1986). Another of the laurels shown in Figure 4,
Cryptocarya pleurosperma, with a somewhat smaller seed (c. 11 g), has an even thicker
endocarp (78% of dry mass) and is commonly taken by Uromys. It has a nitrogen
concentration on the upper edge of the cloud of points for the Lauraceae in Figure 4.
Several non-Lauraceous species in northern Queensland with smaller seeds (1. 3-11 g) have
very thick endocarps which, again, are probably bitten through only by the white-tailed rat,
e.g. Athertonia diversifolia in the Proteaceae (85% of dry mass in endocarp), Fontainea
picrosperma in the Euphorbiaceae (62%), and Pouteria castanosperma (73%) in the
Sapotaceae. Such species (filled squares in Figure 4) are found to have a nitrogen
concentration in the embryo appreciably higher than expected for their mean embryo dry
mass on the trend line for the Lauraceae, and high relative to other members oftheir families
with seeds for which the flesh is the attractant (Niemeyera prunifera, Sapotaceae; Triunia
erythrocarpa, Proteaceae; unfilled squares in Figure 4). It seems as though the seeds
represent a package of especially nutritious food which only very few dispersers can utilize,
and which will be prized items in the diet of the animal concerned. Certainly very high
proportions of all the seeds of the species mentioned as attractive to the white-tailed rat are
consumed by that species (G.N. Harrington & A.K. Irvine, personal communication, April
1993). The seeds of some species with a fairly thick and hard endocarp, and a relatively
high nitrogen concentration in the embryo, e.g. Prunus turneriana (3 g, 40% in endocarp;
Figure 4) are taken less by the white-tailed rat than by the musky rat kangaroo (A. Dennis,
personal communication, April 1993).
We are still in the earliest stages of understanding the evolution and present-day dispersal
of these large seeds (see also Howe 1986), but I suggest that there is enough evidence to
make us think in terms of large seed size and high protein content as being significantly
adaptive in attracting dispersers.
It often seems to be taken for granted that we should expect strongly light-demanding
species to produce large numbers of smalJ seeds - how else will they produce enough to
ensure their arrival (or presence in the seed bank) when a canopy gap is formed? But in fact
it is quite clearly possible for a plant with a large seed (say mean seed dry mass 730-7800
mg) to be an effective colonizer of canopy-gaps. Aleurites cordata at the site studied by
Grubb (1992) and Kohyama & Grubb (1993) in southern Japan, and Aleurites moluccana in
northern Australia (as well as much of Malesia) are striking examples; they behave
essentially as 'pioneers' in the sense of Swaine & Whitmore (1988). The seeds of A.
moluccana can persist and remain viable a long time in the soil, probably much longer than
two years (Hopkins & Graham 1987). Both these species have large seeds containing large
embryos (embryo and endosperm make up 3.1 g out of total 7.8 g in A. moluccana, which
has a hard thick testa). Other notable light-demanding species have large seeds but smalJ
embryos: various Elaeocarpus species such as E. angustifolius in Australia (97% of dry seed
mass as endocarp), and Spondias species in C. America and S.E. Asia (cf. Ng 1991, Roth
1987). In tropical Queensland the Aleurites and the Elaeocarpus are very much sought after
226 PETER 1. GRUBB
by the white-tailed rat, and appear to be scatter-hoarded by them. In South America the
agouti (Dasyprocta punctata) utilizes and disperses Spondias (Roth 1987). It is not easy to
see why large-seeded gap-demanders have not evolved in more families. It is certainly
striking that most of the largest-seeded trees of the temperate forests of the northern
hemisphere (Carya, Castanea, Corylus, Juglans and Quercus) are strongly light-demanding
(Ellenberg 1963; Fowells 1965), not 'pioneers' sensu Swaine & Whitmore (1988) but
light-demanding 'non-pioneers'. Platt & Hermann (1986) give a valuable account of
dispersal, establishment and onward growth for species of Carya and Quercus at a site in the
south-eastern USA.
To summarize the revised perspective on seed size, very wide and overlapping ranges of
seed mass values are found for shade-tolerant and strongly light-demanding species.
Tiny-seeded shade-tolerant species need reliably moist and litter-free sites for establishment.
Large-seeded strongly light-demanding species are effectively dispersed into actual and
potential gaps by large animals, and may be enabled to persist in the soil seed bank by their
thick hard coats. The smallest-seeded light-demanders are mostly characteristic of
nutrient-poor soils, where (it is hypothesized) competition among colonizers is less critical.
The possession of especially large seeds, among both shade-tolerant and strongly
light-demanding species, is seen as being primarily adaptive in attracting dispersers rather
than in enabling the seedlings to be especially resistant to drought or shade. There is
preliminary evidence that the large seeds are especially valuable to potential dispersers
through having high concentrations of protein as well as providing a large amount of energy
per unit picked up. Soil seed banks are composed of species which need canopy gaps.
The definition of soil seed bank species is difficult because of our lack of precise
information on the longevities of seeds of different species in the soil (V azquez-Yanes &
Orozco-Segovia 1993). A theoretical qualification might be that a significant proportion of
the seeds can remain viable in the soil for more than one year. In practice, one goes on the
density of viable seeds inthe soil relative to the probable current seed rain. Where viable
seeds are numerous in the soil,and fruiting adults are rare or absent in the vicinity,
persistence in the soil appears very probable.
It is commonly believed that in all kinds of forest the soil seed bank is composed of
species which need canopy gaps for establishment, and this idea is linked with the idea that
germination is triggered mainly by changes in the red/far red ratio of the radiation incident
on the soil, and by increased daily amplitude of soil-surface temperature (cf. Vazquez-Yanes
& Orozco-Segovia 1984). I have myself contributed to the perpetuation of this idea (Grubb
1988). However, while working in Queensland, I have found that the most abundant species
in the seed bank may display a very wide range of shade-tolerance. In a forest on a strongly
leached basalt-derived soil I have found that in terms of ecological tolerance the seed bank
species range from those found naturally on edges, e.g. Alphitonia petriei, through those
found commonly in the centres of single-treefall gaps as well as on edges, e.g. Polyscias
murrayi, to those found typically in small gaps or at the edges of gaps and occasionally in
extremely small gaps as well as in the centres of gaps, e.g. Dendrocnide moroides and
Solanum dal/achyi. In terms of physiological tolerance I have found that the first two are
hardly able to grow under a shade screen letting through about 2 % of daylight (Osunkoya
1992 found much the same) while the latter two can grow appreciably under those conditions
and show a marked positive response to the addition of mineral nutrients in such deep shade.
Most interestingly Kennedy (1991) has found that the most abundant species in the soil
seed bank of the Danum Valley forest in Sabah include a number with minute seeds, closely
related to species known from the work of D.l. Metcalfe to be shade-tolerant, e.g.
Ptemandra azurea, species of the scandent shrub Dissochaeta (also Melastomataceae), and
species of Urophyllum, notably U. glabrum. Other species, e.g. Cal/icarpa spp.
(Verbenaceae) and Uncaria spp. (Rubiaceae) seem likely to be much more light-demanding.
The picture is thus of a wide spectrum of shade-tolerance among the soil seed bank species,
as in Queensland, but in this case extending to very shade-tolerant and very tiny-seeded;
Urophyllum glabrum probably has a mean seed dry mass in the range 20-40 mg, judging by
the four species of that genus studied in Singapore). D.l. Metcalfe (unpublished) has
similarly found shade-tolerant tiny-seeded species in the soil seed bank in Singapore. Very
significantly Raich & Gong (1990) showed for two species in this group (Pternandra
coerulescens and Urophyllum glabrum) that they can show high percentage germination under
a forest canopy (27-67%) and more there than in a large clearing.
Much more work is needed to establish the full and true position, but it looks as though
the seed bank species in tropical rain forests without frequent risks of marked desiccation
cover about as wide a range of shade-tolerance as the non-seed-bank species - possibly wider
if virtually all the most light-demanding species are present in the seed bank. Among the
species which combine strong shade-tolerance as juveniles and possession of minute seeds
there is evidence of a wide range in requirements for onward growth, some being able to
flower and fruit in deep shade (e.g. species of Urophyllum) while others flower only when
they have reached a considerable size and have direct sunlight on at least a large part of their
crown (e.g. Pternandra echinata). These species whose start in life is geared to litter-gaps
thus parallel the species with large seeds which germinate and establish in shade without the
need for a litter-gap, and which display the whole spectrum of canopy-gap requirement for
persistence and onward growth.
A lesser revision of the paradigm is needed in respect of those soil seed bank species
which do indeed need a sizeable canopy gap for establishment and onward growth. There is
increasing evidence that bare soil open to unfiltered radiation and subject to large daily
fluctuations in temperature is not necessary for germnation. Physical disturbance of the
upper soil, even where the litter layer is replaced, and even under a forest canopy, can lead
to the germination of numerous seeds (Dalling 1992, Hopkins & Graham 1984, Kennedy &
Swaine 1992). Also Raich & Gong (1990), in their trials with fresh seed in western
Malaysia, showed that very few 'gap species' conformed strictly to the supposed
characteristic of a 'pioneer' proposed by Swaine & Whitmore (1988): no germination under
forest shade and abundant germination in unfiltered daylight. One 'gap species', Trema
tomentosa, did conform, showing negligible germination in repeated trials under a forest
canopy but abundant germination in a large clearing, but others formed a continuum to the
'non-pioneer species', as illustrated by the sequence MeLastoma maLabathricum (3-6%
germination in different trials under a forest canopy), Commersonia bartramia (0-14%) and
Porterandia anisophylla (9-31 %). The shade regime is all-important for regeneration.
Compared with the innumerable references to shade effects on plants in tropical rainforest,
very little has been published on the effects of root competition. In the 1930s, in a number
of papers thoroughly summarized by Richards (1952), C. Coster drew attention to the
potential role of root competition, but there have been few attempts to follow up the idea
with appropriate experiments. There are several significant questions that might be asked
about the impact of shortage of water and/or mineral nutrients, induced by the roots of
established trees.
We start by asking whether species found to establish only in sizeable canopy gaps are
excluded from the understorey to a significant extent by root competition as well as by shade.
In temperate forests this idea has received support from the finding that the gap-demanding
species Galeopsis tetrahit and Rubus idaeus invaded plots freed of root competition by
trenching under Fagus sylvatica and Picea abies respectively (Slavikova 1958; V.G. Karpov
1962, cited by Walter & Breckle 1985). In temperate deciduous forest the mass of fine roots
per unit volume of topsoil can drop appreciably in canopy gaps (Wilczynski & Pickett 1993),
and the same was established for lowland rain forest in Costa Rica by Sanford (1989). In
the same forest the water content of the topsoil is greater in canopy gaps (Vitousek &
228 PETER 1. GRUBB
Denslow 1986). Also in a Panamanan moist forest with a pronounced dry season the water
potential of the topsoil was found not to drop as much in gaps as in the understorey (Becker,
Rabenold, Idol & Smith 1988). It seems that the only experimental study showing
desiccation rather than lack of light to be the apparent cause of exclusion from the
understorey is that of Mulkey, Smith & Wright (1991) for the giant herb Pleiostachya
pruinosa (Marantaceae) in the Panamanian forest studied by Becker et al. (1988); survival
of plants in the understorey over 3.5 years was markedly and significantly greater in an
irrigated plot than in a control plot.
The reduction in rooting density in canopy-gaps may lead to significantly greater
availability of nutrients for colonizing plants but that will depend on the relationship between
the mean concentration of a nutrient ion in the soil solution and rooting density. Tilman &
Wedin (1991) have illustrated this relationship for a number of temperate grasses; the mean
nitrate concentration in the soil solution fell very markedly with a modest rise in root mass
per unit volume and then only a little more over a wide range of increasing root mass values.
If the availability of nutrients is increased significantly in gaps, then this is likely to have a
marked positive effect on growth rate and survival of seedlings of gap-demanders. As
explained below, there is commonly a marked positive interaction between irradiance and the
supply of limiting nutrients on growth, and at low irradiance many species are not responsive
to nutrients. Such species of small gaps as Dendrocnide moroides, mentioned above as
responsive to nutrients in about 2 % daylight, seem likely to be kept in gaps by root
competition as well as by shade.
Of special interest is the preferential occurrence of certain 'pioneers' within canopy gaps
on the mounds of soil on tipped-up root systems (Putz 1983); it has been suggested that the
effective supply of limiting mineral nutrients may be greater in root-free subsoil than in
topsoil densely filled with roots (Grubb 1992) and the idea has been justified in a laboratory
study of simulated grassland gaps (Reed & Grubb submitted) but a critical test is needed for
rainforest tip-ups.
A further set of questions concerns the impact of root competition for species which are
markedly shade-tolerant at the stage of establishment. First, does it have any significant
effect on the growth or survival of seedlings and saplings in deep shade? Secondly, does it
affect rates of onward growth in sizeable canopy gaps? Thirdly, do different species of tree
display different combinations of sensivity to shade, and sensitivity to nutrient supply or
water supply, such that different species will be favoured by different combinations of
conditions in different parts of a single gap? The last question represents a special case of
the more general question asked by Orians (1982).
A distinction should be made between the effects of water shortage and those of nutrient
shortage. Very few studies have been made on the combined effects of different degrees of
shade and different degrees of water shortage, but Papavasilliou (1991) found for four
European shrub species no interaction; the proportional deleterious effect of drought on
growth was the same at different irradiances. The same has been found by Burslem (in
press) for some tropical rain forest trees. In contrast, addition of nutrients to plants in pots
or solution culture commonly fails to increase the growth rate in deep shade, even where it
has a marked positive effect in bright light (references summarized by Peace & Grubb 1982;
see also Latham 1992). Turner (1991) found no interaction between irradiance and
nutrient-addition effects on the growth of potted plants of the 'pioneers' Melastoma
malabathricum and Trema tomentosa, grown in the open and in the shade of Eugenia
grandis; an interaction might have been found at lower irradiances. Turner, Brown &
Newton (1993) found no increase in growth of potted Shorea macroptera in the shade of the
same Eugenia, but Sundralingam (1983) did find an increase in growth of Dryobalanops spp.
fertilized in relatively bright light. Failure to respond to nutrient addition in a pot experiment
does not mean that a species would not have its growth inhibited when deprived of nutrients
in the same soil by roots of established plants. However, on the basis of the majority of the
evidence one might expect competition for water from the roots of tall trees always to be
liable to cause a reduction in growth and survival of seedlings, while for most tree species
competition for nutrients might have significant effects only in canopy gaps - the size of gap
needed for an effect depending on the species. I have written 'liable to cause a reduction'
because what actually happens will depend on the sensitivity of seedlings to water status.
Burslem (in press) has found a striking ability of seedlings of a number of tree species in
Singapore to maintain growth rates even though watered infrequently in deep shade - a result
which could be interpreted to show that there has been selection for an ablity to resist
periodic desiccation even in plants of 'aseasonal' rain forest.
What has been found out in experiments on root-trenching and on addition of water and/or
nutrients to plants in different degrees of shade in rainforests? Connell (1970) reported no
effect of root-trenching on survival for spontaneously established seedlings of one species of
strongly shade-tolerant Sapotaceae in two degrees of shade (under a mother tree and away
from it) in northern Queensland, but Fox (1973) reported slight increases in growth rate of
naturally developed dipterocarp seedlings in Sabah. In both cases there were only
pseudo-replicates in the sense of Hurlbert (1984). More recently Denslow, Newell & Ellison
(1991) found no effect of root-trenching on either survival or growth of seedlings of two
shade-tolerant Inga spp. sown into the understorey at La Selva in Costa Rica. Working in
the strongly seasonal moist forest of Panama, Fisher, Howe & Wright (1991) found that
irrigation in the dry season tended to increase survival of first-year seedlings of the nutmeg
Virola surinamensis in the understorey (19/21 vs 16124), but the difference was not
statistically significant. Turner, Brown & Newton (1993) found no effect of adding NPK to
dipterocarps in the understorey in western Malaysia and Sabah. More provocatively
Denslow, Schultz, Vitousek & Strain (1990) found no effect of adding a complete fertilizer
to cuttings of various shrubs (species of Miconia and Piper) in gaps in a Costa Rican
rainforest. Denslow et at. did not establish the primary limiting nutrient for the test plants
in soil densely occupied by roots, and it might well have been nitrogen (cf. Burslem in
press); it is notable that there was no increase in nitrogen concentration in the test plants or
in the soil - perhaps because all the added nitrogen was taken up by competitors.
Clearly there is a need for further, more extensive and more elaborate experiments
covering a range of species, soils and climatic regimes. In the meantime we are left with
the suggestive observational studies of Gentry & Emmons (1987) which showed that there
are markedly fewer species of herb and shrub per unit area in lowland rain forests with a
marked seasonality in rainfall or marked nutrient-deficiency in the soil. This result would
be explained most economically by an increased impact of root competition in a given degree
of shade at sites where water and/or nutrients are in short supply. The experimental work
on the effects of seasonal drought on Pleiostachya in Panama, mentioned above, supports the
explanation, but we are lacking experimental data on the sensitivity of herbs to the relevant
range of nutrient-availabilities.
CONCLUDING REMARKS
In this review I have summarized the evidence that there is a need to revise at least five
commonly quoted paradigms regarding rainforest dynamics, and I have set out suggested
replacement paradigms (Table 2). Critical studies are needed to test their generality.
There is general agreement between the points made in this review and those made in
certain papers which had not reached me when I was writing my lecture for the symposium,
most notably those by Clark & Clark (1992), Ellison et at. (1993) and Kelly & Purvis
(1993). I thoroughly endorse the plea by Clark & Clark (1992) that attention be directed to
230 PETERJ.GRUBB
studies which embrace all life history stages, and that comparisons be based on operational
criteria. I also support the view of Ellison et af. (1993) that in the recent past insufficient
effort has been put into studies of life-forms other than trees that have the potential to reach
the top of the canopy. We are on the verge of obtaining a much better balanced
understanding of rainforest dynamics, which will enable us to interpret more soundly
evolution of life histories, the significance of morphological and physiological differences
between species, and the maintenance of species-richness, while also providing a firmer basis
for the planning of conservation and replanting schemes.
ACKNOWLEDGEMENTS
This paper was written at the Tropical Forest Research Centre of the CSIRO Division of
Wildlife and Ecology at Atherton in northern Queensland. I thank the staff there for making
facilities freely available to me and for many kindnesses; Tony Irvine, in particular,
introduced me to the local rainforests and collected a number of important seeds. I am
indebted to Ian Turner for unstinted hospitality in Singapore in recent years, and for valuable
comments on early drafts of this paper, to David Burslem for important suggestions, and also
to Daniel Metcalfe for permission to quote unpublished results. The Royal Society of
London provided a grant to cover the expenses of the work done by my wife and myself in
Australia.
LITERATURE CITED
BAKER, F.S. 1949. A revised tolerance table. JOUI7U11 of Forestry 47:179-181.
BAZZAZ, F.A. & PICKETT, S.T.A. 1980. Physiological ecology of tropical succession: a comparative review.
Annual Review of Ecology and Systematics 11 :287-310.
BECKER, P., RABENOLD, P.E., IDOL, J.R. & SMITH, A.P. 1988. Water potential gradients for gaps and
slopes in a Panamanian tropical moist forest's dry season. JouI7U11 of Tropical Ecology 4:173-184.
BROKAW, N. V.L. 1985. Treefalls, regrowth, and community structure in tropical forests. Pp. 53-69 in Pickett,
S. T.A. & White, P.S. (eds). The ecology ofnatural disturbance and patch dynamics. Academic Press, Orlando.
BROWN, N,D. 1993. The implications of climate and gap microclimate for seedling growth conditions in a
Bornean lowland rain forest. JouI7U11 o/Tropical Ecology 9:153-168.
BUDOWSKI, G. 1970. The distinction between old secondary and climax species in tropical central American
lowland forests. Tropical Ecology 11 :44-48.
BURSLEM, D.F.R.P., in press. Differential responses to nutrients, shade and drought among tree seedlings of
lowland tropical rain forest in Singapore. Proceedings of the Workshop on Tropical Tree Seedling Ecology.
University of Aberdeen, September 1992.
CANHAM, C.D. 1989. Different responses to gaps among shade-tolerant tree species. Ecology 70:548-550.
CLARK, D.A. & CLARK, D.B. 1992. Life history diversity of canopy and emergent trees in a tropical rain forest.
Ecological Monographs 62:315-344.
CONNELL, J.H. 1971. On the role of natural enemies in preventing competitive exclusion in some marine animals
and in rain forest trees. pp. 298-312 in den Boer, P.I. & Gradwell, G.R. (eds). Dynamics of populations. Centre
for Agricultural Publishing and Documentation, Wagenginen.
DALLING, J.W. 1992. Regeneration on landslides in the Blue Mountains. Jamaica. PhD dissertation, University
of Cambridge.
DENSLOW, J.S. 1987. Tropical rainfrest gaps and tree species diversity. Annual Review of Ecology and
Systematics 18:431-451.
DENSLOW, I.S., NEWELL, E. & ELLISON, A.M. 1991. The effect of understorey palms and cyclanths on the
growth and survival of lnga seedlings. Biotropica 23: 225-234.
DENSLOW, J.S., SCHULTZ, J.C., VITOUSEK, P.M. & STRAIN, B.R. 1990. Growth responses of tropical
shrubs to treefall gap environments. Ecology 71:165-179.
EDWARDS, P.J. 1989. Insect herbivory and plant defence theory. Pp. 275-297 in Grubb P.I. & Whittaker, J.B.
(eds). Toward a TlWre exact ecology. Blackwell Scientific Publications, Oxford.
ELLENBERG, H. 1963. Die Vegetation Mitteleuropas mit den Alpen. Ulmer, Stuttgart.
ELLISON, A.M., DENSLOW, J.S., LOISELLE, B.A. & BRENES M., D. 1993. Seed and seedling ecology of
neotropical Melastomataceae. Ecology 74: 1733-1749.
FISHER, B.L., HOWE, H.F. & WRIGHT, S.J. 1991. Survival and growth of Virola surinnmensis yearlings: water
augmentation in gap and understory. Oecologia 86:292-297.
FORGET, P.·M. 1990. Seed·dispersal of Vouacapoua americann (Caesalpiniaceae) by cavimorph rodents in French
Guiana. Journnl of Tropical Ecology 6:459-468.
FOSTER, S.A. 1986. On the adaptive value of large seeds for tropical moist forest trees: review and synthesis.
Botanical Review 52:260-299.
FOSTER, S.A. & JANSON, C.H. 1985. The relationship between seed size and establishment conditions in tropical
woody plants. Ecology 66:773-780.
FOX, 1.E.D. 1973. Dipterocarp seedling behaviour in Sabah. Malaysian Forester 36:205-214.
FOWELLS, H.A. 1965. Sylvics of forest trees of the United States. Agriculture Handbook 271, US Department
of Agriculture, Washington.
GENTRY, A.H. & EMMONS, L.H. 1987. Geographical variation in fertility, phenology, and composition of the
understory of neotropical forests. Biotropica 19:216-227.
GIVNISH, T.1. 1988. Adaptation to sun and shade: a whole·plant perspective. Pp. 63-91 in 1.R. Evans, S. von
Caemmerer & W.W. Adams (eds). Ecology of photosynthesis in Sun and Shade. CSIRO, Melbourne.
GRUBB, P.L 1977. The maintenance of species richness in plant communities: the importance of the regeneration
niche. Biological Reviews 52:107-145.
GRUBB, P.1. 1986. Problems posed by sparse and patchly distributed species·rich plant communities. Pp.207-225
in Diamond, J. & Case, T.L (eds). Community ecology. Harper & Row, New York.
GRUBB, P.J. 1988. The uncoupling of disturbance and recruitment, two kinds of seed bank, and persistence of
plant populations at the regional and local scales. Annales Zoologici Fennici 25:23-36.
GRUBB, P.J. 1992. A positive distrust in simplicity· lessons from plant defences and from competition among
plants and among animals. Journal of Ecology 80:585-610.
GRUBB, P.J., in press. A paradox of root competition: a possible resolution and implications. JP'hywcoomlogia.
HALLWACHS, W. 1986. Agoutis (Dasyprocta punctata) , the inheritors of guapinol (Hymenaea courbaril:
Leguminosae). Pp. 285-304 in Estrada, A. & Fleming, T.H. (eds). Frugivores and seed dispersal. Junk,
Dordrecht.
HAWTHORNE, W.D. 1993. Forest regeneration after logging in Bia South C.P.R., Ghana. Overseas Development
Agency Forestry Series No.4. N.R.I., Chatham.
HOPKINS, M.S., ASH, 1., GRAHAM, A.W., HEAD, J. & HEWETT, R.K. 1993. Charcoal evidence of the
spatial extent of the Eucalyptus woodland expansions and rainforest contractions in North Queensland during the
late Pleistocene. Journal of Biogeography 20.
HOPKINS, M.S. & GRAHAM, A.W. 1984. The role of soil seed banks in regeneration in canopy gaps in
Australian tropical lowland rain forest· preliminary field experiments. Malaysian Forester 47:146-158.
HOPKINS, M.S. & GRAHAM, A.W. 1987. The viability of seeds of rainforest species after experimental soil
burials under tropical wet lowland forest in north·east Australia. Australian Journal of Ecology 12:97-108.
HOWE, H.F. 1986. Seed dispersal by fruit.eating birds and mammals. Pp. 123-189 in Murray, D.R. (ed.). Seed
di~persal. Academic Press, Sydney.
HUBBELL, S.P. 1979. Tree dispersion, abundance and diversity in a tropical dry forest. Science 203:1299-1309.
HUBBELL, S.P., CONDIT, R. & FOSTER, R.B. 1990. Presence and absence of density·dependence in a
neotropical tree community. Philosophical Transactions of the Royal Society of London B330:269-281.
HUBBELL, S.P. & FOSTER. R.B. 1986. Biology, chance and history in the structure of tropical rain forest tree
communities. Pp. 314-329 in Diamond, 1. & Case. T.1. (eds). Community ecology. Harper & Row, New York.
HURLBERT, S.H. 1984. Pseudoreplication and the design of ecological field experiments. Ecological Monographs
54: 187-211.
HYLAND, B. & WHIFFIN, T.R. 1993. Key to the tropical forest trees of Australia. CSIRO, Melbourne.
JANZEN, D.H. '1970. Herbivores and the number of tree species in tropical forests. American Naturalist
104:501-528.
KELLY, C.K. & PURVIS, A. 1993. Seed size and establishment conditions in tropical trees. On the use of
taxonomic relatedness in determining ecological patterns. Oecologia 94:356-360.
KENNEDY, D.N. 1991. The role of colonizing species in the regeneration of dipterocarp rainforest. Ph.D. thesis,
University of Aberdeen.
KENNEDY, D.N. & SWAINE, M.D. 1992. Germination and growth of colonizing species in artificial gaps of
different sizes in dipterocarp rain forest. Philosophical Transactions of the Royal SOCiety B335:357-367.
KERSHAW, A.P. 1988. Australasia. Pp. 127-306 in Huntley, B. & Webb, T. (eds). Vegetation history. Kluwer,
Dordrecht.
KIEW, R. 1988. Herbaceous flowering plants. Pp. 56-76 in Cranbrook, The Earl of (ed.). Key environments:
Malaysia. Pergamon, Oxford.
232 PETER 1. GRUBB
KOHY AMA. T. & GRUBB, PJ., in press, 1994. Below- and above-ground allometries of shade-tolerant seedlings
in a Japanese warm-temperate rain forest. Functional Ecology 8.
LATHAM, R.E. 1992. Co-occuring tree species change rank in seedling performance with resources varied
experimentally. Ecology 73:2129-2144.
LIEBERMAN, D., LIEBERMAN, M., HARTSHORN, G. & PERALTA, R. 1985. Growth rates and age-size
relationships of lowland tropical wet forest trees in Costa Rica. Journal of Tropical Ecology 1:97-109.
LIEBERMAN, D., HARTSHORN, G.S., LIEBERMAN, M. & PERALTA, R. 1990. Forest dynamics at La Selva
Biological Station 1969-85. Pp. 509-521 in Gentry, A.H. (eds). Four neotropical rainforests. Yale University
Press, New Haven.
LIEBERMAN, M., LIEBERMAN, D. & PERALTA, R. 1989. Forests are not just Swiss cheese: canopy
stereogeometry of non-gaps in tropical forests. Ecology 70:550-552
MARTIN, T.E. 1985. Resource selection by tropical frugivorous birds: integrating mUltiple functions. Oecologia
66:563-573.
MEDINA, E., MOONEY, H.A. & VAZQUEZ-JANES, C. (eds) 1984. Physiological ecology of plants of the wet
tropics. Junk, The Hague.
MERRENS, E.J. & PEART, D.R. 1992. Effects of hurricane damage on individual growth and stand structure in
hardwood forest in New Hampshire, USA. Journal of Ecology 80:787-795.
METCALFE, DJ. & GRUBB, P.J. Seed weights and light requirements for regeneration in south-east Asian rain
forest. Submitted to New Phytologist, February 1994.
MOLOFSKY, J. & AUGSPURGER, C.K. 1992. The effect of litter on early seedling establishment in a tropical
forest. Ecology 73: 68-77.
MULKEY, S.S., SMITH, A.P. & WRIGHT, SJ. 1991. Comparative life history and physiology of two understorey
tropical herbs. Oecologia 88:263-273.
NG, F.S.P. 1978. Strategies of establishment in Malayan forest trees. Pp. 129-162 in Tomlinson, P.B. &
Zimmermann, M.H. (eds). Tropical trees as living systems. Cambridge University Press.
NG, F.S.P. 1991. Manual of forest fruits, seeds and seedlings, Volume 1. Forest Research Institute, Kepong,
Malaysia.
OLDEMAN, R.A.A. 1990. Forests: elements of silvology. Springer, Heidelberg,
OLDEMAN, R.A.A. & DDK, 1. VAN 1991. Diagnosis of the temperament of tropical rain forest trees. pp.21-65
in Gomez-Pompa, A., Whitmore, T.C. & Hadley, M. (eds). Rain forest regeneration and management.
UNESCO, Paris.
ORlANS, G.H. 1982. The influence of tree-falls in tropical forests on tree species richness. Tropical Ecology
23:255-279.
OSUNKOY A, 0.0. 1992. Ecology Of seed survival and seedling establishment in Australian tropical forest. Ph.D.
thesis, Australian National University, Canberra.
PAPAVASILLIOU, S. 1991. Ecophysiological studies on chalkland shrubs. M.Phil. thesis. University of
Cambridge.
PEACE, W.J.H. & GRUBB, P.l. 1982. Interaction ofiight and mineral nutrient supply in the growth of Impatiens
parvif/ora. New Phytologist 90:127-150.
PEARCY, R. W. 1983. The light environment and growth of C, and C. tree species in the understorey of Hawaiian
forest. Oecologia 58: 19-25.
PLATT, W.J. & HERMANN, S.M. 1986. Relationships between dispersal syndrome and characteristics of
populations of trees in a subtropical forest. Pp. 309-321 in Estrada, A. & Fleming, T .H. (eds). Frugivores and
seed dispersal. Junk, Dordrecht.
POPMA, J., BONGERS, F., MARTINEZ-RAMOS, M. & VENENKLAAS, E. 1988. Pioneer species distribution
in treefall gaps in neotropical rain forest; a gap definition and its consequences. Journal of Tropical Ecology
4:77-88.
POULSON, T.L. & PLATT, W.J. 1989. Gap light regimes influence canopy tree diversity. Ecology 70:553-555.
PRIMACK, R.B.' 1992. Tropical community dynamics and conservation biology. BioScience 42: 818-821.
PUTZ, F.E. 1983. Treefall pits and mounds, buried seeds, and the importance of soil disturbance to pioneer trees
on Barro Colorado Island, Panama. Ecology 64:1069-1074.
PUTZ, F.E. & APPANAH, S. 1987. Buried seeds, newly dispersed seeds, and the dynamics of a lowland forest
in Malaysia. Biotropica 19:326-333.
RAICH, J.W. & GONG, W.K. 1990. Effects of canopy openings on tree seed germination in a Malaysian
dipterocarp forest. Journal of Tropical Ecology 6:203-217.
REED, R.A. & GRUBB, PJ. Submitted. Nutrient supply available to species colonizing gaps in chalk grassland.
Submitted to Journal of Vegetation Science, March 1994.
RICHARDS, P.W. 1952. The tropical rainforest. Cambridge University Press.
ROTH, I. 1987. Stratification of a tropical forest as seen in dispersal types. Junk, Dordrecht.
SALISBURY, E.J. 1942. The reproductive capacity of plants. Bell, London.
SANFORD, R.L. 1989. Fine root biomass under a tropical forest light gap opening in Costa Rica. Journal of
SIM, I.W.S., TAN, H.T.W. & TURNER, I.M. 1992. Adimmdra belukar: an anthropogenic heath forest in
Singapore. Vegetatio 102: 125-137.
SLAVIKOV A, J. 1958. Einfluss der Buche (Fagus silvatica L.) als Edifikator auf die Entwicklung der Krautschicht
in Buchenphytozonosen. Preslia 30: 19-42.
STAPANIAN, M.A. 1986. Seed dispersal by birds and squirrels in the deciduous forests of the United States.
Pp.225-236 in Estrada, A. & Fleming, T.H. (eds). Frugivores and fruit dispersal. Junk, Dordrecht.
STEENIS, C.G.G.J. van 1958a. Basic principles of rain forest sociology. pp.159-162inProceedingsoftheKandy
symposium on the study of tropical vegetation. UNESCO, Paris.
STEENIS, C.G.G.I. van 1958b. Rejuvenation as a factor for judging the status of vegetation types: the biologcal
nomad theory. pp. 212-218 in same.
STOCKER, G.C. & IRVINE, A.K. 1983. Seed dispersal by cassowaries (Casuarius casuarius) in north
Queensland's rainforest. Biotropica 15:170-176.
STRUIK, G.I. 1965. Growth patterns of some native annual and perennial herbs in southern Wisconsin. Ecology
46:401-420.
SUNDRALINGHAM, P. 1983. Response of potted seedlings of Dryobalanops aromatica and Dryobalanops
oblongifolia to commercial fertilizers. Malaysian Forester 46:86-92.
SWAINE, M.D. & WHITMORE, T.C. 1988. On the definition of ecological species groups in tropical rain forests.
Vegetatio 75:81-86.
TILMAN, D. & WEEDIN, D. 1991. Plant traits and resource reduction for five grasses growing on a nitrogen
gradient. Ecology 72:685-700.
TOMLINSON, P.B. & ZIMMERMANN, M.H. (eds) 1978. Tropical trees as living systems. Cambridge University
Press.
TURNER, I.M. 1991. The effects of shade and fertilizer addition on the seedlings of two tropical woody pioneer
species. Tropical Ecology 32:24-29.
TURNER, I.M., BROWN, N.D. & NEWTON, A.C. 1993. The effect of fertilizer application on dipterocarp
seedling growth and mycorrhizal infection. Forest Ecology and Management 57:329-337.
VAZQUEZ-YANES, C. & OROZCO-SEGOVIA, A. 1984. Ecophysiology of seed germination in the tropical
humid forest of the world. pp. 37-50 in Medina, E., Mooney, H.A. & Vazquez-Yanes, C. (eds). Physiological
ecology of plants in the wet tropics. Junk, The Hague.
VAZQUEZ-YANES, C. & OROZCO-SEGOVIA, A. 1993. Patterns of seed longevity and germination in the
tropical rainforest. Annual Reviews of Ecology and Systematics 24:69-87.
VITOUSEK, P.M. & DENSLOW, J.S. 1986. Nitrogen and phosphorus availability in treefall gaps of lowland
tropical rainforest. Journal of Ecology 74:1167-1178.
WALTER. H. & BRECKLE, S.-W. 1985. Ecological systems of the biosphere. I. Ecological principles in global
perspective. Springer, Berlin.
WATT, A.S. 1947. Pattern and process in the plant community. Journal of Ecology 35:1-22.
WELLESLEY-WHITEHOUSE, H. 1983. White-tailed rat. Pp. 371-373 in Strahan, R. (ed.). Complete book of
Australian mammals. Angus & Robertson, Sydney.
WHITMORE, T.C. 1975. Tropical rainforests of the Far East. Clarendon Press, Oxford.
WHITMORE, T.C. 1984. Tropical rainforests of the Far East, 2nd edn. Clarendon Press, Oxford.
WILCZYNSKI, C.J. & PICKETT, S.T.A. 1993. Fine root biomass within experimental canopy gaps: evidence for
a below-ground gap. Journal of Vegetation Science 4:571-57.
The interaction of rainfall, tree diameter and sulphuric-acid treatments
on resin yield of Palosapis (Anisoptera thurl/era (Blanco) Blume ssp.
thurl/era)
ARSENIO B. ELLA
Forest Products Research and Development Institute, Department of Science and Technology
College, Laguna 4031, Philippines
ABSTRACT. Thirty-two trees of palo sap is (Anisoptera thuri/era (Blanco) Blwne ssp. thuri/era), naturally
growing in the logging concession of the Eastern Plywood Corporation in Llavac, Real, Quezon, Luzon
Island, Philippines were tapped (boxed) for resin or oil yield. Four diameter at breast height (dbh)
classes: 30-39, 40-49, 50-59 and 60 cm and over were selected. Eight trees were designated for each
of four levels of aqueous sulphuric acid treatment at 0 %, 40 %, 50 % and 60 %. Resin yield per tree was
determined and the influence of rainfall, diameter classes and sulphuric acid concentration was studied.
Months with higher rainfall in most cases gave greater resin yield. The effect of different diameter
classes on resin yield was not statistically significant except in the case of 60 cm and over which gave
the highest yield. The four levels of acid concentration had no significant effect on the resin yield.
KEY WORDS: resin yield, Anisoptera, Dipterocarpaceae.
INTRODUCTION
Tapping or boxing of dipterocarp species for resin provides a livelihood for people in
parts of the Philippines. In Quezon province for instance, many eke out their living by
tapping dipterocarp species while waiting for the harvest of their farm crop. Some of those
whose farm lands are already depleted of soil nutrients become full-time tappers. Boxing of
this species thus provides a means of generating employment.
One of the most common resin producing dipterocarp species in the Philippines is
Anisoptera thurifera (Blanco) Blume ssp. thurifera locally known as palosapis. The resin or
'oil of palosap is' is very similar to 'balau' from apitong (Dipterocarpus grandiflorus Blanco).
It is obtained in the same manner and is used for the same purposes, e.g., in the manufacture
of varnish which makes a brilliant, tough and durable coating (Brown 1921). Clover (1906),
as cited by Brown (1921), said that the oil from palosapis resembles that from apitong, but
dries much faster. In addition, the 'oil of palosapis' is light colored, apparently homogenous
in composition and so viscous that it can scarcely be poured.
The traditional method of obtaining the resin is by boxing and refreshing by burning at
frequent intervals to expose the cut (West & Brown 1921). This method is the same as that
used in the production of dammar (Tongacan 1973). Either by tapping or by accidental
injury (Tanchico et al. 1934), resin is exuded from the resin ducts or gum canals found in
the sapwood. This practice however, is still crude and often results in the entry of wood-
destroying organisms, exposure to wind and rain and/or subsequent death of the tree.
Reports of studies conducted by Parameswaran & Zamuco (1976) based on the
miscroscopy of the bark of four species of the so-called 'Philippine mahogany' and by
Whitmore (1962) on the bark morphology of Malaysian Dipterocarpaceae make no mention
of the presence of resin ducts in the bark. Metcalfe & Chalk (1957), however, reported that
in the young stems of dipterocarps, resin canals are always present in the pith, especially in
235
236 ARSENIO B. ELLA
Figure 1. Initial cut is made but limited only up to the sapwood-heartwood boundary. The initial height of the box
is about 25 mm and is V-shaped, in order to facilitate resin flow. The width of boxing should be 113 the girth of
the tree.
Figure 2. Sulphuric acid solution is applied to the freshly tapped palosapis tree using a one-pint squeeze sprayer.
the perimedullary region, and that they also occur in the phloem and cortex.
Few studies of the boxing of dipterocarp species have been conducted. Tofigacan and
Ordinario (1976) studied the boxing of apitong for balau production and the results revealed
that monthly rainfall was found to have an inverse influence on monthly balau yield. Within
the limits of sulphuric-acid concentrations used in the study, acid treatments were found to
have no effect on balau yield. Nor were there significant effects reported for either
interactions between rainfall and diameter, or rainfall and acid concentrations on resin yield.
The objective of this study was to determine the effects of rainfall, tree diameter and
sulfuric acid treatments in resin yield of boxed palosapis trees.
Resin yields of Paiosapis (Anisoptera thurifera) 237
Figure 3. Photograph showing the correct way to securely attach a polyethylene sheet before sealing with plastic
roofing cement. This prevents seepage of water into the tacked plastic receptacle below.
MATERIALS AND METHODS
The tapping experiment

The tapping or boxing experiment was conducted in the logging concession of Eastern
Plywood Corporation in Barangay Llavac, Real, Quezon, Luzon Island, Philippines for a
duration of 12 months. Thirty-two healthy palosapis trees belonging to four diameter classes
i.e. 30-39,40-49,50-59 and ~60 cm (8 trees each) were included in the study. Four levels
of sulphuric acid concentration vis.: 0%, 40%, 50% and 60% (by volume) were applied
using 2 trees each. The experimental design of this boxing experiment is a 4 x 4 factorial
in completely randomized design with only two replications owing to the scarcity of suitable,
vigorous, healthy trees.
The tapping procedure

Palosapis bark was first cleaned or scraped off, using a sharp bolo or wood chisel. The
depth of tapping conformed to the thickness of sapwood as determined by incremental boring;
the initial height of the box was about 25 mm and its form V-shaped to facilitate resin flow;
the width was 1/3 of the girth of the tree. Subsequent rechippings proceeded vertically
upward at a width of about 20 mm to make sure that a minimum amount of the tissues
destroyed by acid in the previous tapping was removed.
Sulphuric acid solution was immediately applied evenly in a fine mist after cutting the
streaks, using a US standard plastic squeezer of one pint capacity. Plastic bags were used
as receptacles. Each tree boxed was properly labelled. The collection of resin was conducted
weekly. Rechipping was done after each collection and sulphuric acid solution was applied.
The amount of resin gathered was weighed and recorded.
The effect of variation in rainfall on resin yield in relation to different diameter classes
and sulfuric-acid treatment was determined. Monthly rainfall data was obtained from the
Philippine Atmospheric Geophysical and Astronomical Services Station (PAGASA), in
Llavac, Real, Quezon, Luzon Island. The experimental site falls under the second climatic
type, having no dry season but with a very pronounced maximum rain period from
November to January.
238 ARSENIO B. ELLA
Figure 4. Photograph showing a tapped palosapis tree (40-49 cm dbh) all set for next resin collection, rechipping,
and sulfuric acid application.
Table 1. Annual resin yield from different diameter classes. r values correlation coefficients; p values =
probability. 'total yield of 2 trees; "total yield of 8 trees.
Diameter class (cm) Annual resin yield (kg/yr)' Total Yield Average yield
Acid concentration (% v. v) (kg/yr) (kg/tree/yr)
0 40 50 60
30-39 7.51 5.49 8.22 8.74 29.96 3.74
40-49 5.05 9.77 6.75 4.92 26.49 3.31
50-59 7.61 7.19 4.26 7.92 26.98 3.37
60 & up. 10.59 14.68 13.31 8.69 47.27 5.91
r values 0.59 0.74 0.30 0.14 0.58 0.58
p values 0.4098 0.2697 0.7005 0.8561 0.4227 0.4218
Total yield (kg/yr)" 30.76 37.13 32.54 30.27
Average yield (kgltree/yr) 3.845 4.64125 4.0675 3.78375
Effect of diameter of the tree on resin yield

The results are shown in Table 1. Statistical analysis revealed that diameter of the tree
has no significant effect on the resin yield (Table 2). This lack of significance, however, is
Resin yields of Paiosapis (Anisoptera thurifera) 239
Table 2. Analysis of variance of total resin yield. ns = not significant at 5 % level.
SOURCE OF VARIATION d.f SS MS F

Treatment (1) 15 63.224
Diameter 3 35.779 11.933 2.9027ns
Acid concentration 3 3.157 1.052 0.2558ns
Diameter x Acid concentration 9 24.276 2.697 0.6560ns
Residual 16 65.783 4.111
TOTAL 31 129.007
Table 3. Monthly resin yield per tree. r values = correlation coefficient; p values = probability.
Monthly rainfall in mm Monthly resin yield (kg) Total monthly resin yield
(2nd climatic type) (kg)
Acid concentration (% v/v)
0 40 50 60
1978
February 161.4 1.68 2.23 2.8 1.19 7.9
March 77.4 2.66 2.97 2.67 2.1 10.4
April 282.7 2.37 2.19 2.15 2.09 8.8
May 249.2 2.14 2.91 1.93 1.92 8.9
June 323.6 1.68 2.64 2.05 2.74 9.1
July 218.8 1.89 2.85 2.31 2.37 9.42
August 563.9 1.64 2.98 2.39 1.79 8.8
September 366.5 1.97 2.83 1.52 2.29 8.61
October 1397.4 4.14 4.87 2.94 2.86 14.81
November 637.8 2.66 4.04 3.61 3.29 13.6
December 687.4 2.85 3.2 2.97 3.61 12.63
1979
January 148.4 2.65 3.83 4.16 3.75 14.4
r values 0.72 0.74 0.16 0.33 0.58
P values 0.0082 0.0060 0.6247 0.2917 0.0490
probably the result of low population (n) as comparison among the 4 diameter classes showed
trees with diameter of 60 cm and up do have higher resin yield per tree than the 3 other
diameter classes (Table 1). This might be due to the fact that larger trees have larger bark
surface area and wider tapping cut exposure that have direct positive effect on the resin
production. Table 1 also shows that the annual resin yield per tree is correlated to tree
diameter (r = 0.58; p = 0.4218). A similar degree of correlation was shown at 0% acid
treatment (r = 0.59; P = 0.4098) and slightly higher at 40% acid treatment (r = 0.74; p =
0.2697). However, at higher acid treatments (50% and 60%) poor correlation was observed
r = 0.3; p = 0.7005 and r = 0.14; P = 0.8561, respectively). This is consistent with the
yield decrease at 50% acid treatment in 40-49 cm, 50-59 cm diameter classes and at 60%
acid treatment in the 60 cm and greater diameter class. It can be seen that trees with larger
diameter need higher acid concentration to lower their yield.
Effect of rainfall on resin yield

The monthly resin yield is highly correlated with monthly rainfall at 0% and 40% v/v acid
240 ARSENIO B. ELLA
treatment with r values of 0.72 and 0.74 respectively (Table 3). However, such correlation
was not observed at higher acid concentration treatments (50% and 60%) in which lower r
values were obtained (0.16 and 0.33 respectively).
Effect of sulphuric acid on resin yield

The four levels of acid concentration had no significant effect on the resin yield (Table
2). Hence, there is no need to compare the means of the four levels of acid studied.
Although not statistically significant (Table 2) acid concentration did appear to interact with
dbh (i.e. only at low acid concentrations was there a high r of resin yield on dbh, Table 1).
CONCLUSIONS
Monthly rainfall was found to have a direct influence on monthly resin yield.
Specifically, months with higher rainfall in most cases gave more resin yield. Acid
concentration had no significant effect on the resin yield.
Further experimental work using similar variables and tapping techniques but involving
a larger number of replicates should be done in a region of the country with pronounced wet
and dry seasons and also in a region with no dry season to serve as a basis of comparison
of resin yields, and the influence of climatic conditions.
LITERATURE CITED
BROWN, W.H. 1921. Philippine resins. Minor products of Philippine forests. Vol. 2:52-64.
CLOVER, A.M. 1906. Philippines wood oils. Philippine JOUT1Ul1 of Science 1: 181-202.
KRAMER, F.J. & KOZLOWSKI, T.T. 1960. Physiology of trees. McGraw-Hill, New York.
METCALFE, C .R. & CHALK, C. 1957. Anatomy of the dicotyledons. Clarendon Press, Oxford. Vol. 1:212-220.
PARAMESWARAN, N. & ZAMUCO, G.I., JR. 1976. Microscopy of barks of the so- called "Philippine
Mahogany· species. Unpublished report, FPRDI, College, Laguna.
TANCHICO, S.S., WEST, A.P. & FORTANOZA, J. 1934. Philippine Panau (Dipterocarp) resin. Philippine
JOUT1Ul1 of Science 54:75-81.
TONGACAN, A.L. 1973. FPRDI Technical Note No. 136. FPRDI, College, Laguna.
TONGACAN, A.L. & ORDINARIO, F.F. 1976. The interaction of tree diameter rainfall and SUlfuric acid on
"Balau" yield of apitong (Dipterocarpus grandiflorus Blume). Unpublished report, FPRDI, College, Laguna.
WEST, A.P. & BROWN, W.H. 1921. Philippine resins, gums, seed oils and essential oils. Minor Forest Products
of Philippine Forests. Volume 2:5-225.
WHITMORE, T.C. 1962a. Studies in systematic bark morphology. I. Bark morphology in Dipterocarpaceae.
The New Phytologist 62:191-207.
Preliminary studies of altitudinal zonation and nitrogen supply on Bukit
Belalong, Brunei Darussalam
C.A. PENDRY and J. PROCTOR
Department of Biological and Molecular Sciences, University of Stirling, Stirling FK9 4LA,
United Kingdom
ABSTRACT. Forest structure is briefly described from plots at 200 m, 500 m and 850 m on Bukit
Belalong (913 m). Evergreen lowland rain forest is found at 200 m and 500 m with tall lower montane
rain forest at 850 m. Automatic weather stations at the base and the summit suggest that the summit
receives 20% more rainfall and mean temperatures there are 3.9·C lower. A cloud cap does not usually
form and mean daily totals of solar radiation are similar at the base and summit. Total smalllitterfall was
measured for one year and the following values (t ha-' yr-') were recorded: 10.56 (200 m), 10-48 (500
m), and 8.26 (850 m). Leaf litterfall values (t ha-' yr-') were 7.88 (200 m), 7.89 (500 m) and 6.02 (850
m). Nitrogen concentrations (mg g-') in totallitterfall were 14.5 (200 m), 12.7 (500 m), 12.7 (850 m)
and in leaf litterfall 11.4 (200 m), 10.6 (500 m) and 11.3 (850 m). Totallitterfall and leaf litterfall were
significantly lower at the highest altitude (p <0.05), but differences in nitrogen concentrations were not
significant. The following soil nitrification rates (kg ha-' 21 d-') were measured: 4.65 (200 m), 2.60 (500
m) and 3.72 (850 m), with no significant differences between altitudes. A bioassay using rice (Oryza
sativa L.) was carried out near the base and summit of Bukit Belalong using soils from both altitudes.
There was a dramatic response to altitude, a lesser one to soil type and only a slight response to nitrogen
fertilisation. It is concluded that nitrogen is not more limiting in the montane forest compared with the
lowland forest and that forest differences with altitude are most likely to be related to changes in
temperature.
KEY WORDS: Litterfall. nitrogen, nitrification, bioassay, montane tropical rain forest.
INTRODUCTION
Floristic and physiognomic changes with altitude are commonly observed features of tropical
mountains (Whitmore 1984), but the causes are still unclear. They include: lower
temperatures (Grubb 1977), high winds (Sugden 1986), low soil pH (Grubb & Tanner 1976),
nutrient limitation due to reduced decomposition and mineralisation rates (Edwards & Grubb
1977, Heaney & Proctor 1989), low radiation totals (Grubb 1977), periodic water shortage
(Whitmore 1984) and soil toxicity (Bruijnzeel et al. 1993). These factors may operate singly
or in combination and the reasons for the replacement of lowland forest by montane forest
certainly vary from site to site. Attention has recently focussed on the stunting of upper
montane forest (Bruijnzeel et ai. 1993, Veneklaas 1992) which has been associated with the
presence of a cloud cap (Grubb 1977, Whitmore 1984). The cloud causes reduced solar
radiation and higher soil moisture which reduces the nitrogen supply. Nitrogen has been
considered to be the nutrient most likely to limit montane forests (Grubb 1989, Heaney &
Proctor 1989), and in this paper we examine this hypothesis for Bukit Belalong.
This paper examines the forests of Bukit (Mount) Belalong which has relatively tall lower
montane forest at its summit and develops a cloud cap infrequently. We try to explain the
limited altitudinal changes in the forests on Bukit Belalong and contrast it with other
mountains where there is stunted forest and we present some of our data on climate and
nitrogen supply. We report litterfall nitrogen concentrations and soil nitrification at three
altitudes and part of a bioassay experiment using hill rice (Oryza sativa L.). The studies on
Bukit Belalong are still continuing and further data about the site will be reported elsewhere.
241
242 C.A. PENDRY AND J. PROCTOR
STUDY SITE
Bukit Belalong (913 m) is in the Temburong District of Brunei, north western Borneo (Figure
1). The mountain forms part of the Batu Apoi Forest Reserve (4 30' north, 115 10' east),
0 0
and has never been logged. The nearest settlement, Kampong Batang Duri, is 8 km
downstream from the Field Studies Centre (FSC) on the Temburong river, and though
shifting cultivation is practised along the river it appears that no land has been cleared
upstream from Kuala Belalong. Human impact on the forest is limited to hunting and the
infrequent gathering of forest products.
Evergreen lowland rain forest with emergent trees up to 60 m tall is found up to c. 750
m. Above this altitude the forest is lower montane rain forest (sensu Whitmore 1984) in
which the trees are smaller but reach heights of 33 m. Dipterocarps are present at all
altitudes.
Climate
Two automatic weather stations (A WS) were used: one at the summit of Bukit Belalong,
from 7 May 1991 to 24 April 1992; and another on a small cleared hill (60 m) near Kuala
Belalong, from 18 June 1991 to 23 April 1992. Rainfall was also measured in a standard
rain gauge at the FSC from 1 February 1991 to 31 May 1992. The rain gauge was sited
within 3 m of the nearest building, and recorded 6 % more rain than the properly sited AWS
at Kuala Belalong. The mean annual rainfall (1973 to 1992) at Semabat Agricultural Station,
20 km to the north of Kuala Belalong was 4086 mm. There is no dry season, though
February and March and June to October tend to be the drier months. Since the Semabat
values are similar to those at Kuala Belalong for the period of measurement we estimate the
mean annual rainfall at Kuala Belalong to be c. 4100 mm. Whilst the AWS was at the
summit rainfall there was 21 % higher than at Semabat so the mean annual rainfall is
estimated at c. 5000 mm.
Figure 1. The locations of study plots, automatic weather stations (AWS) and the Field Studies Centre, Batu Apoi
Forest Reserve, Brunei Darussalam.
Altitudinal zonation and nitrogen supply 243
The highest daily rainfalls were for 24 September 1991 when 181.5 mm rain fell at Bukit
Belalong and 133 mm at Kuala Belalong. The longest dry periods were 7 days at Kuala
Belalong and 12 days at Bukit Belalong. During February and March 1992 Semabat received
251. 8 mm of rain which is 45 % of the mean rainfall for those months measured between
1973 and 1992. This is the second lowest total recorded for these months (237.2 mm fell
in the 'EI Nino' year 1983) . The rainfall totals for February and March 1992 were 224.0
mm at Kuala Belalong and 327.0 mm at Bukit Belalong. Rainfall data for the entire period
of the study are summarised in Figure 2.
Mean daily temperature (July 1991-March 1992) was 25. 7°e at Kuala Belalong and 21.8°e
at Bukit Belalong . The lapse rate is therefore estimated at 0.44°e per 100 m, which is
typical for a dry adiabatic tropical lapse rate (Bruijnzeel et al. 1993). The ranges of mean
monthly maxima are 4.5°e at Kuala Belalong and 5.5°e at Bukit Belalong, whilst the
corresponding ranges of mean monthly minima are 2.4°e and 3.2°C. The mean daily solar
radiation for this period was 15.63 MJ m-2 at Kuala Belalong and 15.71 MJ m-2 at Bukit
Belalong, but it should be noted that there is greater variability at the summit. Temperature
data are summarised in Figure 3 and solar radiation data in Figure 4.
Geology
The geology of Brunei is dominated by older Tertiary sandstones and shales. Narrow
ridges with steep sides and deep valleys are characteristic of the Belalong area and are a
consequence of Miocene folding and lifting of the Setap shales. Large remnants of sandstone
have been reported from further south in Temburong (Ashton 1964), but on Bukit Belalong
these are small and localised, so the mountain is essentially continuous shale from bottom
to top. Erosional processes are exceeded by rates of tectonic uplift so slope failures are
frequent and are the principal determinant of the area's geomorphology (A. Dykes, personal
communication) .
800
700
600
500
400
300
200
100
o
M J J A SON D J F M A M
Month (from May 199 1 )
Figure 2. Monthly rainfall at Kuala Belalong and Bukit Belalong, May 1991 to May 1992. The unshaded portion
of the column represents rainfall at Kuala Belalong and the shaded portion the extra rainfall at Bukit Belalong (in
February 1992 more rain fell at Kuala Belalong than Bukit Belalong). The values for May and June 1991 and April
and May 1991 are estimated from a rain gauge at Kuala Belalong Field Studies Centre. * represents the mean
monthly rainfall (1973 -1992) at Semabat Agricultural Station.
244 C.A. PENDRY AND 1. PROCTOR
Solar
radiation
Tempera ture °c (MJ m·2)
40 20
30 ·15
20 10
10
o
J A o N D J F M J A SON D J F M
Mon th (from July 1991) Month (from July 1991)
Figure 3. The maximum (closed symbols) and Figure 4. Mean daily totals of solar radiation (MJ
minimum (open symbols) monthly temperatures at m-2) at Kuala Belalong, 60 m (_) and Bukit
Kuala Belalong, 60 m (_) and Bukit Belalong, Belalong, 913 m ( .. ) between July 1991 and
913 m ( .. ) between July 1991 and March 1992. March 1992.
Soils from low altitudes have been classified as orthic acrisols (FAO/UNESCO, S.M.
Ross, personal communication). The soils lack distinct horizons, are acid (mean pHH20 was
3.9 at 200 m and 3.8 at 880 m) base-poor and highly weathered. Soil pits were dug beside
each plot and although the soils from higher altitudes had more organic matter in the upper
horizon the profiles were similar to those at lower altitudes.
METHODS
Plot selection
Three 50 m x 50 m plots were established in mature forest at each of three altitudes: plots
1-3 at c. 200 m, 4-6 at c. 500 m and 7-9 at c. 850 m. Some details of the plots are given
in Table 1. Random plot-location was not possible because of the difficult terrain. None of
the plots had atypical features and they are treated as statistically valid independent samples.
The plots were divided into twenty five 10 m x 10 m subplots which were used as a sample
grid.
Forest description
All trees (~ 10 cm dbh) were enumerated and tagged at 140 cm above the ground. Girths
were measured 10 cm below the tag except where buttresses, prop roots or damage distorted
the trunk, and on these trees girth was measured at 30 cm above the protrusion. A second
nail then marked 10 cm above the point of measurement. Trees with multiple stems were
tagged on the main trunk, and the points of measurement on the other trunks (~ 10 cm dbh
Table \. Mean altitude, aspect and slope of nine 50 m x 50 m plots on Bukit Belalong.
Plot number Altitude Aspect Slope
200 210 23
2 200 240 25
3 210 280 26
4 480 80 16
5 520 130 21
6 510 150 30
7 830 40 33
8 860 130 35
9 880 30 30
only) marked with nails but not tagged. The position of each tree was measured to the
nearest 50 cm.
A 60 m x 7.5 m strip of representative mature forest was selected for a profile diagram
from plots 1, 5 and 9. The heights of all the trees in these plots were measured and the size
of each crown was estimated by measuring the height to the emergence of the lowest branch
and the furthest horizontal projections of the crown in the plane of the diagram.
Small litterfall
Small litterfall was collected from 8 May 1991 to 8 May 1992 in plots 1-6 and 7 May
1991 to 7 May 1992 in plots 7-9. The litter traps were baskets of perforated metal sheeting,
with a horizontal area of 40 cm x 40 cm, mounted on poles 1.2 m above the ground. There
were thirteen litter traps placed in a restricted random design in each plot.
It was difficult to collect litterfall at regular intervals because of the remoteness of the
sites, but collections were usually at 14-20 day intervals. Two collections (9-10 December
1991 and 5-6 April 1992) were after 12 days and two other collections were after 36 days
(26-27 June 1991) and 39 days (17-18 January 1992).
The litter traps were emptied into cotton bags and the litter was air dried within two days
of collection. Alternate collections (eleven out of nineteen) were sorted into four fractions,
following the recommendations of Proctor (1983): leaves (including stipules), small wood
(twigs with diameter < 2 cm, and bark fragments smaller than 2 cm along their longest
dimension), reproductive parts and trash (any plant fragments with a longest dimension
smaller than 2 mm and faeces, insect bodies and resin). The sorted samples were weighed
and combined to give bulked samples of the litterfall fractions for each collection from each
plot. The bulked samples were ground and a 20 g sub-sample dried at 105°C before
chemical analysis. When the total mass of the bulked sample was less than 20 g the entire
sample was oven dried. Unsorted collections were oven dried at 105°C and weighed.
0.4 g of each subsample was digested in 4.4 ml of a sulphuric acid/hydrogen peroxide
mixture (Moore & Chapman 1986). Samples were digested at 100°C for 20 min, followed
by 2 h at 370°C. Nitrogen was determined as ammonium on a Tecator FIAstar 5000 flow
injection analysis system.
It was not practical to analyse all the samples in duplicate so the precision of the method
was tested by analysing five replicates of eight samples. The results were always within 10%
of the mean and usually within 5 %. Analysis of reference material supplied by E. V.J.
Tanner gave results within 3 % of those found by investigators in other laboratories.
Nitrification rates
In April 1992 one random position was located in each of ten of the undisturbed sub-plots
(not used for litterfall collection) in each of the nine plots. At each position the litter layer
was removed and a 5-cm deep soil sample collected. The sample was thoroughly sorted to
remove stones, roots and large fragments of organic matter and crushed by hand till no
particles were bigger than 10 mm. The soil was mixed thoroughly and a c. 2oo-g sub-sample
placed in a plastic bag and replaced at the collection site and covered with litter. The top
of the bag was folded down to allow gaseous exchange, but to prevent the entry of water.
The soil was left to incubate for 21 d before collection and extraction. A second sub-sample
was returned to the FSC for immediate extraction and for determination of the moisture
content (after oven drying at 105°C).
20 g soil (wet weight) was extracted in 100 ml 1.5 M potassium chloride containing 3 mg
I-I mercuric chloride to inhibit bacterial activity. The samples were shaken and left to
equilibrate for 4 h before filtration and bottling. They were then frozen until sent to the UK
where they were stored in the dark at 2°C until analysis. Soil nitrate/nitrite nitrogen was
determined as nitrite after reduction by a copperised cadmium column on the FIA system.
In each of the sub-plots used for the collection of soil samples a second random point was
located for the collection of a soil bulk density sample. At each point the litter layer was
removed and a 10 cm x 10 cm open-ended stainless steel box was inserted to a depth of 5
cm and the soil excavated from within it. The samples were oven dried at 105°C, weighed
and then the stones (> 5 mm) and roots were sorted and weighed separately.
Bioassay
The species used in the bioassay was a variety of hill rice (Oryza sativa L.) obtained from
local farmers. Local tree species might have been preferable to rice for the test plant but this
was not practical and in any case species from a nutrient-poor environment may not
necessarily be responsive to fertilisation (Grubb 1989).
A total of about 6 kg of soil was collected to a depth of 5 cm from six random positions
located around each of plots 3 (bottom soil - BS) and 9 (summit soil - SS). The soil was
sorted to remove stones, roots and larger fragments of organic matter (> 1 cm along any
dimension), crushed by hand to give aggregates no larger than 10 mm, and thoroughly
mixed. The soils were stored in sealed plastic bags for up to 5 days before the experiment
was set up.
The two sites for the experiment were the summit of Bukit Belalong and an open space
at the FSC. At both sites the plants were grown on a table with a screen of green plastic
netting to exclude flying insects and with metal sheeting on the legs to exclude rats and mice.
It is accepted that the netting will have reduced the photosynthetically active radiation
available to the plants. At each site forty-eight 14 cm diameter pots were each filled with
500 cm 3 of BS and a further forty-eight pots filled with a similar quantity of SS and the pots
placed on saucers. Seven pregerminated rice seedlings were placed in each pot and after
emergence the plants were thinned to five per pot. Nitrogen was added to six replicates of
each soil as 100 mg sodium nitrate in 10 ml of water (from the nearest stream) and the six
control pots were watered with 10 ml of stream water only. The fertilisation rate was
adapted from Healey (1989) but reduced by 50 % because of the shorter growing period and
was equivalent to 38 kg nitrogen ha- I .
The nitrogen treatment was part of a larger experiment consisting of eight nutrient
treatments and the pots were laid out in six blocks containing one replicate of each treatment.
The pots were randomised within each block and rerandomised after 7 days at the summit
site and after 8 days and 11 days at the FSC. The plants were watered as required, using
water from the nearest stream. After 14 days the plants were harvested, the soil was washed
from their roots and they were oven dried and the shoots and roots were weighed separately.
Table 2. Basal area, mean tree (;:: 10 cm dbh) height, maximum tree height, tree density and percentage of trees
in the Dipterocarpaceae and Myrtaceae in plots at three altitudes on Bukit Belalong.
Plot
5 9
Altitude (m) 200 520 880

Basal area (m2 ha") 58.8 43.5 32.0
Mean tree height (m) 22.9 22.2 17.8
Maximum tree height (m) 60 45 33
Tree density (ha") 470 610 670
Dipterocarpaceae (%) 16.0 14.5 9.7
Myrtaceae (%) 1.6 4.4 16.1
Statistical analyses
The data were analysed by ANOV A on the Minitab statistical package. Normality was
tested by the Shapiro-Wilk test and non-normal data were log transformed.
RESULTS
Forest description
Table 2 summarises data from plots 1, 5 and 9. The maximum tree height declines with
altitude and is reduced from 60 m in plot 1 to 33 m in plot 9, though the mean tree height
declines only slightly because of the preponderance of understorey trees. The number of
trees per hectare increases by almost half between plots 1 and 9. The profile diagrams
(Figures 5-7) give an overall impression of the forest structure at the three altitudes and
show that emergents are an important feature of the lower plots, with the canopy becoming
more even at the higher altitudes. The proportion of Myrtaceae increases and that of
Dipterocarpaceae decreases with altitude.
On the basis of the reduction in stature and frequency of tall (> 2 m) buttresses (10-11 %
trees at low altitudes, 5% of trees at highest altitude), increase in bryophytes and
disappearance of large emergents, the forest at 850 m can be considered Lower Montane
Rain Forest according to Whitmore's physiognomic classification (1984), whilst the forests
at 200 m and 500 m are Evergreen Lowland Rain Forest.
Litterjall
The values for small litterfall declined from a mean of 10.56 t ha- I yr- I in the low plots to
a mean of 8.26 t ha- I yrl in the top plots (Table 3). All fractions except flowers and fruits
decreased with altitude and ANOV A shows that the reductions in total litterfall and leaf
litterfall in the top plots were significant (p < 0.05). There was a marked peak in leaf
litterfall production in April-May 1992 (Figure 8) which followed the low rainfall during
January-March. The peak was highest in the mid plots and least in the top plots.
Nitrogen concentrations varied in no obvious pattern during the year and the results are
summarised as annual means in Table 4. An ANOV A showed no significant differences in
nitrogen concentrations between altitudes for leaf and small wood litter, but significantly
lower concentrations were found in the reproductive and trash fractions in the mid and top
plots compared with the low plots (p < 0.05). The mean concentration values in Table 4
were multiplied by the values for annual litterfall to give estimates of the annual rate of
addition of nitrogen to the forest floor via each litter fraction at each altitude (Table 5). The
inputs followed a similar temporal pattern to those of small litterfall mass.
60 m
Figure 5. Profile diagram (60 m x 7.5 m) of Plot 1 (200 m), Bukit Belalong. Trees under 6 m excluded.
30 m
o
Table 3. Estimated litterfall (t ha·' yr·') from thirty-nine 40 cm x 40 cm litter traps at each of
three altitudes on Bukit Belalong. Values are means ± 95 % confidence limits.
Litter fraction Altitude

200 m 500 m 850m
Leaves 7.88 ±0.44 7.89 ±0.46 6.02 ±0.41
Wood « 2cm diameter) 1.76 ±0.57 1.60 ±0.40 1.32 ±0.24
Flowers & fruits 0.47 ±0.12 0.53 ±0.14 0.53 ±0.22
Trash 0.44 ±0.06 0.46 ±0.08 0.40 ±O.IO
Total 10.55 ±0.80 10.48 +0.69 8.27 ±0.63
Table 4. Mean nitrogen concentrations (mg g.') in small litterfall from three altitudes on Bukit
Belalong. Values are the means of eleven samples ±95% confidence limits.

200 m 500m 850 m
Leaves 11.4 ±0.41 10.6 ±0.21 11.3 ±0.28
Wood « 2cm diameter) 7.9 ±0.50 7.2 ±0.40 7.6 ±0.41
Flowers & fruits 17.8 ± 1.68 16.0 ± 1.21 13.7 ±1.17
Trash 20.8 ±0.82 17.2 ±1.03 18.1 ±0.69
Total 14.5 12.7 12.7
Litterf_ll
V' .-. d- 1
(a)
M J J .. 0 N b .:I F M A M
(b)
Figure 8. The rate of leaf litterfa11 (g m·2 d·') from
May 1991 to May 1992 at three altitudes on Bukit
Belalong. (a) 200 m, (b) 500 m, (c) 850 m. Leaves
were sorted from eleven collections out of a total of
nineteen and the columns represent the means of the
sorted collections and the error bars are the 95 %
confidence limits.
M J J .. 0 N D J F M .. M
(e)
.'
o N D J P M A M
1991 1992
Table 5. Estimated rate of addition of nitrogen (kg ha-' yr') to the forest floor from smalliitterfall
fractions for three altitudes on Bukit Belalong. Values are the means for three plots at each
altitude with the ranges in parentheses.

200m 500m 850m
Leaves 90.0 83.1 67.8

(81.4-95.5) (73.5-94.9) (63.2-72.3)
Wood « 2cm diameter) 14.4 11.1 10.1
(12.3-17.2) (9.4-12.3) (8.0-12.0)
Flowers & fruits 8.5 8.4 7.7
(7.1-9.5) (6.2-12.2) (5.4-10.1)
Trash 9.2 7.8 7.3
(6.8-10.6) (6.5-9.5) (5.6-9.4)
Total 122.1 110.4 92.9
(114.2-127.8) (102.6-122.2) (90.4-95.2)
The quantity of nitrogen in the litterfall declined with altitude for all fractions, and the total
annual input in the top plots was 76% of the annual total in the bottom plots. The
differences in nitrogen input between the top plots and the low plots were significant for total
(ANOVA p<0.05) and leaf (p<0.01) litterfall but differences between the top and middle
plots were not significant.
Nitrification rates
The soils are more organic in the middle and top plots and hence bulk density lower than
in the low plots (Table 6). The difference in bulk density (defined as mass of soil in unit
volume, excluding stones and roots) between soils at these two altitudes can be attributed to
the greater stoniness of the top plots.
Pigs and termites caused considerable damage to the samples left to incubate in the field
particularly in the top plots where only nine out of thirty bags were undisturbed. Table 7
summarises the results from the soil analyses. The initial concentrations of extractable nitrate
showed no significant differences between altitudes (one way ANOVA) and the results from
the top plots showed the greatest variability. Nitrification rates were between 11.6 p.g N g-l
21 d- 1 in the mid plots and 16.6 p.g N g-l 21 d- 1 in the top plots but the differences were not
significant (one way ANOVA). When the rates were expressed on an area basis the lowest
rate was still found in the middle plots but because of the higher bulk density of the soils in
the low plots the rate was highest there.
Bioassay
Climatic data from the two weather stations for the period of the bioassay are summarised
in Figure 9. Maximum and minimum daily temperatures (Figure 9a) were higher at 60 m
Table 6. Bulk density (g cm-3) and concentration of stones (>5mm) (g 100 cm-3 soil) in ten surface soil samples
(10 cm x 10 cm area and 5 cm deep) at three altitudes on Bukit Belalong. Values are the means ± 95 % confidence
limits of thirty samples.
Altitude Bulk density Stones

(m) (g cm-') (g 100 cm-' soil)
200 0.619 ±0.06 8.62 ±4.96
500 0.448 ±0.06 1.28 ± 1.13
850 0.398 +0.05 12.82 +5.82
Table 7. Concentrations of nitrate/nitrite nitrogen and nitrification rates in soil (0-5 cm) at three
altitudes on Bukit Belalong. Values are the means ±95% confidence limits where applicable.
n=23 at 200 m, 17 at 500 m and 9 at 850 m.
Altitude NO/NO,-N Nitrification rate

(m) (JLg N g") (JLg g'! 21 d'!) (kg ha'! 21 d")
200 5.2 ± 1.2 15.0 ±4.6 4.65
500 6.4 ±2.5 11.6 ±5.4 2.60

850 4.8 ±2.0 16.6 ±1O.7 3.72
and the mean difference was 3. 13°C for maxima and 1.95°c for minima. Solar radiation
(Figure 9b) was generally higher at 913 m and the mean values were 14.47 MJ m· 2 at 60 m
and 17.03 MJ m-2 at 913 m. The Penman transpiration estimate (Figure 9c) followed a
similar pattern to solar radiation and the total values were 51.5 mm at 60 m and 58.2 mm
at 913 m.
..... '.
~
.'i
. ' ... " ~\,/.
:: ~
,l
, ~
:1
, I
10
~r '
Figure 9. (a) Maximum (closed symbols) and minimum (open symbols) daily temperature (,C), (b) daily total solar
radiation (MJ m·2) and (c) Penman evaporation (mm) at Kuala Belalong, 60 m (_) and Bukit Belalong, 913 m ( .. )
between 12 and 30 August 1991.
The results of the bioassay experiment are summarised in Figure 10. Thes,e results are
reported fully in Wilkie (1992). There were substantial differences in the rice plants grown
at the two altitudes. The shoot height (Figure lOa) was significantly reduced (p < 0.001,
three way ANOYA) for plants grown in all the treatments at 913 m compared with those at
61 m. Shoot mass (Figure lOb) showed a different response. In BS the plants grown at 913
m had increased shoot mass compared with those at 61 m whilst in SS the opposite trend was
found.
Shoot
Shoot (b) maS8
(al height (DIg)
(mm)
70
300 60
50 I
40
200
T
30
100 20
10
R.oot
(e) mass
(DIg) (d) Root
ratio
Ishoot
BO 1.60
70
60
1.40
1.20
I
50 1.00
40 0.80
30 0.60
20 0.40
10 0.20
0.00
Treatment Treatment
Figure 10. The response to nitrogen addition of rice seedlings (Oryza sativa) grown in two soils (Base soil, BS and
Summit soil SS) at two altitudes on Bukit Belalong. (a) shoot height, (b) shoot mass, (c) root mass, (d) ratio of root
mass: total mass of plant. Treatments: (I) BS, 61 m. (2) BS +N, 61 m. (3) SS, 61 m. (4) SS +N, 61 m. (5)
BS, 913 m. (6) BS +N, 913 m. (7) SS, 913 m. (8) SS +N, 913 m. The columns represent the mean values for
six replicates and the error bars are the 95 % confidence limits.
The differences between the two altitudes were significant (p < 0.01) only for the plants
grown in SS without the addition of nitrogen (Treatments 3 and 7). Root mass (Figure 10c)
was significantly increased (p<O.OOl) in plants grown in BS at 913 m compared with 61 m.
The increases were not so marked for plants grown in SS and were significant (p < 0.01) only
for the fertilised plants (Treatments 4 and 8). There were large differences in the root:shoot
ratios (Figure lOd) with higher values at 913 m (p<O.OOl, three-way ANOVA).
The addition of nitrogen had some effects, but these were slight compared with the effects
of site microclimate and were only significant for shoot height in SS (p<0.05). Plants
responded differently to fertilisation in the two soil types. Without addition of nitrogen shoot
height was significantly smaller in SS compared with BS at both 61 m (p < 0.001) and 913
m (p < 0.05) but there were no significant differences between fertilised soils. At 61 m
differences in shoot mass between control and fertilised treatments were negligible, but at
913 m fertilised plants had greater shoot mass in BS than SS (p<0.05) whilst there was no
significant difference between shoot mass on unfertilised soils. There were also negligible
differences in root mass between soils and treatments at 61 m but plants responded to
fertilisation differently at 913 m. In the control pots root mass was significantly higher
(p<0.001) in BS compared with SS but differences were no longer significant after
fertilisation.
DISCUSSION
Climate
Gunung Silam, in Sabah, is similar in height (884 m) and latitude (5° N) to Bukit
Belalong, but has very stunted forest at its summit and its climate also shows some
significant differences. The most up-to-date account of the climate on Gunung Silam is
contained in Bruinzeel et al. (1993). Mean temperatures on Bukit Belalong are c. 2°C lower
than at similar altitudes on Gunung Silam and almost twice as much rain falls on Bukit
Belalong as on Gunung Silam (c. 5000 mm vs c. 2700 mm). A cloud cap usually forms on
Gunung Silam early each afternoon and consequently daily solar radiation totals at the
summit are reduced by 34 % compared with the base. A cloud cap does occasionally form
on Bukit Belalong, there is no reduction in the mean solar radiation because on clear days
radiation is higher at the summit than at the base. The cloud cap also reduces evaporation
and leads to much wetter conditions and the stunted forest is characterised by thick growths
of epiphytic and forest floor bryophytes. Occult precipitation is significant on Gunung Silam,
but may also be important on Bukit Belalong. On several rain free evenings in March it was
observed that fog in the early evening deposited moisture. The deposition was heavy enough
to leave the forest floor damp for several hours in the morning. On colder, clear nights dew
formation may also have been an important source of unrecorded precipitation. Epiphytic
bryophytes are more abundant near the summit (though not to the same degree as on Gunung
Silam) and terrestrial herbs are also considerably more abundant there, and this has been
related to a moister environment (Poulsen, this volume).
Litterfall
Table 8 gives data from litterfall studies in other South East Asian primary forests. It can
be seen that the values for the forests up to 500 m on Bukit Belalong are at the upper end
of the range for lowland forests (exceeded only by Alluvial forest at Gunung Mulu) whilst
the lower montane forest has a higher litterfall than any other montane forest from the
region, for which litterfall was measured for a complete year (Proctor et al. (1983b) sampled
litterfall on Gunung Mulu for only 16 weeks). The Belalong forests thus appear to be highly
productive, but the results must be treated with caution since the study lasted for only one
N
Table 8. Leaf and total small litter production (t ha" yr") and litter nitrogen concentrations (mg N g") in South East Asian primary rain forests. I; This study, ..,.
VI
2; Yamada (1976), 3; Bruijnzeel (1984), 4; Lim (1978),5; Edwards (1982), 6; Proctor et al (1989),7; Proctor et al. (1983a), 8; Proctor et al. (1983b).
Location Forest type Altitude Total litter Leaf litter

(m) (t ha" yr") (mg N g") (t ha" yr") (mg N g")
Brunei, Belalong'. Lowland 200 10.7 14.5 7.7 11.4
Lowland 500 10.4 12.7 7.5 10.6 n

Lower montane 850 8.3 12.7 5.7 11.3 >
Java, Cibodas2 • Lower montane 1550 6.0 4.5 '"CI
trl
Java, Pringombo'. Lower montane 650 7.4 5.8 11.2 Z
0
Malaya, Pasoh'. Lowland 100 8.9 11.0 6.4 12.0 ;:t:I
><:
New Guinea, Kerigornna'. Lower montane 2500 7.6 12.0 6.1 13.0
~
Sabah, Silam6 • Lowland 280 6.5 13.0 3.9 12.6 Z
0
Lowland 330 7.4 10.4 4.5 10.8 ~
Lowland 480 5.2 11.5 3.4 11.2 '"CI
;:t:I
Lower montane 610 5.6 10.7 4.1 10.5 0
n
...,
Lower montane 790 5.5 10.8 3.7 10.5
0
Lower montane (stunted) 870 4.8 8.6 3.3 8.3 ;:t:I
Sarawak, Mulu 7• Lowland (alluvial) 50 11.5 9.7 6.6 9.0

Lowland 200 8.8 9.2 5.4 9.5
Heath 200 9.2 6.0 5.6 5.7
Forest over limestone 300 12.0 12.0 7.3 12.0
Sarawak, Mulus. Lowland 500 10.9 10.0 6.7 10.0
Lower montane 1310 11.0 7.8 5.7 8.1
Upper montane 1860 3.6 7.9 2.3 7.4
year and longer studies have shown up to a two-fold variation in Iitterfall from year to year
(Blasco & Tassy, 1975). After the dry period at the beginning of 1992 there was an increase
in leaf litterfall at all altitudes, but especially in the mid plots where the peak value was 2.35
times the mean annual rate compared with 1.85 times higher in the low plots and 1.65 times
higher in the top plots. The mid plots are sited close to a ridge top and may be more
vulnerable to drought than the top plots despite the steeper slopes and shallow, stony soils
of the latter (Table 6).
Litter production, but not its nitrogen concentration, is significantly lower in the top plots.
The absence of changes in nitrogen concentrations with altitude on Bukit Belalong contrasts
with studies on Gunung Mulu, Gunung Silam, VoIcan Barva and in the Luquillo Mountains
reviewed by Heaney & Proctor (1989). On all these mountains decline in forest stature was
correlated with reduced Iitterfall nitrogen concentration and it was concluded that nitrogen
is the element most likely to restrict ecosystem processes on wet tropical mountains. The
data for Bukit Belalong thus reinforce these conclusions since it has a relatively large stature
forest with an apparently adequate supply of nitrogen at the highest altitude. The high
nitrogen concentrations in Iitterfall from the forest at 850 m on Bukit Belalong are striking
when compared with the low concentrations from the stunted forest at 870 m on Gunung
Silam (maximum height of emergent tree 16 m) and they resemble nitrogen concentrations
in the tall montane New Guinean forest at 2500 m studied by Edwards & Grubb (1977)
where the largest trees were 37 m tall.
Compared with other forests in South East Asia (Table 8) the litterfall from the forests at
all altitudes on Bukit Belalong has relatively high nitrogen concentrations. The total small
litterfall at 200 m on Bukit Belalong has the highest concentration of nitrogen in Table 8 and
those at 500 m and 850 m are exceeded only in the forest at 280 m in Sabah (Proctor et al.
1989). Leaf Iitterfall nitrogen concentrations are also relatively high though lower than
several values in Table 8.
Nitrification rates
Marrs et al. (1988) found considerable differences in nitrate availablity and production
along an altitudinal transect on VoIcan Barva, Costa Rica. In that study soil nitrate/nitrite-
nitrogen concentrations decreased from 11.6 to 1.5 Ilg N g-I soil between 100 m and 2600
m, and rates of nitrification also declined with increasing altitude.
On Bukit Belalong however, there was no significant difference in nitrate supply with
altitude which is not unexpected since the climatic differences on Bukit Belalong are much
less pronounced than on VoIcan Barva. Soil moisture content partially regulates the rate of
nitrification (Robertson 1989), with nitrate concentrations decreasing as a soil becomes wetter
(Chandler 1985). An increase in soil moisture due to elevated humidity within a cloud cap
(leading to the accumulation of mor humus or peat) has been recorded on mountains with
very stunted forests (Whitmore 1984, 8ruijnzeel et al. 1993), and the absence of a daily
cloud cap on 8ukit 8elalong is reflected in the similarity between nitrate concentrations at
all three altitudes.
The results for 8ukit 8elalong indicate low rates of nitrification compared with the VoIcan
Barva data (3.0 to 21.8 Ilg N g-I soil 14 d- I), but the results are within the ranges reported
by Vitousek & Matson (1988) for undisturbed forest soils (6 to 25 p.g N g-I soil 10 d- I), and
are similar to rates measured by Chandler (1985) in soils from two Malaysian lowland forests
(6.3 and 14.8 p.g N g-I soil 16 d- I) and one 'sub-montane' forest (5.0 p.g N g-I soil 16 d- I).
Bioassay
Nitrogen limitation of bioassay plants in soils from stunted Jamaican montane forests has
been shown by Healey (1989). Significant differences were found between leaf and root
mass of the grass Melinis multiflora Beauv. grpwn in lowland and montane soils, and the
addition of nitrogen and phosphorus significantly increased growth of plants in soils from
both upper montane and lower montane forests. This was in contrast to the results of Tanner
(1977) whose bioassay on the same soils showed no such limitation. Healey attributed the
conflicting results to flaws in Tanner's experimental design in which the test plants were
grown at a single lowland site but not at a montane site with a climate similar to the forests
from which the soils were collected. The bioassay in this study was replicated at two
altitudes, but provided only slight evidence that rice plants are more nitrogen limited in
montane soil compared with lowland soil, and whilst it would have been desirable to have
extended the length of the experiment to ensure exhaustion of the seedling reserves, and
possibly to have used a higher rate of nitrogen addition, the results are revealing. The large
differences in partitioning between root and shoot at the two altitudes (Figure 10d) may be
at least partly explained by the higher insolation and greater potential evapotranspiration at
913 m (Figures 9b, 9c). The long term data (Figure 4), however indicate no significant
differences in solar radiation between altitudes. We are left with temperature changes as
the most likely cause of the relatively slight differences between the forests at different
altitudes on Bukit Belalong.
CONCLUSION
Litterfall nitrogen concentrations, soil nitrate concentrations and rates of nitrification do not
indicate any reduction in nitrogen supply in the lower montane forest on Bukit Belalong
compared with surrounding lowland forests, and the bioassay experiment gave little evidence
of nitrogen limitation at higher altitude. The marked dry season peak of leaf litterfall in the
mid plots compared with the top plots suggests that the montane forest is not more
susceptible to drought than the lowland forest, and despite the high rainfall at the summit
there are no indications that waterlogging might adversely affect growth. In the absence of
a cloud cap and substantial differences in solar radiation, altitudinal effects on Bukit Belalong
seem more likely to be related to differences in temperature than differences in other factors
including nitrogen.
ACKNOWLEDGEMENTS
The authors would like to thank Universiti Brunei Darussalam and the Royal Geographical
Society for the opportunity to participate in the 1991-1992 Brunei Rainforest Project, and the
Leverhulme Trust and the Royal Geographical Society for financial support. We particularly
thank Mr Peter Wilkie for his assistance with the bioassay and Ramlah anak Sapong for her
help throughout the field work.
LITERATURE CITED
ASHTON, P.S. 1964. Ecological studies in the mixed dipterocarp forests of Brunei State. Oxford Forestry Memoirs
25.
BLASCO, F. & TASSY, B. 1975. Etude d'un ecosysteme forestier montagnard du Sud de I'Jnde. Bulletin
d'Ecologie 6:525-539.
BRUIJNZEEL, L.A. 984. Elemental concentration of litterfall in a lower montane rain forest in Central Java,
Indonesia. Malayan Nature Journal 37:199-208.
BRUIJNZEEL, L.A., WATERLOO, M.J., PROCTOR, J., KUITERS, A.T., & KOTTERINK, B. 1993.
Hydrological observations in montane rain forests on Gunung Silam, Sabah, Malaysia, with special reference to
the Massenerhebung effect. Journal of Ecology 81:145-168.
CHANDLER, G. 1985. Mineralization and nitrification in three Malaysian forest soils. Soil Biology and
Biochemistry 17:347-353.
EDWARDS, P.J. 1982. Studies of mineral cycling in a montane rain forest in New Guinea V. rates of cycling in
throughfall and litterfall. Journal of Ecology 70:807-827.
EDWARDS, P.J. & GRUBB, P.J. 1977. Studies of mineral cycling in a montane rain forest in New Guinea. I. The
distribution of organic matter in the soil and vegetation. Journal of Ecology 65:943-969.
FAO/UNESCO. 1974. Soil map of the world 1: 5000000. I. Legend. UNESCO, Paris.
GRUBB, PJ. 1977. Control of forest growth and distribution on wet tropical mountains: with special reference to
mineral nutrition. Annual Review of Ecology and Systematics 8:83-107.
GRUBB, P.J. & TANNER, P.1. 1976. The montane forests and soils of Jamaica: a reassessment. Journal of the
Arnold Arboretum 57:313-368.
HEALEY, 1.R. 1989. A bioassay study of soils in the Blue Mountains ofJamaica. Pp. 273-283 in Proctor, J. (ed.).
Mineral Nutrients in Tropical Forest and Savanna Ecosystems. Blackwell Scientific Publications, Oxford.
HEANEY, A. & PROCTOR, 1. 1989. Chemical elements in litterfall in forests on Vol can Barva, Costa Rica. pp.
255-271 in Proctor, J. (ed). Mineral Nutrients in Tropical Forest and Savanna Ecosystems. Blackwell Scientific
Publications, Oxford.
LIM, M. T. 1978. Litterfall and mineral nutrient content of litter in Pasoh Forest Reserve. Malayan Nature Journal
30:375-380.
MARRS, R.H., PROCTOR, J., HEANEY, A. & MOUNTFORD, M.D. 1988. Changes in soil nitrogen
mineralization and nitrification along an altitudinal transect in tropical rain forest in Costa Rica. Journal of
Ecology 76:466-482.
MOORE, P.D. & CHAPMAN, S.B. 1986. Methods in Plant Ecology. (2nd edition). Blackwell Scientific
PROCTOR, J. 1983. Tropical forest litter fall I. Problems of data comparison. pp. 267-273 in Sutton, S.L.,
Whitmore, T.C. & Chadwick, A.C. (eds). Tropical rainforest: ecology and management. Blackwell Scientific
PROCTOR, 1., ANDERSON, 1.M., FOGDEN, S.C.L. & VALLACK, H.W. 1983a. Ecological studies in four
contrasting lowland rain forests in Gunung Mulu National Park, Sarawak. II. Litterfall, litter standing crop and
preliminary observations on herbivory. Journal of Ecology 71:261-283.
PROCTOR, J., ANDERSON, J.M. & VALLACK, H.W. 1983b. Comparative studies on forests, soils and litterfall
at four altitudes on Gunung Mulu, Sarawak. The Malaysian Forester 46:60-76.
PROCTOR, 1., LEE, Y.F., LANGLEY, A.M., MUNRO, C.M. & NELSON, T. 1988. Ecological studies on
Gunung Silam, a small ultrabasic mountain in Sabah, Malaysia. I. Environment, forest structure and floristics.
Journal of Ecology 76:320-340.
PROCTOR, 1., PHILLIPS,C., DUFF,G.K., HEANEY,A. & ROBERTSON,F.M. 1989. Ecological studies on
Gunung Silam, a small ultrabasic mountain in Sabah, Malaysia. II. Some forest processes. Journal of Ecology
77:317-331.
ROBERTSON, G.P. 1989. Nitrification and denitrification in tropical forests. pp. 55-69 in Proctor, J. (ed.).
Mineral Nutrients in Tropical Forest and Savanna Ecosystems. Blackwell Scientific Publications, Oxford.
SUGDEN, A.M. 1986. The montane vegetation and flora of Margarita island, Venezuala. Journal of the Arnold
Arboretum 67: 187-232.
TANNER,E. V.J. 1977. Four montane forests of Jamaica: a quantitative characterization of the floristics, the soils
and the foliar mineral levels and a discussion of the interrelations. Journal of Ecology 65:883-918.
VENEKLAAS, E.J. 1992. Litterfall and nutrient fluxes in two montane tropical rain forests, Colombia. Journal
of Tropical Ecology 7:319-336.
VITOUSEK, P.M. & SANFORD, R.L.Jr. 1986. Nutrient cycling in moist tropical forest. Annual Review of Ecology
and Systematics 17:137-167.
VITOUSEK, P.M. & MATSON, P.A. 1988. Nitrogen transformations in a range of tropical soils. Soil Biology and
Biochemistry 20:361-367
WHITMORE, T.C. 1984. Tropical Rain Forests of the Far East, (2nd edition). Clarendon Press, Oxford.
WILKIE, P.W. 1992. A Bioassay to Investigate the Causes of Altitudinal Zonation on Bukit Belalong, Brunei.
Unpublished BSc thesis, University of Stirling, Stirling, U.K.
YAMADA, I. 1976. Forest ecological studies of the montane forest of Mt. Pangrango, West Java. III. Litterfall of
the tropical montane forest near Cibodas. South east Asian Studies 14:193-229.
Soil conditions, erosion and nutrient loss on steep slopes under mixed
dipterocarp forest in Brunei Darussalam
S.M. ROSS' and A. DYKES 2
'Department of Geography, University of Bristol, United Kingdom

2Department of Geography, King's College, London, United Kingdom
ABSTRACT. Soil physical and chemical characteristics are reported under virgin, mixed dipterocarp
forest growing on steep slopes in the Temburong District of Brunei Darussalam. Hydrological,
sedimentological, soil organic matter and nutrient losses from experimental erosion plots are presented
for both wet and dry season conditions. Total annual runoff losses from bare soil surfaces under an intact
rainforest canopy exceed 107 Iitres per hectare, while total annual sediment losses exceed 46 tonnes per
hectare. These values are approximately 6 and 20 times higher respectively than those measured from
the undisturbed forest floor, comprising leaf litter layer and root mat. The combined root mat and litter
layers are more important in protecting the soil surface against sediment loss than they are in reducing
overland flow. These high values indicate how serious runoff and soil erosion can be on such steep
slopes with very low permeability clay textured soils. Losses of soil nitrogen, phosphorus and potassium
in eroded sediments are serious, particularly from the bare soil treatment, where 388 kg N ha'!, 13 kg
P ha'! and 196 kg K ha'! were lost in twelve months.
KEY WORDS: orthic acrisols, soil chemistry, sediment loss, throughfall, sediment yields, Borneo.
INTRODUCTION
Brunei Darussalam is one of the few countries in SE Asia and in the tropics generally where
there has been no widespread felling of forests either for logging or for shifting cultivation.
Around 81 % of the land remains forested, with as much as 59% under primary forest
(Anderson & Marsden 1988, Collins et al. 1991). Under these circumstances, field
experimentation and discussions of geomorphological and soil effects of large scale
deforestation are inappropriate for the study region in the Temburong District of Brunei
which is characterised by dense dipterocarp forest clothing steep river valleys. Instead we
report here the results of a series of field experiments designed to elucidate detail about soil
erosion processes under the forest canopy.
Evidence from previous soil erosion and soil process studies under tropical rain forest
have indicated that the forest canopy, the shrubs, herbs and the leaf litter layer at the soil
surface all act to intercept raindrops and to dissipate their energy (e.g. Wiersum 1984).
Where raindrop energy is dissipated directly at the mineral soil surface, soil particles are
detached and moved downslope. Whether or not soil particles are eroded from the soil
surface during rainstorms is determined by three main factors: (i) rainfall intensity, (ii) soil
surface conditions (iii) soil hydrological conditions. Physical conditions of the surface soil
determine how much water infiltrates into the soil profile and how much water flows away
over the soil surface as overland flow or runoff. Soil texture and soil structure, which
determine soil bulk density, antecedent soil moisture conditions, as well as the presence of
leaf litter at the soil surface, are the main controlling factors at the forest floor.
In our experimental plots on the West Ridge of the Belalong valley, our aim was to study
the processes of soil hydrology and erosion under the intact dipterocarp canopy. In
particular, our objective was to examine the roles of the root mat and leaf litter at the soil
surface in partitioning incoming through fall between overland flow over the soil surface and
259
260 S.M. ROSS AND A. DYKES
infiltration into the soil profile. A subsequent aim was to quantify sediment, organic matter
and nutrient losses in erosion from the three treatments.
In this paper we firstly report details of the soil conditions in the Belalong valley under
mixed dipterocarp forest. Secondly, we describe field experiments designed to aid our
understanding of soil surface hydrology and erosion processes. Finally we compare our
results to those reported elsewhere in rain forests of SE Asia.
SOIL PROFILE CHARACTERISTICS
The steep slopes and ridges of the Temburong District of Brunei are covered by relatively
thin, silty clayey soils, developed over weathered and permeable shales. The entire region
shows much evidence of small and large scale landslides, manifest in the forest by treefall
gaps. These breaks in the canopy allow direct rainfall-fed soil erosion. Soil survey work
by Ashton (1964) and his soil physical and chemical analysis indicate that the soils of the
region are relatively nutrient poor, of varying depth and rather impervious, due to their high
clay content. Ashton (1964) identified four soil types in the Belalong region:
(i) Ridge soils, developed on sandstone, described as yellow clay latosols.
(ii) Upper slope shale lithosols - shallow and stony.
(iii) Lower Hillslope clay lithosols - two types:
a. steep slopes, often just below ridges, little litter, often signs of surface erosion.
b. more gentle slopes, deeper soils, often sticky and humid, with leaf litter at the soil
surface.
Our descriptions of midslope soils of the western and eastern valley slopes are given in
Tables 1a and 2a respectively. Of the 12 soil profiles excavated and described in detail from
various slope locations up to 913 m, all show a lack of organic matter beneath the immediate
soil surface and a relatively friable topsoil containing numerous weathered soft shale clasts.
Both examples reported here would appear to represent Ashton's gentler slope clay lithosol
category, with sticky, moist surface conditions after rainfall, a complete cover of leaf litter
at the soil surface and weathered shale fragments throughout the profile. Soil physical and
chemical characteristics for the two example profiles are given in Tables 1b and 2b
respectively. The downprofile particle size data indicate relatively sandy textured topsoils,
Table la. Soil profile description, Hydrology Pit No.6, West Ridge.
General Description: Altitude 110m above the R. Belalong. 30 slope; aspect 65 Tree canopy 60-70 m, with light
0 0 •
sapling understorey. 100% superficial cover of leaf litter, averaging 2-3 cm depth.
Soil Profile: Two horizons identified.
0-417cm Very fine crumb (granular) structure with tiny fragments ofleaflitter. Silty clay texture. Colour:
between 5YR 3/3 and 5YR 3/2. Visibly porous and friable with numerous small pores < lmm.
Root mat at the soil surface with numerous fine roots up to lmm diameter. No true mottles. A
very few tiny stones <2mm diameter on the soil surface.
Diffuse and regular boundary to:
>7cm Medium angular blocky structure about 5cm diameter. Silty clay texture. Colour 10YR 5/8
(yellowish brown). Few visible pores with no cracks. Frequent medium sized roots « 4mm) to
17cm, but very infrequent below. 3 large roots (5cm diameter) at !Ocm depth. Stones frequent
below 35cm, including large (l0-15cm diameter) weathered shale fragments, which make up 40%
of the profile below 35cm depth.
Soil conditions on steep slopes in rairiforest 261
Table lb. Soil profile characteristics, Hydrology Pit No.6, West Ridge.
Depth (cm) %LOI Particle Size (%) Bulk Density (g cm")

sand silt clay
0-4 35.4 53.9 28.3 17.8 0.703
5-10 18.1 27.2 42.5 30.3 1.180
>25 12.2 6.5 39.4 54.0 1.193
Soil Total Nutrients (mg g")

Depth (cm) N P K Na Ca Mg
0-4 4.65 0.36 9.73 1.86 0.15 0.08

5-10 2.27 0.26 11.2 2.49 0.17 0.23
>25 0.50 0.20 13.57 2.90 0.05 0.03
Table 2a. Soil profile description, Rattan Plot No. DIO, East Ridge.
General Description: Altitude 180m. 27 slope, aspect 275 Tree canopy 60-70 m, with very dense throughstorey.
0 0 •
Understorey of saplings, palms and rattans to a height of 1.5 m. 100% leaf litter cover at the soil surface,
averaging 1-2 cm depth. 20-30% woody litter cover, no bare soil.
Soil Profile: Three horizons identified:
0-2.5cm Very superficial organic horizon, with numerous roots up to I mm diameter. Much very finely
fragmented litter mixed into the soil. Very friable silty clay; matrix colour 10YR 4/4 (dark
yellowish brown). No stones. Occasional larger roots up to 2cm diameter. Diffuse, regular
boundary to:
2.5-35cm Medium angular blocky structure, peds 5cm. Cracks between peds up to Imm wide. Ped faces
not visibly porous. Silty clay loam, matrix colour IOYR 6/6 (brownish yellow). Frequent roots
up to 8mm diameter. No stones. Very diffuse and irregular boundary to:
>35cm Poor angular blocky structure of medium size; ped up to 5cm across. Fairly consolidated silty
clay matrix plastered around numerous shale fragments up to 20cm across. Stones make up
70-80% of the horizon. No crack between peds or between matrix and stones. Not porous.
Orange/brown bands (2-3cm wide) of deposited iron oxides. Matrix colour IOYR 6/6 (brownish
yellow); pale weathered shale IOYR 8/4 (very pale brown); deposited iron oxides 7.5 YR 6/8
(reddish yellow). Infrequent roots to a depth of 50cm.
changing to very clayey textures below 5 cm depth. All loss on ignition percentages are
higher than the visible organic matter content would suggest. This may be interpreted as
weight loss due to interstitial water in weathered secondary clays, or to volatilisation of
carbonates or salts. The very low calcium and magnesium concentrations, however, would
suggest that the soils are not high in these compounds. As expected, our soil chemical
analysis indicates higher total soil nitrogen and phosphorus in the surface 0-5 cm where the
soil organic matter contents are higher. All soils are poor in both calcium and magnesium,
with much higher potassium content, reflecting intense weathering and the presence of
potassium in secondary clay minerals.
According to the FAD-Unesco (1974) world soil classification, the Belalong soils are
classed as orthic acrisols. This means that they are relatively poor in bases and have an
argillic (a clay-enriched) B-horizon. Determinations of cation exchange capacity and
percentage base saturation are in progress and will be used to assess how typical are the
Belalong soils of this class.
Table 2b. Soil profile analytical characteristics, Rattan Plot No. DlO, East Ridge.
Depth (cm) %LOI Particle Size (%) Bulk Density (g cm")
sand silt clay
0-2.5 28.0 59.9 23.5 16.6 na
2.5-10 9.8 14.4 43.9 41.7 na
>35 6.58 25.5 32.3 42.2 na
Soil Total Nutrients (mg g")

Depth (cm) N P K Na Ca Mg
0-2.5 4.75 0.51 12.66 2.80 0.20 0.80
2.5-10 0.98 0.31 12.80 3.48 0.14 1.01
>35 0.03 0.24 15.68 4.22 0.13 1.10
FIELD EXPERIMENTS ON SEDIMENT LOSS
Materials and Methods

(a) Field Site and Methods
Soil hydrology and erosion studies were carried out 0.5 km from the Kuala Belalong Field
Studies Centre (KBFSC) in the Temburong District of Brunei (Figure 1). The plots were
located on a North Easterly facing slope, described elsewhere as the West Ridge.
Hourly rainfall measurements were monitored during the entire study period at Kuala
Belalong (Dykes, this volume), 1.5 km from the forest study site. Throughfall measurements
were made daily using 19 simple funnel collectors, 25 cm in diameter, which were randomly
relocated every week in the soil hydrology study area of 1500 m2 , during three and a half
months between early November 1991 and mid February, 1992, thus including both wet and
dry season conditions. In order to estimate throughfall from measured rainfall, fifty two
through fall collection dates were used to calculate an empirical relationship between rainfall
and through fall. To measure runoff generation and sediment losses under the forest canopy,
three treatment plots were constructed at a height of 110 m above the R. Belalong, on the
west slope of valley, above the KBFSC. In the three treatments, we aimed to create three
different soil surface conditions which would allow us to elucidate the roles of leaf litter and
the root mat in surface soil hydrology and sediment loss processes. The treatments were:
(Tl) intact forest floor, retaining the surface root mat and fallen leaf litter, (TI) soil with
surface root mat intact and all leaf litter removed, and (T3) completely cleared and bare soil
surface. These Gerlach-type plots were 5 m long by 2.5 m wide, bounded by 20 cm timber
edges, with plot slope angles of 30°. Remoteness from easy access, laboratory analysis and
supplies limited our ability to replicate the experiment. Results and discussion acknowledge
the lack of replication and small size of runoff plots. Runoff collection was measured daily
or after major storms at the base of each plot and sediments were collected at the same times
using very fine mesh bags inserted into the runoff collectors. Runoff volumes reported in
this paper are currently uncorrected for direct precipitation into the collection vessels. After
collection, the sediments were sieved into 5 size fractions: < 0.25 mm, 0.25-1 mm, 1-4.75
mm, 4.75-11.2 mm and> 11.2 mm. All fractions were then oven dried at 105°C, weighed
and transported to the UK for chemical analysis.
Runoff plot
1 ~ Soilpil
Tensiometer nesl
N Main loolpath
i Contours al 2m intervals.
heights above river leyel
o
~
10
metres
Figure I. Map showing field study site and field plot locations in the Belalong valley, Temburong District, Brunei
Darussalam.
(b) Laboratory Analysis.

Sediment samples were hand sorted into organic and mineral fractions, then weighed to
determine proportional organic matter contents. Loss on ignition determinations were also
carried out on the sediment samples, using a 24 hour ignition period at 450°C. Total N, P
and K in eroded sediments were determined using a microdigestion in concentrated sulphuric
acid with hydrogen peroxide (Allen 1989), followed by automated colorimetric
determinations for total nitrogen as ammonium (Crooke & Simpson 1971) and phosphorus
(Murphy & Riley 1962), and using atomic emission spectrophotometry for the determination
of potassium.
RESULTS
Rairifall, Interception and Through/all Characteristics

Monthly rainfall totals for Kuala Belalong (Figure 2) indicate that September to December
marked a distinct wet season, despite rather low rainfall in October, while January to April
marked a distinct dry season. Rainfall comparisons with measured through fall amounts
showed a good correlation of r = 0.989. Interception and percentage throughfall
measurements for tropical rainforests have shown great variability in quantity and distribution
of 'net rainfall'. Bruijnzeel (1990) for example comments on more than 70 studies from
natural rain forest sites whose results differ greatly and cites several climatic, vegetative and
procedural reasons for this. Calculated on the basis of the throughfall study period only,
700
600
500
E
§.
Sc: 400
e2-
300
~0
::;
200
100
o
JUN JUL AUG SEP OCT NOV DEC JAN FEB MAR APR MAY
Figure 2. Monthly rainfall totals recorded at the automatic rainstation at Kuala Belalong, June 1991 - May 1992.
interception accounted for 20% of rainfall, 1% was assumed to reach the forest floor in stem
flow, and the remaining 79% of rainfall was collected as throughfall. The throughfall
patterns will be discussed in detail elsewhere. Calculated over the 12 month field project
period, a range of 3-93 % of rainfall was intercepted by the canopy of mixed dipterocarp
forest at Belalong. This wide range of values can be accounted for by variations in the
rainfall amounts and intensities, combined with differences in canopy structure and foliage
throughout the year. We shall consider only throughfall, or net rainfall, as pertinent to our
runoff and erosion studies since the forest canopy above the erosion plots remained
completely intact throughout the study period.
Runoff Generation
The relationship between through fall amounts and runoff volumes for the three plots are
given in Table 3. All three treatments show extremely good r2 values. Very similar
throughflow lrunoff patterns were observed for T2 and T3, both of which showed runoff rates
around six times higher than the intact root mat and litter soil surface. Total annual runoff
quantities, calculated on the basis of six months of data: three months representing the 'wet'
season (November - January) and three months representing the 'dry' season (February -
April), are given for the three treatments in Table 3. Calculated as a percentage of
throughfall, runoff percentages were 0.26-4.97% for Tl, 0.3-7.1 % for T2 and 0.75-6.98%
for T3. These low values agree well with those calculated for runoff generation elsewhere
in Borneo. In undisturbed forest in Sabah, Sinun et al. (1992) reported runoff percentages
Table 3. Relationship between actual throughfall and runoff volumes and between runoff volumes and sediment yield
for three surface soil treatments under Mixed Dipterocarp Forest (based on seven months of measurements: October
1991 - May 1992). Th = Throughfall; Rv = Runoff volume; Sy = Sediment yield.
Treatment Actual Throughfall vs Runoff f Runoff Volume f

Volume vs Sediment Yield
Intact forest floor Rv = -5.76 + 1.01 Th 84.0 Sy = 17.5 + 0.484 Rv 39.3
(root mat + litter) (Tl)
Root mat only (T2) Rv = -38.3 + 5.97 Th 95.3 Sy = 11.1 + 1.79 Rv 71.1
Bare soil (TI) Rv = -24.5 + 6.00 Th 94.6 Sy = -14.3 + 3.37 Rv 62.5
Soil conditions on steep slopes in rainforest 265
of 2-3%; while Douglas et al. (1992), working in the same area, quotes runoff percentages
of 5 %. Only on logging tracks were higher runoff percentages of 55 % reported. Root mat
and bare soil plots showed significantly lower runoff percentages in the very dry season
months of March and April. Seasonal patterns in runoff volumes follow the seasonal rainfall
patterns (Figure 2) very closely. Greatest runoff volumes were generated in the peak wet
season months of November and December 1991, with increasing runoff volumes again
apparent in May 1992 (Figure 3). In all months, T1 showed only small quantities of runoff,
reaching a maximum at 31 litres/m 2 for the month of November 1991, compared to 135 and
151 litres/m 2 for T2 and T3 respectively.
Sediment loss
Total annual mineral and organic sediment losses, calculated on the basis of six months
of data; three months representing the 'wet' season (November - January) and three months
representing the 'dry' season (February - April), are given for the three treatments in Table
4. Combining amounts of mineral and organic material, a calculated total of more than 46
tonnes per hectare of sediments were eroded from bare soil under the forest canopy. This
amounted to 18.6 times more than from the intact forest floor. Regression equations for
sediment yield on runoff volumes are presented in Table 3. Both the removal of leaf litter
(T2) and the removal of the combined root mat and litter (T3) cause incremental increases
in sediment loss. This indicates that both leaf litter and root mat play an important part in
sediment retention. The strength of relationship between runoff volume and sediment loss
over the study period is very good for T2 and T3, showing r2 values of 71.1 % and 62.5%
respectively. The relationship between runoff and sediment loss from the undisturbed forest
floor (T1) is poorer at r2 = 39.3. This can be explained by the same argument given for
variability in canopy interception of rainfall - namely that the variability is due to different
quantities and intensities of net rainfall coupled with different litterfall rates and root growth
dynamics throughout the year. Seasonal patterns of sediment loss follow the seasonal runoff
patterns (Figure 3) very closely. Greatest sediment losses, like runoff, occur in the peak of
the wet season (Figure 4). Increases in sediment losses in May 1992 are not as great as the
increases in runoff volumes. There are several possible explanations for this observation:
(i) that much of the easily detachable surface soil particles had already been eroded, (ii) that
2000 r-~_.
~0
';>
E 1500
.Q
~
~
,s
"0 1000
IJ)
"C
~
IJ)
:J
0
.r:
t:.
500
'5c:
:J
II:
o
Oct 91 Nov 91 Dec 91 Jan 92 Feb 92 Mar 92 Apr 9.2 May 92
• Intact forest floor .Root mat only OBare soil
Figure 3. Monthly runoff totals for treatment plots T1 (intact forest floor), T2 (root mat only) and T3 (bare soil)
(all values in thousands of litres per plot).
Table 4. Total runoff and eroded mineral sediments and organic litter under Mixed Dipterocarp Forest soil surface
treatments (calculated in kg ha-' y-' from six months of field data).
Treatment Runoff (1 ha-') x 106 Mineral Sediment Loss (kg ha-' y-') Organic Litter (kg ha-' y-')
Intact Forest Floor (T1)

(Root mat + Litter) 0.913 1198 1308
Root mat only (T2) 4.715 23070 868
Bare soil (T3) 5.245 45689 936
previous wet season rainsplash followed by dehydration during the dry season caused surface
crusting which enhances runoff but reduces sediment loss, or (iii) that active root growth at
the beginning of the 1992 growing season as the soil is wetting up after the dry conditions
of February, March and April, stabilised the surface soil. The last suggestion seems
untenable since T3 (bare soil surface) also showed little sediment loss and it was not
protected by a root mat.
Leaf Litter Removal

Very clear differences were observed in the amounts of leaf litter eroded from treatment
plots. The actual quantities of leaf litter compared to mineral sediments lost for one example
collection date in each month of the study period are illustrated in Figure 5. The organic
matter data are calculated for hand-sorted litter > 11.2 cm in size. Substantial quantities of
leaf litter fell onto treatment plots T2 and T3, and were eroded before they could be cleared
by hand. These litter losses undoubtedly made significant contributions to nutrient losses
described below. Proportionally larger quantities of leaf litter were found in eroded
sediments in March and April 1992, at the end of the dry season, after the main period of
leaf fall. During these two months, the material eroded from T1 consisted almost entirely
of leaf litter. This is an interesting result since it illustrates that significant loss of soil
organic matter and nutrients bound up in leaf litter can be eroded from steep forested
hillslopes even without any canopy removal.
Loss of Nutrients in Eroded Sediments

Total annual quantities of nitrogen, phosphorus and potassium in eroded sediments and
organic matter, calculated on the basis of six months of data; three months representing the
14
12
E
t:0 10
E
I 8
~
~ 6
t:CD
E
'C 4
CD
'" 2
0
Oct 91
-~ ~ ~ -~ ~ ~. -~
• Intact forest floor ffiE!lRoot mat only DBare soil
Figure 4. Monthly sediment loss totals for treatment plots TI (intact forest floor), T2 (root mat only) and T3 (bare
soil) (all values in kg per plot).
3500
3000
2500
§ _Mineral sediment
'"
.9'"
2000 IiIilOrganic litter
E
"E
'C
1500
"
Ul
1000
500
NOV DEC JAN FEB MAR APR

Figure 5. Quantities of mineral sediment and organic litter eroded from three treatment plots on three example
collection dates selected from each month of the study period. T1 = Intact forest floor, T2 = root mat only, T3
= bare soil (all values in g per plot).
Table 5. Total Nitrogen, Phosphorus and Potassium losses in eroded sediments under Mixed Dipterocarp Forest
soil surface treatments. (Calculated in kg ha·' y-' from six months of field data).
Treatment Nutrient loss (kg ha·' y.')
N P K
Intact Forest Floor (Tl) 29.02 1.454 20.066

(Root mat + Litter)
Root mat only (T2) 131.57 8.545 106.400
Bare soil (T3) 388.324 13.077 195.79
'wet' season (November-January) and three months representing the 'dry' season
(February-April), are given for the three treatments in Table 5. The amounts of all three
nutrients lost from a cleared bare soil surface are very high. For nitrogen, an extremely high
388 kg N ha-! y-! was lost, amounting to 13 times that lost from the intact forest floor. Nine
times more phosphorus and ten times more potassium were lost from the bare soil plot than
from the intact forest floor.
DISCUSSION
The complexity and dynamism of physical, chemical and biological activity in the forest floor
of tropical rain forests is just being realised. Coelho Netto's (1987) description of the rain
forest litter layer also describes our conditions exactly: ..... a surface layer, representing the
current fall of leaves which show little evidence of decay, and a second layer comprising all
partially decomposed material above the mineral soil. Fine roots with variable mat thickness
overlie the mineral soil, being easily peeled back from the soil like a carpet. ..
Leaf litter and roots at the soil surface play several roles in the physical and chemical
processes occurring on the forest floor. Firstly, they provide both absorption and channelling
of incoming throughfall, which can result in higher infiltration rates and lower quantities of
runoff. Secondly, litter prevents rainsplash at the mineral soil surface. Litter and leaf
fragments can, however, become entrained within the surface mineral soil and within
transported slope sediments. Thirdly, organic litter is the main source of recyclable nutrients
for feeding roots. Thus, if litter is removed from the site by erosion, it may result in
significant nutrient depletion.
The exceedingly steep 30 slopes of the Belalong valley provided an unusually extreme
0
testbed for our experiments on the role of soil surface leaf litter and roots in protecting the
soil under mixed dipterocarp rain forest from erosion.
Since several studies confirm the theory that the kinetic energy of rain forest through fall
exceeds that of above-canopy rainfall (e.g. Brandt 1988, Vis 1986), our experiments were
designed to examine the role of protective litter layers in dissipating this energy. Other
authors have alluded to the role of leaf litter as a dissipator of rainsplash energy and have
shown how sediment losses can be enhanced when the litter layer is removed (eg. Ross et
al. 1990). Their Brazilian site, however, had no significant root mat at the soil surface.
Since runoff from T2 and T3 plots at Belalong are very similar, but sediment losses from T3
are significantly higher than from T2, it appears to be the litter rather than the root mat
which provides the main protection for the mineral soil and prevents rainsplash and particle
detachment.
Working in Atlantic rain forest on silty clay latosols near Rio de Janeiro, Coelho Netto
(1987) identified three routes of water movement at the soil surface: (i) over-litter flow, (ii)
through-litter flow and (iii) within the root litter carpet. All three routes are likely to occur
during the heaviest rainstorms at Belalong. Although we did not specifically measure the
flows associated with the litter layers, they must have provided significant transporting power
since as much as 130 g (oven dry weight) of leaf litter per m2 was calculated to have been
removed by erosion processes in a year from the intact forest floor. While no other
quantification of litter transport is available for tropical rain forest conditions Van Zon
(1980), reported very significant leaf transport from hillslopes under temperate beech, pine
and spruce forests in Luxembourg. It is interesting to compare the quantities of litter lost
in erosion to rates of litterfall and litter decay measured elsewhere for similar dipterocarp
forest in the Danum valley in Sabah. Burghouts et al. (1992) report litterfall and litter
disappearance (kL) rates for primary forest ranging from 3 - 11 t haol yol. The 1.3 t haol yol
litter erosion rates recorded in this study indicate that erosion could account for loss of as
much as 40 % of the organic matter on the forest floors of some dipterocarp forests with
lower litterfall rates. Young feeding roots actively grow through and into decaying leaf litter
at the soil surface since it is the main location of nutrient release and recycling in tropical
forests. The roots anchor the litter to the surface in the vicinity of the mother tree, providing
an easily accessible supply of nutrients. Under these conditions, the scale of litter transport
from the intact forest floor even on very steep slopes at Belalong was something of a
surprise. Malaysian dipterocarp litterfall amounts are around the lowest of any tropical
forests (Bray & Gorham 1964). The loss of nutrients in eroded litter undoubtedly contributes
to the already low fertility of soils on these steep slopes.
The sediment yields recorded for intact forest in this study are of the same order of
magnitude as values reported elsewhere in SE Asia, despite different methods of
measurement and calculation. Table 6 lists a range of different sediment loss determinations
for tropical rain forest in SE Asia. The higher sediment yields recorded for bare soil under
the forest canopy at Belalong are more similar to rates reported by Morgan et al. (1982) for
secondary forest, also on steep, 25 slopes in Sarawak, at 35-462 kg haol yol. Morgan et
0
,
al. (1982) describe both primary and secondary forests as having "thin" soil litter layers, so
the litter layer itself may not be the main influencing factor here. Soil organic matter plays
an important role in stabilising the soil structure. When the organic matter is removed,
aggregate disruption makes particles more easily detached by raindrop splash. Wiersum
(1984), summarising the work of many studies on clean weeded forests and forests whose
Soil conditions on steep slopes in rainforest 269
Table 6. Sediment yield in a range of tropical rain forest locations in SE Asia, using three estimation techniques:
(a) Catchment Measurements, (b) Erosion Plot Measurements, (c) Numerical Simulation Modelling. All values in
t ha·!y·!.
Study Location Sediment Yield (t ha'! y'!) Author
(a) Catchment Studies

Sg. Tenom, Cameron Highlands 0.34-0.35 Shallow (1956)
Sg. Gombak, Selangor 0.97 Douglas (1975)
Diu Segama, Sabah 3.12 Douglas et al. (1992)
(b) Erosion Plot Studies
Semongok, Sarawak 0.035-0.0247 Morgan et. al. (1982)
Diu Segama 0.16-0.29 Sinun et. al. (1992)
Belalong, Brunei This study
Intact forest Floor 2.42
Root mat only 24.05
Bare Soil 461.2
(c) Numerical Simulation Studies
Semongok, Sarawak 0.277 Morgan et. al. (1982)
Kalimantan 0.0085 Besler (1987)
litter had been removed or burnt, showed that sediment yields could be increased by around
20-200 times if the treatments were applied repeatedly.
Our very simple experiments at Belalong have indicated that even very minimal
disturbance of the forest system, by altering the forest floor conditions, can have a
deleterious effect on soil quality. These relatively small scale soil surface mass movements
contribute to both slope form and sediment loading to streams. The steep slopes of this part
of Brunei Darussalam show frequent small landslides, controlled in part by shallow soils and
by reduced soil strength under wet conditions. The presence of both leaf litter and
particularly the root mat have an effect on soil strength, through altering surface soil
moisture conditions and by welding soil particles and aggregates together. Further work on
the Belalong sites is elucidating how surface soil conditions affect infiltration capacity and
influence mass movements at all scales.
SUMMARY OF CONCLUSIONS
1. The combined root mat and litter layers are more important in protecting the soil surface
against sediment loss than they are in reducing overland flow.
2. Around 20 times more sediment is eroded from bare soil under a dipterocarp canopy than
from the intact forest floor.
3. The mass:mass ratio of organic matter to mineral sediments lost from the bare soil surface
was 1:49 compared to that of the intact forest floor at 1:0. 9.
4. For low Iitterfall dipterocarp forests, erosion of leaf litter, particularly on steep slopes,
may cause up to 40% depletion of incoming soil organic matter.
5. Almost 400 kg N ha'! y'! and 200 kg K ha'! y.! can be lost from the bare soil surface
under the forest canopy.
ACKNOWLEDGEMENTS
The authors gratefully acknowledge the logistical support of The Royal Geographical Society
and the University of Brunei Darussalam, particularly Dr David Edwards, Catriona Prebble,
Samhan bin Nyawa, Dr Kamariah Abu Salim, Amanda Simpson and John Wills. The
financial support of the Greencard Trust, Lloyds Bank, Royal Brunei Airlines and Brunei
Shell is also gratefully acknowledged. We are particularly indebted to numerous Iban and
British colleagues who assisted both in field construction and sampling, particularly Kuni
Anggat. We also wish to acknowledge the laboratory assistance in the UK of Aysen Brown
and Bob lies. Figure 1 was drawn by the Cartographic Unit of Kings College, Geography
Department, London.
LITERATURE CITED
ALLEN, S.E. 1989. Chemical Analysis of Ecological Materials. Blackwell Scientific Publications, Oxford, UK.
ANDERSON, I.A.R. & MARSDEN, D. 1988. Brunei Forest Resources and Strategic Planning Study. Unpublished
report to the Govermnent of Brunei (quoted by Collins et al. 1991).
ASHTON, P.S 1964. Ecological Studies in the Mixed Dipterocarp Forests of Brunei State. Oxford Forestry
Memoirs No.25. Clarendon Press, Oxford, UK.
BESLER, H. 1987. Slope properties, slope processes and soil erosion risk in the tropical rain forest of Kalimantan
Timur (Indonesian Borneo). Earth Surface Processes and Landforms 12:195-204.
BRANDT, J. 1988. The transformation of rainfall energy by a tropical rain forest canopy in relation to soil erosion.
Journal of Biogeography 15:41-48.
BRAY, J .R. & GORHAM, E. 1964. Litter production in forests of the world. Advances in Ecological Research
2:101-157.
BRUUNZEEL, L.A. 1990. Hydrology of Moist Tropical Forests and Effects of Conversion: A State of Knowledge
Review. Netherlands IHP Conunittee for UNESCO and lAHS.
BURGHOUTS, T., ERNSTING, G., KORTHALS, G. & DE VRIES, T. 1992. Litterfall, leaf litter decomposition
and litter invertebrates in primary and selectively logged dipterocarp forest in Sabah, Malaysia. Philosophical
Transactions of the Royal Society, London, B 335:407-416.
COELHO NETTO, A.L. 1987. Overland flow production in a tropical rainforest catchment: the role of litter.
Catena 14:213-231.
COLLINS, N.M., SAYER, J.A. & WHITMORE, T.C. (eds) 1991. The Conservation Atlas of Tropical Forests:
Asia and the Pacific. MacMillan and the World Conservation Union.
CROOKE, W.M. & SIMPSON, W.E. 1971. Determination of ammonium in kjeldahl digests of crops by an
automated procedure. Journal of the Science of Food and Agriculture 22:9-10.
DOUGLAS, I. 1975. The impact of urbanisation on river systems. Proceedings of the [GU, Regional Conference,
8th New Zealand Geographical Conference pp. 307-317.
DOUGLAS, I., SPENCER, T., GREER, T., BIDIN, K., SINUN, W., & WAI WENG, W. 1992. The impact of
selective commercial logging on stream hydrology, chemistry and sediment loads in the Ulu Segama rain forest,
Sabah, Malaysia. Philosophical Transactions of the Royal Society, London, B 335:397-406.
FAO-UNESCO 1974. Soil Map of the World, 1:5()()()()()(). (18 sheets).
MORGAN, R.C.P., HATCH, T., SDLAIMAN, W. & HARUN, W. 1982. A simple procedure for assessing soil
erosion risk: a case study for Malaysia. Zeitschrift fur Geomorphologie, Suppl.-Bd. 44:69-68.
MURPHY, I. & RILEY, I.P. 1962. A modified single solution method for the determination of phosphate in natural
waters. Analytica Chemica Acta 27:31-36.
ROSS, S.M., THORNES, I.B. & NORTCLIFF, S. 1990. Oil hydrology, nutrient and erosional response to the
clearance of terra firme forest, Maraca Island, Roraima, northern Brazil. The Geographical Journal 156:267-282.
SHALLOW, P .G.D. 1956. River flow in the Cameron Highlands. Hydroelectric Technical Monograph 3, Central
Electricity Board, Kuala Lumpur, Malaysia.
SINUN, W., WAI MENG, W., DOUGLAS, I. & SPENCER, T. 1992. Throughfall, stemflow, overland flow and
throughflow in the Ulu Segama rain forest, Malaysia. Philosophical Transactions, Royal Society London, B
335:389-395.
VAN ZON, H.I.M. 1980. The transport of leaves and sediment over a forest floor. A case study in the Grand
Duchy of Luxembourg. Catena 7:97-110.
VIS, M. 1986. Interception, dropsize distributions and rainfall kinetic energy in four Columbian forest ecosystems.
Earth Surface Processes and Landforms 11:591-603.
WIERSUM, K.F. 1984. Surface erosion under various tropical agroforestry systems. In O'Loughlin, C.L. &
Pearce, A.I. (eds). Proceedings of a Symposium on Effects of Forest Land Use on Erosion and Slope Stability.
IUFRO, Vienna, and East-West Centre, Honolulu, Hawaii.
Construction of a low-cost runoff and sediment monitoring plot
ALEXANDER K. SA YOK 1 and NALAU BINGEDING2
IDepartment of Forestry, PNG University of Technology, Lae, Papua New Guinea

2PNG Forest Products, Bulolo, Papua New Guinea
ABSTRACT. A low-cost runoff and sediment monitoring technique which uses rectangular plots and
collection devices is described. The technique is presently being tested near Hobu Community School,
approximately 15 Ian NE of the Papua New Guinea University of Technology, Lae. Sediment and runoff
are being monitored on four plots of Leucaena leucocephala seedlings, with two of these plots
interplanted with sweet potato (Ipomea balatas). The plots are inexpensive to construct and maintain,
as well as being easy to replicate and control.
KEY WORDS: soil loss, Universal Soil Loss Equation, Papua New Guinea.
INTRODUCTION
lnstream Monitoring Systems

Traditionally, runoff and sediment from a watershed area or tract of land are sampled
directly from the stream in that particular watershed. However, watershed runoff discharge
and sediment transport is a complex process due to diversified topographic configuration,
surface roughness, and drainage systems, in addition to changes in land-use. As such,
instream runoff and sediment transport rates are highly variable both in time and space
(Beschta 1985). A soil particle may be trapped or deposited in a micro-depression at a site
with changes in slope, just beside any debris, or at the bottom of a stream for days, weeks,
or even years before it is monitored at the stream gauging station (Sayok et al. 1993). Thus,
traditional instream sediment monitoring method may not be a sensitive indicator of changes
in watershed conditions (Jackson et al. 1985).
Alternative Monitoring Systems

Among the alternatives to instream sampling for direct watershed monitoring are erosion
net studies and erosion conditions assessments (Clark 1974), retention basin studies (Lusby
1979), erosion transect studies (Blaney & Warrington 1983), simulated rainfall studies
(Meyer 1988), and laboratory erosion studies (Mutchler et al. 1988).
Each of the systems above has advantages as well as disadvantages. Both erosion net
studies and erosion transects are amenable to replication and control, and are easy to
maintain. In this they are similar to the modified Universal Soil Loss Equation (USLE) plot
technique described here. However, they only provide soil-loss data, not runoff data.
Erosion condition assessments which provide relative ratings are best suited only to
inventory-type studies. Small retention basin studies are also similar in concept to the
technique described here, but are less amenable to the design principles of control and
replication. With respect to simulated rainfall studies using plots large enough for soil-loss
assessments which can be carefully designed and controlled, Jackson et ai. (1985) stated that:
"They are labor-intensive as well as expensive. Furthermore, data obtained may not be
representative of natural runoff and erosion rates because simulated storms generally have
little resemblance to natural storms, and temporal variations in infiltration and soil erodibility
are not usually sampled. "
271
272 ALEXANDER SA YOK AND NALAU BINGEDING
Small laboratory plots are best for studying basic erosion phases that are difficult to study
in detail on larger ones (Mutchler et al. 1988, Meyer 1988). On the other hand, their
operation inevitably requires rainfall simulators and disturbed soils, therefore, it inherits all
the problems faced by the simulated rainfall studies above.
USLE Plot System

Erosion research plots large enough to represent the complete process of rill and inter-rill
erosion and equipped with retention boxes have been successfully used to measure long-term
runoff and erosion rates. A good example is the 0.01 ac plot of 72.6 ft (22.1 m) long which
was a standard runoff unit used to develop the Universal Soil Loss Equation (USLE) by the
Agricultural Research Service to estimate water erosion on agricultural land under different
cultivation practices (Wischmeier & Smith 1978). This plot measures erosion from the top
of a slope where runoff begins. It is sufficiently wide to minimize edge or border effects
and large enough for development of downslope rills. Today, the system has been applied
in the United States to estimate soil losses or sediments for construction sites (Wischmeier
& Meyer 1973), irregular slopes (Foster & Wischmeier 1974), forest land (Dissmeyer 1973)
and a surface-mine spoil bank (McKenzie & Studlick 1979). It has also been used in
Canada, Peru, Ivory Coast, and Nigeria (Soil Conservation Society of America 1976). The
system has been modified by various researchers to particular requirements (Jackson et at.
1985, Lal 1981). Jackson et at. (1985) found the modified system to be extremely accurate
when used on grazed and ungrazed plots in Denver, Colorado, as well as being easily
replicated and controlled. In addition, it is cheap to install and maintain. Chang et at.
(1992) stated that the system could be a useful and powerful tool for estimating annual, as
well as single, storm soil losses in forested areas. The system in Papua New Guinea,
described below, is also a modified USLE system. The objective is to determine whether
the technology is applicable to Papua New Guinea where rainfall is over 4 000 mm per year.
THE PAPUA NEW GUINEA EROSION MONITORING SYSTEM
Plot Construction
Four plots were established on an abandoned food garden (sugar cane and sweet potato)
near Hobu Community School compound. These plots are constructed from low cost and
readily available materials. The cost of material per plot including collecting boxes is
approximately US$130.00. The list below summarises the materials required for plot
installation. The time required to construct each plot is approximately three man days.
Material List for Each Runoff-Sediment Plot :
4 sheets 4 mm exterior plywood (for construction of plot borders)
3 sheets 8 mm exterior plywood (for storage boxes)
25 30 em sticks (diameter, 0.50 em)
1/2 kg 2.5 em flat-head galvanized nails
Vi kg 3.75 em flat head galvanized nails

4 3 m PVC rain gutters
10 em black plastic sheet (cut into 4) gutter end caps
4 gutter connecting sleeves
1 gutter corner
1 downspout
1 40 kg bag cement
3 50 kg bag sand and small pebble mix
Approximate cost of materials = US$130.00
The size of each plot is 0.015 ha (15 m long and 10 m wide - Figures 1 and 2). Its side
Sediment monitoring plot construction 273
10.0 ... ~I
10 _5 ..
* · .· '"
..
,. · · .
Tree
~
. ;
~
··
~
c::
~ 'l
· 11-
·· ·· ,.
t<
Food
crop
".
·. It
* ·
'I<
·.*· :t<
··· ··· .. Concrete
Polythene
sheet
Standard Size
Gutter trough
/
Collection Boxes
O.Bll1. X O.Bll1. X 1.2211.
Figure 1. Plan view of runoff and soil erosion plot along with collection devices.
/ ; ,to,
Collection
Boxes
/ /; //
Figure 2. Cross-section view of runoff and erosion plot along with collection devices.
'-----'----"
/ / / /
and upper boundaries are made from 15 em plywood strips set about 8 em into the soil and
7 em sticking above ground. These strips are stabilized with sticks, 17 em long, obtained
from the surrounding bush. At the lower boundary, a standard PVC rain gutter was placed
with its upper edge flush with ground level. To ensure that runoff and sediment will move
through the gutter, a slight drop was maintained. A length of black plastic sheet (60 em
274 ALEXANDER SA YOK AND NALAU BINGEDING
.... 45 elll...-+-!
I
~C:;O-.J_dial= 2.54 ell
,<IL"":.: ___(" . . - -
,. .
:
-:~r I
I
I
.:l-_
: : I
I I
J____ J. ~
Figure 3. Cross-section view of splitter and collection boxes.
wide) was laid on the disturbed soil above the gutter with its lower edge overhanging the
edge of the gutter giving a stable flow of runoff and sediment into the gutter. Cement,
mixed at an appropriate consistency, was laid on the disturbed soil above the gutter. When
set, this cement mix creates a smooth surface. A PVC drain (downspout) was attached at
one corner of the lower boundary of the plot. Runoff and sediment that drop from the plot
into the gutter are delivered through the downspout to a collection box below.
Runoff and Sediment Collection Devices

Each collection box measures 0.61 m by 0.61 m by 1.22 m, large enough to store runoff
for a 50-year 48-hour storm, was installed at the lower end of the downspout. Inside this
box, a small storage box with two compartments was placed for convenience in measuring
smaller runoffs. Both boxes are made from exterior grade plywood.
Automatic Sampling Equipment and Continuous Monitoring

To augment or replace manual sampling, a stage-height recorder and automatic sediment
samplers can be installed. However, due to budget constraints and high incidences of
vandalism in some parts of Papua New Guinea, neither of these two pieces of equipment was
installed. So, only runoff depth for a particular rain event or period (a week) was calculated
based on the amount collected in the box. Sediment delivery rate (mass over time) is
determined by dividing the sediment weight with the length of time since the last observation.
Splitter or Multidivisor
The plot described above collects nearly 100% of sediment and runoff. To obtain only
aliquots or portions of runoff and sediment samples from the plot, a splitter or multidivisor
can be installed. Various types of multidivisor can be improvised such as those used by Lal
(1981) in his studies. In this project, a splitter of 45 cm by 45 cm and 30 cm deep plywood
with twelve 2.54 cm diameter openings (3 on each side) had been designed. A rubber hose
of appropriate diameter was fitted to each opening. All runoff entering the splitter can be
distributed evenly among the 12 openings. Only that runoff (1/12 or 8.33%) which flows
through the opening at the lower side of the splitter will flow via the hose into the collection
box (Figure 3).
CONCLUSION
The runoff and sediment monitoring plots described above are quick and inexpensive to
construct as well as easy to replicate and control. Sample plots were installed in June 1992
Sediment monitoring plot construction 275
near Hobu Community School, approximately 15 km NE of the Papua New Guinea

University of Technology. The plots were set up in four plots of Leucaena leucocephala
seedlings, two of which were interplanted with sweet potato (Ipomea batatas). The plots
have been monitored since 1992. Soil loss details will be available after one year's
monitoring.
ACKNOWLEDGEMENTS
The authors would like to thank the Research Committee, PNG University of Technology,
Lae, for funding this project. Presentation of this paper was made possible through a grant
awarded by the International Tropical Timber Organization (lITO) under its
Fellowship/Small Grant Scheme.
LITERATURE CITED
BESCHTA, R. L. 1985. Conceptual Model of Sediment Transport in Streams. In C. R. Thorne (ed.). Proceeding
of the InteT7Ultionnl Workshop on Problems of Sediment Transport in Gravel-Bed Rivers. Colorado State
University, Colorado.
BLANEY, D. G. & WARRINGTON, G.E. 1983. Estimating Soil Erosion using an Erosion Bridge. US
Department of Agriculture Forest Service. WSDG-TP-OOOO8. 5 pp.
CHANG, M., SAYOK, A.K. & WATTERSTON, K.G. 1992. Comparison of Three Proposed C-Values of the Soil
Loss Equationfor Various Forested Conditions. Proceedings of Conference XXIII. International Erosion Control
Association. Feb 18-21. Reno, Nevada.
CLARK, R. 1974. Erosion Conditions Classification System. US Department of the Interior, Bureau of Land
Management Technical Note 346. BLM Service Center, Denver, Colorado. 47 pp.
DISSMEYER, G. E. 1973. Evaluating the Impact of Individual Forest Management Practices as on Suspended
Sediment, a Presentation at the National Meeting of Soil Conservation. Society of America, Hot Springs,
Arkansas.
FOSTER, G. R. & WISCHMEIER, W.H. 1974. Evaluating Irregular Slopes for Soil Loss Prediction. Transactions
of the American Society of Agricultural Engineers. 17:305-309.
JACKSON, W. L., KNOOP, K., SZALONA, LJ. & HUDSON, S. 1985. A Runoff and Soil-loss Monitoring
Technique using Paired Plots. US Department of the Interior, Bureau of Land Management. Technical Note
368. Denver, Colorado. 20 pp.
LAL, R. 1981. Deforestation of Tropical Rainforest and Hydrologic Problems. In Lal, R. & Russell, E.W. (eds).
Tropical Agricultural Hydrology. John Wiley and Sons Ltd.
LUSBY, G. C. 1979. Effects of Grazing on Runoff and Sediment Yieldfrom Desert Rangeland at Badger Wash in
Western Colorado, 1953-1973. US Geological Survey Water Supply Paper 1532-1. 134 pp.
MCKENZIE, G. D. & STUDLICK, J.R.J. 1979. Erodibility of Surface-mine Spoil Banks in Southeastern Ohio:
An Approximation. JOUT7Ul1 of Soil and Soil Conservation 34:187-190.
MEYER, L. D. 1988. Rainfall Simulators for Soil Conservation Research. Pp. 75-95 in Lal, R. (ed.). Soil Erosion
Research Methods. Soil and Water Conservation Society. Ankeny, Iowa.
MUTCHLER, C. K., MURPHREE, C.E. & MACGREGOR, K.C. 1988. Laboratory and Field Plots for Soil
Erosion Studies. Pp. 9-36 in Lal, R. (ed.). Soil Erosion Research Methods. Soil and Water Conservation
Society. Ankeny, Iowa.
SAYOK, A. K., CHANG, M. & WATTERSTON, K.G. 1993. Forest Clearcutting and Site Preparation on a Saline
Soil in East Texas: Impact on Soil Losses. Paper presented during IAMAP/IAHS'93 Meeting. July 11-23, 1993.
Yokohama, Japan.
SOIL CONSERVATION SOCIETY OF AMERICA 1976. Soil Erosion: Prediction and Control. Special
Publication No. 21. 393 pp.
WISCHMEIER, w. H. & MEYER, L.D. 1973. Soil Erodibility on Construction Areas. Special Report 135.
Highway Research Board. National Academy of Sciences, Washington, D.C. pp. 20-29.
WISCHMEIER, W. H. & SMITH, D.D. 1978. Predicting Rainfall Erosion Losses: A Guide to Conservation
Planning. US Department of Agriculture, Handbook 537.58 pp.
The growth performance of four dipterocarp species under different
terrain and shade conditions
LOK ENG HAl, HASHIM MD. NOOR and FAUZIDAH AHMAD
Forest Research Institute of Malaysia (FRIM), Kepong, 52109 Kuala Lumpur, Malaysia
ABSTRACT. Four dipterocarp species, Shorea curtisii, S. leprosula, S. parvifolia and Dryobalanops
aroTfllltica, were planted on level terrain and on hill sites, with and without the shade of nurse trees
(Gme/ina arborea). Twelve years of growth indicate that trees on level terrain attain greater heights
than those on hill tops. Shade did not significantly influence the performance of the dipterocarps.
Interactions of the three factors, terrain, shade and species, had no influence on height and diameter
growth, but interactions of terrain-shade and terrain-species have a significant affect on survival. Overall
performance in terms of diameter, height and survival of the dipterocarps can be arranged in the order
S. leprosula > D. aroTfllltica > S. parvijolia > S. curtisii.
KEY WORDS: Shorea, Dryobalanops, plantations, tropical forests, Malaysia.
INTRODUCTION
Forest plantations in the tropics play an important role in ensuring a continuous supply of
timber from desired species. Future large-scale forest plantations could supplement the
depleted natural rainforest resources which have been reduced in area due to extensive and
damaging logging activities. In forest plantations the better quality species can be selected,
thus maximising growing space, growth rates and allowing easier manipulation of growing
conditions.
In Malaysia, research into establishing forest plantations was started in FRIM in the early
50's and 70's with the establishment of both exotic and indigeneous species. Based on their
growth data, some commercially important timber species belonging to the Dipterocarpaceae
family have been identified, including species of Dryobalanops, Shorea and Hopea. The
family Dipterocarpaceae comprises about 515 species with 16 genera (Jacobs 1981) and
occurs naturally in the forests of Guyana and Venezuela in South America, as well as Africa
and the Indo-Malayan region (Jacobs 1981, Symington 1941). In Peninsular Malaysia, there
are about 150 species of dipterocarps occurring in lowland forests; the number of species
decreases as altitude increases (Ashton 1964, Symington 1941).
This paper highlights some of the silvicultural requirements of four dipterocarp species
grown in cleared secondary lowland forest.
Location
The experiment was located in Block P, Compartment 10 of Sungai Buloh Forest Reserve,
Selangor, Malaysia (3° 18' N, 101°34'E). The sites were on a south-facing slope which has
annual rainfall of 2613 mm. The lower and upper terrain sites are at an altitude of 50 and
100 m above sea level respectively. The soil is a sandy clay of the Rengam series. The
planting sites were initially clear-felled and burnt in March 1977 but planting was delayed
until November 1977, when suitable planting stock was available; consequently, the prepared
277
278 LOK ENG HAl, HASHIM MD. NOOR AND FAUZIDAH AHMAD
sites were colonised by perennial weeds such as Musa acuminata, Macaranga gigantea,
Mallotus paniculatus, Trema spp. and Mikania micrantha. These weeds were cut down
immedidately prior to planting the dipterocarps.
Experimental design
Planting was based on a square spacing pattern of 2.4 x 2.4 m for the four species of
dipterocarps: S. curtisii, S. leprosula, S. parvifolia and Dyrobalanops aromatica. The
experiment was laid out using a split-split plot design, with terrain (lower and upper slope)
as the main plot, shade (with and without shade tree intercropping) as sub-plots and species
as the sub-sub plot. The design was replicated three times. Fast growing seedlings of
Gmelina arborea were planted between the dipterocarp species. These shade seedlings were
much taller than the dipterocarps when planted in November 1977 and provided some shade
with their sparse crowns.
The initial height of the dipterocarp seedlings was recorded immediately after completion
of planting. Survival and height were monitored quarterly during the first year of
establishment and yearly thereafter. Diameter measurement at breast height (dbh) was
recorded when the trees reached the minimum girth of 3.0 cm. Weeding and climber cutting
was performed regularly.
The growth performance data at age 12 years were analysed using the SAS software
Package. Means of the treatments were compared using Duncan's Multiple Range Test.
Tables 1a and 1b give the summary results of the average diameter, height and survival rate
at both sites and shade conditions. S. leprosula gave the greatest increase in growth at both
sites and shade conditions. The height increments (final-initial heights) of the dipterocarps
were observed to be significantly affected by both terrain and species factors at P = 0.05 and
P = 0.01 respectively but not by shade. There was no differential effect of terrain-species,
shade-species and terrain-shade-species upon the height increment (Table 2).
Table la. Average diameter, height and survival of four Dipterocarp species, 12 years old· level
terrain conditions (lower slope).
Species Diameter (cm) Height (m) Survival (%)

A B A B A B
s. curtisii 13.9 12.7 11.2 12.1 14.6 28.8
s. /eprosula 14.8 15.9 12.9 14.0 62.1 69.6
S. parvifolia 14.7 13.6 10.9 10.6 26.5 39.8
D. aromatica 13.1 13.6 13.1 13.4 57.6 62.9
Table lb. Average diameter, height and survival of four Dipterocarp species, 12 years old - hill
terrain conditions (upper slope).
Species Diameter (cm) Height (m) Survival (%)

A B A B A B
s. curtisii 10.3 13.5 8.7 8.8 35.5 21.0
S. /eprosula 13.1 14.9 10.2 9.7 41.1 31.0
S. parvifolia 10.4 10.8 8.1 7.3 24.4 21.1
D. aromatica 10.1 10.8 9.3 8.8 38.9 41.1
Growth performance of dipterocarps 279
Table 2. Anova result for height increment on four dipterocarps. N.S. = Not Significant; * =
Significant at P = 0.05; ** = Significant at P = 0.0\.
Source of Variation df SS MS F value Remarks

Replicate 2 0.90 0.45 0.18 N.S.
Terrain (n 171.00 171.01 69.39 *
Shade (SH) 1.40 1.40 0.65 N.S.
Species (SP) 3 53.08 17.69 6.51 **
T xSH 8.67 8.67 4.04 N.S.
T x SP 3 3.47 1.16 0.43 N.S.
SH x SP 3 2.78 0.93 0.34 N.S.
T x SH x SP 3 4.94 1.65 0.61 N.S.
Table 3. Anova result for diameter increment on four dipterocarps. N.S. = Not Significant;
* = Significant at P = 0.05.

Replicate 2 0.59 0.29 0.03 N.S.
Terrain (n 97.36 97.36 9.46 N.S.
Shade (SH) 0.25 0.25 0.09 N.S.
Species (SP) 3 73.55 24.52 4.53 *
T x SH 1.09 1.09 0.40 N.S.
T x SP 3 10.33 3.44 0.80 N.S.
SH x SP 3 12.95 4.32 0.80 N.S.
T x SH x SP 3 1.29 0.43 0.80 N.S.
Table 4. Anova result for survival percentage on four dipterocarps. N.S. = Not Significant;
* = Significant at P = 0.05; ** = Significant at P = 0.01; *** = Significant at P = 0.001.

Replicate 2 482.44 241.22 2.41 N.S.
Terrain (n 2073.36 2073.36 7.96 N.S.
Shade (SH) 0.26 0.26 0.00 N.S.
Species (SP) 3 7553.70 2517.90 25.14 ***
T xSH 840.59 840.59 9.13 *
T x SP 3 1932.11 644.04 6.43 **
SH x SP 3 155.46 51.82 0.52 N.S.
T x SH x SP 3 455.03 148.34 1.48 N.S.
Concerning diameter growth, the species had a significant affect at P = 0.05, but terrain
and shade had no significant influence. There were also no differential interactions among
the three factors on the diameter of the trees (Table 3). Survival was significantly influenced
by the species of dipterocarps. However, this was not affected by terrain and shade factors
alone. Nevertheless, the interaction of terrain-shade and terrain-species had a significant
affect on survival of the trees (Table 4). Shade and open plantings made no significant
280 LOK ENG HAl, HASHIM MD. NOOR AND FAUZIDAH AHMAD
Table 5. Duncan new multiple range tests on height, diameter increment and survival percentage of four
dipterocarps. Note: Values in a column with the same letter are not significantly different at P = 0.05.
Height
Species Diameter Survival %
increment (m)
S. leprosula 11.7' 14.7' 50.8'
S. parvifolia 9.3 b 12.4b 26.9b
D. aromatica 11.2' 11.9b 50.0'
S. curtisii 9.5 b 11Ab 23.9b
difference to height or diameter increment. This may have been due to poor growth of the
Gmelina arborea shade trees in the plots, which provided little shade for the dipterocarps.
Of the species tested, S. leprosula and D. aromatica had the greatest height increment,
while S. parvifolia and S. curtisii the least (Table 5). S. leprosula had the greatest diameter
increment, a value which was significantly different from the diameter increments of the
other three species. The outstanding performance of S. leprosula, which is considered as a
late secondary species (Koyama, 1981) was expected since it is a light-demanding fast
growing species and is commonly found in young regenerating forests (Symington 1941).
The species has a very high photosythetic rate (Kayoma 1981, Kira 1982). In fact, Watson
(1935) has reported that S. leprosula was the most vigorous of the light hardwood dipterocarp
species. Watson (1935) recorded an average height of 10.7 m and diameter of 9.7 cm at an
age of seven years for this species, which was better that the mean height recorded in the
present study. He also listed D. aromatica among the fast-growing trees. According to
Walton (1938), the growth of S. parvifolia was slower that S. leprosula but the former was
more shade tolerant. Although S. leprosula surpassed S. parvifolia during the early growth,
the latter is likely to overtake the former after about 25 years (Symington 1941). Both S.
leprosula and S. parvifolia occur naturally on the gentle slopes and in the valleys.
CONCLUSION AND RECOMMENDATIONS
The present study indicated that dipterocarp trees growing on the lower slope gave better
growth and survival than trees growing on the upper hill slope. However, only height
increment was significantly affected. Shade had no significant effect on the growth of the
dipterocarps. There was a significant variation in the performance of the dipterocarp species.
Overall, performance in terms of diameter, height growth and survival of the dipterocarps
can be arranged in the order of S. leprosula > D. aromatica > S. parvifolia > S. curtisii.
Interactions among the investigated factors showed no significant effect on the height and
diameter of the dipterocarps, but interactions of terrain-shade and terrain-species did
influence survival. However, further trials are needed to test the various site effects and
silvicuItural treatments upon the performance of S. leprosula and D. aromatica under
plantation conditions.
ACKNOWLEDGEMENTS
We wish to thank FRIM for sponsorship and help in carrying out the projects. We are also
grateful to our Research Assistant, Mr. M. Markandan, and the staff of the Plantation
Silviculture Unit for their help in plot maintenance and data collection. We would also like
to record our appreciation to Mr. Chew Teng Kok, who initiated the experimental trials, and
Dr. Zakaria Ibrahim for comments on the first draft of this manuscript.
Growth performance of dipterocarps 281
LITERATURE CITED
ASHTON, P.S. 1964. Ecological studies in the mixed dipterocarp forests of Brunei State. Clarendon Press, Oxford.
BURGESS, P.F. 1975. Silviculture in the hillforests of the Malay Peninsula. Forest Reserach Institute, Research
Pamphlet 66. Kepong, Malaysia.
JACOBS, M. 1981. Dipterocarpaceae: The taxonomic and distribution framework. The Malayan Forester
44:168-189.
KAMIS AWANG, MOHO BASRI HAMZAH, TAJUDDIN BAHAROM & SAHIR OTHMAN 1981. Effects of
some environmental factors on the growth of three hill forest dipterocarps. BIOTROP Special Publication
13:9-21.
KAYOMA, H. 1981. Photosynthetic rates in lowland rainforest trees of Peninsular Malaysia. Japanese Journal of
Ecology 31:361-369.
KIRA, T. 1982. Patterns of productivity in tropical trees and forests. Proceedings of the workshop an Ecological
basis for rational resource utilization in the humid tropics of South East Asia. UPM Serdang.
LONDON, F.H. 1948. Planting experiments with Kapur. The Malayan Forester 11:146-153.
SYMINGTON, C.F. 1941. Forester's manual of dipterocarps. Malayan Forest Record 16.
VINCENT, A.J. 1961. A note on the growth of three meranti (Light Heavy Woad Shorea) hill forest species in
naturally and artificially regenerated forest of Malaya. Forest Research Institute - Forest Department Federation
of Malaya Research Pamphlet.
WALTON, A.B. 1938. Some notes on nursery and planting technique. The Malayan Forester 7:122.
WATSON, J.G. 1935. Plantation experiments at Kepong. The Malayan Forester 4:110-119.
WHITTY, R.H. 1932. The use of nursery transplants in planting. The Malayan Forester 1. 228 pp.
Variation in reproductive success of gynodioecious figs (Ficus spp.,
Moraceae) and their pollinators (Hymenoptera: Agaonidae) in New
Guinea
GEORGE D. WEIBLEN
Cape Tribulation Tropical Research Station, P.M.B. 5, Cape Tribulation via Mossman, 4873,
Queensland, Australia
ABSTRACT. Variability in ovary utilization by figs and their obligate wasp pollinators was measured in
middle and lower montane rainforest in New Guinea. Examination of fruit crops from individuals of nine
gynodioecious tree species revealed unprecedented levels of variation in seed set and wasp predation.
Approximately one-third of all fig flowers were preyed on by wasps. In female figs, the proportion of
flowers containing seeds ranged 18 - 97% among the species sampled. In hermaphroditic figs, the
proportion of seeds killed by wasps ranged from 12 - 87 %. Contrary to expectation, results suggest that
fig and wasp reproductive success is positively related in gynodioecious species. The ecological
significance of fig reproductive output to tropical rain forests is discussed.
KEY WORDS: Ficus, Agaonidae, gynodioecy, pollination, seed predation, New Guinea.
INTRODUCTION
The pollination of figs (Ficus Spp., Moraceae) by species-specific wasps (Hymenoptera:

Agaonidae) is arguably the most widely known example of an obligatory mutualism between
plant and pollinator (cf. Bronstein 1992). Evidence of extreme host-specificity of agaonids
to figs (est. 800 species worldwide) suggests that figs and their exclusive wasp pollinators
have evolved in parallel (Wiebes 1979). The persistence of the mutualism in some tropical
rain forests has been identified as critical to community-level conservation (McKey 1987,
Terborgh 1986). Figs are food for a range of frugivorous animals during periods of scarcity,
and this role as possible 'keystone' species stems from the peculiar nature of fig pollination.
The fig/pollinator interaction is thoroughly described in Bronstein (1992). In monoecious
species, flowers are arranged inside a closed receptacle (syconium) which is accessible only
through a bract-lined opening (ostiole). This arrangement of flowers is technically an
inflorescence, although ecologically it is a fruit. In a young fig, hundreds of female flowers
line the internal wall of the syconium. Pollen-carrying female wasps enter syconia through
the ostiole near the time of female flower receptivity. Inside the fig cavity, female wasps
deposit eggs in fig ovaries and simultaneously pollinate the flowers. Developing seeds are
consumed by wasp larvae. As the fig ripens, adult female wasps hatch into the fig cavity
after being fertilized by wingless males, which die shortly after chewing a hole through the
ostiole. Females collect pollen from male flowers usually clustered around the ostiole while
leaving the fig. Pollen-carrying females then find receptive figs in which to oviposit, thus
completing their life cycle and at the same time achieving pollination.
Wasp survival is ensured by the availability of figs at differing stages of development.
Flowering asynchrony at the population-level is therefore common in figs (Janzen 1979a),
although there are exceptions (Kjellberg et al. 1987). Population asynchrony can thus
provide ripe figs for frugivores at times when food is otherwise scarce (Terborgh 1986).
Maintenance of this keystone species interaction obviously depends on continued reproductive
283
D. S. Edwards et aJ. (eds.), Tropical Rainforest Research - Current Issues, 283-289.
284 GEORGE WEIBLEN
success of the fig and its pollinator. However, natural history suggests that a fundamental
conflict of interest imposes reciprocal constraints on the success of each partner. Seeds
containing wasp larvae are destroyed and consequently, wasps are both pollinators and seed
predators. Viable seed production in figs is therefore constrained by the percentage of
ovaries used by the wasps. The conflict of interest is determined by a single floral trait, the
length of the style (Bronstein 1992, Janzen 1979a). The ovaries of short-styled flowers are
accessible to wasp ovipositors while those of long-styled flowers are not. Wasps oviposit in
both types but successfully leave offspring in only the short-styled flowers. Janzen (1979b)
was first to ask: what proportion of fig ovaries are donated to wasps in exchange for seeds?
Bronstein (1992) further defined the tradeoff between seeds and wasps based on the premise
that they alternately use fig ovaries. The expectation that wasp and seed maturation are
negatively related can be tested by measuring reproductive success (i.e. numbers of
offspring) using fig ovaries as 'common currency'.
This approach has not yet been applied to gynodioecious figs, which comprise over 300
species in the Old World tropics (Berg 1987). Gynodioecious populations consist of trees
whose figs contain both male and female flowers (hermaphrodites) and trees with only female
flowers (females). Short- and long-styled flowers occur together in monoecious figs, but
gynodioecious figs are completely heterostylous. Female trees have long-styled flowers and
hermaphroditic trees apparently have only short-styled flowers. Accordingly, females set
seed and hermaphrodites produce mostly wasps (Verkerke 1987). Hermaphrodites producing
few or no seeds are functionally male.
The conflict of interest for functionally dioecious species is different than for their
monoecious relatives. Note that in dioecious species, wasps pollinating female figs leave no
offspring. There should therefore be strong evolutionary pressure for wasps to exploit
hermaphroditic trees while avoiding lethal female trees (Kjellberg et at. 1987). I tested this
possibility by measuring variation in seed set and wasp maturation in pairs of hermaphroditic
and female trees. This comparative approach, using fig ovaries as 'currency', allowed me
to investigate fig and wasp success for nine gynodioecious species in New Guinea.
METHODS
Fruiting trees were sampled in disturbed and undisturbed forests around Wau, Papua New
Guinea during October and November of 1992. Sampling localities included the Wau
Ecology Institute arboretum, the slopes of Mount Kaindi (r 24' S, 146° 44' E), and the
Kuper Range Wildlife Conservation Area (8° 77' S, 146° 80' E). Figs were collected in
lower montane (1100 - 1300 m altitude) and mid-montane forests (1300 - 1800 m altitude).
The climate in this area is highly humid and quite aseasonal, with rainfall distributed
somewhat evenly throughout the year. Lower and mid-montane forests on Mount Kaindi
receive approximately 2000 and 3000 mm of rain per year, respectively. Mean annual
temperature is 21°C at 1200 m and falls to around 16°C at 1800 m altitude. Average daily
temperature during the year is nearly constant.
Natural vegetation in the study area is mixed Lithocarpus-Castanopsis rain forest in the
mid-montane and mixed rain forest dominated by Araucaria in the lower montane zone
(Gress itt & Nadkarni 1978). However, anthropogenic grassland has replaced much of the
lower montane forest due to extensive clearing and burning. Fig trees were especially
common in regrowth forests and abandoned gardens, where they comprised up to 25 % of the
standing trees with diameters exceeding 10 cm (personal observation).
Pairs of hermaphroditic and female trees representing nine gynodioecious species were
sampled. I collected figs when ripe over an 8-week period. Ten figs from each of 18 trees
were stored in 70% ethanol. Figs were later sectioned into quarters along the longitudinal
Gynodioecious figs and their pollinators 285
axis and one quarter was chosen at random for dissection under 15x magnification. There
were three visually distinguishable classes of fig seeds. (1) Mature seeds were white-to-
yellow with hard seed coats. (2) Seeds killed by wasps had conspicuous exit holes in the
seed coats, were dark brown in color, or contained dead wasps. (3) Abortive ovaries were
considerably smaller than (1) and (2). I counted totals for each class and calculated
percentages for each based on the total number of ovaries in the sample.
Table 1 lists collection numbers for fig species employed in the study. Species
identification required considerable effort due to the unresolved status of Ficus taxonomy in
New Guinea. Many of the 138 fig species recorded for New Guinea are known from only
one or a few specimens. Nearly all species selected for this study were common around Wau
and well-represented in the PNG National Herbarium collection. Plant specimens were
identified using the diagnostic keys of Corner & Henty in Johns (1988) and checked against
the reference collection at the National Herbarium, where vouchers are deposited. Pollinator
species (also listed in Table 1) were later checked against Wiebes (1982).
RESULTS
An average of 65 percent of ovaries were occupied by seeds (33%) or wasps (32%), while
the remaining 35 percent were aborted (Table 2). The mean percentage of fig ovaries
producing both seeds and wasps ranged from 21±7 (X±SD) in F. trachypison to 96±3 in
F. complexa. Seed set averaged 66±26 percent in female figs and 7±7 percent in
hermaphroditic figs (Figure 1). Wasp larvae exploited an average of 58±24 percent of
hermaphroditic fig ovaries. The percentage of wasps in hermaphroditic figs and of seeds in
female figs showed wide variation among the sampled species. Female and hermaphroditic
trees in F. complexa produced the highest percentage of seeds and wasps on average (97 ±2
and 87 ±5 percent) and F. trachypison matured the lowest percentage of seeds and wasps
(18±5 and 12±6 percent).
Table 1: Gynodioecious figs (Ficus spp.) and their wasp pollinators (family Agaonidae) collected at Wau, Morobe
Province, Papua New Guinea.
Fig Collection N" Collection N° Pollinator

Female Bisexual
Ficus complexa Corner 074 103 Ceratosolen gressitti Wiebes

F. erythrosperma Miq. 049 072 undescribed
F. mollior F.v. M. ex Benth. 112 071 C. medlerianus Wiebes
F. nodosa Teysm. et Bhm. 002 065 C. nexilis Wiebes
F. robusta Corner 068 067 undescribed

F. septica Burm. f. 085 004 C. bisulcatus Mayr
F. trachypison K. Schum var. 005 108 undescribed

pa/lida Corner
F. virgata Reinw. ex BI. var. 076 101 Liphorrhopalum sessilis Hill

sessilis (Bur.) Corner
F. wassa Roxb. 003 089 Blastophaga wassae Wiebes

286 GEORGE WEIBLEN
Table 2: Mean (±SD) fig diameter in millimetres and total female flowers per fig for nine species. Total flowers
per fig was calculated by multiplying the number of flowers per sample by a factor of 4. Percent ovary use was
calculated by pooling data from hermaphroditic and female trees of the same species (N = 20).
Ficus species diameter flowers per fig % ovary use
complexa 20 (2) 668 (87) 96 (3)

erythrosperma 10 (1) 449 (40) 72 (6)
mollior 19 (I) 828 (119) 36 (4)
nodosa 34 (1) 2648 (231) 63 (8)
robusta 46 (1) 3515 (282) 83 (5)
septica 18 (1) 576 (55) 51 (6)
trachypison 15 (1) 616 (44) 21 (7)
virgata 12 (1) 503 (59) 86 (8)
wassa 12 (2) 513 (138) 79 (9)
Average (N = 180) 21 (11) 1146 (1072) 65 (24)
The mean percentage of aborted fig ovaries was remarkably similar among female and
hermaphroditic figs (34±26 and 35 ±23 percent), despite a highly variable percentage among
species. Vacant ovaries ranged 2-82 percent for females and 5-75 percent for
hermaphrodites.
average ~·ij .~ ....~ ..~ .. ~ ...~.....~ ..~ ... ~ ...•...;....~ ..~ ....~ .. ~·t; ~=i:
... :.;~' ~·:=Jt'." ·......·.····.·M•...... ························1····················
wassa
virgata
trachypison
septica
robusta
nodosa
mollior
erythros perma
complexa ~i
o 10 20 30 40 50 60 70 80 90 100
Percentage ovary use
.. % seeds (females)
m % wasps
o % seeds (hermaphrodites)
Figure I: Levels of seed set and wasp predation for nine gynodioecious figs in New Guinea. Average seed set for
female and hermaphroditic trees for each species is listed separately (N = 10). Error bars indicate 95 % confidence
intervals.
Average percent seed and wasp maturation varied significantly among sampled species
(one-way ANOV As, F = 200.2, P < 0.001 and F = 353.4, P < 0.001). There was a
strong positive correlation between percent seed production in female figs and percent wasp
predation in hermaphroditic figs (r = 0.882, P < 0.001, N = 90). Fig diameter was
obviously correlated with numbers of female flowers per sample (r = 0.674, P < 0.001, n
= 161), but interestingly, female figs were on average significantly larger than
hermaphroditic figs (mean difference = 0.773 mm, ANOVA, F = 33.7, P < 0.001).
DISCUSSION
Although wasp pollinators had access to every ovary within hermaphroditic figs, nearly 35
percent remained unoccupied on average. In some hermaphroditic figs, a fraction of the
unused ovaries produced seed, suggesting that wasps pollinate slightly more flowers than they
occupy. There are several possible reasons why wasps may not successfully exploit all
available ovaries. (1) Wasps may fail to pollinate or oviposit in all flowers (i.e. too few
'foundresses'). (2) Limitation of plant resources may limit the number of wasps a fig can
support (Bronstein 1988a). (3) Predators and parasites of fig wasps may limit maturation
success. Bronstein (1988b) suggested that factors extrinsic to the mutualism, such as ant
predation, can impose constraints on wasp and fig reproductive success. There are presently
not enough data available to conclude exactly which factors determine the relative success
of fig and wasp.
The seemingly logical tradeoff between seed and wasp production draws no support from
the findings for figs in New Guinea. I found that wasp and seed maturation was positively
correlated among nine species sampled. Surprisingly, this result for gynodioecious species
is consistent with the findings of Bronstein (1986) for monoecious F. pertusa. We based our
expectation that wasp and fig success should be negatively related on the premise that the
partners alternately exploit fig ovaries. In the case of gynodioecy, however, it is false to
assume that wasp overexploitation in hermaphroditic figs is disadvantageous. Increased wasp
production enhances male fig reproductive success by increasing the amount of pollen
donated to female figs. I found that nearly 90% of the ovaries were exploited in some
hermaphroditic figs. Hermaphrodites therefore appear to donate the bulk of their ovaries to
wasps (and lose female fitness) in exchange for increased male fitness.
I found that hermaphrodites occasionally set a few seeds and that females often contain
the remains of aborted anthers clustered around the ostiole. The presence of relictual
'maleness' and 'femaleness' in both sexes conforms to the taxonomists' belief that
gynodioecious figs evolved from monoecious ancestors (Wiebes 1979). A simple change in
style length could facilitate the transition to dioecy, which is thought to have occurred at least
twice (Berg 1987). The evolution of dioecy would have replaced the potential for a
fig/pollinator tradeoff in ovary utilization with a different conflict of interest.
Conflict between gynodioecious figs and their pollinators results from the separation of
seed and wasp production among two sexes. Wasps leave no offspring in female figs, but
have access to every ovary in hermaphroditic figs. Thus, the seed-producing (female) trees
are evolutionarily lethal to their pollinators. It seems logical to suppose that strong selection
against pollinators of female trees would favor wasps with a means of discriminating between
the sexes. Individual wasps visiting only hermaphroditic figs would achieve higher
reproductive success. I found that ovary utilization among the sexes were remarkably similar
across nine species (Figure 1). Surprisingly, there were more seeds on average in female
figs than wasps in hermaphrodites. These results suggest that wasps are efficient pollinators
of both sexes, although pollination efficiency for females appears slightly higher than for
hermaphrodites. Considering the obvious disadvantage to individual pollinators of female
288 GEORGE WEIBLEN
figs, what explains the apparent stability of gynodioecious fig pollination?

The persistence of over 300 gynodioecious species and their pollinators is possibly
explained by flowering phenology at the population level. Kjellberg et al. (1987) showed
for gynodioecious F. carica that the majority of wasps emerged from hermaphroditic figs
when the female figs were receptive. Asynchronous flowering among the sexes can thus
eliminate selection against wasps which pollinate females. Kjellberg et al. (1987) suggest
that sexes flower at different times in seasonal environments, where the relative costs of wasp
and seed production vary throughout the year. However, this explanation does not account
for dioecious figs in aseasonal environments. The New Guinea highlands climate, for
example, may not offer the necessary environmental cues for flowering synchrony. Answers
to this question for figs in aseasonal climates will remain a matter of speculation until new
information is available. Only a single study at present has detailed the wasp associations
of a fig species in New Guinea (Godfray 1988) and there are no published data on flowering
phenology. Fruit production in gynodioecious F. variegata in nearby tropical North
Queensland, however, appears to be highly seasonal (H. Spencer, unpublished data).
Fig reproductive output is of critical importance to conservation in some tropical forests
where figs are a keystone resource for frugivorous animals (Terborgh 1986). The
observation that female figs are on average slightly larger and qualitatively sweeter than
hermaphroditic figs is consistent with a hypothesis that seed-bearing female figs are more
attractive to animal dispersers (H. Spencer, personal communication). Frugivores in New
Guinea include megabats, marsupials, parrots, fruit doves, and birds of paradise. The
production of female figs as food for these animals of high conservation value appears to
depend on the successful exploitation of hermaphroditic trees by pollinating wasps. Whether
this food resource is available continuously or patchily during the year has yet to be
determined. .
In conclusion, this study suggests that for gynodioecious species: (1) wasps do not
successfully exploit all available fig ovaries; (2) figs do not show an evolved means of
lowering wasp success in overexploited syconia, and (3) fig and wasp maturation success are
positively correlated. These results contribute to a collection of data which has failed to find
evidence for a simple trade-off between figs and their pollinators. The persistence of the fig-
pollinator interaction in New Guinea rain forests currently under threat could decide the fate
of a host of other important species.
ACKNOWLEDGMENTS
The Institute of Papua New Guinea Studies, Wau Ecology Institute and PNG National
Herbarium made the study possible. Many thanks to Harry Sakulas and Robert Kiapranis.
I am grateful for criticism and discussion provided by Yves Bassett and Robert Bino. I am
also indebted to 'papa bilong graun' who allowed me access to their lands and Deren and
Roy who climbed the tall trees. The study was supported by a post-graduate fellowship from
the Thomas Watson Foundation.
LITERATURE CITED
BERG, C. C. 1987. Classification and distribution of Ficus. Experientia 45:605-611.

BRONSTEIN, J. L. 1986. Coevolution and constraints in a neotropical jig-pollinator wasp mutualism. Ph.D.
Dissertation, University of Michigan, Ann Arbor.
BRONSTEIN, J. L. 1988a. Limits to fruit production in a monoecious fig: consequences of an obligate mutualism.
Ecology 69:207-214.
BRONSTEIN, 1. L. 1988b. Predators of fig wasps. Biotropica 20: 215-219.
BRONSTEIN, J. L. 1992. Seed predators as mutualists: ecology and evolution of the fig/pollinator interaction. Pp.
1-47 in Bernays, E. A. (ed.). Insect-Plant Interactions Vol. 4. CRC Press, Boca Raton, Florida.
GODFRAY, H. C. J. 1988. Virginity in haplodiploid populations: a study on fig wasps. Ecological Entomology
13:283-29l.
GRESSITT, J.L. & NADKARNl, N. 1978. Guide to Mt. Kaindi: Background to Montane New Guinea Ecology.
Wau Ecology Institute Handbook No.5. Wau, Papua New Guinea, 135 pp.
JANZEN, D. H. 1979a. How to be a fig. Annual Review of Ecology and Systematics 10:13-5l.
JANZEN, D. H. 1979b. How many babies do figs pay for babies? Biotropica 11:48-50.
JOHNS, R.1. 1988. The Flowering Plants of Papuasia: Hamamellidae. P.N.G. University of Technology, Lae,
Papua New Guinea, 139 pp.
KJELLBERG, F., GOUYON, P. H., IBRAHIM, M., RAYMOND, M., & VALDEYRON, G. 1987. The stability
of the symbiosis between dioecious figs and their pollinators: a study of Ficus carica L. and Blastophaga psenes
L. Evolution 41:693-70l.
McKEY, D. 1987. Population biology of figs: applications for conservation. Experientia 45:661-673.
VERKERKE, W. 1987. Structure and function of the fig. Experientia 45:612-62l.
WIEBES, 1. T. 1982. Fig wasps (Hymenoptera). pp. 735-755 in Gressitt, J. L. (ed.). Biogeography and Ecology
of New Guinea. Vol. 2. W. Junk Publishers, The Hague.
WIEBES, J. T. 1979. Co-evolution of figs and their insect pollinators. Annual Review of Ecology and Systematics
10:1-12.
Interactions between lycaenid butterflies and ants in Peninsular Malaysia
KONRAD FIEDLER
Zoologisches lnstitut II, Universitiit Wurzburg, Am Rubland, D-97074 Wiirzburg, Federal

Republic of Germany
ABSTRACT. Interactions between Iycaenid butterflies and ants were studied in West Malaysia. The
majority of species observed so far are facultative myrmecophiles that interact unspecifica1ly with a broad
range of ant taxa. A few cases of obligate myrmecophily have been detected, including the polyphagous
herbivore Rapala dieneces and two aphytophagous members of the genus Logania (Miletinae).
Myrmecoxeny, i.e. lack of ant-association, is rare among Oriental Iycaenids, but was confirmed for two
species. Together with data from available literature, these findings suggest that in tropical rainforests
the proportion of obligate myrmecophiles is not much higher than in the northern temperate zone.
Therefore, myrmecophily has probably contributed little to the speciation events themselves, but its role
in maintaining species diversity through resource partitioning and niche segregation remains to be
discovered.
KEY WORDS: Ants, butterflies, evolution, Lycaenidae, Malaysia, mutualism, myrmecophily, rainforest,
species diversity, zoogeography.
INTRODUCTION
Interspecific, and in particular mutualistic, interactions between organisms of very different

kinds are extremely widespread and highly important in tropical ecosystems. Such symbiotic
interactions are thought to playa significant, but still insufficiently understood role in the
evolution and maintenance of species diversity (e.g. contributions in Huxley & Cutler 1991).
Ants are well-known for their involvement in various interactions with plants, Homoptera and
a large guild of myrmecophilous arthropods (H6lldobler & Wilson 1990). An interesting
paradigm for ecological and evolutionary studies on myrmecophily is provided by the
butterfly family Lycaenidae. Many lycaenid species are myrmecophilous, i.e. they live in
association with ants. In general, myrmecophily occurs in the later larval stages, but in
certain taxa interactions with ants also involve the pupae and/or adult butterflies. The larvae
and pupae of lycaenids possess a variety of epidermal glands the secretions of which
manipulate ant behaviour and thereby enable the caterpillars to enter into persistent
associations with ants (Cottrell 1984, Fiedler 1991a, Malicky 1969, Pierce 1989).
It has been suggested by various authors that myrmecophily has made an important
contribution to the evolution and diversification of lycaenid butterflies (e.g. Atsatt 1981).
Specifically, Pierce (1984) described two possible ways in which ants could influence the
evolution of diversity in the Lycaenidae: through increase of 'oviposition mistakes' in relation
to ant-dependent host plant selection, and through local fragmentation of populations that
require both specific host plants and host ants. Both mechanisms only apply to obligatorily
myrmecophilous Iycaenids, i.e. butterfly species that depend on specific host ants for survival
(see Fiedler 1991a). A zoogeographic survey by Pierce (1987) provided supporting evidence
that cases of very close, obligate myrmecophily are particularly common on the southern
hemisphere. Her sample mostly covered subtropical faunal regions (India, Australia, South
Africa) with limited representation of tropical rainforests.
It therefore remains to be tested whether myrmecophily has played an equally important
role in the evolution of lycaenids in tropical rainforests. This biome supports by far the
291
292 KONRAD FIEDLER
largest number of species. From West Malaysia, for example, 395 Lycaenidae species are
recorded (excluding the Riodinidae; Eliot 1992), and Seki et al. (1991) list 379 species for
the island of Borneo alone. If the scenario of Pierce (1984, 1987) is true for the centre of
the Oriental faunal region, then one should postulate that the majority of Malaysian lycaenid
butterflies are myrmecophilous. Certainly a greater proportion of obligate myrmecophiles
is to be expected in comparison with, for example, the northern temperate Holarctic region.
In addition, many species should either be local endemics or should occur only in scattered,
fragmented populations.
A complete analysis of this question is still out of reach, because life-history data are as
yet available for only a fraction of Malaysian lycaenid butterflies. During the past decade,
however, intensive efforts have been made to collect larger sets of pertinent biological data
in Peninsular Malaysia. In this paper I will firstly outline some of the most recent findings,
then discuss these observations in terms of the importance of myrmecophily and species
diversity. Questions of endemism and population structure will not be addressed in this
paper.
My observations were conducted in the vicinity of the Ulu Gombak Field Studies Centre
of the Universiti Malaya (20 km north of Kuala Lumpur), in an advanced secondary
dipterocarp forest at 200-400 m elevation.
OBSERV ATIONS
Facultative myrmecophiles
Lycaenidae species whose larvae are unspecifically visited by various ant genera are
termed facultative myrmecophiles. Such species are not fully dependent on ant-association
for their survival, although tending ants can provide substantial fitness benefits (Pierce &
Easteal 1986). Moreover, ant-associations are not necessarily universal or permanent in
facultative myrmecophiles. A fraction of larvae can usually be found without attendant ants.
Among facultative myrmecophiles, the intensity of ant-interactions varies from very loose and
casual associations to almost persistent tending. Facultative interactions with ants are the
dominant type of myrmecophily among Holarctic Lycaenidae (Fiedler 1991b).
During my fieldwork in West Malaysia I documented a number offacultative, non-specific
ant-associations. These included Actyolepis puspa (Horsfield 1828), 7 species of lamides,
Prosotas dubiosa (Semper 1879), Surendra florimel Doherty 1889, Arhopala buddha
Bethune-Baker 1903 and A. metamuta (Hewitson 1863), Eooxylides tharis (Geyer 1837),
Drupadia ravindra (Horsfield 1828), Hypolycaena othona (Hewitson 1865), Rapala iarbus
(Fabricius 1787) and R. pheretima (Hewitson 1863) (cf. Fiedler 1992a, b & in press a, b).
Within the genus lamides, there is no evidence that myrmecophily has played any role in
speciation. In all lamides species with documented life-histories, the caterpillars possess a
full set of myrmecophily organs and are facultatively tended by an overlapping range of ant
taxa (especially Crematogaster, Rhoptromyrmex, Technomyrmex, and Camponotus). No
preferences of particular lamides species for certain ants could be observed during
oviposition or in the larval stage.
Weak and casual ant-associations occur in Prosotas dubiosa and Hypolycaena othona.
Contacts with ants observed in the field were usually short, although in captivity more or less
stable tending behaviour could be induced in ants such as Crematogaster or Meranoplus
mucronatus .
Obligate myrmecophiles
Only a few species with an obligate symbiosis with ants have yet been found in West
Malaysia. Prominent and well-documented examples include Anthene emolus (Godart 1824)
Interactions between lycaenid butterflies and ants 293
and Hypolycaena erylus (Godart 1824) both of which are exclusively associated with the
weaver ant, OecophyUa smaragdina (Jacobson 1912, Fiedler & Maschwitz 1989a). In both
species, oviposition occurs on a variety of plants, but always in the presence of the host ant.
The larvae are constantly tended by weaver ants throughout their development, and the ants
carry the caterpillars to feeding places without harming them. A new case of obligate
myrmecophily was observed in Rapala dieneces (Hewitson 1878). Caterpillars of this species
feed on a variety of host plants (young trees and shrubs in the Fabaceae, Mimosaceae,
Caesalpiniaceae, Annonaceae, Myrtaceae, Sapindaceae and further families), where they are
invariably visited by 10-40 worker ants of an Iridomyrmex species (Dolichoderinae). If
disturbed, these ants are highly aggressive and might be able to ward off potential enemies
of the caterpillars (in spite of this, tachinid flies have repeatedly been bred from
field-collected larvae). A first instar larva was once observed being carried by its host ants
as if it was ant brood, and larvae as well as pupae can be found in the galleries of
Iridomyrmex ants (A. G. Orr, personal communication). Another new obligate myrmecophile
record with a similar life-history is the polyphagous Drupadia theda (C. & R. Felder 1862)
whose larvae are specifically associated with a black Crematogaster species.
Myrmecoxenous species
Lycaenid larvae that do not possess a dorsal nectary organ have been termed
myrmecoxenes by Kitching & Luke (1985). Myrmecoxenous caterpillars are hardly ever
found in stable associations with ants. Hence, such larvae are not protected by ants against
enemies. Whilst myrmecoxeny is associated with .a number of temperate zone lycaenids,
very few confirmed records of tropical myrmecoxenes are available. In Malaysia, I have
observed Cheritra Jreja (Fabricius 1793) and two Curetis species (santana (Moore 1858),
bulis (Westwood 1851» whose legume-feeding larvae are not tended by ants. In addition,
the Rhamnaceae-feeding Caleta roxus (Godart 1824) has turned out to be myrmecoxenous;
the larvae have neither a dorsal nectary organ nor eversible tentacle organs, and they are
never tended by ants. Even in captivity, they were constantly ignored by Iridomyrmex and
only very briefly antennated by Rhoptromyrmex wroughtonii (Fiedler in press a). For
systematic reasons, myrmecoxeny in Ch. Jreja and C. roxus must be seen as a secondary
character. In closely related genera like Drupadia or Discoiampa, caterpillars are
myrmecophilous.
The selective forces that have repeatedly promoted the loss of myrmecophily in a number
of lycaenid lineages are still poorly understood. This phenomenon of secondary
myrmecoxeny is particularly unexpected in tropical rainforests where ants are virtually
omnipresent in large numbers.
Carnivorous species
In the almost exclusively paleotropical subfamily Miletinae, the caterpillars feed on
sap-sucking Homoptera (aphids, coccids, or membracids). Some species even drink
honeydew or feed on ant brood (Cottrell 1984). Certain Miletinae larvae are not tended by
ants, but are largely ignored, even when feeding in trophobiotic associations of homopterans
with ants (e.g. some Allotinus species, Spaigis, Taraka). All known Miletinae larvae lack
a nectary organ and hence are morphologically myrmecoxenous. Some Miletinae taxa,
however, are true myrmecophiles and may even be intensively tended.
Two novel examples of obligate myrmecophily in the subfamily Miletinae have been
discovered. In the Oriental genus Logania, the caterpillars as well as the pupae live in
permanent association with specific host ants (Fiedler 1993). The ants constantly harvest
secretions from specialized epidermal organs. These organs occur on the prothorax and the
seventh abdominal segment and consist of pairs of lateral hemispherical bulges equipped with
dense clusters of pore cupola organs. In Logania malayica Distant 1884, young larvae are
294 KONRAD FIEDLER
occasionally fed by the host ant, Rhoptromyrmex wroughtonii, with trophallactic

regurgitations. Older larvae drink extrafloral plant nectar as well as honeydew from aphids
and membracids. They also prey upon these homopterans. The adult butterflies of L.
mnlayica spend their whole life, including feeding, mating and oviposition, in association
with these host ants. Adults drink extrafloral nectar and homopteran honeydew, but never
visit flowers. Like the caterpillars, the butterflies are never attacked by the ants.
A closely related species, Logania mnrmorata (Moore 1884), has a very similar life cycle,
but utilizes a different, unrelated host ant (HypocZinea thoracicus [Dolichoderinae]).
Associations with specific ants have also been found in certain Allotinus species (with
Anoplolepis or Myrmicaria) and in Miletus (with HypocZinea; see Fiedler & Maschwitz
1989b; Maschwitz et al. 1985, 1988). Therefore, specializations towards particular host ants
may have contributed to the radiation of Miletinae butterflies in South East Asia.
DISCUSSION
The observations summarized here indicate that in oriental rainforests facultative

myrmecophily is prevalent among the Lycaenidae butterflies. This is consistent with available
life-history data extracted from the literature, covering 112 of the 395 Lycaenidae species of
Peninsular Malaysia (28.4%; see Eliot 1992, Fiedler 1991a). Of these 112 species, 63
(56.3%) are facultative myrmecophiles, 15 (13.4%) are myrmecoxenous and only 21 (18.8%)
are obligatorily associated with specific ants. For 10 additional species the degree of
myrmecophily is unknown, but they are probably facultative myrmecophiles as are their
congeners. Three species are so poorly known that not even a tentative scoring of their
relationship with ants is possible. In all, 75-85 % of the species with recorded life-histories
are associated with ants, but less than one fifth are specifically dependent on ants.
If only the most species-rich subfamily Lycaeninae (sensu Scott & Wright 1990, as
adopted by Eliot 1992) is considered, the pattern becomes even more obvious. Of 93 species
for which information is available, 61 (65.6%) are facultative myrmecophiles, 8 (8.6%) are
myrmecoxenous, and 11 (11. 8 %) are obligatorily myrmecophilous. Again 10 additional
species with insufficiently well-recorded life-histories are most likely to be facultative
myrmecophiles, while the status of 3 species must be left open. These figures are very
similar to the results of an analysis of the Palaearctic fauna (Fiedler 1991b), where the
proportion of all myrmecophiles was about 80%, and obligate myrmecophiles accounted for
roughly 10% of the entire Lycaenidae fauna.
There are two alternative explanations for this remarkable similarity in the proportions of
myrmecophiles when comparing the Oriental and Palaearctic regions:
1) The present knowledge of Oriental lycaenids is too poor and strongly biased towards
species with facultative ant-associations, possibly because of a still scanty representation
of canopy species in the available data base. If so, one must then postulate that further
intensive fieldwork will elucidate a considerable, and proportionally higher, number of
additional obligate myrmecophilous systems.
2) The pattern is real, indicating quite a low proportion of obligate myrmecophiles in
tropical rainforests in contrast to the semi-arid biotas of South Africa and Australia.
At present, it is impossible to definitively decide between these alternatives. Three lines of
evidence, however, support the second view.
a) Obligate myrmecophiles are particularly attractive and apparent to students of natural
history, hence such interactions are more likely to be reported, at least anecdotally,
whereas even in well-surveyed areas like Europe or North America numerous facultative
ant-associations have gone undetected for a long time.
b) Obligate and specific interactions with ants are much more likely to evolve in relation
Interactions between Lycaenid butterflies and ants 295
to abundant and dominant ant species that form large, long-lived colonies (Holldobler &
Wilson 1990). In the extremely complex and dynamic ant mosaics of tropical rainforests,
only a few ant species, such as OecophyLLa smaragdina, are sufficiently dominant or occur
predictably to foster the establishment of stable, specific symbioses.
c) A comparative survey of myrmecophily data from other tropical rainforest areas
(Africa, South America) revealed that in these regions, as well, the proportions of obligate
myrmecophiles among lycaenid butterflies appear to be quite low (Fiedler 1991a).
Symbioses with ants might therefore have been less important in the radiation of tropical
rainforest lycaenid butterflies than believed by some authors. Nevertheless, specific
myrmecophily systems have repeatedly evolved in certain taxonomic subgroups. In the
subfamily Miletinae there is increasing evidence that specific ant-associations are an
important facet of their life-history patterns, and investigations into the role of
ant-associations for resource partitioning and niche segregation among tropicallycaenids have
just started. If myrmecophily has played only a limited role in speciation and radiation (i.e.
in the origin of species diversity), this does not necessarily imply it has a minor role in the
maintenance of species richness.
As a next step, a more complete inventory of individual life-histories is required,
especially in the highly diverse faunas of tropical rainforests. In parallel, detailed case
studies should focus on the mechanisms that allow coexistence of large species numbers with
similar myrmecophily systems. In combination with other studies on interspecific
interactions (e.g. ant-plant symbioses), investigations into lycaenid myrmecophily may then
aid in the understanding of the general significance of mutualistic interactions in terms of
the evolution and maintenance of tropical biodiversity.
ACKNOWLEDGEMENTS
My sincere thanks go to Dr. Azarae H. Idris and Prof. Dr. Yong Hoi Sen (Universiti
Malaya, Kuala Lumpur) for their hospitality and help during my visits to the Ulu Gombak
Field Studies Centre. Prof. Dr. Ulrich Maschwitz (University of Frankfurt) first introduced
me to the biology oftropicallycaenids, and Lt. Col. John N. Eliot (Taunton, UK) graciously
provided his unsurpassed expertise in the systematics of oriental lycaenids. Thanks also to
Jutta Klein for her steady assistance during my studies. My fieldwork was generously
supported by a post-doctoral fellowship from the Leibniz Award of the Deutsche
Forschungsgemeinschaft (DFG) to Prof. Dr. Bert Holldobler (University of Wiirzburg) as
well as through DFG grant Fi 547/1-1.
LITERATURE CITED
ATSATT, P. R. 1981a. Lycaenid butterflies and ants: selection for enemy-free space. American Naturalist
118:638-654.
COTTRELL, C. B. 1984. Aphytophagy in butterflies: its relationship to myrmecophily. Zoological Journal of the
Linnean SOciety 79: 1-57.
ELIOT, J. N. 1992. The butterflies of the Malay Peninsula. Fourth edition, Malayan Nature Society, Kuala Lumpur.
x + 595 pp., 69 piS.
FIEDLER, K. 1991a. Systematic, evolutionary, and ecological implications ofmyrmecophily within the Lycaenidae
(Insecta: Lepidoptera: Papilionoidea). Bonner zoologische Monographien 31:1-210.
FIEDLER, K. 1991b. European and Northwest African Lycaenidae (Lepidoptera) and their associations with ants.
Journal of Research on the Lepidoptera 28:239-257.
FIEDLER, K. 1992a. Notes on the biology of Hypo/ycaena othona (Lepidoptera: Lycaenidae) in West Malaysia.
Nachrichten des entomniogischen Vereins Apollo, Franlifurt. N. F. 13 :65-92.
FIEDLER, K. 1992b. The life-history of Surendra fiorimel Doherty 1889 (Lepidoptera: Lycaenidae) in West
296 KONRAD FIEDLER
Malaysia. Nachrichten des entomoiogischen Vereins Apollo, Frankfurt, N. F. 13:107-135.
FIEDLER, K. 1993. The remarkable biology of two Malaysian lycaenid butterflies. Nature Malaysiana 18:35-43.
FIEDLER, K. in press a. The life-history of Caleta roxus (Lepidoptera: Lycaenidae). Nachrichten des
entomologischen Vereins Apollo, Frankfurt, N. F. 14.
FIEDLER, K. in press b. Observations on the biology of Eooxylides tharis (Lepidoptera: Lycaenidae). Nachrichten
des entomologischen Vereins Apollo, Frankfurt, N. F. 14.
FIEDLER, K. & MASCHWITZ, U. 1989a. The symbiosis between the weaver ant, OecopJryl1a smaragdina, and
Anthene emolus, an obligate myrmecophilous lycaenid butterfly. Journal of Natural History 23:833-846.
FIEDLER, K. & MASCHWITZ, U. 1989b. Adult myrmecophily: the role of the ant Anoplolepis longipes in the
feeding and oviposition behaviour of Allotinus unicolor (Lepidoptera: Lycaenidae). Tyo to Ga 40:241-251.
HOLLDOBLER, B. & WILSON, E. O. 1990. The ants. Harvard University Press, Cambridge (Massachusetts). xii
+ 732 pp.
HUXLEY, C. R. & CUTLER, D. F. (eds) 1991. Ant-plant interactions. Oxford University Press, Oxford. xviii +
601 pp.
JACOBSON, E. 1912. Symbiose zwischen der Raupe von Hypolycaena erylus Godart und Oecophylla smaragdina
Fab. Tijdschrift voor Entomologie 55:9-14.
KITCHING, R. L. & LUKE, B. 1985. The myrmecophilous organs of the larvae of some British Lycaenidae
(Lepidoptera): a comparative study. Journal of Natural History 19:259-276.
MALICKY, H. 1969. Versuch einer Analyse der iikologischen Beziehungen zwischen Lycaeniden (Lepidoptera) und
Formiciden (Hymenoptera). Tijdschrift voor Entomologie 112:213-298.
MASCHWITZ, U., DUMPERT, K. & SEBASTIAN, P. 1985. Morphological and behavioural adaptations of
homopterophagous blues (Lepidoptera: Lycaenidae). Entomologia Generalis 11:85-90.
MASCHWITZ, U., NASSIG, W. A., DUMPERT, K. & FIEDLER, K. 1988. Larval carnivory and myrmecoxeny,
and imaginal myrmecophily in Miletine lycaenids (Lepidoptera: Lycaenidae) on the Malay Peninsula. Tyo to Ga
39: 167-181.
PIERCE, N. E. 1984. Amplified species diversity: a case study of an Australian Iycaenid butterfly and its attendant
ants. Symposium of the Royal Entomological Society of London 11 (The biology of butterflies): 197-200.
PIERCE, N. E. 1987. The evolution and biogeography of associations between Iycaenid butterflies and ants. Oxford
Surveys in Evolutionary Biology 4:89-116.
PIERCE, N. E. 1989. Butterfly-ant mutualisms. Pp 299-324 in P. 1. Grubb & J. Whittaker (eds). Towards a more
exact ecology. Oxford: Blackwell Science Publishers.
PIERCE, N. E. & EASTEAL, S. 1986. The selective advantage of attendant ants for the larvae of a Iycaenid
butterfly, Glaucopsyche lygdamus. Journal of Animal Ecology 55:451-462.
SCOTT, 1. A. & WRIGHT, D. M. 1990. Butterfly phylogeny and fossils. Pp 152-208 in O. Kudrna (ed).
Butterflies of Europe. Volume 2: Introduction to lepidopterology. Wiesbaden: Aula-Verlag.
SEKI, Y., TAKANAMI, Y. & OTSUKA, K. 1991. ButterfliesofBomeo, Volume 2, no. 1: Lycaenidae. Tobishima
Corp., Tokyo, x + 113 pp. (English section), 72 pis.
A quantitative survey of the termite assemblage and its consumption of
food in lowland mixed dipterocarp forest of Brunei Darussalam
DAVID T. JONES!
Department of Biology, Universiti Brunei Darussalam, Bandar Seri Begawan 2028, Brunei
Darussalam
ABSTRACT. Preliminary results of sampling the termite assemblage in the lowland mixed dipterocarp
forest of Batu Apoi Forest Reserve in Temburong District, Brunei Darussalam, are presented. Numbers
of epigeal and arboreal nests in a one hectare plot were recorded and the populations of termites in soil
and dead wood were measured. Termites were initially separated into four taxonomic groups: the family
Rhinotermitidae, and the Termitidae subfamilies Termitinae, Macrotermitinae and Nasutitermitinae. The
percentage of total density represented by each taxon was 9.2 %,28.8%,33.9% and 28.1 %, respectively.
Density above-ground (513 m,2), in dead wood (320 m,2) and in soil (616 m,2) suggests that the total
termite population exceeds 1500 m,2. Generalized rates of consumption by each taxon were used to
estimate that total consumption of organic matter by termites was 35.2 g (dry weight) m,2 yr'), which
represents approximately 3 % of the annual litter production. Odontotermes spp. (Macrotermitinae) were
responsible for consuming 67 % of the total amount of organic matter taken by termites. Comparisons
were made with data from Mulu National Park, Sarawak, and Pasoh Forest Reserve, Peninsular Malaysia.
The combined data suggest that the density ofMacrotermitinae (fungus-growers) increases with decreasing
annual rainfall.
KEY WORDS: termites, density, rain forest, organic matter, consumption, Macrotermitinae, Borneo.
INTRODUCTION
Termites are a major invertebrate group in the tropics and can have a dramatic influence on
the environment (see the reviews of Collins 1989, Wood & Sands 1978). Stork (this volume)
recognizes termites as one of several key animal groups in tropical forests and emphasizes
their contribution to decomposition processes and forest dynamics. In many tropical forests
termites are present at very high densities and are the dominant arthropod detritivores. For
example, in the lowland rain forests of Malaysia, termite populations can exceed 38 million
individuals per hectare (Abe & Matsumoto 1979) and can represent 60 % of individuals
among the soil macro-invertebrate assemblage and 42 % of its biomass (Collins 1979a). The
ecological impact of termites is derived from this vast abundance.
Since approximately 63 % of the net primary production in tropical forests enters directly
into the decomposer system (Kukihara & Kikkawa 1986), termite feeding habits (reviewed
by Wood 1978) playa critical role in the release of nutrients and carbon via the comminution
and mineralization of dead plant material. In addition, termites modify the physical structure
of the environment by translocating organic matter away from feeding sites and by building
nests and galleries. Their role in soil genesis, altering physical and chemical properties of
the soil, and the effect of their subterranean tunnels on water infiltration has been reviewed
by Anderson (1988) and Lobry de Bruyn & Conacher (1990).
Given the cryptic habits of most termite species plus their heterogenous spatial
distribution, appropriate sampling techniques are required to ensure their numbers are not
lpresent address: Department of Entomology, Natural History Museum, Cromwell Road, London SW7 5BD, United
Kingdom.
297
298 DAVID T. JONES
severely underestimated, as occurred in the survey by Stork & Brendell (1993). To date,
only two studies have attempted to quantify the abundance and distribution of the termite
fauna in Southeast Asian rain forests; Abe & Matsumoto (1979) worked in Pasoh Forest
Reserve, Peninsular Malaysia, while Collins (1983) worked in Mulu National Park, Sarawak,
east Malaysia.
This paper reports some general findings based on field work designed to assess the
density and diversity of termites in the lowland rain forest of Brunei Darussalam, and makes
estimates of their consumption of organic matter. Termites were separated initially into four
taxa (the family Rhinotermitidae, and three Termitidae subfamilies i.e. Termitinae,
Macrotermitinae and Nasutitermitinae) to allow direct comparisons with the results from
Mulu and Pasoh.
Study site
Field work was conducted in the Batu Apoi Forest Reserve, Temburong District, based
at the Kuala Belalong Field Studies Centre. The terrain is extremely rugged, with steep
sided, v-shaped valleys separated by sharp, narrow ridges. Hillslopes of 35° are common
and occasionally exceed 5SO. Geomorphological processes in the Belalong valley have been
described by Dykes (this volume). The relatively thin soils, classified as orthic acrisols, are
mainly silty clays derived from the eroded shale bedrock (Ross 1994). The vegetation was
described as mixed dipterocarp forest by Ashton (1964). The Kuala Belalong area receives
between 4000 - 4500 mm rainfall per annum (Dykes, this volume), with 4135 mm being
recorded in the first year of measurement (Poulsen, this volume).
Termite studies were concentrated in a one hectare research plot (100 m x 100 m) on
which a 5 m x 5 m grid had been marked using aluminium poles and string. The plot is
situated on the north east facing slope of a spur leading from the West Ridge of the Belalong
valley (115°1O'W, 4°32'N). The ground surface is steeply sloping and the middle of the plot
lies at an altitude of approximately 250 m. The location and topography of the plot is shown
in Poulsen (this volume, referred to as Plot A in Figure 1). The plot is designated as a
permanent research plot and efforts were made to minimise disturbance. Therefore, all soil
pits and 'dead wood quadrats' (described below) were located outside the hectare plot within
40 m of its perimeter.
Density of epigeal and arboreal termite nests

The plot was censused systematically for epigeal mounds and arboreal nests. When
searching, special attention was paid to the base of trees, buttresses, and the ground beside
fallen trunks. All vegetation and the surface of tree trunks were inspected for arboreal nests
to a height of 3 m. Mounds and nests were opened with minimal disturbance to the colony
and termites collected for identification. Termites in the forest canopy were excluded from
this survey.
The estimates of mean nest population size employed by Collins (1983) were used to
produce figures for the total number of termites in epigeal and arboreal nests. As the
relationship between nest volume and population size was unknown it was assumed that all
nests contained the quoted mean population. The total arboreal and epigeal populations were
combined to give the 'above-ground' population size within the hectare plot and the above-
ground density of termites per m2 •
Density of termites in dead wood

Ten 'dead wood quadrats', each 5 m x 5 m in area, were randomly located. Within each
Termites in Mixed Dipterocarp Forests 299
quadrat all items of dead wood with a diameter greater than 1 cm were examined. Earlier
qualitative examination revealed that twigs with a diameter less than 1 cm were rarely
occupied by termites. The diameter and length of each item of dead wood was recorded
before being opened with a machete to search for termites. The number of termites in each
item was measured using a method of visual estimation developed by Collins (1983). As the
item of dead wood was successively broken open and the termites revealed, the number of
termites seen was counted at an appropriate order of magnitude (i.e. 10, 1()2 or 103).
Specimens from each item were retained for identification. The total population of each
taxon within the dead wood quadrats was converted to an abundance in dead wood per m2
of forest floor.
Density of subterranean termites

Ten 'soil quadrats', each 0.25 m2 (50 cm x 50 cm) in surface area were randomly located.
Loose leaf litter was removed and the soil in each quadrat was then excavated to a depth of
30 cm. Soil and leaf litter was hand-sorted in trays on site and all termites preserved for
identification and enumeration. Soil was removed as rapidly as was practicable in order to
prevent termites escaping from the disturbed area.
Another four randomly positioned soil pits were dug in order to search for subterranean
termite nests. Three of these additional pits were 3 m x 3 m in surface area and the fourth
was I m x 3 m. In each case the soil was dug out to a depth of 30 cm to search for nests.
RESULTS
The density of termites

A total of 76 epigeal and arboreal nests were found in the hectare plot (Table 1), of which
46 (60% of nests) were mounds of Dicuspiditermes nemorosus. The above-ground density
of individual termites totalled 513 m·2 (Table 1), of which 89% was represented by three
species: D. nemorosus (41 % of above-ground density), Hospitalitermes medioflavus (35%)
and Longipeditermes iongipes (12 %).
A total of 204 items of dead wood were examined for the presence of termites. Of these,
31 % contained termites, with some items having more than one species. The total density
in dead wood was 320 m- 2 (Table 2) of which the Rhinotermitidae, which feed mainly on
sound or rotten wood, constituted the largest proportion at 36%. This sampling method
possibly gives figures on the low side as isolated populations in larger logs and branches may
have been overlooked.
The density of subterranean termites was 616 m-2 (Table 3). This figure was derived from
two sources, firstly 592 m-2 averaged from the ten 0.25 m2 soil pits, and secondly, an
additional 24 m-2 due to eight subterranean nests found during the hectare plot census (listed
in Table 1). The Macrotermitinae constituted 78% of soil termites, and all of these were
Odontotermes spp. Very few individual termites were found in the leaf litter samples and
these were included with the termites from the soil pit from above which the leaf litter was
collected.
The use of a greater number of soil pits would have been preferable (see Baroni-Urbani
et ai. 1978) but limitations of time and man-power prevented this. Wood et ai. (1977)
reported that 78 % - 92 % of termites were collected by hand-sorting soil. It is believed that
due to careful hand-sorting an equally high rate of extraction was achieved in this survey.
Undoubtedly, some nests which did not protrude far above the soil surface will have been
overlooked in the census of the hectare plot. In Pasoh, 25 % of Homallotermes foraminifer
nests went unrecorded during an initial census of a 200 m2 area and were only found when
the entire area was dug up (Abe & Matsumoto 1979).
300 DAVID T. JONES
Table I. Termite nest densities and population densities estimated from the census of the hectare plot at Batu Apoi.
Above-ground and subterranean nests are listed separately.
Above-ground Population Density

nest density per nest (No. m'')
(No. ha· l ) (x 104)
Rhinotermitidae
Parrhinotermes spp. 2 3'1 6
Termitinae
Microcerotermes spp. 2 3'1 6
Termes spp. 6 3'2 18
Homnllotermes spp. 3 1.26'3 4
Dicuspiditermes nemorosus 46 4.6'3 212

Subtotal 59 246
Nasutitermitinae
Bulbitermes spp. 5 4.24'3 21
Longopeditermes longipes 8 7.87'3 63
Hospitalitermes medioj1£lvus 3 60'4 180
Subulitermes spp. 3'1 3
Subtotal 17 267
Grand Total 76 513
Subterranean Population Abundance

nest density per nest (No. m·2)
(No. ha· l ) (x 104)
Termitinae
Prohamitermes spp. 5 3'1 15
Labritermes spp. 3'1 3
Procapritermes spp. 3'2 3
Subtotal 7 21
Nasutitermitinae
Subulitermes spp. 3'1 3
Grand Total 8 24
'I Conservative estimate of nest population.

'2 from Collins (1983)
'3 from Abe & Matsumoto (1979)
'4 from Collins (1979b)
Very few termites were found below a depth of 20 cm. This is consistent with the
findings of Abe & Matsumoto (1979) who observed at Pasoh that most termites occurred
in the top 15 cm of the soil profile and were very scarce below 30 cm. During the
excavation of the ten 0.25 m2 pits and the four larger pits (total surface area 32.5 m2)
accumulations of termites were found but no hard subterranean nest structures (made of
hardened soil or faecal material) were unearthed.
Table 2: Estimate of density of termites in dead wood (ten 5 m x 5 m quadrats sampled) in the forest at Batu
Apoi. Number of woody items examined = 204; Number of items attacked by termites = 64 (31 %).
Density of termites in dead wood

(No. m·2 of forest floor ± S.E.M.)
Rhinotermitidae 116 ± 52
Termitinae 98 ± 59
Macrotermitinae 11 ± 7
Nasutitermitinae 95 ± 45
Total 320
Table 3. Estimate of the density of termites in soils (from hand-sorting 10 soil pits 50 cm x 50 cm x 30 cm
depth, and the number of subterranean nests (see Table I}) at Batu Apoi.
Density of termites
(No. m·2 ± S.E.M.)
Soil pits Soil nests Total
Rhinotermitidae 11± 10 0 11
Termitinae 58± 23 21 79
Macrotermitinae 481 ± 295 0 481
Nasutitermitinae 42 ± 40 3 45
Total abundance 592 24 616
The total density of 1449 m,2 (Table 4) is an underestimate because termites infesting
living trees and those inhabiting the canopy were ignored. For example, Coptotermes spp.
often nest inside mature trees (Collins 1984) and colony populations can exceed 5 x lOS
(Hadlington 1987). Therefore, the total density of termites in the forest at Batu Apoi will
probably exceed 1500 m,2. Of the four taxonomic groups none were recorded from either
all ten 'dead wood quadrats' or from all ten 0.25 m2 soil pits, reflecting the highly
heterogenous distribution of termites.
Consumption of organic matter

Table 4 gives the total density and biomass of the four taxa (combining the data from
Tables 1-3) and estimations of their consumption of organic matter. Termites are
estimated to consume 35.2 g (d.w.) m,2 yr,l of organic matter, of which 67% is accounted
for by Macrotermitinae. As all the Macrotermitinae collected were Odontotermes spp. a
rate of consumption specific to this genus was used to estimate the consumption
Table 4. Estimates of total density and biomass of termites and their consumption of organic matter in the
lowland mixed dipterocarp forest of Batu Apoi. *' from Collins (1983); *2 from Abe (1979); *' mean (alcohol
w.w.) of 300 Odontotermes individuals collected from soil pits at Batu Apoi (this study); *4 rate for
Macrotermitinae is mean of two figures for Odontotermes spp. given in Collins (1983).
Total density biomass of individual Total biomass Rate of consumption Total consumption
(No. m·2) (mg w.w.) (g m,2) (mg d.w. g" day") (g d.w. m·2 yr")
Rhinotermitidae 133 3.0 *' 0.399 20 *' 2.91

Termitinae 417 1.674 *2 0.681 30 *' 7.46
Macrotermitinae 492 2.855 *' 1.405 46.2 *4 23.69
Nasutitermitinae 407 0.758 *2 0.309 10 *' 1.13
All termites 1449 2.794 35.19
302 DAVID T. JONES
by Macrotermitinae (see Table 4). The biomass (alcohol wet weight) of Odontotermes was
measured as the mean of 300 individuals from the soil pits (after specimens had been
blotted on filter paper to remove surface liquid). All other values of biomass and rates
of consumption are from Collins (1983).
The annual small litter (leaves, twigs, fruits, flowers and trash) production at Batu
Apoi is 10.67 t ha- 1 yr- 1 (C.A. Pendry, personal communicaion) but the production of
large woody litter (> 2 cm diameter) is not known. Assuming that the ratio of woody
litter to small litter at Batu Apoi is the same as that at Mulu (Proctor et al. 1983) then the
total annual litter production would be 11.56 t ha- 1 • In this case termites would be
responsible for consuming 3.0% of the annual litter produced in the lowland forests of
Batu Apoi.
DISCUSSION
Comparison of termite density between sites

Table 5 compares the density at Batu Apoi with that of Mulu and Pasoh. The
Nasutitermitinae, Termitinae and Rhinotermitidae are all less abundant in Batu Apoi than
at Mulu but overall termite density is similar due to the greater number of
Macrotermitinae at Batu Apoi. The estimates for all taxa enumerated at Pasoh are higher
than at the other two sites and the overall total would have been even larger had the
Rhinotermitidae been estimated.
A total of 17 epigeal and arboreal nests were recorded from a one hectare plot at Mulu
(Collins 1983). The resulting above-ground density of 249 m-2 was only half that recorded
at Batu Apoi and much of this difference is due to the relatively low density of
Dicuspiditermes nemorosus mounds at Mulu (2 ha- 1 compared with 46 ha- 1 at Batu Apoi).
However, striking differences are seen when Batu Apoi is compared with Pasoh. A
hectare plot at Pasoh had 249 epigeal and arboreal nests with a total above-ground density
of 870 - 1520 m- 2 (Abe & Matsumoto 1979). Of these nests, 131 were mounds of D.
nemorosus, while 85 and 20 were mounds of Homallotermes foraminifer and Macrotermes
carbonarius respectively. Only three nests of Homallotermes spp. were found at Batu
Apoi and none at Mulu. Nests of Macrotermes spp. were not found in the plot at either
Batu Apoi or Mulu.
At Mulu the density of termites in dead wood was 153 m- 2 (Collins 1983), only half
that of Batu Apoi. However, while the minimum diameter for items examined at Batu
Apoi was 1 cm, it was only 2 cm at Mulu. Of the 46 items from Batu Apoi with a
diameter in the range 1 - 2 cm only three contained termites.
Table 5. Comparisons of annual rainfall, the density of termites and their conswnption of organic matter in the
forests of Batu Apoi (Brunei Darussalam), Mulu (Sarawak) and Pasoh (peninsular Malaysia).
Mulu Batu Apoi Pasoh

Source of data Collins (1983) This study Abe & Matswnoto (1979)
Rainfall (mm yr') 5\07 4000 - 4500 2000
Density Conswnption Density Conswnption Density Conswnption
(No. m-2) (g m-2 yr-') (No. m-2) (g m-2 yr-') (No. m-2) (g m-2 yr-')
Rhinotermitidae 293 6.42 133 2.91 Not estimated
Termitinae 612 11.21 417 7.46 1280-1730 23.76-31.32
Macrotermitinae 5 0.70 492 23.69 934-984 129.43-138.41
Nasutitermitinae 616 1.70 407 1.13 943-\093 2.23-3.47
Total 1526 20.03 1449 35.19 3157-3807 155.42-173.20
Excluding these 46 items has a negligible effect upon abundance since the populations
in the three occupied twigs were very small. The higher annual rainfall at Mulu compared
with Batu Apoi (see Table 5) may increase the initial rate of fungal and microbial attack
on freshly fallen wood and thus help to make a greater proportion of the dead wood more
palatable to termites and therefore more susceptible to invasion. The Rhinotermitidae are
usually associated with wood which is already in a state of decay due to the presence of
fungi (Sands 1969). The density of termites in dead wood at Pasoh was not estimated.
The density of subterranean termites at Batu Apoi is only about half that of Mulu (1125
m 2 ; Collins 1979a) and only one third that of Pasoh (1898 m- 2 ; Abe & Matsumoto 1979).
The absence of hard subterranean nests in 32.5 m2 of soil pits at Batu Apoi indicates that
their density may be low compared with Mulu and Pasoh. For example, Collins (1983)
searched a 50 m2 area at Mulu to a soil depth of 15 cm and recorded 15 nests. Abe &
Matsumoto (1979) reported 4.5 nests of Procapritermes sp. in a soil pit of only 16 m2 at
Pasoh.
Subterranean nests may be relatively more vulnerable in steep areas experiencing soil
instability, erosion and a higher level of water runoff. Ashton's (1964) description of the
hillside soils in the Belalong valley states that "evidence of surface erosion and profile
truncation was rarely absent". If physical conditions at Batu Apoi are less favourable for
soil-nesting termites then epigeal mounds may also be exposed to similar destructive
forces. Very few of the D. nemorosus mounds at Batu Apoi are free standing but instead
are supported on the base of tree trunks or around the basal stems of palms and saplings,
which may reflect the need for anchorage on steeper and more unstable slopes.
Comparison of consumption by termites between sites

The estimates in Table 5 indicate that consumption of organic matter by termites at
Pasoh is approximately 8 times higher than at Mulu and about 4.5 times that at Batu Apoi.
At Pasoh and Batu Apoi the Macrotermitinae are the dominant taxa in terms of
consumption, being responsible for 80 % - 83 % and 67 % of the total consumption at these
respective sites. At Mulu the Macrotermitinae are less important and consume only 3.5 %
of the organic matter taken by termites. In terms of their overall consumption, termites
consume an estimated 2.1 % of the total annual litter production at Mulu (Collins 1983),
3.0% at Batu Apoi, and 14.7% - 16.3% at Pasoh (Matsumoto & Abe 1979). The high
total consumption at Pasoh relative to the other two sites (Table 5) is a result of the
greater overall density of termites, but more importantly, it is due to the dominance of the
Macrotermitinae which have relatively higher weight-specific rates of consumption
(Collins 1989). However, as Collins (1983) states, the use of generalized rates of
consumption to estimate total consumption by termites is only approximate and figures
should be regarded as minimum values.
The Macrotermitinae cultivate mutualistic fungi on faecal combs in their nests, a
feeding habit unique amongst termites. This enables them to feed on fresh litter while
other groups of termites prefer litter which has already been partially decayed. The mixed
dipterocarp forests of Mulu and Pasoh lie at opposite ends of a climatic spectrum (see
rainfall figures in Table 5). Collins (1980, 1983) speculates that as a result the
Macrotermitinae are more prevalent in drier forests such as Pasoh where they dominate
the decomposition process but in the much wetter forests of Mulu, where the activity of
free-living fungi and microbes is far greater, they lose their ecological advantage and are
virtually absent. The intermediate level of rainfall at Batu Apoi may have given rise to
the relatively high density of Odontotermes compared with Mulu, but it is still too wet for
the mound-building Macrotermes spp. favoured at Pasoh. In the Western Ghats of India,
Ferry (1992) reported that the Macrotermitinae are the dominant Isoptera taxa in drier
forests but become less common at sites with increasing rainfall. Wood & Sands (1978)
304 DAVID T. JONES
suggested that the Macrotermitinae's share of the organic matter consumed by termites in
West African savannas was lower in those areas with higher climatic humidity.
The experiments of Matsumoto & Abe (1979) showed that litter removal at Pasoh was
dominated by invertebrate feeding activities, in particular the termite species Macrotermes
carbonarius and Longipeditermes longipes which both feed on leaf litter. In contrast, at
Mulu, termites consume less and playa smaller role in the initial decomposition process
(Anderson et at. 1983, Collins 1983). With its higher density of Odontotermes (which
have been observed feeding on leaf litter at Batu Apoi; unpublished data, D. T. Jones) and
L. longipes, termites at Batu Apoi will have a greater role in the consumption and
comminution of fresh litter when compared with Mulu but less than at Pasoh.
Although termites may only consume a small proportion of the annual litter production
in South east Asian ever-wet rain forests such as Batu Apoi and Mulu, their impact on the
forest ecosystem is still highly significant through indirect means. Termites introduce
microorganisms into their feeding and nest sites (Collins 1988, Sands 1969) while
comminution of litter increases surface area thereby encouraging additional microbial and
fungal activity and thus enhancing overall rates of decomposition. The feeding and
nesting habits of several species can also affect forest dynamics by killing mature trees
and saplings. For example, Tho (1982) found that Microcerotermes dubius kills a range
of tree species and is involved in canopy gap formation in the primary forests of
Peninsular Malaysia. Similar behaviour by M. dubius has been observed in the forests
of Brunei Darussalam (unpublished data, D.T. Jones).
ACKNOWLEDGEMENTS
This work was funded by a research grant from the Universiti Brunei Darussalam (grant
No. UBD/T6/RG.02). The author is grateful for assistance in the field from the
Venturers and staff of Raleigh International. Thanks are given to Bert Orr, Paul Eggleton
and an anonymous referee for their helpful comments on the manuscript.
LITERATURE CITED
ABE, T. 1979. Studies on the distribution and ecological role of termites in a lowland rain forest of West
Malaysia. 2. Food and feeding habits of termites in Pasoh Forest Reserve. Japanese Journal of Ecology
29: 121-135.
ABE, T. & MATSUMOTO, T. 1979. Studies on the distribution and ecological role of termites in a lowland
rain forest of West Malaysia. 3. Distribution and abundance of termites in Pasoh Forest Reserve. Japanese
Journal of Ecology 29:337-351.
ASHTON, P.S. 1964. Ecological studies in the mixed dipterocarp forests of Brunei State. Oxford Forestry
Memoirs No. 25. Clarendon Press, Oxford.
ANDERSON, I.M. 1988. Invertebrate-mediated transport processes in soils. Agriculture, Ecosystems and
Environment 24:5-19.
ANDERSON, J.M., PROCTOR, J. & VALLACK, H.W. 1983. Ecological studies in four contrasting lowland
rain forests in Gunung Mulu National Park, Sarawak. 3. Decomposition processes and nutrient losses from
leaf litter. Journal of Ecology 71:503-527.
BARONI-URBANI, C., JOSENS, G. & PEAKIN, G.I. 1978. Empirical data and demographic parameters. pp.
5-44 in Brian, M.V. (ed.). Production ecology of ants and termites. mp Vol. 13. Cambridge University
Press, Cambridge.
COLLINS, N .M. 1979a. A comparison of the soil macrofauna of three lowland forest types in Sarawak. Sarawak
Museum Journal 27:267-281.
COLLINS, N .M. 1979b. Observations on the foraging activity of Hospitalitermes umbrinus (Haviland), (lsoptera:
Termitidae) in Gunung Mulu National Park, Sarawak. Ecological Entomology 4:231-238.
Termi~s in Mixed Dipterocarp Forests 305
COLLINS, N.M. 1980. The effect of logging on termite (Isoptera) diversity and decomposition processes in
lowland dipterocarp forests. pp. 1\3-121 in Furtado, J.1. (ed.). Tropical ecology and development.
International Society of Tropical Ecology, Kuala Lumpur.
COLLINS, N.M. 1983. Termite populations and their role in litter removal in Malaysian rain forests. Pp. 311-
325 in Sutton, S.L., Whittnore, T.C. & Chadwick A.C. (eds). Tropical rain forest: Ecology and
Management. Blackwell, Oxford.
COLLINS, N.M. 1988. Termites. Pp. 196-211 in Earl of Cranbrook (ed.). Key environments - Malaysia.
Pergamon, Oxford.
COLLINS, N.M. 1989. Termites. Pp. 455-471 in Lieth & Werger (eds). Tropical rain forest ecosystems:
Biogeography and ecological studies. Elsevier, Oxford.
DYKES, A. (this volume). Analysis of factors contributing to the stability of steep hillslopes in the tropical
rainforest of Temburong, Brunei Darussalam.
FERRY, B. 1992. Distribution of the important litter decomposing termites (Isoptera) in the Western Ghats
forests of Karnataka (India). Pedobiologia 36: 193-211.
HADLINGTON, P. W. 1987. Australian termites. New South Wales University Press, Kensington, NSW.
KUKIHARA, Y. & KIKKAWA, J. 1986. Trophic relations of decomposers. pp. 127-160 in Kikkawa, J. &
Anderson, D.J. (eds). Community ecology: Pattern and process. Blackwell, Oxford.
LOBRY DE BRUYN, L.A. & CONACHER, A.J. 1990. The role of termites and ants in soil modification: A
review. Australian Journal of Soil Research 28:55-93.
MATSUMOTO, T. & ABE, T. 1979. The role of termites in an equatorial rain forest ecosystem of West
Malaysia. 2. Leaf litter consumption on the forest floor. Oecologia 38:261-274.
POULSEN, A.D. (this volume). The herbaceous ground flora of the Batu Apoi Forest Reserve, Brunei
Darussalam.
PROCTOR, J., ANDERSON, I.M., FOGDEN, S.C.L. & VALLACK, H.W. 1983. Ecological studies in four
contrasting lowland rain forests in Gunung Mulu National Park, Sarawak. 2. Litterfall, litter standing crop
and preliminary observations on herbivory. Journal of Ecology 71 :261-283.
ROSS, S.M. 1994. Soils and hydrology. Chapter 4 in Cranbrook, Earl of, & Edwards, D.S. Belalong, A tropical
rainforest. Sun Tree Publishing, Singapore.
SANDS, W.A. 1969. The association of termites and fungi. Pp. 495-524 in Krishna K. & Weesner, F.M. (eds).
Biology of termites. Vo\. I. Academic Press, New York & London.
STORK, N. (this volume). Tropical forest dynamics: the faunal components.
STORK, N.E. & BRENDELL, M.J.D. 1993. Arthropod diversity studies in lowland rainforest of Seram,
Indonesia. Pp. 115-130 in Edwards, I.H., Macdonald, A.A. & Proctor, J. (eds). The natural history of
Seram. Intercept, Andover.
THO, Y.P. 1982. Gap formation by the termite Microcerotermes dubius in lowland forests of Peninsular
Malaysia. Malaysian Forester 45:184-192.
WOOD, T.G. 1978. Food and feeding habits of termites. Pp. 245-292 in Brian, M.V. (ed). Production ecology
of ants and termites. IBP Vo\. 13. Cambridge University Press, Cambridge.
WOOD, T.G., JOHNSON, R.A., OHIAGU, C.E., COLLINS, N.M. & LONGHURST, G. 1977. Ecology and
importance of termites in crops and pastures in northern Nigeria. Project report. Centre for Overseas Pest
Research, London.
WOOD, T.G. & SANDS, W.A. 1978. The role of termites in ecosystems. In Brian, M.V. (ed.). Production
ecology of ants and termites. IBP Vo\. 13. Cambridge University Press, Cambridge.
Resource use in a freshwater fish community of a tropical rainforest
stream in northern Borneo
SATISH C. CHOY\ SALWANA A. LATIF and YONG N. YUNG
Universiti Brunei Darussalam, Bandar Sen Begawan 2028, Brunei Darussalam
ABSTRACT. Partitioning of key environmental resources (food, space and time) was studied in
a freshwater fish community in the Belalong River, Brunei Darussalam, northern Borneo. Of a
total of twenty-five species recorded, eight were confined to the lentic (slow-flowing) zone and
seven to the lotic (fast-flowing) zone of the stream. An additional ten species were found in both
these zones. In the lentic zone, fish species partitioned the habitat vertically into surface, pelagic,
demersal, benthic and substratum niche types. In the lotic zone this stratification was not observed
as the vertical component was compressed. Here, the abundance of all but two species of fish was
low. Ecological indices (prey diversity, prey dominance, niche breadth and niche overlap) were
used to analyse the data on food items ingested by each species. Although there was some overlap
in food niche, the fish species were ecologically separated by differences in morphological features
related to foraging tactics, feeding behaviour and space utilisation. Species that were close
trophically either occupied different micro-habitats or foraged at different times. Results indicate
that differences in the utilisation of space, time and food play an important role in the co-existence
of species in the freshwater communities.
KEY WORDS: fish, lentic, lotic, resource partitioning, space, food, time, Brunei Darussalam.
INTRODUCTION
Although tropical communities are considered to be complex and highly structured, with the
presence of many specialised species (Goulding 1980, Lowe-McConnell 1975), extreme
specialisation need not be necessary for the existence of high species diversity (Connell 1978;
Sale 1982). Differential use of environmental resources by closely-related sympatric species
is thought to be a mechanism for their co-existence within a community (Pianka 1981,
Schoener 1986). Fishes, like other animals, partition resources along habitat, time and food
axes (Helfman 1978, Lowe-McConnell 1987, Wooton 1992). Although habitat and food
resource use and partitioning in tropical freshwater fishes have been extensively studied
(Angermeier & Karr 1983, Balon & Stewart 1983, Bishop 1973, Inger & Chin 1962, Moyle
& Senanayake 1984, Watson & Balon 1984a, 1984b), little consideration has been
simultaneously given to all three resource dimensions: space, food and time.
The aim of the present study was to determine patterns of resource partitioning in a
tropical freshwater fish community and see if such partitioning could explain the co-existence
of sympatric species.
Fieldwork was carried out in a slow-flowing (lentic) zone and a fast-flowing (lotic) zone of
the Belalong River within the Batu Apoi Forest Reserve, close to the Kuala Belalong Field
Studies Centre (KBFSC) in Temburong district, Brunei Darussalam. The Belalong River is
'Present address: DPI Water Resources Technical Centre, 1345 Ipswich Road, Rockiea, Qld 4106, Australia.
307
D. S. Edwards et al. (eds.). Tropical Rainforest Research - Current Issues, 307-314.
308 SATISH CHOY, SALWANA LATIF AND YONG N. YUNG
about 5-30 m wide and 0.3-5 m deep and runs through primary mixed dipterocarp forest.
Details of its hydrology and water quality are given in Dykes (this volume) and Choy (this
volume), respectively.
In the study area the deep, slow-flowing channel (lentic zone) was 3-7 m wide, 1.0-3.0
m deep and flowed at about 0.1 m S-I. The shallow, fast-flowing (lotic zone) was 5-15 m
wide, 0.2-1.0 m deep and flowed at 0.5-1.0 m S-I. The fish species composition in each of
these zones is different, but with some overlap (Choy, this volume).
Observations on microhabitat utilisation were made visually using mask and snorkeL
Investigations relating to time utilisation were based on visual observations and on diurnal
variation in fish catches, using hand nets and gill nets. To study food utilisation, the stomach
contents and certain morphological features of the fish were analysed. The morphological
attributes measured were biomass, total length, standard length, head length, head width,
gape width, cleft length, eye diameter, gut weight, gut length and gut volume. The
morphology of the head, in terms of the orientation of the mouth, its protrusibility and the
presence or absence of barbels was also examined. These attributes are considered to be
those most likely to reflect habitat utilisation, diet and foraging tactics of the fish (Watson
& Balon 1984a). Description of the morphological features, their measurement and
ecological significance can be found in Gatz (1979a, 1979b) and Watson & Balon (1984a).
For stomach content analysis, all fish caught were either dissected while fresh or preserved
in 10% formalin. Stomach contents were examined under the microscope and the food items
were identified to the lowest possible taxon. Gut volume was calculated based on its
cylindrical dimension and the volume of each food item was calculated from its proportion
in the gut multiplied by the total volume of the gut.
To measure the prey diversity in each species, the Shannon-Wiener diversity index was
used:
where Pi is the volume or frequency of occurrence of a food item (ith) in the stomach.
Prey evenness within a species was calculated using: E = H'/Hmax, where Hmax = In N,
N being the total number of food items consumed by the species. The degree of dominance
of a particular food item in each species was calculated using the Berger-Parker Index: d =
dlll1lx lN, where N is the total volume of all food items and dmax is the volume of the most
abundant food item. Since d is obtained as a fraction, it has been expressed in its reciprocal
form_ Niche breadths of each sympatric species were estimated using the Brillouin Index:
HB= InN! -L In (n i )
N
where ni is the volume of each food item and N is the total volume of all food items in each
species. For N > 10, the Sterling's Slacker approximation (InN! = nlnN - N) was used.
Niche overlap between the fish species was calculated using Pianka's Index:
where Pij and Pik are volumes of the l-th food item used byJ-th and It' species. The value of this
index varies between 0 (no overlap) and 1 (total overlap). For justification, strengths and
shortcomings of the above-mentioned indices see Magurran (1988)_
Resource use in a freshwater fish community 309
RESULTS
Species composition and microhabitat utilisation

A total of twenty five species of fish were recorded from the Belalong River near KBFSC
(Table 1). Of these, eight species were confined to the lentic (slow-flowing, deep) zones and
seven to the (fast-flowing, shallow) lotic zones. An additional ten species were found in both
these zones. Most of these latter were daytime lentic dwellers which made nocturnal
foraging excursions into the lotic zone. For details on distribution and ecological zonation
of these fish see Choy (this volume).
The majority of the species in the lentic zone belong to the family Cyprinidae while those
in the lotic zone belong to the family Balitoridae, particularly the subfamily
Gastromyzontinae. Fish of the lentic zone are mainly surface, pelagic and demersal
dwellers, only one is strictly benthic (Table 1). Juveniles of many pelagic species were
observed near the banks and in flood pools while those of benthic species were found in flood
pools and in the accumulated plant litter on the beds of streams.
Fishes in the lotic zone are mainly demersal and benthic species, with the exclusive
members of this community (the gastromyzontins) belonging to the latter group (Table 1).
Table I. Fish species composition (numbers caught) and microhabitat utilisation. L, lentic zone; Lo, lotic zone; DI,
diurnal; N, nocturnal; S, surface; P, pelagic; D, demersal; B, benthic.
Species Composition Microhabitat Activity

L Lo DI N
Nematabramis steindachneri 21 0 S 21 4
Rasbora argyrotaenia 17 0 S 17 0
Barbodes collingwoodi 11 0 P, D 5 6
Hampala bimaculata 8 2 D 3 7
Tor douronensis 6 2 P, D 2 6
Lobocheilus bo 5 2 D 2 5
Osteochilus spilurus 5 0 P, D 3 2
Puntius binotatus 4 0 P,D 0 4
Rasbora sumatrana 3 0 S, P 0
Glyptothorax major 2 D,B 2
Rasbora sp. 0 S, P 0
Pseudogobiopsis sp. 0 B 1 0
Anguilla mamwrata 1 D,B 0 2
Mystus baramensis D,B 0 2
Clarias teijsmanni D,B 0 2
Pangio mariarum D,B
Macrognathus maculatus D,B
Silurichthys hasselti D,B 1
Paracrossochilus acerus 0 12 B 12 0
Gastromyzon bomeensis 0 8 B 8 0
Neogastromyzon nieuwenhuisi 0 3 B 3 0
Gastromyzon monticola 0 3 B 3 0
Gastromyzon lepidogaster 0 2 B 2 0
Gastromyzon ridens 0 B 0
Protomyzon whiteheadi 0 B 0
Time utilisation: activity patterns

In the lentic zone five species (Nematabramis steindachneri, Rasbora argyrotaenia, R.
sumatrana, Rasbora sp. and Pseudogobiopsis sp.) were mainly diurnally active while eight
species were mainly nocturnally active (Table 1). The remaining five species seemed to be
active at both times. The seven species confined only to the lotic zone are all diurnal. The
ten migratory species are bottom-dwelling and are either nocturnal or active during both
times. However, their presence in the lotic zone was mainly during nocturnal hours.
Fish morphology in relation to micro-habitat and diet

Fishes confined to the lotic zone (the gastromyzontins and Paracrossocheilus acerus) are
generally smaller and more dorso-ventrally flattened (low CI and IVF values) than those of
the lentic zone (predominantly cyprinids) (Table 2). Most of the former are benthic and have
fins which are modified for clinging on to the substrate. These and other benthic species
(such as Lobocheilus bo, Mystus baramensis and Leiocassis micropogon) have their eyes
placed more dorsally than demersal, pelagic or surface species. Their mouths are ventrally
placed, enabling them to feed on items found on the substrate. Surface and pelagic species
are more laterally compressed (deep-bodied) than demersal or benthic ones. Surface dwellers
such as Nematabramis steindachneri and Rasbora argyrotaenia are small and their mouths
are placed more dorsally than species of the other strata. This allows them to feed on items
(such as insects) falling on the surface of the water.
The relative gut lengths of the herbivores are greater than those of the carnivores (Table
2). In general, the gut of most species is relatively long, reflecting their omnivorous diets
(see next section).
Diet
Items identified in the stomachs of the fishes were plant material, insect parts, fish bones
and scales, nematodes and sediment (sand and small pebbles). Unidentifiable material was
clumped and greenish, probably comprising partly digested food. Many specimens had either
empty stomachs or had in them only unidentifiable material; these have been excluded from
the data (Table 3). Similarly, nematodes which were ubiquitous and are thought to be part
of the stomach fauna have not been included.
Most species had a predominance of either plant or insect material. Three species
(Nematabramis steindachneri, Puntius binotatus and Mystus baramensis) did not have any
plant material in their stomachs. Insects contained in the stomachs of the surface dwellers
(N. steindachneri and R. argyrotaenia) belonged mainly to the terrestrial orders Diptera,
Hemiptera, Homoptera, Hymenoptera, Coleoptera, Orthoptera and Lepidoptera whereas those
contained in the stomachs of bottom dwellers (M. baramensis, Leiocassis micropogon, P.
acerus and N. nieuwenhuisO belonged to the aquatic larvae of the orders Ephemeroptera,
Trichoptera, Odonata and Plectoptera. Pelagic species had a combination of all these orders
in their stomachs. Plant material in the stomachs of the lotic dwellers (P. acerus and the
gastromyzontins) was mainly algal whereas those in the stomachs of the lentic dwellers
(Barbodes collingwoodi, Hampala bimaculata, Tor douronensis and Osteochilus spilurus)
were remains of vascular plants. Only H. bimaculata and B. collingwoodi had remains of
fish in their stomachs. No crustaceans or molluscs were identified in the stomach of any
fish.
Large lentic-dwelling omnivores such as H. bimaculata, O. spilurus, T. douronensis and
L. bo, had greater prey diversity and prey evenness than the food specialists and their smaller
lotic counterparts (P.acerus and N. nieuwenhuisO. As expected, the niche breadths of the
food generalists were greater than those of the food specialists. Insects were the dominant
prey for all but two (0. spilurus and L. bo) lentic species whereas they were dominant in
Table 2. Morphological characteristics of the seventeen predominant fish species. The last four species are found
exclusively in the lotic zone. The wet weights (yNt) and standard lengths (SL) are mean values. RGL, relative gut
length (GLlSL); RED, relative eye diameter (ED/SL); cr, compression index (greatest body depth/greatest body
width); NF, index of ventral flattening; M, position of mouth; EP, position of eyes; B, barbels; L, lateral; D,
dorsal; LID dorso-Iateral; T, terminal to oblique; V, ventral; A, absent; P, present.
Species Wt SL RGL RED CI NF M EP B

(g) (cm)
Nematabramis steindachneri 3.0 5.7 11.48 0.059 2.33 0.70 D L P

Rasbora argyrotaenia 3.7 5.8 11.04 0.072 2.00 0.67 D,T L A
R. sumatrana 3.9 6.1 11.13 0.070 1.98 0.64 D,T L A
Rasbora sp. 2.9 4.6 10.96 0.069 2.03 0.66 D,T L A
Barbodes collingwoodi 29.0 10.7 11.51 0.071 2.24 0.52 T L P
Hampala bimaculata 71.2 15.2 11.55 0.054 2.01 0.48 T L P
Tor douronensis 38.8 15.0 6.35 0.057 1.71 0.51 T L P
Lobocheilus bo 45.5 12.7 7.69 0.066 1.75 0.51 T L,D P
Osteochilus spilurus 35.1 11.2 8.65 0.063 2.03 0.51 T L P
Puntius binotatus 33.5 15.5 9.03 0.039 2.20 0.51 T L P
Mystus baramensis 48.1 14.4 11.81 0.035 1.20 0.55 T L,D P
Macrognathus maculatus 40.9 24.0 3.75 0.025 1.69 0.57 T L A
Leiocassis micropogon 33.9 13.3 9.61 0.032 1.18 0.52 T L,D P
Paracrossochilus acerus 1.9 4.3 5.63 0.041 1.24 0.52 V L,D P
Gastromyzon bomeensis 3.3 4.0 12.19 0.030 0.71 0.31 V D P
Neogastromyzon nieuwenhuisi 0.9 4.6 10.75 0.054 0.93 0.37 V D P
Gastromyzon monticola 1.6 3.2 12.38 0.029 0.70 0.30 V D P
only one (N. nieuwenhuisl) lotic species. A cladogram of niche overlap, based on diet is
shown in Figure 1. Although there were some large niche overlaps amongst lentic species,
most of them had either different micro-habitats or foraged at different times (see Table 1).
Resource partitioning
Based on resource use by each fish species, a model of resource partitioning was constructed
(Figure 2). Diet for species which had empty stomachs was obtained from Inger & Chin (1962)
and Bishop (1973). As a result of resource partitioning, very few species occupied the same
niche. Those that did were either the migratory carnivores or the small gastromyrontin
herbivores. The abundance of these was low and the former could be expected to be in different
areas at any time while the latter had an abundance of food and space resource.
DISCUSSION
Fish assemblages in tropical rainforest streams are highly structured (Moyle & Senanayake 1984,
Watson & Balon 1984a, 1984b). However, the degree of ecological segregation can differ
amongst streams (Moyle & Senanayake 1984). In the Belalong River, it is clear that the degree
of ecological segregation is very marked when all three niche dimensions (space, time and diet)
are considered. Longitudinal ronation alone accounts for 60% segregation: the overlap between
the lotic and lentic rone is only 40%. Those species confined to the lotic rone are
morphologically well adapted to living under fast-flowing conditions while species in the lentic
zone are not. Within each of these rones, there is further habitat partitioning, particularly in the
Table 3. Stomach contents of some freshwater fish in the Belalong River. V, mean percentage volwne; F, percentage
frequency of occurrence.
Species Plants Insects Fish Unident.
V F V F V F V F
Nematabramis stei1lli1lchneri 0 0 79.3 100 0 0 20.7 37.9

Rasbora argyrotaenia 1.5 8.3 98.0 100 0 0 0.9 16.7
Barbodes collingwoodi 7.1 33.3 45.9 77.8 1.0 11.1 45.9 100
Hampala bimawlata 15.4 61.5 16.5 38.5 1.6 5.5 68.0 67.0
Tor douronensis 29.3 16.7 66.7 50.0 0 0 3.8 50.0
Lobocheilus bo 24.5 69.2 3.6 15.4 0 0 71.8 84.6
Osteochilus spilurus 52.6 83.3 9.4 25.0 0 0 37.9 51.3
Puntius binotoJus 0 0 80.2 100 0 0 19.8 100
Mystus baramensis 0 0 87.5 100 0 0 12.5 100
Leiocassis micropogon 1.3 25.0 90.6 100 0 0 8.0 75.0
Paracrossochilus acerus 59.6 30.8 13.2 23.1 0 0 26.3 23.1
Gastromyzon borneensis 42.9 33.3 0 0 0 0 57.1 25.0
Neogastromyzon nieuwenhuisi 9.8 25.0 81.3 50.0 0 0 8.9 75.0
Gastromyzon monticola 10.0 33.3 0 0 0 0 90.0 100
lentic rone. Fish species diversity is correlated with habitat complexity (Gorman & Karr 1978),
with most concentrated into deep pools (Angermeir & Karr 1983).
The diversity of body shape and morphological specialisation associated with habitat and
feeding in the Belalong River is typical of tropical streams elsewhere (Bishop 1973, Gatz 1979,
Moyle & Senanayake 1984, Watson & Balon 1984a). It has been suggested that morphological
diversity reflected ecological diversity, implying co-evolution to reduce competition for limited
resources (Gatz 1979). DISTANCE
0.000 0.500
G. bomeensis 9
L. bo 13
O. spilurus
P. acerus 10
5
l
f--
H. bimaculata 1
M. baramensis 8
P. col1ingwoodi 2 r-
P. binotatus 12
N. nieuwenhuisi 11
R. argyrotaenia 3
L. micropogon 7
N. steindachneri 4-
T. douronensis 6
Figure 1. Cladogram of the thirteen predominant fish species, based on diet similarity. Scale is euclidean distance obtained
using single linkage method. For full names of species see Table 1.
Carnivore
Ns
Omnivore
Ra
Surro-pelagic
Omnivore
(2) Rs Rsp
Benthic
Omnivore
(1) P.
Omnivore (H)
0.
Pelago-demersal
(2) Omnivore (C)
Be
(1) Pelago-demersal
(1) Omnivore
Pb
Omnivore
Hb
Demersal
(1)
Carnivore
Am Mb
Bentho-demersal
(3) Omnivore
Demersal (1) Ct
Omnivore
Pelago-demersal (1) Lb
(25) Omnivore
Demerso-benthic (1) Td
Omnivore
Gmj
All-active Carnivore
Mm Sh
(3)
Omnivore
Pm
(7) Herbivore Gb Gr
Diurnal
GmGI
(7) Omnivore (C)
Pw Nn
Omnivore (H)
Pa
Figure 2. A model of ecological segregation based on partitioning of space, time and diet. Lentic and lotic refer to species
resident in this zone; migratory refers to species that move between these two zones; diurnal and nocturnal refer to times
of activity; all-active, active at all times; omnivore (H) and omnivore (C) refer to omnivorous diets but with herbivorous
and carnivorous tendencies, respectively; numbers in brackets indicate the number of species in that category; acronyms
are of species listed in Table 1.
Most studies of the feeding habits of tropical stream fish have found a large degree of dietary
overlap and a heavy reliance on allochthonous food, mainly terrestrial arthropods and detritus
of terrestrial origin (Lowe-McConnell 1975). However, Moyle & Senanayake (1984) found that
overlap was low and that there was heavy reliance on autochthonous material, mainly algae and
aquatic invertebrates. Methodology and time of study have been suggested as possible reasons
for the different findings (Moyle & Senanayake 1984). The present study suggests that reliance
on autochthonous or allochthonous material depended on whether the fish inhabited lotic or lentic
wnes and, in the case of the latter, whether a particular species was surface or bottom dwelling.
The degree of dietary overlap between species can be either high or low depending on factors
such as food abundance, seasonality and the ability of species to utilise different diets.
Furthermore, this study suggests that even though dietary overlap may be high, times of foraging
activity may be different, resulting in a decrease in the level of competition.
This study indicates that differences in space, time and food utilisation play an important role
in the co-existence of freshwater fish species. It also supports the suggestion that fish community
structure is determined by the range of available resources and the associated specialisation of
the co-existing species (Watson & Balon 1984a).
LITERATURE CITED
ANGERMEIER, P.L. & KARR, J.K. 1983. Fish communities along enviroIllIlentai gradients in a system of tropical
streams. Envirorunental Biology of Fishes 9:117-135.
BALON, E.K. & STEWART, D.L 1983. Fish assemblages in a river with unusual gradient (Luongo - Zaire System,
Africa), reflections on river zonation, and description of another new species. Envirorunental Biology of Fishes
9:225-252.
BISHOP, I.E. 1973. Limnology of a small Malaysian river, Sungai Gombak. Monographiae Biologicae 22, Junk, The
Hague.
CHOY, S.C. & CHIN, P.K. 1994. Freshwater fishes from the headwaters of the Belalong-Temburong river system,
Brunei Darussalam, Borneo. R4fJles Bulletin of Zoology 42 (4):757-774.
CHOY, S.C., this volume. Distributional ecology of freshwater fishes in the Belalong-Temburong river system, Brunei
Darussalam.
CONNELL, J.H. 1978. Diversity in tropical rain forests and coral reefs. Sdence 199:1302-1310.
DYKES, A., this volume. Analysis of factors contributing to the stability of steep hillslopes in the tropical rainforest of
Temburong, Brunei Darussalam.
GATZ, AJ. 1979. Community organization in fishes as indicated by morphological features. Ecology 60:711-718.
GORMAN, D.T. & KARR, J.R. 1978. Habitat structure and stream fish communities. Ecology 59:507-515.
GOULDING, M. 1980. The fishes and theforest. University of California Press, Berkeley.
HELFMAN, G.S. 1978. Patterns of community structure in fishes: summary and overview. Envirorunental Biology of
Fishes 3:128-148.
INGER, R.F. & CHIN, P.K. 1%2. The freshwater fishes of Northern Borneo. Fieldiana: Zoology 45:1-268.
LOWE-McCONNELL, R.H. 1975. Fish communities in tropiruljreshwaters. Longman, London.
LOWE-McCONNELL, R.H. 1987. Ecologirul studies in tropirul fish communities. Cambridge University Press,
Cambridge.
MAGURRAN, A.E. 1988. Ecologirul diversity and its measurement. Croom Helm, London.
MOYLE, P.B & SENANA YAKE, F.R. 1984. Resource partitioning among the fishes of rainforest streams in Sri Lanka.
Journal of Zoology, London 202:195-223.
PIANKA, E.R. 1981. Resource acquisition and allocation among animals. pp. 300-314 in Townsend, C.R. & Calow, P.
(eds). Physiologirul ecology: an evolutionary approach to resource use. Blackwell Scientific, Oxford.
SALE, P.F. 1982. Stock-recruitment relationships and regional coexistence in a lottery competitive system: a simulation
study. Amerirun Naturalist 120:139-159.
SCHOENER, T.W. 1986. Resource partitioning. pp. 91-126 in Kikawa, J. & Anderson, DJ. (eds). Community ecology:
patterns and processes. Blackwell Scientific, Melbourne.
WATSON, DJ. & BALON, E.K. 1984a. Ecomorphological analyses of fish taxocenes in rainforest streams of northern
Borneo. Journal of Fish Biology 25:371-384.
WATSON, DJ. & BALON, E.K. 1984b. Structure and production of fish taxocenes in tropical rainforest streams of
northern Borneo. Canadian Journal of Zoology 62:927-940.
WOOTON, R.J. 1992. Fish ecology. Thompson, New York.
Spatio-temporal resource utilization by a Bornean rainforest
herpetofauna: Preliminary results
INDRANEIL DAS i
Department of Biology, Universiti Brunei Darussalam, Bandar Seri Begawan 2028, Brunei
Darussalam
ABSTRACT. Field work was conducted on the community organization of the herpetofauna at Batu Apoi
Forest Reserve, a lowland mixed dipterocarp forest in Brunei Darussalam. The present communication
is a preliminary analysis of the patterns of utilisation of microhabitat and diel-time resources by the more
abundant species of amphibians and reptiles at the study site. A total of fifty-two species of amphibians
and forty-six species of reptiles were recorded, both faunas composed of a species-poor riparian and a
species-rich non-riparian assemblage. Several species recorded at the Reserve are frequently human
commensals, and therefore indicators of disturbed habitats. Human activities tend to create new niches
or expand existing ones that are rapidly filled by these weed species associated with man. Analysis of the
microhabitat and diel-time reveal that the non-riparian assemblage has more space and time-specialists
than the riparian one. It is concluded that high diversity is maintained in the more climatically-stable
forested zones by species specializing in the use of both microhabitat and diel-time.
KEY WORDS: Brunei Darussalam, amphibians, reptiles, rainforest, diversity, community structure,
microhabitat, diel-time, resource use, human commensals.
INTRODUCTION
The organization of communities and assemblages of tropical vertebrates continue to be

poorly-understood, although a plethora of recent studies in the past two decades have shown
that various factors, including competition (Pianka 1986) and predation (Lowe-McConnell
1975) play important roles in shaping natural communities. Holt (1984) suggested that both
competition and predation may complement each other in structuring communities.
I commenced long-term field studies on the herpetological community at Batu Apoi,
Brunei Darussalam, Borneo, in January, 1991. The two goals of this research were to
prepare an inventory of the amphibians and reptiles of the Batu Apoi Forest Reserve, soon
to be gazetted a National Park, and to examine the organization of the herpetological
assemblages. This paper presents preliminary results of investigations of the organisation of
amphibian and reptile species in time and space at Batu Apoi.
STUDY AREA AND METHODS
Field work was conducted at Batu Apoi, a lowland rainforest site in Temburong District,
Brunei Darussalam, in north-western Borneo. A general description of the site, including
topography and vegetation, has been given by Edwards (1992). In January, 1991, a ca. 6
hectare area on the West Ridge of the Belalong Valley, in the vicinity of the Kuala Belalong
Field Studies Centre, was selected for intensive field work. This area included both riparian
and non-riparian habitats and covered an altitudinal range from ca. 60-80 m (the riparian
lpresent address: Centre for Herpetology, Madras Crocodile Bank Trust, Post bag 4, Mamallapuram 603104, Tamil
Nadu, India
315
D. S. Edwards et al. (eds.), Tropical Rainforest Research - Current Issues. 315-323.
316 INDRANEIL DAS
zones, where the rivers Sg. Belalong and Sg. Temburong are > 10 m in width) to 80-110 m
(the maximum height of the non-riparian zone that was sampled, which included streams < 2
m in width). Of the species listed in Tables 1 and 2, only two (Megophrys baluensis and
Ptychozoon horsfieldil) were taken outside this altitudinal range, both from the East Ridge
of the Belalong Valley at an altitude of ca. 440 m. Four transects were opened, three along
forest streams, one through a mixed dipterocarp forest. Over one thousand four hundred
hours were spent sampling the fauna. The effective observation area was ca. 2 m off either
side of the transect line and about 3.0 m off the ground, and observations were thus over an
area of 206,500 m2 , the sampling volume being 619,500 m3 of the study area. These figures
refer to the total area and volume sampled, the same transects being repeatedly sampled
during all months of the year and during all times of the day. In addition to hand collecting
along the transects, pit-fall traps in association with drift fences were employed, as were
blow-pipes and rat-glue on tree-trunks, the last two techniques with limited success. The
collection has been deposited at the Department of Biology, Universiti Brunei Darussalam,
Bandar Seri Begawan, with a voucher series being deposited at the Zoological Reference
Collection, National University of Singapore, Singapore. All visual and many acoustic
observations were noted. Data recorded for each individual specimen observed and/or
collected include time and date of observation, vertical position, microhabitat and behaviour.
General locations of collection were noted and the data subsequently entered into a
Geographical Information System (GIS) database, maintained by the Universiti Brunei
Darussalam (Wills, this volume). To prepare a more complete inventory of the species of
amphibians and reptiles of the Reserve, species that were collected by colleagues have also
been included. Resource data matrices (m x n) were created for microhabitat and diel-time
resources separately for the riparian and non-riparian assemblages. The microhabitat
categories were based on subjective evaluations (e.g., saplings 0-30 cm above the forest
floor, under fallen logs, tree-holes), the diel-time categories were one-hour partitions of the
24-hour diel cycle. Only the relatively more abundant species (n = 5 and above) were
utilised for studies on the use of time and space.
Niche overlaps were calculated using a symmetrical version (fide Pianka 1986) of
MacArthur and Levins' (1967) equation:
o = E:=1 Pij Pile

j1c /L Pi? Pu?-
where, Pij and Pi.< are the proportions of the ith resource state used by the jth and kth species.
Cluster analyses on the overlap data were conducted using SYSTAT, Version 5.03
(Wilkinson 1990).
RESULTS
Species diversity
Tables 1 and 2 list the identified amphibian and reptile species recorded from the Batu
Apoi Forest Reserve. A total of 446 individual animals were collected and/or observed, not
including a large series of eggs and larvae that were also collected. Table 3 shows the
number of species represented by numbers of individuals.
The inventory of the Reserve's herpetofauna is far from complete: Reading (1992)
conducted an eight-week extensive survey of the frogs of Batu Apoi in April-May, 1992 and
reported 28 species. His work, combined with my more intensive long-term sampling from
a smaller area (mostly the West Ridge) have resulted in a better knowledge of the amphibian
Bomean Rainforest Herpeto!auna 317
Table 1. Checklist of amphibians verified at the Batu Apoi Forest Reserve.
GYMNOPlllONA
FAMILY: ICHTHYOPlllIDAE
1. Yellow-black caecilian Caudacaedlia nigroflava (Taylor, 1960)
ANURA
FAMILY: PELOBAT1DAE
2. Large-eyed litter frog Leptobrachium montanum Fischer, 1885
3. Slender litter frog Leptola/ax gracilis (Giinther, 1872)
4. Bornean horned toad Megophrys nasuta (Schlegel, 1858)
5. Kinabalu horned toad Megophrys baluensis (Boulenger, 1899)
6. Cricket litter frog Nesobia mjobergi (Smith, 1925)
FAMILY: BUFONIDAE
7. White-eared stream toad Ansonia albomacu/ata Inger, 1960
8. Slender-footed stream toad Ansonia leptopus (Giinther, 1872)
9. Long-fingered stream toad Ansonia longidigita Inger, 1960
10. Spiny stream toad Ansonia spinulijer (Mocquard, 1890)
11. Malayan giant toad Bufo asper Gravenhorst, 1829
12. Giant river toad Bufo juxtasper Inger, 1964
13. Double-ridged toad Bufo divergens Peters, 1871
14. Brown tree toad Pedostibes hosei (Boulenger, 1892)
15. Green tree toad Pedostibes rugosus Inger, 1958
16. Short-legged toadlet Pelophryne brevipes (peters, 1867)
FAMILY: MICROHYLIDAE
17. Saffron-bellied frog Chaperina fusca Mocquard, 1892
18. Intermediate sticky frog Kalophrynus intermedius Inger, 1966
19. Red-sided sticky frog Kalophrynus pleurostigma Tschudi, 1838
20. Bornean tree-hole frog Metaphrynella sundana (peters, 1867)
21. Rufous froglet Microhyla bomeensis Parker, 1926
22. Stony froglet Microhyla petrigena Inger & Frogner, 1979
FAMILY: RANIDAE
23. Blotch-thighed stream frog Meristogenys poecilus (Inger & Gritis, 1983)
24. Dwarf mountain frog Micrixalus baluensis (Boulenger, 1896)
25. Seep frog Occidozyga baluensis (Boulenger, 1896)
26. Yellow-bellied puddle frog Occidozyga laevis (Giinther, 1858)
27. Blyth's frog Rana blythi (Boulenger, 1920)
28. White-lipped frog Rana chalconota (Schlegel, 1837)
29. Paddyfield frog Rana erythraea (Schlegel, 1837)
30. Poisonous rock frog Rana hosei Boulenger, 1891
31. Ridged river frog Rana ibanorum Inger, 1964
32. Kuhl's creek frog Rana kuhli Dumeril & Bibron, 1841
33. Smooth guardian frog Rana palavanensis Boulenger, 1894
34. Spotted stream frog Rana signata (Giinther, 1872)
35. Rock skipper Staurois latopalmatus (Boulenger, 1887)
36. Black-spotted rock frog Staurois natator (Giinther, 1858)
37. Green-spotted rock frog Staurois tuberilinguis Boulenger, 1918
FAMILY: RHACOPHORIDAE
38. Painted tree frog Nyctixalus pictus (peters, 1871)
39. Obscure bush frog Philautus tectus Dring, 1987
40. Collett's tree frog Polypedates colletti (Boulenger, 1890)
41. Five-lined tree frog Polypedates leucomystax (Gravenhorst, 1829)
42. Dark-eared tree frog Polypedates macrotis (Boulenger, 1894)
43. File-eared tree frog Polypedates otilophus (Boulenger, 1893)
44. Jade tree frog Rhacophorus dulitensis Boulenger, 1892
45. Mossy tree frog Rhacophorus everetti Boulenger, 1894
46. Harrisson's tree frog Rhacophorus harrissoni Inger & Haile, 1959
47. Emerald tree frog Rhacophorus kajau Dring, 1983
48. Wallace's flying frog Rhacophorus nigropalmatus Boulenger, 1895
49. Harlequin tree frog Rhacophorus pardalis Giinther, 1858
fauna than that currently exists for the reptiles. The snake and lizard lists appended identify
perhaps 70% and 90%, respectively, of the total fauna of the Reserve.
Inger and Voris (1992) reported that the maximum number of amphibian species collected
from a single site on Borneo (Nanga Tekalit, in Sarawak, East Malaysia) was 57. The 52
amphibians and 46 reptiles represent 36 % and 15.1 %, respectively, of the total Bornean
amphibian fauna (currently 145 amphibian and 305 reptile species; Das, unpublished). Thus
the amphibian fauna of Batu Apoi must be described as relatively diverse for Borneo.
318 INDRANEIL DAS
Table 2. Checklist of reptiles verified at the Batu Apoi Forest Reserve.
SAURIA
FANULY: GEKKONIDAE
1. Four-clawed gecko Gehyra mutilatus (Wiegman, 1835)
2. Barking gecko Gekko smithii Gray, 1841
3. Kinabalu bent-toed gecko Gonydactylus baluensis (Mocquard, 1890)
4. Peters' bent-toed gecko Gonydactylus consobrinus (peters, 1871)
5. Inger's bent-toed gecko Gonydactylus ingeri (Hikida, 1990)
6. Grooved bent-toed gecko Gonydactylus pubisulcus (Inger, 1958)
7. Asian house gecko Hemidactylus frenatus Durneril and Bibron, 1836
8. Dwarf gecko Hemiphyllodactylus typus Bleeker, 1860
9. Horsfield's flying gecko Ptychozoon horsjieldii (Gray, 1827)
FANULY: AGAMIDAE
10. Green crested lizard Bronchocela cristatella (Kuhl, 1820)
11. Crested flying lizard Draco cristatellus Gunther, 1872
12. Black-bearded flying lizard Draco melanopogon Boulenger, 1887
13. Six-lined flying lizard Draco quinquefasciatus Hardwicke & Gray, 1827
14. Bornean angle-headed lizard Gonocephalus bomeensis (Schlegel, 1848)
FANULY: SCINCIDAE
15. Common tree skink Apterygodon vittatum Edeling, 1864
16. Three-keeled ground skink Mabuya rudis Boulenger, 1887
17. Blue-bellied litter skink Sphenomorphus cyanolaemus Inger & Hosmer, 1965
18. Beccari's water skink Tropidophorus beccarii (peters, 1871)
19. Brooke's water skink Tropidophorus brookei (Gray, 1945)
FAMILY: VARANIDAE
20. Water monitor lizard Varanus salvator (Laurenti, 1768)
SERPENTES
FAMILY: BOIDAE
21. Reticulated python Python reticulata (Schneider, 1801)
FANULY: COLUBRIDAE
22. Speckle-headed whip snake Ahaetullafasciolatus (Fischer, 1885)
23. Oriental whip snake Ahaetulla prasinus (Boie, 1827)
24. White-fronted water snake Amphiesma flavifrons (Mocquard, 1887)
25. Blunt-headed tree snake Aplopeltura boa (Boie, 1828)
26. White-spotted catsnake Boiga drapiezii (Boie, 1827)
27. Twin-barred flying snake Chrysopelea pelias (Linnaeus, 1758)
28. Striped bronzeback tree snake Dendrelaphis caudolineatus Gray, 1834
29. Spotted water snake Enhydris punctata (Gray, 1849)
30. Yellow-striped water snake Hydrablabes periops (Gunther, 1872)
31. Orange-bellied snake Liopeltis baliodeirus (Boie, 1827)
32. Three-coloured ringnecked snake Liopeltis tricolor (Schlegel, 1837)
33. Malayan banded wolf snake Lycodon subcinctus Boie, 1827
34. Annulated kukri snake Oligodon annulifer (BouIenger, 1893)
35. Sinister water snake Opisthotropis typica (Mocquard, 1890)
36. Painted mock viper Psammodynastes pictus Gunther, 1858
37. White-collared reed snake Pseudorabdion albonuchalis (Gunther, 1896)
38. Dwarf reed snake Pseudorabdion longiceps (Mocquard, 1892)
FAMILY: ELAPIDAE
39. Blue coral snake Maticora bivirgatus (Boie, 1827)
40. Malayan striped coral snake Maticora intestinalis (Laurenti, 1768)
41. King cobra Ophiophagus hannah (Cantor, 1836)
FANULY: VIPERIDAE
42. Bornean flat-nosed pit viper Trimeresurus bomeensis (peters, 1871)
43. Wagler's pit viper Tropidolaemus wagleri (Boie, 1827)
TESTUDINES
FANULY: BATAGURlDAE
44. Malayan flat-shelled turtle Notochelys platynota (Gray, 1834)
45. Spiny turtle Heosemys spinosa (Gray, 1831)
FANULY: TRIONYCHIDAE
46. Malayan softshell turtle Dogania subp1ana (Geoffroy Saint-Hillaire, 1809)
Included in the fauna from this primarily lowland site is Microhyla petrigena, a species
considered to be restricted to montane zones (> 800 m) by Inger & Stuebing (1992).
Correlates of species diversity are poorly-known, and, besides historical components, may
include litter production (Inger 1980) and annual moisture regime (Duellman 1978). Scott
Bomean Rainforest Herpeto!auna 319
Table 3. Number of specimens and of species of amphibians and reptiles recorded at Batu Apoi. Figures in
parentheses are percentages of total species count (N).
No. of No. of amphibian No. of reptile

specimens species species
1 10 (22.7) 23 (57.5)
2 7 (15.9) 6 (15.0)
3 3 (6.8) 2 (5.0)
4 0(0) 0(0)
5 4 (9.1) 3 (7.5)
6 3 (6.8) 1 (2.5)
7 1 (2.3) 1 (2.5)
8 2 (4.5) 1 (2.5)
9 2 (4.5) 1 (2.5)
10 0(0) 1 (2.5)
11 1 (2.3) 1 (2.5)
12 0(0) 0(0)
13 1 (2.3) 0(0)
14 1 (2.3) 0(0)
15 2 (4.5) 0(0)
16 2 (4.5) 0(0)
17 2 (4.5) 0(0)
18 1 (2.3) 0(0)
19 0(0) 0(0)
20 1 (2.3) 0(0)
21 0(0) 1 (2.8)
22 0(0) 0(0)
23 0(0) 0(0)
24 1 (2.3) 0(0)
N = 44 (100) N = 40 (100)
(1976) found a negative correlation between the length of the dry season and species richness
of rainforest litter herpetofauna.
Diel-temporal resource use

Fifty-two species of amphibians were recorded from the study site (Table 1 lists only the
verified species). The amphibian fauna can be divided into three components: aquatic or
aquatic-margin dwelling (23 species), terrestrial and/or fossorial (14 species) and arboreal
(15 species). All species of amphibians, at least as adults, at the site are nocturnal, except
Rana palavanensis, which was collected both by day and at night (but see below). Forty-six
species of reptiles were recorded from Batu Apoi (Table 2). Three turtles were recorded, one
(Dogania subplana) from the main river, one from the forest streams (Notochelys platynota)
and one from the dipterocarp forest floor (Heosemys spinosa). No tortoises have been
recorded (only one species of the family Testudinidae, Manouria emys, occurs on Borneo,
fide Iverson 1992). Freshwater turtles in general are scarce in the study area, possibly a
result of hunting in the past by Iban tribesmen, and on account of their rarity, no further
mention of turtles will be made in the subsequent discussion on community structure.
The squamates include 12 riparian lizards (two, Sphenomorphus cyanolaemus and
Gonydactylus consobrinus, also occurring on the non-riparian sites), 11 riparian snakes (two,
Trimeresurus bomeensis and Tropidolaemus wagleri, also occurring on the non-riparian
sites); 13 non-riparian lizards, 14 non-riparian snakes. The lizard fauna is composed of a
greater number of diurnal species (12) than nocturnal ones (9); the snake fauna has equal
number of diurnal and nocturnal species (13). Classified by substrates used, 15 lizards are
arboreal, while six are terrestrial or with limited scansorial capabilities. Among snakes,
eight are arboreal, ten largely terrestriallfossorial and four primarily aquatic.
All nocturnal lizards (nine species) are arboreal. Among the snakes, the diurnal species
have six terrestriallfossorial and four arboreal members, while the nocturnallcrepuscular ones
320 INDRANEIL DAS
have five terrestriallfossorial and five arboreal forms. Aquatic snakes at Batu Apoi include
both diurnal (two species) and nocturnal (two species) forms.
Classifying species according to activity patterns is beset with at least one problem: some
species show no clear-cut pattern. The pit vipers, Tropidolaemus wagleri and Trimeresurus
borneensis were found abroad both by day and at night, but have been considered
nocturnal/crepuscular here, based on data available from other sites. Rand and Myers (1990)
mentioned that ambush hunters at a New World tropical forest site show similar diel-activity
patterns. A possible explanation for this observation is that the sedentary foraging mode is
energetically less demanding, and predators practicing it can afford (and probably need) to
forage for longer periods, according to the optimum foraging theory. Juveniles of the aquatic
Rana kuhlii were also taken both by day and at night, the day time activity of these size-
classes possibly due to the need to feed continuously in order to grow out of the more
vulnerable smaller body size. Excluding these problematic species, a single frog appear to
show bimodal peaks of diel-activity: Rana palavanensis, the adults of which were found on
the leaf litter of the forest floor during the afternoons and during the late evenings.
Microhabitat use
Two assemblages of amphibians and reptiles are readily differentiated, with little overlap
in terms of species composition, the assemblages being spatially separated along a
topographical gradient: the riparian and the non-riparian assemblages. The first is composed
of 18 medium to large amphibians, including four bufonids, nine ranids, three rhacophorids,
one microhylid and one ichthyophid. The non-riparian assemblage is much more diverse (38
species) but its members are smaller in mean body size, and include six pelobatids, six
bufonids, five microhylids, nine ranids, twelve rhacophorids and one ichthyophid.
The riparian assemblage of reptiles is composed of 23 species, including one trionychid,
five gekkonids, two agamids, three scincids, one varanid, one boid, seven colubrids, one
elapid and two viperids. The non-riparian assemblage is slightly more speciose (27 species),
and is composed of two batagurids, five gekkonids, three agamids, three scincids, ten
colubrids, two elapids and two viperids.
Three amphibians (Rana blythi, Staurois natator and Rhacophorus kajau) and four reptiles
(Gonydactylus consobrinus, Sphenomorphus cyanolaemus, Trimeresurus borneensis and
Tropidolaemus wagleri) were found in both riparian as well as non-riparian habitats, the taxa
thus shared between the two zones being 5.8% for amphibians and 8.7% for reptiles. Rana
blythi is a large ranid that utilises forest streams, but may also be found along large rivers,
while Rhacophorus kajau, usually a forest species, has been recorded breeding in the riparian
zone. Staurois natator has been collected from riparian habitats as well as from non-riparian
ones, far from any permanent waterbodies. Both Gonydactylus consobrinus and
Sphenomorphus cyanoiaemus are associated with large trees, the first utilising tree-trunks,
the second a tree-buttress specialist. Both are comparatively rare in the riparian zones.
Trimeresurus borneensis and Tropidolaemus wagleri have been found in both the riparian and
non-riparian zones, microhabitats associated with this species being low saplings as well as
from the litter, and on account of their comparative rarity, no generalisation of habitat
preference will be made. The non-riparian amphibian (38) and reptile (27) faunas were thus
found to be more diverse than the riparian ones (18 and 23, respectively), a pattern similar
to that reported from the Western Ghats, south-western India by Inger et al. (1987). Another
similar trend observed is the scarcity of animals under logs: a single Sphenomorphus
cyanolaemus was found in such a microhabitat.
Structure of space and time guilds

Cluster analysis provides a visual representation of guild structure based on the niche
overlap data. A general observation of the figures produced using cluster analyses (which will
Bomean Rainforest Herpetofauna 321
be presented in a later paper on resource use when the studies currently in progress are
complete) of the niche overlap data on diel-time and microhabitat use separately for the
riparian and non-riparian assemblages is that the relatively more speciose non-riparian
assemblage has a larger cluster with larger guilds than the depauperate riparian one. A few
congeneric species that are phylogenetically (and therefore morphologically) close fall far out
ecologically, including the species-pairs Rana kuhlii - R. palavanensis, Ansonia albomaculata
- A. spinulifer and Leptolalax gracilis - Leptobrachium montanum. Inexplicably, the above
pattern is thus similar to that observed by Inger & Colwell (1977) at a seasonal locality in
Thailand, but at variance with the observations of Inger (1969) at a non-seasonal site in
Borneo. Frogs and lizards are part of the same assemblage, if microhabitat use is
considered. These two groups also utilised similar trophic resources (studies on their
foraging ecology are now in progress), although to recognise functional guilds (sensu Root
1967), a broader taxonomic range of organisms needs to be studied. For instance, the
diurnal, arboreal, insectivorous lizard Gonocephalus bomeensis may be a member of a forest
guild along with understorey birds. The cluster showing the microhabitat-association of
riparian species shows a tightly constrained group of two true frogs (Rana chalconota and
R. signata) and a toad (Bufo juxtasper), all of which utilise the banks of rivers. The rest of
the riparian species are loosely associated, the species clearly set apart from its sympatrics
in its use of space being the agamid Bronchocela cristatella, which perches on low vegetation
and on tree trunks along river banks. In its use of temporal resources, the agamid is once
again far apart from the rest of the members of its assemblage, being diurnal, similar to the
monitor lizard Varanus salvator. However, the latter species utilises river-edges and river
banks, and is thus spatially separated. Rana kuhlii shows another remarkable pattern as it
falls out from the rest of the non-riparian species in both the use of time and space. Adults
of the species are nocturnal, newly-metamorphosed juveniles being taken both by day and
at night. The juveniles of Occidozyga baluensis show this pattern too, but to a lesser degree.
Relatively fewer species-pairs in the non-riparian habitats show high overlaps (qk > 0.06)
compared to riparian ones for both diel-time (13.9% versus 14.3%) and microhabitat (0%
versus 17.9%) resources. This indicates the presence of many more time and space
specialists in the non-riparian zones.
New niches or better opportunities?

Several species normally associated with humans have been recorded from within the
boundaries of the Reserve several months after the establishment of the Kuala Belalong Field
Studies Centre at Batu Apoi and are thus considered human commensals. These include the
frogs Rana erythraea and Polypedates leucomystax and the lizard Hemidactylus frenatus.
These co-exist with man and are therefore indicators of disturbed habitats. The gecko was
transported with the logs used in the construction of the Field Studies Centre (Earl of
Cranbrook, personal communication 1992), but the frogs were recorded approximately one
year later. Human activities tend to create new niches that are rapidly occupied by these
weed species which appear to follow man, but may also persist in low densities at the edges
(such as riparian zones) of undisturbed forests. All three newcomers have been recorded only
from the riparian areas. Any effects of these edge species on the local herpetofauna are at
present unknown.
DISCUSSION
Diversity and inventory methods

Preparing inventories of rainforest herpetofaunas is an extremely labour-intensive activity,
and rapid assessment methods are likely to include only the more abundant species. Over
322 INDRANEIL DAS
45 percent of the amphibians and 77 percent of the reptiles are represented by three or less
specimens (Table 3), although collections spanned more than a year, and included both dry
and wet periods. In the case of at least the tree frogs, the low numbers obtained may be
attributed to their use of arboreal (and therefore difficult to sample) habitats. The same
explanation may hold for the eight arboreal snakes that are represented in the collection by
a single specimen. However, the scarcity of many other groups of amphibians and reptiles,
including species from 'accessible' habitats, such as true toads, caecilians, litter skinks and
terrestrial and aquatic snakes, may not be a sampling bias, and appears to represent their true
rarity in the Reserve. Therefore, rapid assessment methods to identify areas of high
biological diversity for conservation purposes may be unsuitable for amphibians and reptiles
in rainforest habitats, where species diversity is apparently offset by low densities of many
individual species. Short-term surveys thus would tend to be biased towards the more
abundant, r-selected species, such as weed species of the forest edges, rather than the true
rainforest endemics.
Patterns of spatia-temporal resource use

The interactions of the habitat and species is evident from the faunal list: the non-riparian
zone has six species of pelobatids (litter frogs) that live on the leaf litter, the riparian zone
has none; the non-riparian zone also has twelve species ofrhacophorids (tree frogs), only one
of which co-occurs in the riparian zone. Another family of anurans well represented in the
first zone, with five species (all ant and small beetle specialists), is the Microhylidae, which
has a single species in the riparian areas. The non-riparian zone's greater species diversity,
relative to that of the riparian zone, is attributed to its greater habitat diversity, complexity
(including the thick accumulated layer of leaf litter and mature dipterocarp trees over 30 m
in height) as well as less stable environmental factors in the riparian areas (such as the river,
whose level shows considerable variation on a temporal scale).
The presence of more time- and space-specialists at the non-riparian sites is in
concordance with Margalef's (1963) prediction that as relationships between species become
more specialised, stenophagy and other precise relationships increase, leading to reduced
niche overlap in rainforest communities. Since the non-riparian assemblage is much more
species-diverse than the riparian one, with species showing relatively less resource sharing
habits, specialization in the use of time and space resources appears to be a factor allowing
high species diversity in the more climatically stable environment of the non-riparian zones.
ACKNOWLEDGEMENTS
A large number of people contributed to this research, from arranging logistics to keeping
an eye open for specimens. A partial list includes: Webber Booth, Joe Charles, The Earl
ofCranbrook, Satish Choy, David Edwards, David Jones, Chris Kofron, Ruth Levy, Hanafi
Hj Manaf, Samhan Nyawa, Bert Orr, Catriona Prebble, Kamariah Abu Salim and John Wills.
Robert F. Inger and David Jones read earlier drafts of the manuscript. The project was
supported by a grant (UBD/T6/RG 03) from Universiti Brunei Darussalam.
LITERATURE CITED
DUELLMAN, W. W. 1978. The biology of an equatorial herpetofauna in Amazonian Ecuador. Miscellaneous
Publications of the Museum of Natural History, University of Kansas 65:1-352.
EDWARDS, D.S. 1992. The riches of Batu Apoi. Muhibbah Jan/Feb 1992:24-26.
HOLT, R. D. 1984. Spatial heterogeneity, indirect interactions, and the coexistence of prey species. American
Bornean Rainforest Herpeto!auna 323
INGER, R. F. 1969. Organization of communities of frogs along small rain forest streams in Sarawak. Journal
of Animnl Ecology 38:123-148.
INGER, R. F. 1980. Densities of floor-dwelling frogs and lizards in lowland forests in southeast Asia and Central
America. American Naturalist 115:761-770.
INGER, R. F. & COLWELL, R. K. 1977. Organization of contiguous communities of amphibians and reptiles in
Thailand. Ecological Monographs 47:229-253.
INGER, R. F., SHAFFER, H. B., KOSHY, M. & BAKDE, R. 1987. Ecological structure of a herpetological
assemblage in South India. Amphibia-Reptilia 8: 189-202.
INGER, R. F. & STUEBING, R. B. 1992. The montane amphibian fauna of" northwestern Borneo. Malayan
Nauture Journal 46:41-51.
INGER, R. F. & VORIS, H. K. 1992. Variation in communities of amphibians in Bornean forests. pp. 494-495 in
Ismail, G., Mohamed, M. & Omar S. (eds). Forest biology and conservation in Borneo. Proceedings of the
International Conference on Forest Biology and Conservation in Borneo. July 30-August 3, 1990, Kota Kinabalu.
Yayasan Sabah, Kota Kinabalu.
IVERSON, J. B. 1992. A reVised checklist with distributional mnps of the turtles of thf' world. Privately printed,
Richmond.
LOWE-McCONNELL, R. H. 1975. Fish communities in tropical freshwaters: their distribution, ecology and
evolution. Longman, New York.
MACARTHUR, R. H. & LEVINS, R. 1967. The limiting similarity, convergence, and divergence of coexisting
species. American Naturalist 101:377-385.
MARGALEF, R. 1963. On certain unifying principles in ecology. American Naturalist 97:357-374.
PIANKA, E. R. 1986. Ecology and natural history of desert lizards. Princeton University Press, Princeton, New
Jersey.
RAND, A. S. & MYERS, C. W. 1990. The herpetofauna of Barro Colorado Island, Panama: An ecological
summary. pp. 386-409 in Gentry, A.H. (ed.). Four Neotropical rainforests. Yale University Press, New Haven
& London.
READING, C. J. 1992. Frogs found within the Batu Apoi Forest Reserve in Brunei. Unpublished Final Report,
UBD-RGS Rainforest Project, 1991-1992.
ROOT, R. 1967. The niche exploitation pattern of the blue-gray gnat-catcher. Ecological Monographs 37:317-350.
SCOTT, N. J. 1976. The abundance and diversity of the herpetofaunas of tropical forest litter. Biotropica 8:41-58.
WILKINSON, L. 1990. SYSTAT: The system for statistics. SYSTAT, Inc., Evanston, Illinois.
WILLS. J. T. (this volume). The creation of a Geographical Information System for th" Kuala Belalong Field
Studies Centre.
The altitudinal distribution of forest tree families on the western slopes
of Mount Banahao, Luzon, Philippines
EUSTAQUIO G. ARAGONES
Forest Products Research and Development Institute, Department of Science and Technology,
College, Laguna 4031, Philippines
ABSTRACT. The vertical distribution of forest tree species represented by families was sampled in 0.42
ha quadrats at different altitudes between 750 m and 2100 m on the western slopes of Mount Banahao,
Luzon, Philippines. Species and family relative occurrence at different elevational sampling sites was
assessed. Some plant groups such as the MeJiaceae, Lauraceae and Myrtaceae displayed wider distribution
vertically, being represented from 750 m to 1500 m. Other groups were confined to the summit zone,
e. g. Podocarpaceae, Taxaceae, Ericaceae and Winteraceae. Families typical of lowland rainforest, such
as Anacardiaceae, Euphorbiacae, Clusiacaeae, Myristicaceae, and the Nyctaginaceae did not extend
beyond 750 m. The distribution pattern observed is discussed in relation to plant adaptation, dispersal and
the effect of elevation.
KEY WORDS: ecological response, megatherrn, microtherm, elevation effects.
INTRODUCTION
On tropical mountains both woody and herbaceous plants at all altitudes, according to Van
Steenis (1961, 1962), are derived from autochthonous (locally derived) and allochthonous
sources (from plants that are not native to the area). Autochthonous families or genera such
as the Annonaceae, Bombacaceae, Dipterocarpaceae, Meliaceae, Rhizophoraceae,
Marantaceae and Ficus mostly find their uppermost limit in the colline subzone below 1000
m. Normally, they fade rapidly away between 1000 m and 2000 m and only a few are
represented at higher elevations. Conifers, tree ferns and some angiosperm families familiar
to temperate botanists may exist in the tropics only on mountains. Some families, however,
e. g. Labiatae, Oleaceae, Scrophulariaceae and Theaceae have wide altitudinal distribution
in the tropics and wide latitudinal distribution over the globe. Certain species of oaks and
chestnuts, Ajuga, Daphne, Podocarpus, Rhododendron, and Vaccinium occur even at sea
level in Malesia. It is accepted that these genera or families have wide ecology and are able
to tolerate a wide range of temperature. For these, Van Steenis (1961) used the term
megatherm for heat-adapted and microtherm for cold-adapted species.
The effect of elevation influences various ecological factors, the most important of which
are temperature, soil, autecology of the species, dispersal biology and man's activities.
This paper presents and discusses the composition of forest tree species and families in
six 0.42 ha quadrats positioned along an altitudinal gradient on the western slopes of Mt.
Banahao, Luzon, Philippines.
METHODS
The Study Area

The study was carried out on the western slopes of Mt. Banahao, Luzon (14.5° N, 121.5°
325
326 EUSTAQUIO G. ARAGONES
E) from the low-lying Barangay (Bgy) Kinabuhayan, Dolores, Quezon, some 550 m above
sea level, to the summit at 2100 m.
From Bgy Santa Lucia to Bgy Kinabuhayan, the steep slopes and the general ruggedness
of Mt. Banahao are evident. Huge basaltic and andesitic rock outcrops are prominent within
upland farms and in the forest itself, to an elevation of 750 m.
The climate map of Quezon province, Luzon (1975 Soil Survey Report) indicates a Type
II climate for Dolores Quezon, that is with a pronounced maximum rain period from August
to December and without a dry season. Annual rainfall averages 4000 mm. The mountain,
especially at higher altitudes, may have different patterns of rainfall and temperature from
the areas lying below owing to its vegetation and relief.
Sampling Procedure
Natural forest communities encountered at different altitudes (750,950, 1200, 1500, 1800
and 2100 m) were sampled by a modified quadrat method as employed by Whittaker &
Niering (1975). Two contiguous quadrats each 30 m x 70 m were laid upslope of each other
at each of the six altitudes. A systematic sampling interval of 200 m elevation, as used by
Khan (1969) in Mt. Makiling, Luzon, was adopted up to 1200 m elevation, beyond which
it was adjusted to 300 m in succeeding sites when slopes became much steeper and vertical
cliffs were encountered more often.
In each of the two quadrats at each altitude, all trees greater than 10 cm diameter at breast
height (DBH) were censused. Height and diameter were measured with a Haga altimeter and
diameter tape respectively and herbarium specimens collected from each species tallied.
The vertical range of species distribution based on their observed occurrence at each
altitudinal sampling site is presented in Tables 1 and 2.
It is evident from the data that certain plant groups such as the Myrtaceae, Lauraceae,
Fagaceae and Meliaceae have a wide vertical distribution (Figure 1). The Meliaceae were
represented from 750 m to 1500 m altitude while the Fagaceae, which were not encountered
until 950 m, peaked at 1500 m elevation then declined at 1800 m. The Lauraceae were
present in plots from 750 m to 1500 m. The Myrtaceae were present in plots from 750 m
to the summit (2100 m). Members of the Melastomataceae, Theaceae and Symplocaceae on
the other hand, appeared to be high elevation families as they were observed not lower than
950 m altitude and extended vertically to higher sites close to the summit. Other families
such as the Myristicaceae, Nyctaginaceae, Staphyleaceae, and Rutaceae were observed only
at 750 m.
Such patterns of distribution, particularly along gradients of elevation, seem to indicate
major differences in the adaptation of different plant groups to varying niche and habitat
types as influenced by relief and soil conditions (Whittaker 1965, Whittaker et al. 1968). The
existence of what Van Steenis (1934, 1961, 1962) called microtherm taxa (e. g. Drimys,
Rhododendron, Podocarpus and Taxus) at the summit zone of Mt. Banahao's western slopes
(1800 - 2100 m) may well be considered an ecological response to the much lower
temperature. Van Steenis (1934, 1961, 1962) further stressed that these groups of temperate
origin may only be present at high altitudes in the tropics.
The absence of high altitude microclimatological data for most Philippine mountains makes
it difficult to determine exact differences in climate, particularly of temperature, from low
to high elevations. Nonetheless, if an adiabatic lapse rate of 0.67° C is assumed per 100 m
rise in elevation, the summit of Mt. Banahao (2100 m) would have a temperature of about
13 - 14° C, compared to a mean of about 26 - 27° C at sea level. This approximates the
Altitudinal distribution of forest tree families 327
Table 1. Species observed at two or more altitudinal sampling sites on the western slopes of M. Banahao, Luzon.
Altitude (m) Species Family

550-750 Pipturus arborescens Urticaceae
Ficus Ilota Moraceae
550-950 Mallotus multiglandulosus Euphorbiaceae
Ficus septica vaT. septica Moraceae
750-950 Parashorea malaallonall Dipterocarpaceae
Ficus gul Moraceae
Aglaia llallosiana Meliaceae
Ficus millahassae Moraceae
Elaeocarpus calomala Elaeocarpaceae
Saurauia latibractea Actinidiaceae
Cillnamomum mercadoi Lauraceae
Dracolltomeloll edule Anacardiaceae
Syzygium polycephaloides Myrtaceae
Palaquium philippellse Sapotaceae
Ficus ampelas vaT. ampelas Moraceae
Neonauclea media Rubiaceae
750-1200 Euphoria didyma Sapindaceae
Pedicellia juscescells Sapindaceae
Cryptocarya ampla Lauraceae
Acmella acuminatissima Myrtaceae
Acalypha stipulacea Euphorbiaceae
Palaquium merrillii Sapotaceae
Palaquium foxworthii Sapotaceae
750-1500 Aglaia luzolliellsis Meliaceae
Syzygium Ilitidum Myrtaceae
950-1200 Ficus balete Moraceae
Dysoxylum altissimum Meliaceae
Aglaia dijfusa Meliaceae
Palaquium luzolliellsis Sapotaceae
Syzygium simile Myrtaceae
Eusideroxyloll zwageri Lauraceae
950-1500 Cleistocalyx operculata Myrtaceae
950-1800 Lithocarpus spp. Fagaceae
950-2100 Symplocos spp. Symplocaceae
1200-1500 Syzygium maillitense Myrtaceae
Memeryloll spp. Melastomataceae
Syzygium decipiells Myrtaceae
1500-2100 Eurya sp. Theaceae
record of Brown (1919) from the northeastern slopes of Mt. Banahao, and that of Mohr &
Van Baren (1954) in the mountains of Indonesia.
The higher number of species occurring at low and medium altitudes (Tables 1 and 2)
could attest to a preponderance of generalist species in the tropics which have wide ecological
tolerance. In contrast, the most distinctive specialist group in this study is confined to the
summit zone (Table 3), where temperature is much lower, making it difficult for the natural
megatherms to adapt (Cockburn 1978, Corner 1978, Whitmore 1975). Along this line of
reasoning, the altitudinal distribution observed of families (Figure 1) and even that of genera
Table 2. Species observed at only one altitudinal sampling site.
Altitude (m) Species Family

750 Goniothalamus amuyon Annonaceae
Neotrewia cumingii Euphorbiaceae
Artocarpus ovatus Moraceae
Pisonia umbelliferae Nyctaginaceae
Pterospermum niveum Sterculiaceae
Diplodiscus paniculatus Ulmaceae
Octomeles sumatrana Datiscaceae
Pagiantha plumerilifolia Flacourtiaceae
Ficus botryocarpa var. botryocarpa Moraceae
Mallotus philippinensis Euphorbiaceae
Myristica philippinensis Myristicaceae
Grewia multiflora Tiliaceae
Ficus pubinervis Moraceae
Alstonia scholaris Apocyanaceae
Gorcinia binucao Clusiaceae
Homalanthus sp. Euphorbiaceae
Aglaia iloilo Meliaceae
Diospyros philippense Ebenaceae
Toona calantas Meliaceae
Chisocheton pentandrus Meliaceae
Dysoxylon arborescens Meliaceae
Laportea luzonensis Urticaceae
Melicope triphylla Rutaceae
Ficus congesta Moraceae
Allantus integrifolia Simaroubaceae
Litsea perrotetii Lauraceae
Pterocymbium macro cater Sterculiaceae
Palaquium sp. Sapotaceae
Cratoxylum sumatranum ssp. sumatranum Hypericaceae
Ervatamia sp. Apocyanaceae
Canarium asperum Burseraceae
Lepisanthes schizolepis Sapindaceae
Ahernia glandulosa Flacourtiaceae
Phaleria capitata Thymelaeaceae
Casearia fuligunosa Flacourtiaceae
Terminalia foetidissima Combretaceae
Epicahris cumingiana Meliaceae
Myristica elliptica var. simiarum Myristicaceae
Bischojia javanica Bischofiaceae
950 Nephelium sp. Sapindaceae
Aphanamixis perroUetiana Meliaceae
Leucosyke capitellata Urticaceae
Duabanga moluccana Sonneratiaceae
Celtis luzonica Ulmaceae
Cratoxylum arboreum Hypericaceae
Trichadenia macrantha Flacourtiaceae
Cratoxylum formosum ssp. formosum Hypericaeae
Diospyros copelandii Ebenaceae
Ficus variegata var. variegata Moraceae
1200 Aglaia longipes Meliaceae
Ardisia sp. M yrsinaceae
1500 A/seodaphne longipes Lauraceae
Cryptocarya sp. Lauraceae
Lansium dubium Meliaceae
Neolitsea sp. Lauraceae
Syzygium decipiens Myrtaceae
Altitudinal distribution of forest tree families 329
Table 3. Species observed exclusively at the swnmit zone (1800 - 2100 m).
Species Family
Dacrycarpus cumingianus Podocarpaceae
Podocarpus rotundus Podocarpaceae
Taxus sumatrana Taxaceae
Clethra sp. Clethraceae
Rhododendron sp. Ericaceae
Drimys piperita Winteraceae
50
X
45
40
1/1 35
iii
'0
:::l
X
...... ... _x~o 0
\
'>
'/\
30
'0
c
25
~
x___
0
...GI 20 \0..
I~
..
.D
~
E
15 X~x
, I
10 0
,
I
5 0.. ,
,
I
0
~o
.... , .... I
"0
550 750 950 1200 1500 1800 2100

Altitude (m)
Figure 1. Altitudinal distribution of Meliaceae, Lauraceae, Myrtaceae and Fagaceae on the western slopes of Mt.
Banahao, Luzon. x - - x Meliaceae; x-------x Lauraceae; ()----------{) Myrtaceae; 0-------0 Fagaceae.
and species could be purely ecological in context and the effect of elevation obscured (Van
Steenis 1934, 1962). Even so, the elevation effect may still fall within the bounds of
ecological preferences as the latter takes into consideration and gives emphasis to the
significance of dispersal or dissemination of propagules (Hengeveld 1985).
Altitudinal Distribution in relation to elevation effect

To illustrate the elevation effect in the sense used by Van Steenis (1934, 1961, 1962), the
vertical range of distribution of the four most important families of trees in the western
slopes of Mt. Banahao (Figure 1) is further assessed.
The Myrtaceae had a peak density at 1500 m altitude and extended up to the summit zone,
where only one species, Leptospermum flavescens, was represented. Members of the
Meliaceae were abundant at 750 m altitude but abruptly declined beyond this elevation and
faded out beyond 1500 m. The Lauraceae was abundant at 950 to 1200 m, less abundant at
1500 m and finally disappeared beyond there. The Fagaceae was absent from plots below 950
m, was abundant at 1500 m, diminished at 1800 m and was finally absent at the summit plot
(2100 m).
According to Van Steenis (1934, 1961, 1962) and Backhuys (1969) the mountain flora
could have a permanent zone of establishment above a certain level and on both sides above
and below this, upper and lower zones of temporary establishment could occur. Both authors
stressed the significance of downward dispersal ofpropagules from the permanent zone above
and also beyond it for successful development of upper and lower temporary zones. For
example, the Fagaceae could well have 1500 m altitude as their permanent zone of
establishment. However, as Hengeveld (1985) warns, drawing such conclusions from
relatively few samples and a small segment of the mountain could be hazardous.
SUMMARY AND CONCLUSIONS
Some so-called microtherm families, e.g. Winteraceae, Ericaceae and Taxaceae, are confined
to the summit zone of Mt. Banahao's western slopes where their occurrence is probably
influenced greatly by the low temperatures to which they are adapted. The generalist plant
groups like the Meliaceae, Myrtaceae, Lauraceae, Fagaceae and others proved to be of wider
distribution vertically as they adapt more to varying niche and habitat type with change in
altitude.
The elevation effect as conceived by Van Steenis (1934) may apply to the different forest
communities and plant groups observed on the western slopes of Mt. Banahao, Luzon. In
mountains, especially of great height like that of Mt. Banahao, each plant group apparently
displays permanent and temporary zones of establishment depending on source and
availability of propagules and manner of dispersal.
LITERATURE CITED
BACKHUYS, W. 1969. Der elevations-effect bei siningen alpenplanzen der scherreiz. Blumea 16:273-320.
BROWN, W.H. 1919. Vegetation of Philippine Mountains. Government Printing Bureau, Manila.
COCKBURN, R.F. 1978. The flora. In Luping, D.M., Chin, W. & Dingley, E.R. (eds). Kinabalu, Summit of
Borneo. Sabah Society, Kota Kinabalu.
CORNER, E.J.H. 1978. The plant life. In Luping, D.M., Chin, W. & Dingley, E.R. (eds). Kinabalu, Summit of
Borneo. Sabah Society, Kota Kinabalu.
HENGEVELD, R. 1985. On the explanation of the elevation effect by a dynamic interpretation of species
distribution along altitudinal gradients. Blumea 30:353-361.
KHAN, M. 1969. Some physical and chemical properties of the forested soils of Mt. Makiling and its vegetation.
Ph. D. thesis, University of the Philippines, Los Banos College, Laguna.
MOHR, E.C.I. & VAN BAREN, F.A. 1954. Tropical soils. Interscience Publishers, The Hague.
VAN STEENIS, C.G.G.J. 1934. On the origin of the Malaysian mountain flora. Part 2. Altitudinal zones, general
considerations and renewed statement of the problem. Bulletin du Jardin Botanique Buitenzorg III :289-518.
VAN STEENIS, C.G.GJ. 1961. An attempt towards an explanation of the effect of mountain mass elevation.
Proceedings of the Royal Academy of Sciences of the Netherlands C64:435-442.
VAN STEENIS, C.G.G.J. 1962. The mountain flora of the Malaysian tropics. Endeavour 21:183-193.
WHITMORE, T.C. 1975. Tropical rainforests of the far east. Clarendon Press, Oxford.
WHITTAKER, R.H. 1965. Dominance and diversity in land plant communities. Science 147:250-260.
WHITTAKER, R.H. 1972. Ordination and classification of communities. Junk:, The Hague.
WHITTAKER R.H. & NIERING, W.A. 1975. Vegetation of the Santa Catalina Mountains, Arizona. Journal of
Ecology 56:523-544.
WHITTAKER, W.H., BUOL, S.W. & NIERING, W.A. 1968. A soil and vegetation pattern in the Santa Catalina
Mountains, Arizona. Soil Science 105:440-450.
Interspecific colony dispersion and niche relations of three large tropical
rain forest ant species
RUTH LEVY
Department of Zoology, University of Oxford, South Parks Road, Oxford OX] 3PS, United
Kingdom.
ABSTRACT. Three large tropical rain forest ant species (Camponotus gigas, Diacamma intricatum and
Odontoponera transversa) in primary lowland dipterocarp forest, Brunei Darussalam, were not found to
be significantly negatively associated spatially or in occurrences at baits. Food overlap was found to be
high, and time overlap quite low, but not in the species of high food overlap. No evidence of resource
partitioning was found, although each species had food preferences. Positive associations are likely to
be microhabitat mediated.
KEY WORDS: resource partitioning, Camponotus, Diacamma, Odontoponera, Borneo, Brunei

Darussalam.
INTRODUCTION
Since the pioneering work of Lotka (1932) and Volterra (1926) on predictive models of
coexistence of competing organisms, and competitive exclusion experiments by Gause (1934),
competition and resource partitioning have frequently been regarded as major community
structuring forces. These concepts have been investigated in three major ways: workers have
searched for 'substantial differences' in the niches of supposedly competing species (Schoener
1986); empirical data on niche distributions have been compared with null models in which
niches are distributed at random (May 1984); and theoretical studies have been used to
predict exactly how much overlap is permissible for species to coexist (Macarthur & Levins
1967).
Most workers have adhered to Hutchinson's definition of the niche as a 'multi-dimensional
hypervolume' of space within the limits of tolerance of factors that affect an organism's
survival (Hutchinson 1957); Pianka (1981) summarises the definitions. Today, the niche is
commonly defined as a 'utilisation distribution' (Arthur 1987, Giller 1984, Schoener 1986),
or frequency histogram, of food size classes, for example.
The relationship between niche overlap and competition between species has been quite
confused in the past (Schoener 1974), authors recently concluding that the magnitude of
overlap is a bad indicator of the level of competition (Keddy 1987, Schoener 1986).
The sorts of resources which have been found to be significantly partitioned in
communities include habitat or micro-habitat, food type or size, and time, in that order
(Pianka 1969, Schoener, 1986). This is because altering spatial utilisation has the least cost!
benefit ratio of all resource types. Not feeding at any time costs most of all. Schoener
(1986) notes that organisms most likely to be partitioned by time include predators and
ectotherms. This is less likely as species number increases. Those with a relatively high
prey size/body size ratio are often partitioned by prey type.
Workers usually choose to investigate niche relations in species of the same 'guild' (Root
1967). These are groups of species, the members of which are similar in their trophic roles,
and which interact weakly with other guilds (Pianka 1983).
However, despite the interest in resource partitioning as a mechanism enabling coexistence
331
332 RUTH LEVY
in communities, Arthur (1987) reminds us that there are at least five methods of enabling
stable coexistence in communities, and only one of these is resource partitioning.
Lynch et al. (1980) examined ecological segregation in a guild of surface foraging,
generalist ants in north American temperate forests; they found that the less aggressive
species overlapped less in their resources with the most aggressive species, and that the
weakest competitors had broader niches, an assumption of the competitive hierarchy model
of community structure (Keddy 1989). However, Culver (1974) found no relationship
between niche width and aggressiveness. Lynch (1981) found that habitat was a major factor
contributing to community diversity. Differential uses of food sizes and type have been
found to be an effect of differential desiccation resistance, body size, nest site and activity,
and to promote coexistence in a grassland ant community (Hansen 1978). Briese (1982)
found that food type and time overlaps fluctuated greatly in response to environmental
changes in semi-arid Australia, and that there was little evidence for exploitative competition.
Marsh (1985) found the same patterns in Namibia. Davidson (1977) found that differences
in foraging strategy allowed coexistence of ant species in north American deserts.
In this study, the niche relations and interspecific dispersion of a guild of three large,
generalist, solitary foraging, ground nesting ants, Camponotus gigas (subfamily Formicinae),
Diacamma intricatum and Odontoponera transversa (both Ponerinae) in a hectare plot in
Bruneian primary dipterocarp rain forest, were examined. These species were common at
all primary lowland dipterocarp rainforest sites, at all times of year, at baits, in pitfall traps
and in litter samples. The following questions were posed:
1) Is there any evidence of negative interspecific interactions spatially or at baits?
2) Do this guild of species' niches overlap substantially?
3) Are resource overlap and interspecific interactions related?
We might expect that low niche overlap would promote coexistence between the species,
and that any negative interactions between them might be evidenced by relative nest
distributions and associations at baits: if avoiding each other, nearest neighbour nests of
different species would be situated further away from each other than expected randomly,
and the species would forage less together.
This study was also intended as a descriptive one, as we have very little idea of what
tropical rain forest ant species do, and especially what they eat (Naomi Pierce, personal
communication) .
The resources used in the analysis were those of food, space and time (Schoener 1974),
namely, food type, food size, relative occurrence on the trees and the ground (space), and
numbers of entrances to and exits from the nest entrance during 24 hours (time).
METHODS
Fieldwork
The fieldwork took place in June-December 1992 in plot AAU1, on the West Ridge of
the Kuala Belalong Field Centre Reserve, Temburong, Brunei Darussalam. Nest entrance
distributions of the three species of ant were mapped by following sardine baited foragers to
nest entrances in the hectare, and noting locations of colony entrances on a map of the plot.
If multiple nest entrances were found (only in the case of C. gigas), the geometric centre of
these was noted as the colony location. The whole 100 x 100 m was mapped for C. gigas
(45 colonies), a 60 x 60 m area for D. intricatum (62 colonies) and a 40 x 40 m area for O.
transversa (367 colonies). It was not feasible to map the whole hectare for all species,
particularly O. transversa. As the frequencies of all three species were found to be even
over the hectare at baits, I considered it valid to implement the randomisation procedure
Niche relations of three rainforest ant species 333
described below on a section of the whole hectare's ant colonies. Interspecific nearest
neighbour nest distances of 45 pairs of Camponotus-Diacamma, Camponotus-Odontoponera
and Diacamma-Odontoponera were recorded.
Food samples were collected as follows:- food items carried by returning foragers were
collected at the nest entrance of the three species (at least six colonies of each species were
sampled), during the day and after dark, and preserved for later examination in 80% alcohol.
Samples were measured and identified in the laboratory.
Use of space was assessed by placing protein and carbohydrate baits at 100 points at two
levels in the hectare plot, on the litter and at eye level on tree trunks, and recording
occurrences of these species at the baits, day and night. This was repeated three times.
Each bait was visited only once in the day or in the night, in order to keep each observations
independent over time. Diel cycles were recorded by observing ant activity at the nest
entrance throughout four 24 hour cycles. All entrances and exits were recorded for four
different colonies of each species.
ANALYSES
Interspecific Nest Entrance Dispersions

It was not possible to assess the interspecific association of the three species in the same
way as other authors i.e. by using a G-test of contingency, as numbers of colonies of each
species per unit area were so vastly different, and mapping an entire hectare of O.
transversa, around 3000 nest entrances, was not feasible. I therefore decided to approach
the problem in a different way.
The median values of observed interspecific nest entrance distances of 45 nest pairs for
each species combination were compared with 1000 randomised replicates of interspecific
nest distances. For each replicate, the larger (bottom) map (c. gigas or D. intricatum) was
replicated ninefold, and the smaller (top) map (D. intricatum or O. transversa) was
effectively moved around relative to it, always keeping at least a small fraction of the smaller
map inside the centre of the nine larger map replicates. Nearest neighbour nest distances for
the pair of species were then calculated. The ninefold replication which was used as the
overlaid map was liable to go over the edge of the larger map when moved, and drastically
alter inter-nest distances. It ensured that edge effects of the larger map were not a problem
(Ryti & Case 1992), as the distance from the edge of the smaller to the edge of the larger
map replicates was always at least 20 m. Edge effects caused by proximity of the larger map
species to the edge of the smaller map were resolved by taking distances of nests only where
the colony of the larger species fell inside a prescribed boundary area (Figure 1). Edge
effects of this sort are liable to create spuriously high values for inter-nest distances. The
width of this boundary was 2 m for O. transversa and 7.5 m for D. intricatum. Colonies of
the larger map species within the internal boundary of the smaller map species were selected
and interspecific nearest neighbour distances calculated. This method also kept the colony
distributions intact during replication. This was necessary as nest distributions could not be
assumed to be random. This was repeated for each replicate (see Figure 2).
The frequency distributions of 1000 randomised replicate medians were used to find
directly the probability of the observed median result happening by chance. The number of
randomised medians smaller or larger than the observed value were counted, and a p-value
was calculated directly from this number. A negative association of nests (overdispersion)
is shown when most randomised replicates are smaller than the observed value, and a
positive association (clumping) when most randomised values are larger. The 0.05 % level
of significance (50 replicates) was used.
Figure 1. Close up view of nearest neighbour relations in randomised replicates of two species nest distributions.
The species of smaller area (open circles) is overlaid on the nine larger species replicates (closed circles), e.g. D.
intricatum overlaid onto C. gigas, or O. transversa on to D. intricatum. Valid internest distances are shown by lines
joining the nests concerned. Invalid distances are shown by scored lines. This occurs if the larger species occurs
outside the internal boundary area set for D. intricatum and O. transversa.
a a a
~a b:!l
a a a a
a m a a a a a
=
a a
a a a a a a a ~ a
Figure 2. Successive replications of randomised nearest neighbour distance calculations for two ant species (a and
b). Nearest neighbour distances between colonies of species a (larger ninefold replicated map) and b (smaller map)
are calculated for colonies of species a occurring within the internal boundaries of the map of b.
Species Association at Baits

In order to assess interspecific association between the three species of interest at baits,
contingency tests of association (Ludwig & Reynolds 1988) were performed on presencel
absence data for the three two species interactions.
For each two species interaction, the following contingency table was constructed:
species 1
oresent absent
present a b a+b
species 2
absent c d c +d
a + c b +d a+b+c+d
where: expected values (e.g. Era) for cell a)

(a +b)(a+c)
(Equation 1)
a+b+c+d
No correction for small sample size was necessary (Ludwig & Reynolds 1988), as numbers
were large.
Niche Breadths and Overlaps

It was assumed that resources were limited in availability. Niche width and resource
overlaps for food size, food type, use of vertical foraging space and use of diel time were
calculated. The Levins index of niche breadth was used (Krebs, 1989):
1 (Equation 2)
B.I =
where Bi = Levins niche breadth of species i; Pai = proportion of use of resource class
a by species i and n = number of resource classes.
As the index is maximum when all resource classes are used in equal proportion, the
maximum possible niche breadths for each resource for each species were calculated. From
this, an overall mean percentage niche breadth for each species was found.
The percentage overlap was calculated for each two species interaction (Krebs 1989):
Pij = [~(minU Pai' pQj)] 100

(Equation 3)
where Pij = percentage overlap of species i and j; Pai = proportion of use of resource class
a by species i and n = number of resource classes.
Food size niche parameters were calculated from the probability distribution function
generated from the mean and standard deviation of the logged food size data, in order to
facilitate smoothing the distribution. Food type overlap was calculated from the numbers of
prey in each of the classes for each species. Time overlap was calculated for entrances and
exits, and reduced to two classes of activity in the day and night (6 am - 6 pm and vice
versa). Space utilisation (foraging height) was calculated from the relative numbers of
occurrences of each species at tree and ground baits, day and night.
Overall percentage overlap for each pair of species was calculated in the following way:
as spatial utilisation data were dependent on day and night (time), overall percentage overlap
in time and space was calculated as a single parameter for each species. The percentage
overlap of each of four classes, ground bait/day, ground bait/night, tree bait/day and tree
bait/night were calculated and added together to give this parameter (May 1975). This
percentage was then multiplied by those for food type and food size overlap to give an
overall multiplicative percentage overlap. No inter-dependence of other resource axes were
found for any of the species.
RESULTS
Interspecific Nest Entrance Dispersions

Table 1 shows the results of 1000 replicated randomised tests on the medians of the
interspecific nest entrance distances, and the kinds of interactions going on. There are no
significant negative interactions, but the nests of D. intricatum and O. transversa are
significantly positively associated.
Species Associations at Baits

Table 2 shows the probabilities of observed occurrences of species with each other, and
the kind of interactions, positive or negative, occurring.
336 RUTH LEVY
Table 1. Observed median inter-nest distances for three species pairs, nwnbers of randomised inter-nest median
distances smaller than the observed median from 1000 replicates, probability and type of interaction happening
between measured nests.
species pair observed nwnber of p association of observed

median randomised replicates nests
C. gigaslD. intricatum 4.10 857 0.143 negative
C. gigas/O. transversa 1.00 350 0.350 none
D. intricatum/O. transversa 0.65 4 0.004 positive
Table 2. Results of test of interspecific association of three species with each other.
Species p interaction
C. gigas/ D. intricatum 0.29 NS none
C. gigas/ O. transversa 22.10 <0.001 +
D. intricatum/ O. transversa 0.67 NS none
C. gigas and o. transversa are significantly positively associated at baits.
Niche Breadths and Overlaps

Figure 3 shows the percentage utilisation of the three commonest food resources, ants,
termites and fungus by the three species. Each species differs in its favoured food resource.
Other food types utilised by the species included: arthropod parts (usually Orthoptera
legs), droppings (mostly lizard), bees, termites, spiders, Orthoptera, insect pupae and larvae
(occasionally of each other's species), seeds and seed parts (frequently dipterocarp seeds).
I also observed these species respond to fresh vertebrate carrion.
Figure 4 shows the probability distribution of log food size classes consumed by the three
species. Resource widths and overlaps were calculated from this distribution. Figures 5-7
show the combined diel schedules for each species.
40.- - -~
30
I:
11/
~ • C. gigas
"o 20 Gl D. intricatum
...o
-
~
o
10
Em O. transversa
o
Termites Ants Fungus
food type
Figure 3. Percentage food type utilisation of the three commonest food types by the three species.
.9
.8 0 C. gigas
0 D. intricatum
~ .7
~ 0. transversa
j .6
.5
'5
~ .4
i .3
t .2
.1
0
-.1
0 .5 1.5 2 2.5 3 3.5 4
log food size interval
Figure 4. Probability distribution of food size classes collected by the three ant species,
constructed from the means and Standard Deviations of the raw data.
-2~
o 2.5 5 7.5 10 12.5 15 17.5 20 22.5 25
hour
Figure 5. Camponotus gigas diel schedule. This ant is primarily nocturnaL
12
-2:~
o 2.5 5 7.5 10 12.5 15 17.5 20 22.5 25
hour
Figure 6_ Diacamma intricatum diel schedule. This ant is nocturnal and diurnaL
338 RUTH LEVY
:;
0 o in
:5 o out
111
~
.
CD
::.
c:: 8
~
c:i 6
"of!. 4
2
0
-2
0 2.5 5 7.5 10 12.5 15 17.5 20 22.5 25
hour
Figure 7. Odontoponera transversa diel schedule. This ant is primarily diurnal.
Table 3 shows Levins niche breadths for each species: C. gigas has the narrowest niche,
and D. intricatum the widest.
Table 4 shows percentage resource overlaps, mean percentage overlaps and overall
multiplicative overlaps for the three two species interactions. Greatest niche overlap is
between C. gigas and D. intricatum.
DISCUSSION
I found that observed nearest neighbour nest distances were overwhelmingly larger than
expected in the case of D. intricatum and O. transversa, indicating positive association,
clumping, of their colonies. No other significant relationships were found. The only
significant relationship at baits, between C. gigas and O. transversa, was also positive: this
may be because resource overlap is lowest for C. gigas and O. transversa, and highest for
C. gigas and D. intricatum. These two species' nest distributions tended non-significantly
Table 3. Levins niche breadths of three species, and mean percentage of total possible niche breadth.
Species food type food size foraging level: time mean % of total
day/night niche breadth
C. gigas 4.1 6.8 1.1, 1.4 1.2 39.6
D. intricatum 6.8 10.9 1.1, 1.2 1.9 52.0
O. transversa 7.1 9.9 1.0, 1.0 1.4 44.2
Table 4. Percentage resource overlaps for three species pairs, with mean overlaps and overall multiplicative overlaps
for each species pair.
species pair food food size foraging level: time: mean % overall %
t~e da}:/ ni~t in, out overlaI! overlaI!
C. gigas/ D. intricatum 54.7 76.2 99.4, 91.8 72.7,69.3 77.4 29.48
C. gigas/ O. transversa 44.5 50.28 96.8, 84.4 23.4,24.3 53.9 5.33
D. intricatum/ O. 70.4 71.5 97.5,92.6 50.7,55.0 73.0 26.12
towards avoidance. The high resource overlap between these species might be viewed as
surprising as they are in different subfamilies. The evidence for any pattern of
associationcorrelating with niche overlaps in these species is slight. Positive association
relationships may reflect habitat preferences in foraging area, or responses to competitive
pressure from other aggressive species.
The resource creating most difference in species overall resource overlap in this study is
time. Time is known to be an important factor allowing ecological segregation in ant
communities (Holldobler & Wilson 1990), often depending on physiological tolerance limits
of species involved. Time segregation is likely to reduce interference competition between
species, as opposed to food type or size, which reduces exploitative competition (Putman
1994). C. gigas' primarily nocturnal habit reflects that of many other Camponotus species
(personal observations).
All three species are relatively large, compared with other ants in the community, but C.
gigas is the largest and O. transversa the smallest. Mean food size collected is distributed
as expected from rank of body size.
Although negative interspecific competitive interactions have usually been found between
ant species of higher abundance and social organisation, these three species apparently form
an ecologically distinct guild of their own, that of large, ground dwelling, generalist solitary
foragers. It is not clear whether such a group of species would be expected to compete
interspecifically, but as Savolainen & Vepslilliinen (1988) note, within the broad hierarchy
of the Finnish taiga ant community, each tier is further subdivided into further hierarchical
tiers which compete amongst themselves.
The old problem that species differ in their resource utilisation whether they are under
competitive pressure or not (Giller 1984), occurs in this study. A null model would be
needed to test whether the realised niches of these species are random or not. An
improvement would be to have heeded May (1975) and to have calculated multidimensional
resource overlap directly, by creating a single resource spectrum of all combinations of food
type, food size, time and space variables. It would then be possible to avoid the problems
of estimating multidimensional overlap indices indirectly, by adding or multiplying
unidirectional measures, as I have done. However, in the complicated environment of a
tropical rain forest, this is easier said than done. It might be possible with the use of
specially designed baits, administered at prescribed times and places (e.g. Lynch et al. 1980),
but this might also lose important qualitative information.
Putman (1994) discusses the fundamental problems with studies like the present one. He
notes that it must be shown first that resources are in fact in limited supply. This is
notoriously difficult to establish in field studies, and is usually assumed, as I have done.
Another problem is non-independence of resource axes. I found, on analysing data collected
for this study, dependence only of space and time utilisation, and combined these into a
single parameter. Resource type may also segregate species in different ways: differential
utilisation of certain resources may influence potential for interference competition (e.g.space
and time), or for exploitation competition (e.g. food type and size): models of resource
partitioning are invariably assembled to examine exploitative competition. Diffuse
competition from the rest of the community may also alter the effect individual species have
on each other. Putman concludes that characterisation of niche parameters is useful
primarily as a descriptive tool of an organism's role in a community (its original function),
and in fact has little predictive power.
All three species are frequently found in the community: they are distributed quite evenly
over the hectare and at all other sites I visited. They are obviously able to coexist adequately
with this magnitude of resource overlap. The next step in understanding the dynamics of the
relationships between these species would be to perform resource and colony density
340 RUTH LEVY
manipulation experiments, and to look at resource utilisation responses and fitness of the
species as a result.
LITERATURE CITED
ARTHUR, W. 1987. The Niche in Competition and Evolution. John Wiley & Sons, Chichester.
BRIESE, D. T. 1982. Partitioning of resources amongst seed-harvesters in an ant community in semi-arid Australia.
Australian Journal of Ecology 7:299-307.
CULVER, D. C. 1974. Species packing in Caribbean and north temperate ant communities. Ecology 55:974-988.
DAVIDSON, D. W. 1977. Species diversity and community organisation in desert seed-eating ants. Ecology
58:711-724.
GAUSE, G. 1934. The Struggle for Existence. Haefner, New York.
GILLER, P. S. 1984. Community Structure and the Niche. Outline studies in ecology. Chapman and Hall, London.
HANSEN, S. R. 1978. Resource utilisation and coexistence of three species of Pogonomyrmex ants in an Upper
Sonoran grassland community. Oecologia (Berlin) 35: 109-117.
HOLLDOBLER, B. & WILSON, E.O. 1990. The Ants. Belknap of Harvard University Press, Cambridge, Mass.
HUTCHINSON, G. E. 1957. Concluding Remarks. Cold Spring Harbor Symposiun on Quantitative Biology 22:415-
427.
KEDDY, P. A. 1987. Competition. Population and Community Biology Series. Chapman and Hall, London.
KREBS, C. S. 1989. Ecological Methodology. Harper, New York.
LOTKA, A. J. 1932. The growth of mixed popUlations: two species competing for a common food supply. Journal
of the Washington Academy of Science 22:461-469.
LUDWIG, J. A. & REYNOLDS, J.F. 1988. Statistical Ecology. Wiley, New York.
LYNCH, J. F. 1981. Seasonal, successional and vertical segregation in a Maryland ant community. Oikos 37:183-
198.
LYNCH, J. F., BALINSKY, E.C. & VAIL, S.G. 1980. Foraging patterns in three sympatric forest ant species.
Ecological Entomology 5:353-371.
MACARTHUR, R. H. & LEVINS, R. 1967. The limiting similarity, convergence and divergence of coexisting
species. American Naturalist 101:377-385.
MARSH, A. C. 1985. Forager abundance and dietary relationships in a Namib desert ant community. S.-Afr.
Tydskr. Dierk 20:197-203.
MAY, R. M. 1975. Some notes on estimating the competition matrix, alpha. Ecology 56:737-741.
MAY, R. M. 1984. Real and apparent patterns in community structure. pp. 3-18 in Strong, D.R., Simberloff, D.,
Abele, L.G. & Thistle A.B. (eds). Ecological Communities: Conceptual Issues and the Evidence. Princeton
University Press, Princeton, New Jersey ..
PIANKA, E. R. 1969. Sympatry of desert lizards (Ctenotus) in Western Australia. Ecology 50: 1012-1030.
PIANKA, E. R. 1981. Competition and Niche Theory. pp. 167-196 in May, R.M. (ed.). Theoretical Ecology.
Blackwell Scientific, Oxford.
PIANKA, E. R. 1983. Evolutionary Ecology. Harper and Row, New York.
PUTMAN, R. J. 1994. Community Ecology. Chapman and Hall, London.
ROOT, R. B. 1967. The niche exploitation pattern of the blue-gray gnatcatcher. Ecological Monographs 37:317-350.
RYTl, T. R. & CASE, T.J. 1992. The role of neighbourhood competition in the spacing and diversity of ant
communities. The American Naturalist 139:355-374.
SAVOLAINEN, R. & VEPSALAlNEN, K. 1988. A competition hierarchy among boreal ants: impact on resource
partitioning and community structure. Oikos 51:135-155.
SCHOENER, T. W. 1974. Resource partitioning in ecological communities. Science 185:27-39.
SCHOENER, T. W. 1986. Resource Partitioning. pp. 91-126 in Kikkawa, J. & Anderson, D.J. (eds). Community
Ecology: Pattern and Process. Blackwell Scientific, Victoria, Australia.
VOLTERRA, V. 1926. Variations and fluctuations of the number of individuals in animal species living together.
Reprinted in: Animal Ecology, Chapman, R N. (ed.). McGraw-Hill, New York.
Immediate effects on wildlife of selective logging in a hill dipterocarp
forest in Sarawak: mammals
ZAINUDDIN DAHABAN I , M. NORDIN I and ELIZABETH L. BENNETT.
IDepartment of Zoology, Universiti Kebangsaan Malaysia, 43650 Bangi, Selangor Darul

Ehsan, Malaysia
2Wildlife Conservation International, 7 lalan Ridgeway, 93200 Kuching, Sarawak, Malaysia
ABSTRACT. The effects of selective timber extraction on wildlife communities were studied in detail
in a mixed hill dipterocarp forest of the Kapit Division, Sarawak. The main study area was systematically
surveyed for one year before it was logged and for another year immediately after logging. This paper
highlights the changes in diversity and density of mammals in the area following logging, with special
emphasis on the population of primates and squirrels. Comparisons are made to the results of similar
studies carried out in Peninsular Malaysia and Sabah. As a result of logging, some species decline in
number or become locally extinct, some remain relatively stable, others increase, and a few new species
come into the logged area. A system of totally protected areas and large undisturbed areas within logging
concessions are needed to ensure continued survival of the fauna.
KEY WORDS: Borneo, density, diversity, hill dipterocarp forest, logging, primates, squirrels.
INTRODUCTION
In earlier studies of this type, Johns (1983, 1989) studied the effects on wildlife of logging
hill dipterocarp forests in Peninsular Malaysia and in Sabah. These areas were on good
soils, with high initial densities of wildlife, and the virtual absence of hunting. On the other
hand, Sarawak has notoriously poor soils and exceptionally high rainfall, low natural
densities of wildlife (Bennett 1992), and a high level of hunting which almost invariably
accompanies logging (Caldecott 1986). These were likely to make the impact of logging on
wildlife in Sarawak more severe.
With 60% of the Sarawak's land area licensed for logging (Kavanagh et al. 1989) and 4.1
million hectares of production forest logged by 1990 (Mok et al. 1991), there was an urgent
need to determine its effect on wildlife populations. In this paper, we present preliminary
data collected by ZD for two years on the responses of wildlife to logging. At around the
same time, the effects of hunting and shifting cultivation on wildlife of Sarawak were studied
by Bennett (Bennett 1992, Bennett & Dahaban 1992).
The main objective of this work was to document the effects on wildlife of logging a hill
dipterocarp forest in Sarawak. It is essential to know the true impacts of logging on wildlife
and the environment so that effective management programs can be formulated to benefit
both wildlife and humans.
STUDY AREA
The study was carried out in a hill mixed dipterocarp forest logging concession near Kapit.
The 100 ha main study area (MSA) at 01"39'N, 113"24'E, comprised steep terrain between
250 m and 380 m altitude, and the soils were red-yellow podzols. The forest vegetation was
dominated by trees of the Dipterocarpaceae, especially Shorea species. For a full description
341
342 ZAINUDDIN DAHABAN, M. NORDIN AND ELIZABETH BENNET
of this habitat type, see Whitmore (1984, 1992).

Rainfall data were gathered by the logging company at their camp about 65 km north of
the MSA since September 1988. There was no distinct seasonal pattern. Mean annual
rainfall was 5550 mm while monthly totals ranged from 195 mm to 731 mm. Ambient
temperature ranged from 22°C to 32°C.
The site was chosen mainly because of a highly organised infrastructure within the logging
company and its predictable cutting rate. This meant that an area due to be logged halfway
through the study period could be accurately delimited.
Data collection started in December 1989 when the MSA was primary forest, and about
7 km from the nearest access road or any disturbance. From December 1990 to March
1991, the area was selectively logged. Data collection continued until the end of November
1991, and a further survey was done in April 1992, a year after the felling had finished.
In addition, surveys were conducted in two other parts of the same concession area. One,
the two-year old logged forest study area, had been logged in December 1988, and was
surveyed in April and November 1991. The other, the four-year old logged forest study area,
had been logged in January 1987, and was surveyed in September and November 1991.
In all three areas, logging was done using similar methods. As elsewhere throughout
Sarawak's hill forests, tractor logging was employed. Trees were cut using chainsaws,
pulled out by bulldozer to a loading site, loaded onto a truck and transported by road to a
log pond. Timber extraction rates depended on the available volume. In the MSA, 54 % of
4039 trees of at least 30 cm gbh tagged before logging were removed or destroyed. This
treatment was likely to have been similar in all three sites. All three were part of the same
forest block, the maximum distance between them was only 35 km, and the terrain and slope
were similar in all. Thus the flora and fauna prior to logging were likely to have been
similar at all sites.
Prior to logging, the whole area was remote and about 40 km from the nearest human
settlement; hunting would have been negligible. Once access roads had been built and during
the logging process, some hunting occurred. It ceased in the MSA two months after logging
finished because a bridge along the access road collapsed. Even during logging, hunting in
the MSA was lighter than elsewhere because ZD's presence acted as a deterrent. In general,
most hunting was from vehicles along the road. Any animal seen was usually shot at,
including primates, sun bear (Helarctos malayanus), clouded leopard (Neofelis nebulosa), and
all species of ungulate. Hunting at night using vehicles and spotlights occurred in all areas
with road access, and was especially heavy in weeks preceding major local festivals.
METHODS
The diversity of mammals in the study area was monitored throughout the study period. Any
species of animal seen at any time in all three sites was noted.
The basic method to assess animal abundance used was the line transect method
(Brockel man & Ali 1987, Burnham et at. 1980, Davies & Payne 1982, Marsh & Wilson
1981). At each study site, one survey trail two km long was cut. For the post logging
surveys, new trails were cut; old logging roads were not used for surveys since they give a
biased sample of the forest and its wildlife.
Survey sessions were divided by months. In each month, about 10 mornings, equivalent
to 20 km of survey trail were walked. Data from different months were pooled in the
analysis, resulting in a single data block each for the MSA before logging, the MSA after
logging, the two-year old logged forest, and the four-year old logged forest.
Trapping of small mammals was not feasible due to time constraints. Hence, no data was
available on the effect of logging on small mammals.
Effects of selective logging on wildlife 343
RESULTS
Diversity of primates and squirrels

The total number of species of all diurnal primates and squirrels seen in the MSA declined
from 13 before logging to 12 after logging. Those seen in the two-year old logged forest and
four-year old logged forest were 11 and 12 species, respectively. Analysis at species level
showed the composition changed greatly. Some species seen before logging disappeared after
logging and vice versa. Table 1 summarises the changes in species occurrence of primates
and squirrels at all study sites.
Diversity of other mammals

Sixteen other species of mammals were recorded in the MSA before logging and 14
species after. Among those that were seen before but not after logging were clouded
leopard, marbled cat (Felis marmorata), sun bear, thick-spined porcupine (Thecurus
crassispinis), and oriental small-clawed otter (Aonyx cinerea). Conversely, the lesser
treeshrew (Tupaia minor), slender treeshrew (T. gracilis), and collared mongoose (Herpestes
semitorquatus) appeared after logging.
The diversity of other mammals observed in the two-year old and four-year old logged
forests was extremely low. Only six species were seen in the two-year old logged forest,
namely the common treeshrew (Tupaia glis), lesser treeshrew, sambar deer (Cervus
unicolor), barking deer (Muntiacus muntjac), lesser mousedeer (Tragulus javanicus), and
bearded pig (Sus barbatus). In the four-year old logged forest, the five species seen were
common treeshrew, striped treeshrew (Tupaia dorsalis), barking deer, bearded pig, and
malay weasel (Mustela nudipes).
Density of primates and squirrels

The immediate effect of logging on the primates (Table 2) proved to be severe. Bornean
gibbons (Hylobates muelleri) in the MSA decreased more than 70% in group density after
Table 1. Presence and absence of primates and squirrels in main study area (MSA) before logging and after logging,
in two-year old logged forest, and in four-year old logged forest. P=present, A=absent.
Species MSA before MSA after Two-year old Four-year old

logging logging logged forest logged forest
Hylobates muelleri p P P P
Presbytis frontata P P P P
Presbytis rubicunda P P P P
Macaca fascicularis P A A P
Macaca nemestrina A P A P
Ratufa ajJinis P P P P
Callosciurus prevostii P P P P
Callosciurus notatus A P P P
Rheithrosciurus macrotis P P P P
Sundasciurus hippurus P A A P
Sunliasciurus tenuis P P P P
Sunliasciurus lowi P P P A
Lariscus hosei P A A A
Nannosciurus melatonis P P P A
Exilisciurus e.xilis P P P P
Table 2. Density ofprirnates (groups per sq kIn) in main study area (MSA) before logging and after logging, in two-
year old logged forest, and in four-year old logged forest. P = present but not seen during transect surveys.
MSA before MSA after Two-year old Four-year old

Species
Hylobates muelleri 4.73 1.36 2.08 2.00
Presbytis rubicunda 0.38 0.21 1.25 0.05
Presbytis frontata 0.87 0.57 P 1.50
Table 3. Density of squirrels, treeshrews, and barking deer (individuals per sq kIn) in main study area (MSA) before
logging and after logging, in two-year old logged forest, and in four-year old logged forest. P = present but not
seen during transect surveys.
Species MSA before MSA after Two-year old Four-year old

Ratufa ajJinis 5.18 1.48 3.75 1.00
Other squirrels 15.82 24.08 103.50 19.00
All tree shrews 10.28 4.92 10.04 38.45
Muntiacus muntjac 2.91 0.79 10.22 P
logging. The group density of maroon langurs (Presby tis rubicunda) in the MSA dropped
about 45 % after logging while that of the white-fronted langurs (P. frontata) decreased by
35%.
Giant squirrels (Ratufa ajJinis) in the MSA showed a reduction of 71 % in individual
density after logging. On the other hand, the smaller squirrels in the MSA increased by 52 %
in combined individual density after logging. Table 3 shows the squirrel densities at the
three study sites.
Density of other mammals

The combined treeshrews' individual density in the MSA decreased by 52 % after logging.
The individual density of barking deer in the MSA declined by 73% after logging (Table 3).
DISCUSSION
Some species never seen in the MSA before logging appeared after logging. Those that
appeared are all abundant in secondary forests and gardens; for example the pig-tailed
macaque (Macaca nemestrina), plaintain squirrel (Callosciurus notatus), and shrews (Tupaia
spp.). Those animals which disappeared after logging were all endangered or totally
protected species or both.
The logged-over forest supported less than half the number of gibbons than did primary
forest. White-fronted and maroon langurs declined significantly. Giant squirrels were greatly
affected after cutting. On the other hand, smaller squirrels increased in number greatly. The
number of treeshrews and barking deer was less than half immediately after logging.
The purpose of surveying the two logged forest areas was to determine the abundance of
animals there. This was to see if colonisation processes could take place from the old logged
forest to the recently logged forest. From the current study, only the two langurs, the
smaller squirrels, the treeshrews, and the ungulates survived well after logging. It is difficult
to draw any firm conclusions on the fate of animals that vanished after logging and were not
seen in the two older logged forest study sites.
Effects of selective logging on wildlife 345
In previous studies of similar nature in Malaysia (e.g. Johns 1983, 1989, 1992, Johns &
Skorupa 1987), wildlife survival in general was apparently greater than that observed in the
present study. One of the contributing factors could have been that the original densities of
wildlife were lower in the current study area than in Johns' study sites (Tekam, Peninsular
Malaysia and Ulu Segama, Sabah). Lower natural densities might mean that animals are
more vulnerable to disturbance.
All of the studies, however, show that as a result of logging, some species decline in
number or become locally extinct, some remain relatively stable, others increase and a few
new species come into the area. The trend is for edge or coloniser species to replace
primary forest species. Thus, even though the overall species diversity is similar with
disturbance, the change is not necessarily beneficial from the conservation point-of-view.
The species which are most threatened in Sarawak are those which depend on tall forests.
Animals coming into disturbed sites such as plaintain squirrels are common in gardens
throughout South-east Asia and are not likely to become extinct. As disturbance increases,
they take over from species which depend on undisturbed forest and which cannot survive
elsewhere.
The aim of conservation is not to maintain or increase the total number of species in a
small area, but within a country or region. It is well known that localised diversity can often
be increased by disturbing the habitat to create more edges (Murphy 1989). However, this
does not increase biodiversity in the whole of Sarawak. In the long run, it decreases it.
Loss of animals from a forest is not just a conservation problem for their own sake, or
even for the sake of maintaining a protein supply for rural people. Animals are an integral
part of a tropical forest. Without them, the forest itself is likely to suffer a loss of diversity
because of the animals' role as pollinators, dispersers, and as grazers stimulating plant
diversity.
It is clear that the long-term effects of logging on wildlife are not yet fully understood.
At the same time, the resilience of tropical forests to logging is still largely unknown. At
this stage, an effective system of totally protected areas is desperately needed, to act as
reservoirs during logging and to allow ample opportunity for animals to recolonise the
disturbed areas. In addition, large primary forest areas should be retained within logging
concession for the same purpose.
ACKNOWLEDGEMENTS
Studies such as this are rarely successful without a great deal of help from many people, too
many to mention here. We would like to thank the then Sarawak State Secretary, Tan Sri
Datuk Haji Bujang Nor for permission to conduct the study. Members of the Sarawak Forest
Department were invaluable in their support in so many ways, especially Datuk Leo Chai,
Abang Haji Kassim Abang Morshidi, Philip Ngau Jalong, Ngui Siew Kong, Francis Gombek,
and Haji Othman Ismawi. The study could not have taken place without the constant help
from B. Ross Ibbotson, David Manjah, Francis Goh, Patrick Braganza and other staff of
Batang Baleh Forest Enterprises Sdn. Bhd. Others of enormous help include Dr Mikaail
Kavanagh, Cynthia Low, Dr Mary Pearl, Martha Schwartz, and Abdullah Che Sahamat. The
work was funded by Wildlife Conservation International of the New York Zoological Society
and World Wide Fund for Nature Malaysia.
LITERATURE CITED
BENNETT, E.L. 1992. A wildlife survey of Sarawak. Wildlife Conservation International of the New York
Zoological Society and WWF Malaysia.
BENNETT, E.L. & DAHABAN, Z. 1992. Responses of wildlife to different types of disturbance in Sarawa/c, and
implications for forest management. Paper presented at the International Workshop on Ecology, Conservation
and Management of South-east Asian Rainforest. Kuching, Sarawak, Oct 12-14, 1992. Proceedings in press.
BROCKELMAN, W.Y. & ALI, R. 1987. Methods of surveying and sampling forest primates populations. pp.
23-62 in Marsh, C.W. & Mittermeier, R.A. (eds). Primate Conservation in the Tropical Rain Forest. Alan R.
Liss Inc., New York.
BURNHAM, K.P., ANDERSON, D.R., & LAAKE, J.L. 1980. Estimation of density from line transect sampling
of biological populations. Wildlife Monographs 72.
CALDECOTT, 1.0. 1986. Hunting and wildlife management in Sarawak. World Wildlife Fund Malaysia and
National Parks and Wildlife Office, Kuching.
DAVIES, A.G. & PAYNE, LB. 1982. A faunal survey ofSabah. WWF Malaysia, Kuala Lumpur.
JOHNS, A.D. 1983. Ecological effects of selective logging in a West Malaysian rain-forest. D. Phil. dissertation,
University of Cambridge.
JOHNS, A.D. 1989. Timber, the environment and wildlife in Malaysian rainforests. Unpublished report, Institute
of South-east Asian Biology, University of Aberdeen.
JOHNS, A.D. 1992. Species conservation in managed tropical forests. pp. 15-53 in Whiunore, T.C. & Sayer, J.A.
(eds.). Tropical Deforestation and Species Extinction. Chapman and Hall, London.
JOHNS, A.D. & Skorupa, LP. 1987. Responses of rain forest primates to habitat disturbance: a review.
International Journal of Primatology 8: 157-191.
KAVANAGH, M., ABDULLAH ABDUL RAHIM & HAILS, C.J. 1989. Rainforest Conservation in Sarawak, An
International Policy for WWF. WWF Malaysia and WWF International, Kuala Lumpur and Gland.
MARSH, C.W. & WILSON, W.L. 1981. A survey of primates in peninsular Malaysian forest. Final report to
Universiti Kebangsaan Malaysia, Kuala Lumpur.
MOK SIAN TUAN, ALIAS ABDUL JALIL & DAWEND JIWAN, 1991. A WWF Strategy for Tropical Forest
in Sarawak. WWF Malaysia and WWF International, Kuala Lumpur and Gland.
MURPHY, D.D. 1989. Conservation and confusion: wrong species, wrong scale, wrong conclusions.
Conservation Biology 3: 82-84.
WHITMORE, T.C. 1984. Tropical Rain Forests of the Far East, 2nd Edition. Clarendon, Oxford.
WHITMORE, T.C. 1992. An Introduction to Tropical Rainforests. Clarendon Press, Oxford.
Growth performance of a naturally regenerated stand of Kapur
(Dryobalanops aromatica Gaertn. f.) 37 years after a heavy low thinning
treatment
ABD. RAHMAN, K.l, SHAHRULZAMAN, 1. 1 andWEINLAND, G. 2
IForest Research Institute of Malaysia (FRIM), Kepong 52109, Kuala Lumpur

2Technical Advisor, German Technical Cooperation (GTZ), Malaysian-German Forestry
Research Project, Forest Research Institute, Malaysia
ABSTRACT. Growth increment and forest structure within a naturally regenerated Kapur Forest after
37 years following a heavy low thinning in Compartment 15, Kanching Forest Reserve, Malaysia,is
described. Two sample plots (SP 85 and SP 86), each consisting of a thinned subplot (A) and an
unthinned SUbplot (B), were analysed. Increments in diameter, basal area and volume of the dominant
trees in the treated stands were generally higher than in the untreated stands. The mean annual volume
increment for the four subplots was between 7.1 m3 and 9.7 m3 • Considering the abundance of
regeneration in all plots, research on natural regeneration systems for Dryobalanops aromatica is
recommended.
KEY WORDS: Dipterocarpaceae, low thinning, increment, Malaysia.
INTRODUCTION
Kapur forests are named after the dominant species, Dryobalanops aromatica Gaertn. f., or
Kapur as in the vernacular. In Peninsular Malaysia, the Kapur forests are confined to two
large blocks at the east coast (in the south of Trengganu to north of Pahang and south of
Pahang to south, southeast and east of lohor) and one in two small pocke[s at the west coast
in Bukit Lagong F.R and Kanching F.R. (see Figure 1). The Kapur stands at the west coast
have obviously been introduced sometime in the past, when camphor was traded between
the east coast and the west coast of the Peninsula (Wyatt-Smith 1963).
Kapur is regarded as the most suitable species for forest management. The reasons are:
(i) satisfactory and regular seed supply, (ii) high germination rates, (iii) easy nursery
handling, (iv) good survival rates of planted seedlings even under adverse conditions.
Barnard (1954) placed Kapur at the top of his silviculturallist of trees for plantations. Kapur
seedlings and saplings are vigorous in growth. The young plant requires shade and
responds later readily to canopy openings. It recovers quickly from mechanical injuries (Lee
1967, Foxworthy 1932).
In its natural habitat, Kapur tends to dominate the stands, so much so that a forest type
has been named after the species. The stands in the Kanching Forest Reserve are generally
pure stands. In such natural stands Kapur is strongly represented in all size classes, a feature
which is not a characteristic of other emergent species (Wyatt-Smith 1963). Lim (1992)
reported that litter production of a 70 year old Kapur stand in Compartment 15 of Kanching
F.R. was 12.2 t/ha/year, while litter accumulation on the forest floor was 9.5 t/ha, giving
a turnover rate of 1.3 which is fairly typical of many tropical forests. The high production
of litter suggests that the biological productivity of the tree is high (Lim 1992).
Kapur was probably the most widely studied dipterocarp species in Peninsular Malaysia
before the 1960's. During that period, a lot of effort was placed on research to develop
establishment and management procedures for Kapur. Apart from the Kapur plantations
347
© 1996 Kluwel' Academic Publishers.
348 K. ABD. RAHMAN, I. SHAHRULZAMAN AND G. WEINLAND
South China Sea
o
Suma'lra
Figure 1. Location of Study Site and Distribution of Kapur Forest

(shaded, after Wyatt-Smith 1963) in Peninsular Malaysia.
within the FRIM area which received some early thinning, only three thinning trials on
a probing scale had been set up. These were in Panti Forest Reserve, Mersing Forest Reserve
and Kanching Forest Reserve. This paper presents results of the thinning trial in Kanching
Forest Reserve.
The plots are located in Compartment 15, Kanching Forest Reserve, Ulu Selangor (see
Figure 1) about 20 kilometres north of Kuala Lumpur. The topography is very steep to
moderately steep with an elevation of approximately 150 mas!. SP 85 is facing south, while
SP 86 varies from a western to a southern aspect. Substrate parent material is granite. Soil
pits were dug in the vicinity of the plots when the trial was established. Soil texture is of
a clay-loam with a higher content of sand on the upper slope.
At the time of establishment of the trial, the Kapur trees formed a more or less even-aged
pole-size forest with individuals of other species such as Meranti tembaga (Shorea leprosula),
Pulai (Alstonia spp.), Terentang daun besar (Campnospermum spp.) and bamboo. There
were dense patches of Kapur regeneration with heights to about 4 m. Saplings of other
species were also present. The forest was well stocked, containing some 5-10 per cent
forked stems (Vincent 1963). The forest, in which both sample plots are located, has been
treated since 1910, the treatment following the normal system of improvement fellings used
in Selangor. This consisted of two improvement fellings. The first included a release of the
Growth peiformance of Kapur (Dryobalanops aromatica) 349
crowns of the superior trees from competition of inferior trees, and a cutting of bertam
palms (Eugeissona tnste) and creepers if they were smothering or threatening to smother the
crop trees. The treatment was repeated in the second improvement felling after about two
years. In 1914 all trees interfering with the selected crop trees were felled and all bertam
palms cut back. From 1914 until 1934, mature Kapur trees were removed from the stands
and the final felling carried out in 1934. The majority of the mature Kapur trees and
unclassified trees were removed between 1919 and 1923.
The trial consists of the two FRIM Sample Plots, SP 85 and SP 86. Each sample plot is
divided into two subplots, a thinned (subplot A) and an unthinned (subplot B). The size of
each subplot of SP 85 is 0.36 ha (40 m x 90 m) and of SP 86 is 0.40 ha (40 m x 100 m).
The goal of the trial was to determine the effect of a thinning treatment on volume
increment. The crop age was unknown when the trial was established in 1954, but was
assumed to be about 33 years (± 5 years). The mean annual increment is calculated based
on the crop age of 70 years in 1991 assuming that all regeneration was established in 1921.
In both sample plots, subplot B was left as an unthinned control. In subplot A, about half
of the trees, including dominant stems and defective trees, were felled. Volume before
thinning, removal and remaining volume, both for stand total and the tree average, are
presented in Table 1. All data are converted from imperial to metric units.
As can be seen from the ratio v,/v" the thinning method applied was in both cases a low
thinning. For low thinning the average size of the remaining trees shifts to a higher value,
because mostly trees smaller than the average size are removed. This shift in average size
can be expressed by the ratio of the average size of thinned trees (vJ to the average size of
the remaining trees (v r). If this ratio is equal to unity (v/vr = 1.0), trees were removed
systematically favouring no size class. If the ratio is between 0.5 and 1.0, a low thinning
was applied with the removal of mostly smaller trees. A ratio of v/Vr L 1.0 indicates that
a crown thinning was applied, favouring selected trees by removing competing dominant and
co-dominant trees.
Both treatments were distinct low thinnings. How the removal might have appeared is
visualised in Figure 2. Figure 3 shows the stem-diameter distributions for the thinned plots.
The majority of the trees have been removed from the left arm of the distribution, which is
typical for a low thinning.
Unfortunately, no individual measurement data were recorded, only the averages and
totals, including averages and totals of stand, averages and totals of population of the 25
largest and 50 largest trees per acre. As a result, no comparison of individual tree growth
was possible. With two replicates only, no reasonable statistical comparisons could be made.
Table I. Stand volume and average tree volume before and after thinning at commencement of the trial in 1954.
v, = average tree volume of the trees remaining; v, = average tree volume of the felled trees.
Plots N/ha dg G/ha V/ha V/tree

85A
Before thinning 181 40.4 23.20 260 1.4365
Removal 86 36.9 9.20 100 1.1765
After thinning 95 43.3 14.00 160 1.6842
86A
Before thinning 220 36.8 23.40 240 1.0909
Removal III 32.0 8.94 90 0.8108
After thinning 109 41.1 14.46 150 1.3762
The ratios for the thinning plots were: SP 85A : v/v, = 1.1765: 1.6842 = 0.6989
SP 86A : v/v, = 0.8108 : 1.3762 = 0.5892
D !)
.
' .. ....' .'. ~
..
.. , ' , "
Figure 2. Visualisation of how the heavy low thinning in the Kapur thinning trial could have looked (drawing and
description of tree classes modified from Evans 1982). Legend: D - Dominant trees, tallest and most vigorous trees,
most of the crown free; CD - Upper canopy trees, below the crown level of the dominants; SD - Not in the upper
canopy, leaders still have free access to overhead light; S - Branches do not reach up into the branches of the upper
canopy, or leading shoot under some part of the crown of another tree; M - Dying or dead.
The height curves were fitted by the method of least squares using a quadratic function:
h = bo + b l * d + b2 * d2
where: h = dependent variable, height, bO• I •2 = regression coefficients, d = independent
variable, diameter. From the individual height curves the average height of the dominant
stems was calculated. The height-diameter scatter discloses peculiarities of the canopy
structure.
The comparability of the plots depends on the achieved top height at a certain age, which
is regarded as an index for site quality. This top height, which is defined as the average
height of the 100 largest trees per hectare, is usually taken as a basis for comparison. In our
case, the average height of the 62 largest trees per hectare was taken instead, because
individual tree heights from which the average height of the 100 largest could have been
calculated were not available. However, the average height of the 62 trees/ha, which is here
referred to as dominant height, can be regarded as a close approximation to the top height.
SP85A SP B6A
~
j _u-~=rml
Figure 3. Stem-diameter distributions for the thinned plots, SP85A and SP 86A at stand age of 30 years (1954)
before and after thinning.
Growth performance of Kapur (Oryobalanops aromatica) 351
Table 2. Dominant stand height at stand age of 33 years and 70 years.
Subplot Height (m) Increment (m)

1954 1991
85A 36.5 51.7 15.2
85B 35.5 49.6 14.1
86A 33.1 45.7 13.6
86B 33.4 50.0 16.6
Closer inspection of the height values indicate that the subplots in each sample plot obviously
belong to different sites. Table 2 contains the average height of the dominant trees at
commencement of the trial and at stand age of 70 years.
When considering the achieved dominant height at stand age of 70 years, it would be
justified to compare the results of subplot 85A with those of subplot 86B and the results of
subplot 86A with those of subplot 85B. The results are arranged accordingly.
The average diameter is calculated as the diameter of the mean basal area. Since no
double-entry volume table was available for Dryobalanops aromatica, stand volume was
determined approximately by: V = G x Bole Length x 0.6, where the term 0.6 is a reduction
factor for stem taper. The height curves were calculated with ST ATGRAPHICS software.
Stand and stock tables

Table 3 contains the stand and stock tables of the four experimental thinning plots. Except
for stem number (N), it was not possible to retrieve information on the mortality during the
period 1954-1991.
In 1954 the dominant heights ranged from 33.1 m to 36.5 m with the highest values for
the subplots of SP85. In 1991, however, the largest dominant heights were achieved in the
unthinned subplot of sample plot 85 (51. 7 m) and in the thinned subplot of sample plot 86
(50.0 m). The average stand height was highest for the unthinned subplot of sample plot 85
(50.2 m). The average stand height of the remaining three subplots differed little (44.8 m
- 46.3 m). The difference in clean bole height among the subplots was also small, in 1954
as well as in 1991. At the last measurement clean bole height ranged from 24.0 m to 28.5
m.
Larger differences were found for average stand diameter. The thinned stand of SP 85A
achieved the highest average diameter (64.9 cm), the unthinned stand of SP 85B the smallest
diameter (52.7 cm). Overall, the stand diameter reacted as we might expect following
thinning: in both groups the stand diameter of the thinned plots is distinctly higher than the
diameter of the unthinned plots. In sample plot 85 the highest stand basal area and stand
volume was achieved in the thinned stand. In sample plot 86 the unthinned stand had higher
basal area and volume than the thinned stand. Even in sample plot 85 the superiority of the
thinned plot in stand basal area and stand volume over the unthinned plot is not pronounced.
Unlike stand averages the stand totals such as stand basal area and stand volume depend on
the number of stems per unit area as well. Thinning effects might not become evident until
the average stand increment is inspected. In sample plot 85 the increment for all stand
averages and stand totals is higher in the thinned plots than in the unthinned plots. In sample
plot 86, however, the thinned plot shows a higher value only for the average diameter (Og).
Since the average diameter in the unthinned plot is considerably smaller than the average
Table 3: Stand and stock tables for four experimental thinning plots of Dryobaianops aromatica
Year/period N H.om (m) H, (m) Hoot, (m) D. (cm) G(m~ y' (m')
Group 1
Plot SP 85A - thinned
1954 stocking 181 36.5 34.5 18.4 40.4 23.20 260
1954 removal 86 33.2 17.5 36.9 9.20 100
1954 remaining 95 36.5 35.0 18.7 43.3 14.00 160
1954-1991 mortality
1991 stocking 95 51.7 50.2 28.5 64.9 31.43 540
1954-1991 increment 21.6 17.43 380
Plot SP 86B - unthinned
1954 stocking 190 33.4 31.0 17.2 39.0 22.70 270
1954-1991 mortality 15
1991 stocking 175 50.0 46.9 27.5 54.8 41.37 680
1954-1991 increment 15.8 18.67 410
Group 2
Plot SP 86A - thinned
1954 stocking 220 33.1 30.6 17.1 36.8 23.40 240
1954 removal 111 29.8 16.6 32.0 8.94 90
1954 remaining 109 33.1 31.4 17.3 41.1 14.46 150
1954-1991 mortality
1991 stocking 109 45.7 44.0 24.0 58.4 29.20 420
1954-1991 increment 17.3 14.74 270
Plot SP 85B - unthinned
1954 stocking 165 35.5 34.1 19.1 39.3 20.04 230
1954-1991 mortality 30
1991 stocking 135 49.6 47.0 28.2 52.7 29.48 500
1954-1991 increment 13.4 9.44 270
* No double-entry volume tables were available for Dryobaianops aromatica. The approximative volume is
determined by: G x Hoot, x 0.6, where the term 0.6 is assumed as a reduction factor for stem taper. H.om: Average
height of the dominant trees. H,: Average stand height. Hoot,: Average clean bole height. D.: Diameter of the basal
area mean stem. G: Stand basal area. Y: Stand volume.
diameter in the thinned plot, the larger stand basal area of the unthinned plot is obviously
brought about by the distinctly higher stand density (stem numbers). The main crop in all
plots achieved harvestable size.
Dominant trees
In both groups, the thinned plot had a distinctly higher diameter, basal area and volume
increment than the unthinned plot (Table 4). Periodic height increments, however, were, as
expected, not different. The dominant height, in general, is not affected by the treatments.
The increment differences between unthinned and thinned plots are more pronounced,
when the increment percentages are compared. The increment percent (ip %) is the ratio of
the increment value of a growth period to the growth value at the beginning of the period.
The use of ip % is recommended, especially when the initial growth values of the plots being
compared are different.
The ip % for basal area were 119 % and 113 % in the thinned plots and 82 % and 77 % in
the unthinned plots. The ip % for the volume were 233 % and 200 % in the thinned plots and
Growth peiformance of Kapur (Dryobalanops aromatica) 353
Table 4. Standing stock and periodic increment of the group of dominant trees (62 largest trees/ha). 4: periodic
increment; ip%: increment percent; *The volume was determined approximately by: Glha x Bole Height x 0.6;
** Average height in 1954 was calculated based on 28 trees per hectare for SP 85 and 25 trees per hectare for SP
86. For 1991, average height for both sample plots were based on the 62 largest treeslha.
Plot Year Height** Diameter Basal Area Volume*

(m) (cm) (m2) (m')
85A 1954 36.5 47.1 10.8 120
1991 51.7 69.7 23.66 400
ip 15.2 22.6 12.86 280
ip% 119 233
86B 1954 33.4 50.0 12.17 130
1991 50.0 67.4 22.12 360
ip 16.6 17.4 9.95 230
ip% 82 177
86A 1954 33.1 44.5 9.62 100
1991 45.7 64.9 20.51 300
ip 12.6 20.4 10.89 200
ip% 113 200
85B 1954 35.5 47.1 10.44 120
1991 49.6 61.7 18.53 310
ip 14.1 14.6 8.09 210
ip% 77 175
177% and 175% in the unthinned plots. We may assume that there was a stimulating effect
on the increment of the dominant trees following thinning. However, the thinning effect
could not be confirmed as being statistically significant owing to lack of replicates.
Mean annual volume increment

The total stocking of all Kapur trees at the age of 70 including the thinning volume, which
accrued in 33 years, is given in Table 5. The volume from mortality for the unthinned plots,
which has to be known for an exact calculation of the mean annual volume increment, could
not be included because no such information was available. However an approximate, albeit
too low, mean annual increment can be derived.
The approximate mean annual volume increment (mai) at a stand age of 70 years is
between 7.1 m3 and 9.7 m3 • These values are similar to those of broad-leaved species
stands of the same age on good sites in Europe (Anonymous 1990).
The mai figures of Group 1 are larger than those of Group 2 and in Group 1 the mai of
the unthinned plot is larger than that of the thinned plot. Positive increment effects following
a thinning intervention may not be visible in the mean annual increment. On the contrary,
Table 5. Approximate mean annual volume increment for the total stand at the age of 70 years. *Volume is
determined by: Glha x Bole height x 0.6. **mai does not contain mortality (self-thinning) volume.
Plot Age Volumelha * Mean annual volume

(years) (m') increment (m')
Group 1
SP 85A 70 640 9.1
SP 86B 70 '" 680 '" 9.7 **
Group 2
SP 86A 70 510 7.3
SP 85B 70 ",500 '" 7.1 **
354 K. ABO. RAHMAN, I. SHAHRULZAMAN AND G. WEINLANO
if growing stock is removed beyond a certain limit, there will be no positive effect on the
maL The margins are small within which the mai can be raised through thinnings (Assmann
1970).
Structure
The height-diameter curves (Figure 4) cover a height range of about 35 m to 55 m. The
height range of 35 m to 40 m is represented by only a small number of trees. Most of the
trees are in the height range of 40 m to 50 m, which corresponds to the visual impression
of a distinctly mono-layered upper canopy in all the four plots.
The understorey of Dryoba/anops aromatica (trees> 5 cm diameter) ranges in height from
about 4 m to 32 m (Table 6).
However, most of the understorey trees are in the range 10 m to 15 m, which gives the
impression of distinctly two-layered stands. The density in this height class ranges from 70
stems/ha to 141 stems/ha. Altogether, the number of stems/ha of understorey Dryoba/anops
aromatica ranges from 116 to 261. The trees smaller than 5 cm diameter have not been
counted. However, all plots contain plentiful seedlings and saplings. Naturally regenerating
stands of Dryobalanops aromatica should not pose major problems as far as recruitment is
concerned.
All Plots
60r-----------------------------------,
5t-~
50~-
.....""...... 'JtE
I
~ 45i-~L
OJ
I
40~-
30+-~,_r
30 40 50 60 70 80 90 100
Diameter (em)
J--+- SP 65A -+- SP 858 •. ~ .. SP 66A •. JIE-- SP 868 I
Figure 4. Height-diameter curves (1991) for Sample Plot 85 and 86.
Table 6. Stem-height distribution of understorey Dryobalanops aromatica trees in Sample Plot 85 and 86.
Height-Class Stemslha
SP 85A SP 85B SP 86A SP 86B
0- 5 m 2
5 - 10 m 22 14 25 49
10 - 15 m 91 110 74 141
15 - 20 m 50 16 10 57
20 - 25 m 10 12 5 5
25 - 30 m 3 9 7
30 - 35 m 2 2
Total 177 163 116 261
Growth peljormance of Kapur (Dryobalanops aromatica) 355
CONCLUSIONS
The thinning trial in the Kanching Forest Reserve showed that Dryobalanops aromatica
stands seem to react favourably to heavy low thinnings. Even better results could be
expected from crown thinnings. Considering the ease with which the species establishes
abundant seedlings under the canopy of the parent trees, research on regeneration systems
for Dryobalanops aromatica stands would be a profitable exercise.
ACKNOWLEDGEMENTS
We wish to thank all officers and staff of the Forest Research Institute (FRI) and Forestry
Department of Selangor who initiated the study, established, monitored and maintained the
Sample Plots. Thanks are also due to Dr. Wan Razali and Dr. Appanah who read through
the final draft and made valuable comments on the manuscript. The assistance of En. Ahmad
Zuhaidi, who provided his field crew for control measurements, is gratefully acknowledged.
Many thanks are extended to En. Norazmi Yaakub and En. Muzaid Hj. Salleh and other field
staff for the assistance rendered in the data capturing and handling. Lastly we thank the
Director-General of the Forest Research Institute of Malaysia for allowing the authors to
present this paper.
LITERATURE CITED
ANONYMOUS 1990. Hiifstafeln for die Forsteinrichtung. Bayerisches Staatsministeriwn fur Erniihrung,
Landwirtschaft und Forsten.
ASSMANN, E. 1970. The Principles of Forest Yield Study. Pergamon Press, Oxford.
BARNARD, R.C. 1954. A TrUlnual of Malayan silviculture for inland forests. Malayan Forest Research Pamphlet
No. 14.
EVANS, 1982. Plantation in the tropics. Clarendon Press Oxford. 472 pp.
FOXWORTHY, F.W. 1932. Dipterocarpaceae of the Malay Peninsula. Malayan Forest Records No. 10:105-109.
LEE, P.C. 1967. Ecological studies on Dryobalanops aromatica Gaertn. F. Unpublished Ph.D. thesis.
University of Malaya. Kuala Lwnpur.
LIM, M.T 1992. Ecol<.o5ical studies on Dryobalanops aroTrUltica. Proceedings of the NationalIRPA (Intensification
of Research Priorities Areas) Seminar (Agricultural Sector). 6-11 January 1992. Kuala Lwnpur. Volwne II,
p.431.
VINCENT, A.J. 1961. A survey of the Kapur silvicultural treatment research plots in naturally and artificially
regenerated forest, Malaya. Research Pamphlet No. 36. FRIM, Kepong. 96 pp.
WYATT-SMITH, J. 1963. Manual of Malayan silviculture for inland forests. Malayan Forest Records 23.
A Comparative Study of the Ant Fauna in a Primary and Secondary
Forest in Sabah, Malaysia
ARTHUR Y.c. CHUNG & MARYATI MOHAMED
Biology Department, Science & Natural Resources Faculty, Universiti Kebangsaan Malaysia,
Sabah Campus, Locked Bag No. 62, 88996 Kota Kinabalu, Sabah
ABSTRACT. The diversity and ecology of ants in a primary and secondary selectively logged forest was
compared over a 12 month period at Danum Valley, Sabah, Malaysia. Species richness and diversity were
higher in the primary forest. A total of 192 species from seven subfamilies were recorded. The
Myrmicinae was the most common subfamily in species and abundance, for both habitats. Rainfall had
some negative effects on subterranean fauna in primary and secondary forest while high temperature
diminished the number of arboreal species collected in the primary forest.
KEY WORDS: ant diversity, ecology, Myrmicinae, Formicinae, Ponerinae.
INTRODUCTION
Ants are extremely abundant, species-rich, and ecologically important in tropical lowland
rain forests (Atkin & Proctor 1988, Burghouts et al. 1992, Levings 1983). Measurements
suggest that about one third of the entire animal biomass of the Amazonian terra firme rain
forest is composed of ants and termites, with each hectare of soil containing in excess of
eight million ants and one million termites (Fittkau & Klinge 1973). These insects, along
with bees and wasps, account for more than 75 percent of the total insect biomass in that
forest. The diversity of ants is substantial, far exceeding that of other social insects and
reflecting the manner in which ant species have evolved to saturate a wide range of feeding
niches in soil and vegetation (Holldobler & Wilson 1990). In a lowland rain forest in Papua
New Guinea, Wilson (1959) collected 172 species of ants belonging to 59 genera in an area
of about one square mile (2.6 km2). Olson (1991) recorded 41 genera and 122 species from
three primary forest sites in Costa Rica, while Levings (1983) collected 49 genera and 127
species in a tropical moist forest in Panama. Bolton (in Room 1971) recorded 219 species
in 63 genera in a square mile of a cocoa plantation and forest in Ghana.
Tropical rain forests are the richest ecosystems that the world has ever known (Marshall
1992). However, it has been estimated that 11 million hectares of mature tropical forests, of
which 7.5 million hectares are rain forests, are converted each year to other uses. The great
majority is changed to non-forest uses, only 600,000 hectares becoming timber plantations
(Gomez-Pompa & Burley, in Marshall 1992), resulting in a loss of global biodiversity.
Ants, one of the most diverse and abundant group of organisms in the rainforest, have great
potential as indicator species for studying the effects of deforestation. The aims of this
research were to describe the ant fauna in the rain forest, to study the effects of logging and
to look at the effects of biotic and abiotic factors on the diversity and ecology of ants.
Study Area and the Environment

A 12-month study was conducted in Danum Valley Conservation Area, Sabah, Malaysia
(40 58' Nand 1170 48' E), which comprises 43,800 hectares of tropical lowland evergreen
357
D. S. Edwards et al. (eds.), Tropical Rainforest Research - Current Issues, 357-366,
358 ARTHUR Y.c. CHUNG AND MARYATI MOHAMED
temperature (ee) rainfall (mm)

30 . 600
25
300
200
100
Jan Feb Mar Apr May Jun Jul Agu Sap Oct Nov Dec
month (1992)
- temperature ~ rainfall
Figure L Mean temperature and monthly rainfall in Danum Valley Field Centre (1992),
dipterocarp rain forest. This area is situated in eastern Sabah at the upper reaches of the
Segama River, approximately 70 km from the coast.
The total rainfall recorded at Danum Valley Field Centre (DVFC) for 1992 was 2314 mm,
which was lower than the annual rainfall average of 2822 mm (based on 5 years of records).
It was dry at the beginning of the year until April and two rainfall peaks in June (302.7 mm)
and October (428.1 mm) were observed. This pattern (Figure 1) is influenced by the edge
effects of two monsoons; the wetter northeast monsoon from November to March and the
drier but more consistent southwestern monsoon in June and July (Marsh & Greer 1992).
Mean annual temperature at DVFC in 1992 was 26.7°C. The highest mean monthly
temperature recorded was 28.3°C in April, and the lowest was 25.6°C in December.
Two permanent 100 x 150 m plots were used to study the ant fauna; one each in a
primary and secondary forest. The altitude of the plots was 180-230 m. In both plots, three
line transects (150 m in length and 50 m apart) were established to collect ants. The primary
forest plot was located at 0.6 km from DVFC, and was composed primarily of dipterocarps
with a high canopy at about 40 m. The secondary forest plot was located within a three year
old selectively logged forest, 2.0 km from DVFC.
Sampling was done once a month along the transects in each plot. Three methods were used
to collect ants at the arboreal, terrestrial and subterranean levels.
Manual Sampling
A sweep net, an aspirator and forceps were used to collect arboreal ants. The maximum
height that could be reached using this method was about 3 m above the ground. Sampling
was carried out in the morning (0830-1130 hrs), by collecting ants found on vegetation along
the transects.
Baited Pitfall Sampling

Plastic cups with a diameter of 7 cm and a depth of 12 cm were planted in the soil, every
15 m along the transects. Eleven pitfall traps were set up in each transect, resulting in a total
of 33 for one plot. Tuna flakes in oil were used as bait. The bait was wrapped in a piece of
Ant diversity in primary and secondary tropical forests 359
cloth which was hung across the cup with a small stick. The traps were lI8-filled with water
and liquid detergent was added to reduce water tension. They were left overnight in the
forest and emptied the following day. This method was used for collecting nocturnal and
terrestrial ants.
Winkler's Sampling
Leaf litter and soil to about 5 cm depth were collected and sifted from eleven 0.25 m2
quadrats chosen at random along one transect. The sifted litter and soil were transferred to
debris bags in the field. When the bags were brought back to DVFC, the contents were
transferred to several flat mesh bags which were hung inside an outer cloth sack (Winkler's
bag). Soil organisms worked their way out of the litter and dropped into a container which
was placed at the base of the Winkler's bag. Specimens were collected and sorted the next
day.
Sorting and Identification of Ants

Most of the ants were identified to generic level using Bolton (1990) and Murphy (1973).
Specimens are deposited with the Entomology Section, Universiti Kebangsaan Malaysia,
Sabah Campus.
Statistical Analysis
For the analyses of data, the Wilcoxon paired-sample test, which is a non parametric
analogue to the paired sample t-test, was used as the data was not normally distributed.
Shannon's Index (H') was used to evaluate species diversity. This provides the average
degree of uncertainty in predicting what species an individual chosen at random from a
sample will belong to. Hutcheson '1' test was applied to test for the difference between two
diversity indices in each month (Zar 1984).
Hill's Modified Index (E5) was used to calculate the degree of evenness of ant species
distribution. The value approaches zero as a single species becomes more and more dominant
in a community. This index was recommended as it is the least ambiguous and easiest to
interpret. It does not require an estimate of the number of species in the community, a
feature which is affected by sample size (Ludwig & Reynolds 1988).
Multiple correlation was used to detect relationships among variables throughout the year.
The variables were number of species and individuals collected with different methods, total
species, abundance, diversity, evenness, rainfall and temperature.
Species Richness
A total of 192 ant species from more than 50 genera and seven subfamilies were recorded
(see Appendix). However, some taxa were identified only to subfamily level. The percentage
of species and abundance recorded for each subfamily in the primary and secondary plot are
shown in Figure 2(i) and (ii). Myrmicinae was the most abundant and diverse subfamily in
both plots, followed by Formicinae and Ponerinae. Percentage of species and abundance of
the myrmicine ants were relatively higher in the secondary plot. The abundance of formicine
and ponerine ants were higher in the primary plot. Dorylinae, Pseudomyrmecinae and
Leptanillinae were represented by very few species and individuals in both plots.
Throughout the year, species richness was higher in the primary plot (Wilcoxon test,
P<O.OI). 164 (86%) species (out of the total of 192) were recorded in the primary plot,
while 155 (81 %) species were found in the secondary plot. 19.3% of the total species were
found exclusively in the primary plot and 14.6% were recorded exclusively in the secondary
plot. The monthly overlap of species was 33-44 %. The annual overlap of species was 66 %.
360 ARTHUR Y.C. CHUNG AND MARYATI MOHAMED
subfamily
Formlclnae
Myrmlclnae
Dollchoderlnae
Dorytlnae
Ponerlnae
Leptanlllinae
Pseudomyrmeclnae
o 10 20, 30 40 60 60
percentage of species
o primary ~ secondary
Figure 2 (i). Percentage of species in each subfamily, in primary and secondary plots.
subfamily
r~;$ f'" ,
Formlclnae
Uyrmlclnae I
H
, , , , 96.3
Dollchoderlnae
i " : I
Dorytlnae
Ponerlnae ::::::::::::::::::::::1,1."
Leptanlllinae
Paeudomyrmeclnae t '
o 6 10 15 20 26
percentage of abundance
o primary ~ secondary
Figure 2 (ii). Percentage of ant abundance in each subfamily, in primary and secondary plots.
Diversity
The Shannon's diversity indices were significantly higher in the primary plot with the ratio
of 10:2 for the 12 months' sampling (Hutcheson 't' test, P<O.OS). The average value of
Shannon's index (formula based on In) was 1.83 (SD=0.48), compared to 1.47 (SD=0.37)
in the secondary plot. The higher value in diversity index suggests that either one or both
species richness and evenness were higher in the primary plot because diversity indices
incorporate species richness and evenness into a single value.
Evenness
The average value of the evenness index (ES) was fairly low; 0.40 (SD=0.08) in the
primary plot and 0.39 (SD=0.08) in the secondary plot. This suggests that dominance of
certain ant species occurred in both plots and different ant-species had very different colony
sizes. It also showed that higher diversity in the primary plot was due to species richness and
the low value of evenness had indirectly affected the diversity values in both plots.
Table I. Mean total number of species and individuals (±SD) collected each month with different methods in
primary and secondary plots throughout the year.
Pitfall Manual Winkler's

Species
primary plot 21.5±3.5 20.8±3.1 33.9±3.8
secondary plot 20.7±2.6 22.6±2.6 23.8±3.4
Individuals
primary plot 5498±2682 131±34 337±147
secondary plot 6818±3445 156±29 218+67
Table 2. Ant species that achieved more than 5% in the total abundance in respective plots throughout the year.
Species code % of abundance

Primary plot
Pheidole sp. DV83 49.0
Euprenoiepis sp. DV121 8.3
Pheidoiogiton sp. DV45 7.8
Pheidole sp. DV14 6.5
Secondary plot
Pheidoie sp. DV83 66.0
Pheidoiogiton sp' DV45 8.6
Abundance
Although more ants were collected from the secondary plot, the difference was not
significant. Abundance in both plots varied and fluctuated throughout the year. The
inconsistency in ant abundance may be the result of the methods used to collect ants. Baited
pitfall sampling collected the most ants (Table 1), thus it was this method that primarily
affected the total abundance of ants collected. At times, thousands of ants from one species
were found in one pitfall trap because the trap was placed within the foraging trail of the
species. Catches of any ant species depends on the size and abundance of its foraging
population, the dispersion of colonies, levels of activity, the size of individual ants, and the
thoroughness with which foragers cover their territory (Southwood 1966). The abundances
in both plots were dominated by a myrmicine species, Pheidole sp. (DV83), large colonies
of which were collected, mainly by pitfall sampling. Table 2 shows that four ant species from
the primary plot and only two from the secondary plot achieved more than 5 % of the total
abundance in the respective plots.
Frequency of Occurrence
The number of individuals collected in an area is not a good indication of how common
a species is, because of differences in foraging behaviour of different ant species (e.g,
solitary or in colonies). The frequency of occurrence seems to give a more comprehensive
picture. In this context, 'common' refers to the chances of being encountered. Figure 3
shows the frequency of occurrence of ant species at 12 sampling times. High percentages of
species that occurred once only throughout the year in both plots suggests that a large number
of ant species migrate from one place to another for food and shelter. 9.1 % of the ant species
in the primary plot and 8.4% in the secondary plot achieved 100% occurrence (Table 3).
362 ARTHUR Y.c. CHUNG AND MARYATI MOHAMED
Table 3. Ant species that were present lhroughout the year in primary and secondary plot.
Primary plot Secondary plot

Species code Species code
Odontoponera sp. DVI Odontoponera sp. DVI
CampOflOtus gigas DV3 Camponotus gigas DV3
Diacarnma rugosum DVIO Diacarnma rugosum DVIO
Pheidole sp. DV14 Pheidole sp. DV14
Pheidole sp. DV83 Pheidole sp. DV83
POflera sp. DV23 Ponera sp. DV23
Hypoponera sp. DV87 Hypoponera sp. DV87
Odontomachus rixosus DVll Polyrhachis sp. DV7
Pheidole sp. DV20 Crematogaster sp. DV18
Pheidole sp. DVl18 Polyrhachis ypsilon DV32
Tapinoma sp. DV21 Pheidologiton sp. DV45
Colobopsis sp. DV30 Strumigenys sp. DV48
Polyrhachis bihamata DV31 Pheidole sp. DV1l3
Discothyrea sp. DV74
Eugrenolegis sp. DV121
3~r-.
2 3 4 5 6 7 8 9 10 11 12
frequency of occurence
EJ primary I§S3 secondary
Figure 3. Frequency of occurrence in primary and secondary plots lhroughout the year.
Comparison of Species and Individuals Collected with Different Methods in Primary and
Secondary Forest
Table 1 shows the mean total number of species and individuals collected each month.
Number of species (Wilcoxon test, P<O.OOl) and number of individuals (P <0.05) collected
with the Winkler's method were significantly lower in the secondary plot. Korthals (1990),
in his research in Danum Valley, reported that the soil in the secondary forest contained
more sand, less clay and seemed to be more compact than the primary forest. Conductivity,
percentage of carbon, nitrogen, cation exchange capacity, and total and soluble phosphorus
were lower in the secondary forest. The soil in the primary forest is less disturbed and has
a higher nutrient status, thus making it more suitable for the soil fauna.
The number of individuals collected by manual sampling were higher (Wilcoxon test,
P<0.05) in the secondary plot. Manual sampling collected mainly arboreal ants. In general,
10~-:=
80~-·"
60 ..... ;;7~ .
40
I
201-·
oL-~ __ L-~ __ ~ __ ~ __ ~ __ ~ __ ~ __ ~
Jan Feb Mar Apr May Jun Jut Agu Sep Oct Nov Dec
month (1992)
- - primary -+- secondary
Figure 4. Cumulative percentage of species in both plots throughout the year.
trees in the secondary forest are lower with their canopy not as dense as those in the primary
forest. Consequently, more light reaches the lower strata of the forest, resulting in an
undergrowth which is relatively dense in certain parts of the plot. Climbers and shrubs like
Eupatorium, Jacquemontia, Callicarpa and Merremia are abundant in such vegetation. Many
of these plants produce extrafloral nectaries and food bodies which are a source of food for
the ants. In return, the ants protect the plant from herbivore attack. Some plants, such as
Macaranga hypoleuca and Korthalsia furtadoana, also provide shelter for ants. Ants stay in
the stem and branch cavities of M. hypoleuca and in the ochreas (apical leaf sheath) of K.
furtadoana. In primary forest, ground vegetation is scarce and most of the arboreal ants are
found in the canopy (Fittkau & Klinge 1973) which was beyond the limit of the manual
sampling.
Figure 4 estimates that to obtain 50% of the ant species recorded in this study, sampling
using the three different methods should be carried out twice; to obtain 80% of the species,
sampling should be carried out five times (sampling once a month).
Effects of Rainfall and Temperature

The number of species collected with the Winkler's method in the secondary plot and the
number of individuals in the primary plot, collected by the same method, showed negative
correlation with monthly rainfall (P<0.05). In the rain forest, 80.7% of the annual rainfall
in 1990 reached the forest floor as through fall and 1. 9 % as stemflow, giving an interception
loss of 17.4% (Sinun et at. 1992). This percentage of through fall affected the soil fauna.
Rainfall reduced population density in the primary forest but did not affect the popUlation
density in the secondary forest. However, it caused a reduction in the number of species
observed in the secondary forest, suggesting that while some species were disappearing, the
remaining species were thriving after rainfall.
Ant species collected by manual sampling in the primary plot correlated negatively with
high monthly mean temperature (P<O.OOl) and high daily temperature (P<0.05)
(temperature taken at 0800hrs of the sampling day) whereas no correlation was detected in
the secondary plot. This suggests that arboreal ant species in the primary forest have a lower
tolerance towards high temperatures, compared with those of the secondary forest. Possibly
temperature affected only the arboreal ants because microhabitats of arboreal ants are more
exposed to changes of temperature, when compared to terrestrial and subterranean species.
Terrestrial and subterranean microhabitats are often surrounded by soil and leaf litter where
the humidity is constantly high. Torres (1984), in his research on upland tropical forest,
grassland and agricultural land in Puerto Rico, reported that forest ants had a lower tolerance
to high temperatures than ants from non-forested areas.
ACKNOWLEDGEMENTS
We would like to thank Danum Valley Management Committee for permission to work
at DVFC. We are grateful to the Danum researchers, especially Dr. J. Willot, Dr. P.M.
Costa and R. Nussbaum, for comments on an earlier version of this manuscript. This study
was funded by a M.Sc. Scholarship from Sabah Shell and partly from IRPA grant 04-07-03-
07, and supported by the Royal Society, UK.
LITERATURE CITED
ATKIN, L. & PROCTOR, J. 1988. Invertebrates in the litter and soil on Volcan Barva, Costa Rica. Joul7U1lof
BOLTON, B. 1990. Key to the ant genera of Oriental region. pp. 46-53 in Holldobler, B. & Wilson E.O. (eds).
The ants. Belknap Press of Harvard University Press, Cambridge, Massachusetts.
BURGHOUTS, T., KORTHALS, G. & VRIES, T.D. 1992. Litterfall, leaf litter decomposition and litter
invertebrates in primary and selectively logged dipterocarp forest in Sabah, Malaysia. pp. 407-416 in Marshall,
A.G. & Swaine, M.D. (eds). Tropical rainforest: disturbance and recovery. The Royal Society, London.
FITTKAU, E.J. & KLINGE, H. 1973. On biomass and trophic structure of the Central Amazonian rain forest
ecosystem. Biotropica 5:2-14.
HOLLDOBLER, B. & WILSON, E.O. 1990. The ants. Belknap Press of Harvard University Press, Cambridge,
Massachusetts. 732 pp.
KORTHALS, G.W. 1990. A description of the litter dwelling arthropodfauna in aprirnary and a secontinry rain
forest in North-East Borneo. Ph.D. Thesis, Free University, Amsterdam.
LEVINGS, S.C. 1983. Seasonal, annual and among site variation in the ground ant community of a deciduous
tropical forest: some causes of patchy species distributions. Ecological Monographs 53:435-455.
LUDWIG, J.A. & REYNOLDS, J.F. 1988. Statistical ecology. John Wiley & sons, Inc. New York. 337 pp.
MARSH, C.W. & GREER, A.G. 1992. Forest land-use in Sabah, Malaysia: an introduction to Danum Valley.
Pp. 331-340 in Marshall, A.G. & Swaine, M.D. (eds). Tropical rainforest: disturbance and recovery. The
Royal Society, London.
MARSHALL, A.G. 1992. The Royal Society'S South-east Asia rain forest research programme: an introduction.
pp. 327-330 in Marshall, A.G. & Swaine, M.D. (eds). Tropical rain forest: disturbance and recovery. The
Royal Society, London.
MURPHY, D.H. 1973. Preliminary key to genera of Malaysian ants, based on workers. Bulcit Tirnah Survey:
Supplement 1.
OLSON, D.M. 1991. A comparison of the efficacy of litter and pitfall traps for sampling leaf litter ants
(Hymenoptera: Formicidae) in a tropical wet forest, Costa Rica. Biotropica 23:166-172.
ROOM, P.M. 1971. The relative distributions of ants species in Ghana's cocoa farms. Journal of Animal Ecology
40:735-751.
SINUN, W., WONG, W.M., DOUGLAS, I. & SPENCER, T. 1992. Throughfall, stemflow, overland flow and
throughflow in the Ulu Segama rain forest, Sabah, Malaysia. Pp. 389-396 in Marshall, A.G. & Swaine, M.D.
(eds). Tropical rainforest: disturbance and recovery. The Royal Society, London.
SOUTHWOOD, T.R.E. 1966. Ecological methods, with particular reference to the study of insect populations.
Chapman & Hall, London. 542 pp.
WILSON, E.O. 1959. Some ecological characteristics of ants in New Guinea rain forests. Ecology 40:437-447.
ZAR, J.H. 1984. Biostatistical analysis. Prentice Hall, New Jersey. 718 pp.
Appendix. Ant species recorded in Danwn Valley Conservation Area (1972). Species not in italics are identified to
subfamily. sf (Subfamily) - Myrmicinae (M.), Formicinae (Fl, Ponerinae (P), Dolicherodinae (D), Dorylinae (DO),
Pseudomyrmicinae (PS), Leptanillinae (L). Size in mm. Method - Manual (M), Pitfall (Pl, Winkler (W).
No. Genus/Species Code s.f. Size Method No. Genus/Species Code s.f. Size Method
1 Acanthomyrmex A DV 71 M 3 M,W 57 Eurhopa/othrix A DV 99 M 3 W
2 Acanthomyrmex B DV 164 M 3 W 58 Eurhopalothrix B DV 97 M 2 W
3 Acropyga A DV 12 F 4 M 59 Eurhopa/othrix C DV 143 M 3 W
4 Aenictus A DV 16 DO 4 M 60 Formicine A DV 104 F 4 M
5 Aenictus B DV 66 DO 4 M 61 Formicine B DV 108 F 5 M
6 Aenictus C DV 151 DO 3 M 62 Formicine C DV 181 F 4 M
7 Anochetus A DV 17 P 4 W 63 Formicine D DV 187 F 3 P
8 Anochetus B DV 95 P 2 W 64 Formicine E DV 127 F 13 M
9 Anochetus C DV 186 P 6 W 65 Formicine F DV 169 F 1 W
10 Bothroponera A DV 165 P 7 W 66 Gnamptogenys A DV 57 P 6 M
II Calyptomyrmex A DV 50 M 3 W 67 Gnamptogenys B DV 98 P 5 W
12 Calyptomyrmex B DV 93 M 2 W 68 Gnamptogenys C DV 100 P 3 W
13 Camponotus A DV 70 F 5 M,P 69 Gnamptogenys D DV 148 P 5 M
14 Camponotus B DV 79 F 3 W 70 Gnamptogenys E DV 177 P 2 W
15 Camponotus C DV 107 F 12 M 71 Hypoponera A DV 26 P 7 W
16 Camponotus gigas DV 3 F 25 M,P 72 Hypoponera B DV 73 P 2 W
17 Camponotus variegatus DV 22 F 7 P,M 73 Hypoponera C DV 75 P 1 W
18 Catauiacus A DV 115 M 4 M 74 Hypoponera D DV 87 P 3 W
19 Ceraphachys A DV 135 P 2 W 75 Hypoponera D DV 77 P 2 W
20 Ceraphachys B DV 136 P 2 W 76 Hypoponera E DV 176 P 6 W
21 Colobopsis A DV 30 F 5 M 77 Labidogenys A DV 49 M 2 W
22 Colobopsis B DV 59 F 6 M 78 Labidogenys B DV 142 M 2 W
23 Colobopsis C DV 43 F 6 M 79 Leptanilia A DV 91 L 2 W
24 Colobopsis D DV 179 F 7 M 80 Leptanilia B? DV 140 L 4 W
25 Cremntogaster A DV 5 M 4 M 81 Leptogenys A DV 2 P 8 M,P,W
26 Cremntogaster B DV 18 M 2 M,P,W 82 Leptogenys B DV 144 P 6 P
27 Crematogaster C DV 34 M 4 M 83 Leptogenys C DV 6 P 9 M,P,W
28 Crematogaster D DV 62 M 3 M 84 Leptogenys D DV 25 P 3 M
29 Crematogaster difformis DV 28 M 5 M 85 Liometopum A? DV 114 D 5 M
30 Cremntogaster E DV 102 M 2 M 86 Liomyrmex A? DV 124 M 3 P
31 Cremntogaste.r F DV 184 M 2 M 87 Mesoponera A DV 53 P 6 M
32 Crematogaster G DV 105 M 2 M 88 Mesoponera B DV 90 P 7 W
33 Crematogaster H DV 156 M 2 W 89 Monomorium A DV 68 M M
34 Crematogaster I DV 192 M 3 P 90 Myrmicaria A DV 9 M 6 M,P
35 Cremntogaster infIata DV 33 M 5 M 91 Myrmicaria B DV 110 M 5 M
36 Diacamma A DV 162 P 9 M 92 Myrmicaria C DV 155 M 5 M
37 Diacamma rugosum DV 10 P 12 M,P 93 Myrmicine A DV 188 M 2 W
38 Discothyrea A DV 74 P 2 W 94 Myrmicine B DV 191 M W
39 Discothyrea B DV 133 P 3 W 95 Myrmicine C DV 63 M 4 M
40 Dolichoderine A DV 55 D 2 P 96 Myrmicine D DV 85 M 4 M,W
41 Dolichoderine B DV 123 D 5 P 97 Myrmicine E DV 86 M 3 W
42 Dolichoderine C DV 152 D 5 M 98 Myrmicine F DV 92 M 5 W
43 Dolichoderus A DV 4 D 5 M 99 Myrmicine H DV 96 M 1 W
44 Dolichoderus B DV 38 D 4 M 100 Myrmicine I DV 111 M 4 M
45 Dolichoderus C DV 67 D 5 M 101 Myrmicine J DV 119 M 2 P
46 Dolichoderus D DV 167 D 3 M 102 Myrmicine K DV 125 M 2 P
47 Dolichoderus E DV 58 D 5 M 103 Myrmicine L DV 131 M 5 W
48 Dolichoderus F DV 126 D 8 P,M 104 Myrmicine M DV 132 M 3 W
49 Dolichoderus G DV 44 D 6 M 105 Myrmicine N DV 145 M 3 P
50 Dolichoderus H DV 64 D 6 M 106 Myrmicine 0 DV 149 M 1 M
51 Dolichoderus I DV 190 D 3 M 107 Myrmicine P DV 157 M 3 W
52 Dolichoderus thoracicus DV 60 D 4 M,P 108 Myrmicine Q DV 158 M 1 W
53 Echinopla A DV 35 F 6 M 109 Myrmicine R DV 159 M 3 W
54 Epitritus A DV 94 M W 110 Myrmicine S DV 160 3 W
55 Euprenolepis A DV 121 F 5 P 111 Myrmicine T DV 163 2 W
56 Euerenoleeis B DV 13 F 3 M,P 112 Myrmicine U DV 171 3 P
Appendix (continued). Ant species recorded in Danurn Valley Conservation area (1972). Species not in italics are
identified to subfamily. s.f. (subfamily) - Myrmicinae (M), Formicinae (F), Ponerinae (P), Dorylinae (DO),
Pseudomyrmicinae (PS), Leptanillinae (L). Size in mm. Method - Manual (M), Pitfall (P), Winkler (W).
No. Genus/S2ecies Code s.f. Size Method No. Genus/S2ecies Code s.f. Size Method
113 Myrmicine V DV 180 M 2 W 153 Polyrhachis L DV 154 F 8 M
114 Mymwteras A DV 27 F 4 W 154 Polyrhachis M DV 54 F 7 M
115 Mymwteras B DV 134 F 4 W 155 Polyrhachis N DV 168 F 7 M
116 Mystrium A DV 84 M 4 W 156 Polyrhachis 0 DV 174 F 10 M
117 Odontomachus rixosus DV 11 P 12 M,P,W 157 Polyrhachis P DV 175 F 7 M
118 Odontoponera transversa DV 1 P 10 M,P,W 158 Polyrhachis Q DV 182 F 5 M
119 Oligomyrmex A? DV 41 M 2 M 159 Polyrhachis rastellata DV 37 F 5 M
120 Pachycondyla A DV 47 P 11 W 160 Polyrhachis ypsilon DV 32 F 12 M
121 Pachycondyla B DV 52 P 8 M,W 161 Ponera A DV 23 P 3 P,W
122 Pachycondyla C DV 128 P 5 M 162 Ponerine A DV 138 P 2 W
123 Paratrechina A DV 15 F 3 M,P,W 163 Ponerine B DV 172 P 6 M
124 Paratrechina B DV 103 F 2 M 164 Ponerine C DV 183 P 2 W
125 Pheidole A DV 14 M 2 M,P,W 165 Ponerine D DV 189 P 4 M
126 Pheidole B DV 42 M 4 M 166 Proatta A DV 141 M 2 W
127 Pheidole C DV 69 M 1 M,P 167 Proceratium A DV 80 P 4 W
128 Pheidole D DV 83 M 3 M,P,W 168 Proceratium B DV 101 P 3 W
129 Pheidole E DV 113 M 2 M 169 Proceratium C DV 161 P 2 W
130 Pheidole F DV 118 M P 170 Proceratium D DV 185 P 4 W
131 Pheidole G DV 116 M 3 P 171 Pseudolasius A? DV 139 F 3 W
132 Pheidole H DV 122 M 1 P 172 Smithistruma A? DV 56 M 2 P
133 Pheidole I DV 20 M 3 M,P 173 Strumigenys A DV 48 M W
134 Pheidole J DV 129 M 3 M 174 Strumigenys B DV 76 M 1 W
135 Pheidole K DV 178 M W 175 Strumigenys C DV 170 M 2 W
136 Pheidologiton A DV 45 M 2 M,P 176 Strumigenys D DV 24 M 3 W
137 Pheidologiton B DV 88 M 2 W 177 Strumigenys E DV 82 M 3 W
138 Pheidologiton C DV 146 M 3 P 178 Tapinoma A DV 166 F 2 M
139 Polyrhachis A DV 7 F 6 M 179 Tapinoma B DV 21 F 4 M,P,W
140 Polyrhachis armata DV 39 F 7 M,P 180 Tapinoma C DV 29 F 1 M
141 Polyrhachis B DV 8 F 4 M 181 Tetramorium A DV 51 M 2 W
142 Polyrhachis bihamata DV 31 F 11 M 182 Tetramorium B DV 72 M 4 M,P
143 Polyrhachis C DV 36 F 6 M 183 Tetramorium C DV 81 M 3 W
144 Polyrhachis calybea DV 173 F 9 M 184 Tetramorium D DV 89 M W
145 Polyrhachis D DV 40 F 5 M 185 Tetramorium E DV 112 M 4 M
146 Polyrhachis E DV 46 F 6 M 186 Tetramorium F DV 117 M 2 P
147 Polyrhachis F DV 61 F 5 M 187 Tetramorium G DV 137 M 3 W
148 Polyrhachis G DV 105 F 5 M 188 Tetramorium H DV 150 M 2 M
149 Polyrhachis H DV 109 F 7 M 189 Tetramorium I DV 78 M 2 W
150 Polyrhachis I DV 120 F 9 M 190 Tetraponera A DV 147 PS 7 M
151 Polyrhachis J DV 130 F 4 M 191 Trigonogaster A DV 19 M 2 P
152 Polyrhachis K DV 153 F 7 M 192 Vollenhovia A? DV 65 M 4 M
Comparative morphology and physiology of gap and understorey seedling
leaves from a mixed dipterocarp forest
WEBBER E. BOOTH
Biology Department, Universiti Brunei Darussalam, Bandar Seri Begawan 2028, Brunei
Darussalam
ABSTRACT. Leaf morphology and rates of stomatal condllctance, net photosynthesis and dark
respiration were determined for seedlings associated with a gap-understorey situation in the mixed
dipterocarp rainforest of the Batu Apoi Forest Reserve in the Temburong District of Brunei Darussalam.
The species investigated included four emergent tree species, Shorea falciferoides Foxw., Shorea
leprosula Miq., Parashorea macrophylla Wyatt-Sm. ex Ashton, and Dryobalanops lanceolata Burck.,
two canopy species Aporosa lunata (Miq.) Kurz and Drypetes sp. and two pioneer species, Macaranga
trachyphylla Airy Shaw and Musa sp. The pioneer species were confined to the gap while all other
species were present in both the gap and understorey. Comparative morphological studies indicated that
leaves of understorey seedlings are thinner, have a lower stomatal density and higher specific leaf area
than leaves of the same species growing in the gap. Stomatal conductances for leaves of gap seedlings
fall into two categories; high rates for the pioneers and more conservative rates for the non-pioneers. In
the understorey low photosynthetically active photon flux density values are associated with relatively high
stomatal conductance. Rates of light saturated net photosynthesis were significantly higher for pioneer
species than non-pioneers. Based on apparent quantum yields, understorey seedlings tended to be
photosynthetically more efficient than gap plants; also, light compensation points and rates of dark
respiration were lower for understorey plants than for those of the same species growing in the gap.
These features may be important adaptations which allow seedlings to establish in the very low light of
the understorey.
KEY WORDS: tropical rainforest, seedlings, gap, understorey, leaf morphology, photosynthesis,
respiration, Borneo.
INTRODUCTION
The death of an individual tree or group produces a gap in the canopy into which other trees
grow. These in turn reach maturity and perhaps senescence, then die (Whitmore 1975). As
a result of such natural disturbance the tropical rainforest is in a state of dynamic equilibrium
being a mosaic of patches in different stages of growth (Chazdon & Fetcher 1984,
Martinez-Ramos et al. 1988, Whitmore 1989). Based on the understanding that the
successional status of a species is determined by the ability of its seedlings to exploit the
available photosynthetically active photon flux density (PPFD) and the validity of the
observation that in Asian forests the seedlings of canopy dominants and emergent trees often
establish in the understorey then later are dependent on the formation of a gap for further
development (Whitmore 1975, 1978), then it must follow that these seedlings are able to
initially exploit the very low PPFD of the understorey and subsequently the high PPFD of
the gap. Tropical rainforest seedlings therefore require a high plasticity in terms of their
physiology and morphology as they encounter contrasting environments (Canham 1989,
Osunkoya & Ash 1991, Popma & Bongers 1991). The capacity to tolerate different light
regimes involves not only adjustments of the photosynthetic apparatus but also of the leaf
structure (Bjorkman 1981, Loach 1970, Riddoch et al. 1991). In contrast to the canopy and
emergent species, short-lived pioneer plants establish only after the gap is formed and
therefore need only an enhanced capacity to utilize the high PPFD of the gap (Bazzaz &
Pickett 1980, Martinez-Ramos 1988, Whitmore 1989). They can therefore be expected to
367
368 WEBBER BOOTH
have a distinctive physiology.

Differences in structure and physiology of leaves of seedlings from gap and understorey
situations as suggested above, are widely accepted as being applicable in the tropical
rainforest situation but few supporting in situ data exist. It was therefore the objective of this
study to gather such data from a gap-understorey situation within the lowland mixed
dipterocarp rainforest in the Temburong District of Brunei Darussalam for a selection of
emergent, canopy and pioneer seedlings.
Site
The study site was located beside the Belalong river approximately 2 km upstream from
the Kuala Belalong Field Studies Centre in the Batu Apoi Forest Reserve. It consisted of a
gap approximately 20 x 100 m with the longitudinal axis oriented E-Wand extending across
the river so providing exposure to high irradiance from about 0900 to 1500 hours. For
information on the geology and geomorphology of this area refer to Ross & Dykes (this
volume).
Species
The species studied included four emergent trees, Shorea falciferoides Foxw., Shorea
leprosula Miq., Parashorea macrophylla Wyatt-Sm. ex Ashton and Dryobalanops lanceolata
Burck; two canopy trees, Aporosa lunata (Miq.) Kurz and Drypetes sp. and two pioneer
species, Macaranga trachyphylla Airey Shaw, a tree, and Musa sp., a tall, robust herb. The
two pioneers were confined to the gap while the six other species were present both in the
gap and understorey. Only seedlings 0.5 to 2.0 m tall were included in this study.
Leaf morphology
Recently expanded leaves from five seedlings for each species represented in the gap and
understorey were selected for stomatal density and leaf thickness determinations. Total leaf
thickness and the thicknesses of the adaxial epidermis and cuticle, palisade mesophyl\, spongy
mesophyll and abaxial epidermis and cuticle were determined from examination of hand-cut
transverse sections from the mid-lamina region of the leaves using a calibrated ocular
micrometer at 400x.
Stomatal densities were determined from examination of nail-varnish impressions of ad-
and abaxial leaf surfaces using a calibrated ocular grid at 400x. Specific leaf area (SLA)
determinations were based on leaf area/weight measurements for recently expanded leaves
from ten seedlings for each species in each situation. Areas were calculated from leaf
tracings on even density paper and weights from specimens dried to constant weight at 80°C
in a forced draught oven.
Gas exchange
Measurements of rates of carbon dioxide and water vapour exchange were made using a
LI-COR LI-6200 transportable photosynthetic system. The IRGA component was calibrated
against gas of known carbon dioxide content and the RH sensor against a salt-cup and
subsequently factory calibrated LI-1600 steady state porometer. Recently expanded leaves
from three seedlings for each species present in the gap and understorey were tagged and gas
exchange rates were determined for them at one to two hourly intervals. Because of
functional limitations of the LI-6200 system, measurements at ambient RH above 95 % were
avoided. Only two sets of diurnal records are presented in the results, one for a comparison
of the eight gap species measured on 24 August 1991, and the other a gap-understorey
Seedling leaf morphology and physiology 369
comparison using Shoreafalciferoides on 29 July 1991. These, together with results from
other occasions from July 1991 through to February 1992, have allowed comparison of rates
of light saturated carbon assimilation for gap plants and apparent quantum yields and light
compensation points to be determined for both gap and understorey plants. Values influenced
by obvious signs of water stress are not included in these comparisons. Dark respiration
rates were determined from measurements conducted on darkened leaves from five seedlings
for each species in each situation except for Musa sp. which, because of the form of its leaf,
was not able to be satisfactorily darkened.
RESULTS
Leaf morphology
As shown in Table 1, leaf morphologies of gap and understorey seedlings differed in a
number of respects. Total leaf thickness was significantly greater for gap plants than
understorey plants. All leaf tissues contributed to this 22-35% difference. However, the
greatest contribution was from the spongy mesophyll, except in the case of Shorea
falciferoides and Parashorea macrophylla, where the palisade mesophyll contributed most.
The gap species Macaranga trachyphylla and Musa sp. contrasted in their leaf thickness the
former having a thin leaf and the latter a thick leaf. Musa sp. had, within its spongy
mesophyll, distinctive substomatal, tunnel-like cavities. Stomatal densities were substantially
higher (15-52 %) for gap plants than those of the understorey. The gap plant, Musa sp. was
distinctive in that it was amphistomatal, with a very low density of large raised stomata on
the adaxial leaf surface (l7±2 mm-2) and moderate density on the abaxial surface. Specific
leaf areas for the understorey plants were 31-68% higher than those of the gap. The pioneer
species, Macaranga trachyphylla, had a typically low SLA, whereas Musa sp. had a high
SLA, greater than any of the understorey seedlings.
Gas exchange
Diurnal patterns of stomatal conductance (Figure la) and net photosynthesis (Figure Ib)
were recorded for all gap species on 24 August 1991, during a period of relatively dry
weather. Rapid changes in both PPFD and RH at 0900 hours make detailed comparisons
between species impossible at this time although the pioneers are clearly set apart with very
high rates of stomatal conductance and high net photosynthesis. Rates at 1000 hours and
1200 hours are associated with saturating PPFD, low RH and relatively high temperatures.
The two pioneer species maintained high rates of gaseous exchange at 1000 hours while for
all other species there was little difference between them except in the case of Drypetes sp.
which displayed relatively low rates of net photosynthesis, a feature maintained at the 1200
hours sampling. Net photosynthesis and stomatal conductance rates for Shorea leprosula
were high relative to the other non-pioneer plants at the midday sampling. At 1400 hours,
with a substantial drop in PPFD, there was a reduction in rates of a gas exchange for the
pioneers but little change for the other species. All species showed a marked reduction in
net photosynthesis and stomatal conductance with substantially reduced PPFD at 1600 hours.
High RH precluded further, late afternoon readings.
Figure 2 shows the diurnal patterns of net photosynthesis and stomatal conductance for
Shoreafalciferoides both in the gap (Figure 2a) and the understorey (Figure 2b) as measured
on 29 July, 1991, during a dry period. In the morning the PPFD and, as a result, the rates
of net photosynthesis in the understorey were extremely low compared to those of the gap.
Stomatal conductance values, however, did not differ between the two environments to the
same extent; in the understorey low PPFD values were associated with relatively high
Table I. Properties of gap (G) and understorey (U) seedling leaves. Mean ± SO; n = 5-10. w
~
Shorea S. leprosula Parashorea Dryobalanops Aporosa lunata Drypetes sp. Macaranga Musa sp.
jalciferoides macrophylla lanceolata trachyphylla
G U G U G U G U G U G U G G
Total leaf thickness I'm 133±12 98±3 108±11 83±9 124±4 102±7 243±15 179±24 255±23 206±24 261±25 208±11 82±13 241±23
Adaxial epidermis + cuticle I'm 17±2 15±2 20±2 19±2 16±1 14±1 26±1 25±2 27±2 26±2 23±1 19±2 II± I 53±9
Palisade mesophyll I'm 46±6 29±3 42±6 33±4 49±2 36±4 48±9 33±5 87±11 77±6 6H21 57±8 35±9 50±9
Spongy mesophyU I'm 54±5 42±5 38±8 20±5 4H3 41±7 153± 10 IOH28 129± 12 91±20 155±40 118± 16 27±6 122± 18
Abaxial epidermis + cuticle I'm II ± I 1O±1 II ± I 1O±0 II±I 9±1 IH2 14±2 II±I 1O±2 15±2 13±1 8±1 21±3
Abaxial stomatal density mm,2 441 ±44 384±39 404±50 265±34 525±S8 390±S2 263± 10 189±32 438±38 295±28 181 ±29 130±24 174±23 III± 10
~
Specific leaf area cm2 g'l. _'_ 140±19_12.7±27 164±13 215±18 125± 8 196±16 119±1O 164±18 103±13 150±20 109± 8 183+23 109+8 290+ IS tTl
t:!:l
t:!:l
tTl
Table 2. Gas exchange characteristics for gap (G) and understorey (U) seedling leaves. Mean ± SO. :;0
t:!:l
0
Shorea S. leprosula Parashorea Dryobalanops Aporosa lunata Drypetes sp. Macaranga Musa sp. 0...,
jaiciferoides macrophylla lanceolata trachyphyila ::t
Net photosynthetic rate G 7.82±2.08 7.73±2.21 S.68± 1.20 6.22±1.32 7.29± 1.64 S.02± 1.41 12.84± 1.86 11.82± 1.66
I' mol CO2 m,2 S'I n = 21 n = 13 n=8 n=8 n = 8 n = 10 n = 12 n = 14
PPFD > I,OOOJLmol m,2 S'I
G 0.029±.00S 0.026±.006 0,031 ±.004 0.026±.003 O.Q3S±.008 0.024±.OO4 0.040±.OO8 0.038±.011
Apparent quantum yield
n = 21 n = 20 n=S n = 11 n=6 n=8 n = 10 n=8
mol CO2 mol photons'l
PPFD 20-100JLmol m,2 S'I U 0.038±.013 0.040±.018 0.044±.016 0.039±.013 0.036±.00S 0.044±.019 NA NA
n = 18 n = 24 n = 13 n = 17 n = 13 n = 12
Light compensation G 11.9±1.2 18.6± 1.1 17.4± 1.3 16.7±1.6 IS. 1 ±0.2 IS.H 1.6 16.9±2.0 18.1 ±0.6
JLmol photons m'2 S'I
U 8.H 1.5 9.H2.1 13.1±1.6 9.3± 1.6 8.6±0.8 8.0± I.S NA NA
n=3
Dark respiration G -0.93±.18 -0.99± .32 -1.26± .47 -.63±.12 -0.98± .33 -0.69±.15 -1.35±.32 -0.93±.26
JLmol CO2 m,2 S'I
U -0.23±.12 -0.40±.20 -0.33 ± .18 -0.21±.11 -0.33± .07 -0.25±.08 NA NA
n = S
0.8
0.6
'"e 0••
11
~
~
""
0.2 " '-
9 10 12 14 16
Time, hours
Figure la. Diurnal patterns of stomatal conductance for seedlings of the gap on 24 August 1991. Points mean
values ± SO, n = 3. Musa sp. (closed stars), Macaranga trachyphy/la (open stars), Shorea falciferoides (closed
squares), Shorea leprosula (open squares), Parashorea macrophylla (closed circles), Aporosa lunata (open circles),
Dryobalanops lanceolala (closed triangles), Drypetes sp. (open triangles).
14
12
10
H -- --- -----
--------,
8
~
.,.'ss 6
~
0:
'" 4
0
9 10 12 14 16
Time, hours
PPFO 100-1000 ,1000 ,1000 "'SOD ",60
RH 13-95 10 65 10 90
'c 29 33 31 33 28
Figure lb. Diurnal patterns of net photosynthesis for seedlings of the gap on 24 August 1991. Symbols as above.
Points mean values ± SO, n = 3.
372 WEBBER BOOTH
1_
8
6
1---+-- 0.20
0.15
~ 1_
e
'e
~
0.10 c;
-
:z:
=< 4
2
e
~
.
0.05
0
8 10 12 14 16
Time, hours
PPFD 63 116 589 1220 1382 244 138 \18 41 49
RH 98 91 93 84 85 85 90 91 96 98
t 25 26 29 35 35 33 30 29 29 29
Figure 2a. Diurnal pattern of net photosynthesis (solid line) and stomatal conductance (broken line) for Shorea
falciferoides seedlings in the gap on 29 July 1991. Points mean values ± SO, n = 3.
,t',
8 0.20
t
~e
1_
~
..:
:z:
=<
6
4
(' "'" t /~\
, ,/
"
'
\
1
,'-
0.15
0.10
~e
c;
e
.:
2 0.05
0
8
t--.
10
+
12
• 14
-+
16
Time, hours
PPFD 26t20 12tl 15t3 15t4 15tl 24tl1 412
RH 91 90 83 11 90 93 91
't 21 29 31 30 H 29 29
Figure 2b. Diurnal pattern of net photosynthesis (solid line) and stomatal conductance (broken line) for Shorea
falciferoides seedlings in the understorey on 29 July 1991. Points mean values ± SO, n = 3.
stomatal conductances, similar in magnitude to those of the gap. At midday, in the gap,
there was a dramatic reduction in stomatal conductance and subsequently a drop in net
photosynthesis, a sequence of events caused by water stress, a feature that may also be
observed, albeit more weakly, in the understorey. A slight recovery was evident in the
mid-afternoon.
Table 2, line 1, based on a large number of measurements of light saturated net
photosynthesis for all gap species, indicates that the pioneers have higher rates than the
non-pioneers. Although not differing significantly from each other, mean values for the
non-pioneers suggest slightly higher rates may be associated with Shorea leprosula and
Shorea falciferoides and a lower rate with Drypetes sp., a pattern also evident in the diurnal
curve of Figure 1.
In an attempt to determine if there were differences in photosynthetic efficiency between
gap and understorey seedlings apparent quantum yields were calculated for plants in both
situations Table 2 line 2. It was assumed that photosynthetic rate plotted against PPFD
values between 20 and 100 j.tmol m- 2 S-1 would fall within the initial slope of the P vs. I curve
for all species in each situation. The results, although not statistically significantly different,
suggest that for each species the apparent quantum yields for the understorey plants are
greater than those for the gap plants.
Table 2, line 3, presents light compensation point PPFD values for all species in the gap
and understorey. These values are based on three measurements where there was little
or no change in the CO2 level within the chamber of the photosynthetic system over a period
of three minutes. For each species the gap value was significantly higher than that associated
with the understorey seedlings.
A comparison of rates of dark respiration between gap and understorey seedling leaves
(Table 2, line 4) indicate that, for each species, understorey plants have substantially lower
rates of respiration than gap plants.
DISCUSSION
The results of the comparison of the morphological features of gap and understorey seedling
leaves suggest typical sun/shade associated patterns as described by Boardman (1977). The
leaves of the plants grown in the understorey are thinner, have a higher SLA and a lower
stomatal density than those of the gap seedlings. Oberbauer & Strain (1985), in their
investigations into the effects of different light regimes on seedling growth reported a similar
set of associated features. Langenheim et al. (1984) showed differences in SLA (converted
from SLW) close to 50% higher for understorey seedlings compared to those of the gap for
all except one of the Australian and Amazonian rainforest tree species studied. Pearcy
(1987) showed the same trend with a 19% difference in SLA between understorey and gap
plants for Argyrodendron peralatum.
Stomatal conductance and net photosynthesis for the range of seedlings in the gap indicate
that the pioneer species have higher rates for both of these parameters than any of the canopy
or emergent species, a feature also reported by Oberbauer & Strain (1984) from their
comparative studies in the Costa Rica rainforest. Musa sp. and Macaranga trachyphylla net
photosynthetic rates are close to the 12.6-15.8 j.tmol m- 2 S-1 given for the typical pioneer
trees, Mallotus, Macaranga and Glochidion by Bazzaz & Pickett (1980), but are considerably
lower than the 27.7 j.tmol m- 2 S-1 reported for the pioneer Ochroma lagopus of Costa Rica by
Oberbauer & Strain (1984). The light saturated net photosynthetic rates of between 5 and
10 j.tmol m- 2 S-1 for the seedlings of canopy and emergent tree species in this study are of
similar magnitude to those reported by Kwesiga et al. (1986), Langenheim et al. (1984) and
Pearcy (1987) and match the shade type upper canopy category of Koyama (1981) based on
his studies of lowland rainforest trees in Peninsular Malaysia. They range from values
that Bazzaz & Pickett (1980), in their review, associate with early successional trees (8.9
j.tmol m- 2 S-I) to close to those of canopy trees (4.4 j.tmol m- 2 S-I).
The diurnal patterns of stomatal conductance and net photosynthesis for the seedlings of
the gap (Figure 1) are based only on a few points spread over a limited part of the day and
therefore do not include such extensive detailed information as presented in Roberts et at.
(1990); they do, however, give a broad outline of responses which are similar in pattern and
magnitude to those presented by Pearcy (1987) and discussed by Tenhunen et al. (1987).
374 WEBBER BOOTH
The stomatal conductances for the pioneers, although not unique for tropical trees (see
Roberts et at. 1990), are high (they exceeded 0.5 mol m· 2 S-1 in the early part of the day) and
such rates may in part be associated with temperature control for these large leaved species.
The diurnal patterns of gas exchange for Shorea falciferoides seedlings in the gap and
understorey (Figures 2a & b) indicate that the midmorning stomatal conductance rates do not
differ greatly between the two environments in spite of big differences in PPFD and
associated net photosynthesis. This result contrasts with those obtained by Pearcy (1987) and
Roberts et al. (1990) who recorded substantially lower stomatal conductance values for plants
of the same species at low PPFD. In fact, in this present study, all species surveyed
displayed similar patterns of relatively higher stomatal conductance/PPFD ratios at lower
irradiances compared to high, both for plants in the gap and in the understorey. The midday
stomatal closure and resultant depression of stomatal conductance and net photosynthesis for
the gap seedlings in Figure 2a illustrate dramatically the effects of water stress that can occur
in wet tropical rain forest during dry periods. Schulze et at. (1987) discuss direct stomatal
response to soil drought, the probable sequence of events involved in this present situation.
The photosynthetic response as a function of PPFD for leaves of understorey seedlings
showed lower compensation points and higher apparent quantum yields (based on mean
values only) than those of the gap, patterns consistent with maximizing use of the limited
light resources of the understorey. The apparent quantum yield differences were, however,
not statistically significant, and in this respect there is agreement with the observations of
Ramos & Grace (1990) and Sims & Pearcy (1989) that there is surprising insensitivity to the
light regime. Apparent quantum yield values obtained were of similar magnitude to those
of Oberbauer & Strain (1985) and Ramos & Grace (1990) but low when compared with some
of those reported by Kwesiga et al. (1986), Langenheim et al. (1984) and Pearcy (1987).
The light compensation points for understorey seedlings were always lower than those
for gap seedlings, an adaptive feature associated with growth in a low light environment
(Bjorkman 1981). The gap associated values which ranged from 12 to 18 ",mol m-2 S-1 and
understorey values of 8 to 13 ",mol m-2 S-1 are similar to those reported for many tropical
forest seedlings grown under such contrasting conditions of PPFD (K wesiga et at. 1986,
Langenheim et al. 1984, Oberbauer & Strain 1985, Ramos & Grace 1990).
Gap seedlings displayed relatively high rates of dark respiration which, as suggested by
Ramos & Grace (1990), is probably a consequence of their enhanced growth rates. These
respiration rates which ranged from -0.63 to -1.35 ",mol CO2 m-2 S-1 are of similar magnitude
to those reported by Kwesiga et al. (1986) and Ramos & Grace (1990) but higher than the
-0.14 to -0.48 ",mol CO2 m-2 S-1 values of Langenheim et al. (1984), Oberbauer & Strain
(1985) and Pearcy (1987). The low levels of under storey dark respiration which ranged from
-0.21 to -0.33 ",mol CO2 m-2 S-1 correspond closely to most published records for tropical
rainforest seedlings.
In conclusion, the sets of morphological and physiological data obtained for seedlings
grown in the contrasting gap and understorey environments conform to those expected of
plants from respectively high and low light regimes. The understorey plant strategy is one
of maintenance at very low PPFD, with maximized light harvest and minimized energy
expenditure, while that of the gap plant is rapid growth which is made possible by relatively
high rates of photosynthesis and, inevitably, high respiration. Pioneer plants have an
enhanced capacity to use the high PPFD of the gap.
ACKNOWLEDGEMENTS
This study was supported by a research grant from Universiti Brunei Darussalam. I would
like to thank the members of staff of KBFSC and the Biology Department for their support
during this study.
LITERATURE CITED
BAZZAZ, F.A. & PICKETT, S.T.A. 1980. Physiological ecology of tropical succession: a comparative review.
Annual Review of Ecology and Systematics II:287-31O.
BJORKMAN, O. 1981. Responses to different quantum flux densities. Pp. 57-108 in Lange, O.L., Nobel, P.S.,
Osmond, C.B. & Ziegler, H. (eds). Encyclopedia of plant physiology, new series. Physiological plant ecology
I, Volume 12A. Springer Verlag, Berlin. 625 pp.
BOARDMAN, N.K. 1977. Comparative photosynthesis of sun and shade plants. Annual Review of Plant
Physiology 28:355-377.
CANHAM, C.D. 1989. Different responses to gaps among shade-tolerant tree species. Ecology 70:548-550.
CHAZDON, R.L. & FETCHER, N. 1984. Photosynthetic light environments in a lowland tropical rain forest in
Costa Rica. Journal of Ecology 72:553-564.
KOY AMA, H. 1981. Photosynthetic rates in lowland rain forest trees of Peninsular Malaysia. Japanese Journal
of Ecology 31:361-369.
KWESIGA, F.R., GRACE, J., & SANDFORD, A.P. 1986. Some photosynthetic characteristics of tropical timber
trees as affected by the light regime during growth. Annals of Botany 58:23-32.
LANGENHEIM, J.H., OSMOND, C.B., BROOKS, A. & FERRAR, PJ. 1984. Photosynthetic responses to light
in seedlings of selected Amazonian and Australian rainforest tree species. Oecologia (Berlin) 63:215-224.
LOACH, K. 1970. Shade tolerance in tree seedlings II: Growth analysis of plants raised under artificial shade. New
Phytologist 69:273-286.
MARTINEZ-RAMOS, M., SARUKHAN, J. & PINERO, D. 1988. The demography of tropical trees in the context
of forest gap dynamics: the case of Astrocaryum mexicanum at Los Tuxtlas tropical rain forest. Pp. 293-313 in
Davy, D.J., Hutchings, M.J. & Watkinson, A.R. (eds). Plant population ecology. Blackwell, Oxford.
OBERBAUER, S.F. & STRAIN, B.R. 1984. Photosynthesis and successional status of Costa Rican rain forest
trees. Photosynthetic Research 5:227-232.
OBERBAUER, S.F. & STRAIN, B.R. 1985. Effects of light regime on the growth and physiology ofPentadethra
macroloba (Mimosaceae) in Costa Rica. Journal of Tropical Ecology 1:303-320.
OSUNKOYA, 0.0 & ASH, J.E. 1991. Acclimation to a change in light regime in seedlings of six Australian
rainforest tree species. Australian Journal of Botany 39:591-605.
PEARCY, R. W. 1987. Photosynthetic gas exchange responses of Australian tropical forest trees in canopy, gap and
understorey micro-environments. Functional Ecology I: 169-178.
POPMA, 1. & BONGERS, F. 1991. Acclimation of seedlings of three Mexican tropical rain forest tree species to
a change in light availability. Journal of Tropical Ecology 7:85-97.
RAMOS, J. & GRACE, 1. 1990. The effects of shade on the gas exchange of seedlings of four tropical trees from
Mexico. Functional Ecology 4:667-677.
RIDDOCH, I., LEHTO, T. & GRACE, 1. 1991. Photosynthesis of tropical tree seedlings in relation to light and
nutrient supply. New Phytologist 119: 137 -147.
ROBERTS, 1., CABRAL, O.M.R. & DE AGUIAR, L.F. 1990. Stomatal and boundary-layers conductances in an
Amazonian terra firme rain forest. Journal of Applied EcolOgy 27:336-353.
ROSS, S.M, & DYKES, A. (this volume). Soil conditions, erosion and nutrient loss on steep slopes under mixed
dipterocarp forest in Brunei.
SCHULZE, E.D., TURNER, N.C., GOLLAN, T. & SHACKEL, K.A. 1987. Stomatal responses to air humidity
and to soil drought. Pp. 311-321 in Zeiger, E., Farquhar, G.D. & Cowan, I.R. (eds). Stomatal Function.
Stanford University Press.
TENHUNEN, J.D., PEARCY, R.W. & LANGE, O.L. 1987. Diurnal variations in leaf conductance and gas
exchange in natural environments. Pp. 323-351 in Zeiger, E., Farquhar, G.D. & Cowan, I.R. (eds). Stomatal
Function. Stanford University Press.
WHITMORE, T.C. 1975. Tropical Rain Forests of the Far East. Clarendon Press, Oxford, 352 pp.
WHITMORE, T.C. 1978. Gaps in the forest canopy. Pp. 639-655 in Tomlinson, P.B. & Zimmermann, M.H. (eds).
Tropical Trees as Living Systems. Cambridge University Press, Cambridge.
WHITMORE, T.C. 1989. Canopy gaps and the two major groups of forest trees. Ecology 70:536-538.
Distributional ecology of freshwater fishes in tropical rainforest streams
of Borneo
SATISH C. CHOyl
Biology Department, Universiti Brunei Darussalam, Bandar Seri Begawan 2028, Brunei
Darussalam
ABSTRACT. A survey was conducted to study the distribution of freshwater fishes in the middle
and headwater streams of the Belalong-Temburong watershed, Brunei Darussalam, northern
Borneo. The watershed is covered by a mixed dipterocarp primary rainforest. The number of
species found at anyone locality ranged from zero to twenty five; the number of localities
occupied by anyone species ranged from one to twenty six. Distribution and abundance of fish
were not related to water quality, which was good at all the sampled sites. Drainage area,
elevation, gradient and stream order were all significantly correlated with the number of species
collected at a site. Species richness was highest in large watersheds, at low elevation, low
gradient and in streams of high order. In most streams, species addition occurred with increasing
drainage area, increasing stream order, decreasing altitude and decreasing gradient, but species
replacement was frequent as well, particularly when waterfalls were encountered. Species
belonging to the family Balitoridae dominated channels above waterfalls, whereas those belonging
to the family Cyprinidae were predominant in streams below waterfalls. In deep channels,
different fish species occupied different zones of the vertical stratum. Shallowing of channels
corresponded to the compression of this stratum, resulting in lower species richness.
KEY WORDS: species richness, stream order, altitude, gradient, drainage area, Brunei
Darussalam .
INTRODUCTION
Biogeographic distributional patterns of freshwater fishes of Southeast Asia have been

discussed by Burridge (1992), Cranbrook (1981), Inger & Chin (1962), Kottelat (1989),
Rainboth (1991) and Zakaria-Ismail (1991). There seems to be general agreement that the
fish fauna within Borneo is divided into several distinct biogeographic zones (Inger & Chin
1962, Rainboth 1991) but the demarcations are not clear.
On a smaller scale, some quantitative information on the distribution and abundance of
the freshwater fish fauna within watersheds of Borneo have been provided by Inger (1955),
Inger & Chin (1962), Lee et af. (1990), Samat (1990a, b), Taylor (1989) and Watson &
Balon (1984a, b). However, only Samat (1990a) and Watson & Balon (1984b) quantitatively
analysed for correlates between physico-chemical characteristics of streams and species
diversity.
This paper is based on some quantitative analyses between fish species richness and
physical attributes such as drainage area, stream order, elevation and gradient within the
Belalong-Temburong watershed. The objective of the study was to see if any significant
correlation could be found and, if so, how it may be used to estimate or predict species
richness in other watersheds of the region.
lpresent Address: DPI Water Resources Technical Centre, 1345 Ipswich Rd., Rocklea, Qld. 4106, Australia
377
378 SATISH C. CHOY

Sampling
Fish were caught from the two main rivers, Sungai Belalong and Sg. Temburong as well
as from most of their tributaries (Figure 1). The watershed is covered by a mixed
dipterocarp primary rainforest. A list of the sampling stations and their physiographic
features are given in Table 1. Fishing techniques utilized were gill netting, cast netting,
seining, and hand netting. For the present study, fishing effort was standardised to 1 hour
per sampling station. Fish collected from the different micro-habitats were enumerated and
recorded. They were later fixed in 10% formalin and then preserved in 70% ethanol. The
specimens are deposited in the UBD/RGS Reference Collection at the Department of Biology,
Universiti Brunei Darussalam and in the Reference Collection of the Fisheries Department
of Sabah, Likas, Kota Kinabalu, Sabah under catalogue numbers F/1-F/146. Water samples
collected from different localities were analysed by the Chemistry Department of Universiti
Brunei Darussalam and the Soils Laboratory of Kilanas Agricultural Station.
Analysis
Stream order, altitude, gradient and drainage areas were determined from 1:50,000
topographic maps. For stream order, the Horton-Strahler stream classification was used. First
order streams are headwater streams with no tributaries. Stream order increases by one each
time two streams of the same order flow together. Gradient at a sampling locality was
determined by dividing the altitudinal difference between points 0.5 km downstream and 0.5
km upstream of the sampling site and then converted to a percentage. Jaccard's coefficient
of similarity (Southwood 1978) was used to access species similarity in the different micro-
habitats. Statistical procedures follow Schefler (1980) and Daniel (1990).
,
t
N I
,I
I
I
\
\
"
\
\
'- - - - /'}-.-- .. o 2
km
Figure 1. Locality map of part of the Belalong-Temburong watershed showing the UBD-RGS Rainforest Project
boundary (----) and sampling stations 1-39, as listed in Table 1. Station 40 (not shown) is about 10 km southeast of
Bukit (Mt.) Bela10ng ( .. ).
Distributional ecology of freshwater fishes 379
Table I. Sampling station numbers, localities, stream order, altitude, gradient and the number of fish species for
each locality. Stream order, altitude, gradient and drainage areas were determined from I : 50,000 topographic
maps. a.w., above waterfall areas; b.w., below waterfall areas; KBFSC, Kuala Belalong Field Studies Centre; S,
sungai (river).
Station Locality Stream Altitude Gradient Number of

Order (m) (%) fish species
I. S. Anak Esu, b.w. 60 41 5
2. S. Anak Esu, a.w. 152 21 0
3. S. Esu, b.w. 60 15.5 6
4. S. Esu, a.w. 183 12.4 2
5. S. Engkabang, b.w. 60 7.8 7
6. S. Engkabang, a.w. 152 15.5 2
7. Lower S. Sitam, b.w. 3 67 4.3 8
8. Tributary of S. Sitam, a.w. 1 183 30.5 0
9. Middle S. Sitam, b.w. 3 92 3 9
10. Upper S. Sitam, below Pondok Busiri, b.w. 2 131 3.3 9
11. Bathing stream, below Pondok Busiri. a.w. 1 335 30.5 0
12. Lower S. Engkiang, b.w. 3 76 2.5 8
13. Middle S. Engkiang, b.w. 3 107 2.7 8
14. Upper S. Engkiang, b.w. 2 128 3.8 9
15. Upper right trib. S. Engkiang, a.w. 1 134 5 3
16. S. Split, b.w. 2 152 6 7
17. Bathing stream, near KBFSC, b.w. 2 60 10 6
18. Bathing stream, near KBFSC, a.w. 1 122 20 3
19. S. Belalong, near KBFSC, b.w. 4 55 0.7 25
20. S. Belalong, btw. S. Engkabang & S. Sitam, b.w. 4 67 0.7 21
21. S. Belalong, near S. Engkiang, b.w. 4 70 0.5 13
22. S. Belalong, btw. S. Engkiang & S. Split, b.w. 3 100 0.8 10
23. S. Belalong, near S. Split, b.w. 3 137 0.6 6
24. Confluence of S. Belalong & S. Temburong, b.w. 5 49 0.7 13
25. S. Temburong, near S. Babi, b.w. 4 61 0.9 24
26. S. Temburong, near S. T-Machang, b.w. 4 79 0.6 12
27. Lower S. Babi, b.w. 3 92 5 15
28. Trib. Lower S. Babi, a.w. 1 152 12.4 0
29. Upper S. Babi, b.w. 3 214 6 10
30. Upper S. Babi, above runs, b.w. 2 275 10 11
31. Trib. Upper S. Babi, a.w. 2 335 15 3
32. Lower S. T-Machang, b.w. 3 100 1.5 19
33. Trib. S. T-Machang, below Bt. Belalong, a.w. 610 15.5 1
34. Trib. Upper S. T-Machang, a.w. 579 15 1
35. Trib. Upper S. T-Machang, a.w. 457 10 3
36. S. inging, b.w. 3 70 1.9 11
37. Trib. S. inging, a.w. 1 140 31 0
38. S. Apan, b.w. 2 61 10 12
39. S. Apan, a.w. 137 15 4
40. Stream below Bukit Pa~on 1400 4
380 SATISH C. CHOY
RESULTS
Forty four species of fish, belonging to ten families and thirty genera, from a total of forty
sites were collected. The numbers of species collected at each sample site are shown in Table
1 and a checklist of species with the sites of their occurrence are given in Table 2. Species
belonging to the families Balitoridae and Cyprinidae dominated the catch, with fifteen and
thirteen species, respectively. Most members of the latter family were found in the lower,
slow-flowing channels of streams while those of the former family were mainly found
Table 2. Species checklist and sites of occurrence.
Family Anguillidae
Anguilla marmorata Quoy & Gaimard, 1924 - Stations 19,20,25,27,30 & 32.
Family Bagridae
Mystus baramensis (Regan, 1906) - Stations 5, 19, & 25.
Mystus sp. - Stations 24 & 25.
Leiocassis micropogon (Bleeker, 1852) - Stations 19 & 24.
Family Balitoridae
Gastromyzon bomeensis (Gunther, 1874) - Stations 15, 19-21,25,26,29 & 30.
Gastromyzon!asciatus Inger & Chin, 1961 - Stations 19 & 35.
Gastromyzon monticola (Vaillant, 1902) - Stations 7, 10, 14, 19, 20, 25, 27 & 32.
Gastromyzon lepidogaster (Roberts, 1982) - Stations 3, 4, 9, 10, 15, 18-20, 25, 29, 30, 32, 35 & 38
Gastromyzon punctulatus Inger & Chin, 1962 - Stations 6, 19, 25 & 27.
Gastromyzon ridens (Roberts, 1982) - Stations 19,24 & 25.
Glaniopsis gossei - Station 40.
Glaniopsis hanitschi Boulenger, 1899 - Stations 31, 33 & 34.
Nemacheilus sp. - Station 31
Neogastromyzon nieuwenhuisi Popta, 1905 - Stations 3-6,9,10,12-15,17-27,35,38 & 39.
Neogastromyzon sp. - Station 40.
Protomyzon whiteheadi (Vaillant, 1893) - Stations 16, 18-20, 24-26, 29-32 & 39.
Protomyzon griswoldi (Hora & Jayaram, 1951) - Station 32.
Protomyzon sp. - Station 39.
Sundareonectes sabanus Chin, 1990 - Station 40
Family Clariidae
Clarias leiacanthas (Bleeker, 1851) - Stations 19,25 & 32.
Clarias teijsmanni Bleeker, 1857 - Station 19.
Family Cobitidae
Pangio mariarum (Inger & Chin, 1962) - Stations 19-21, 27, 32 & 36.
Family Cyprinidae
Barbodes collingwoodi (Gunther, 1868) - Stations 19-22,27, 29, 30, 32 & 36.
Chela sp. - Station 23
Hampala bimaculata (Popta, 1905) - Stations 7,9, 10, 12-14, 16, 19-27,29,30,32,36 & 38
Lobocheilus bo (popta, 1904) - Stations 7,9, 10, 12-14, 16, 19-22,25-27,32 & 38.
Nematabramis steiluIachneri (Popta, 1905) - Stations 1,3,5,7,9-10, 12-14, 16-17,
19-27,29,30,32,36 & 38.
Osteochilus spilurus (Bleeker, 1851) - Stations 19, 24 & 36.
Paracrossochilus acerus Inger & Chin, 1962 - Stations 5, 10, 12-14, 19-22, 25-27, 29, 30, 32, 36 & 38
Paracrossochilus vittatus (Boulenger, 1894) - Station 36.
Puntius binotatus (Valenciennes, 1842) - Stations 19 & 24
Rasbora argyrotaenia (Bleeker, 1850) - Stations 1,3,5,7,9-10, 12-14, 16-17, 19-27,29,30,32,36 & 38.
Rasbora sumatranfl (Bleeker, 1852) - Stations 1,3,5,7,9, 12-14, 16, 19-26,29,30,32 & 36.
Rasbora sp. - Stations 19-21, 25, 29 & 38.
Tor douronensis (Valenciennes, 1842) - Stations 1,3,5,7,9, 10, 12, 13, 16, 19-27,29,32 & 36.
Family Gobiidae
Parawaous sp. - Stations 1, 7, 9, 19,20,24,25,27,32 & 38.
Pseudogobiopsis cf. oligactus (Bleeker, 1875) - Stations 1, 19 & 38.
Family Mastacembelidae
Macrognathus maculatus (Cuvier, 1831) - Stations 19-27,32,36-38.
Mastacembelus unicolor Cuvier, 1831 - Stations 24 & 36.
Family Siluridae
Silurichthys hasselti Bleeker, 1858 - Stations 17, 19 & 25.
Ompak sp. - Station 19.
Kryptopterus sp. - Station 32.
Silurus sp. - Station 32.
Family Sisorodae
Glyptothorax major (Boulenger, 1894) - Stations 17, 19,24,25 & 27.
in the upper, faster flowing channels. No Cyprinidae was found above any waterfall greater
than Sm high.
The number of fish species collected at a site ranged from 0 to 25, with the main channels
of Sungai Belalong and Sg. Temburong having the highest species richness and diversity
(Table 1). The number of localities where a species was found ranged from 1 to 26 (Table
3).
Table 3. Number of occurrences of fish species in different stream orders, total number of individuals per species
(N) and its percentage frequency (%).
Stream Order 2 3 4 5 Total N %

Number of stations 17 7 10 5 40
Species
Anguilla momwrata 0 2 2 3 0 7 8 0.8
Mystus baramensis I 0 0 2 0 3 6 0.6
Mystus sp. 0 0 0 0 1 1 0.1
Leiocassis micropogon 0 0 0 1 2 18 1.8
Glarias leiacanthus 0 0 1 1 3 3 0.3
Glarias teijsmonni 0 0 0 0 1 1 0.1
Pangia moriarum 0 0 3 3 0 6 8 0.8
Sundareonectes sabanus 1 0 0 0 0 1 2 0.2
Nemocheilus sp. 0 0 0 0 4 0.4
Chela sp. 0 I 0 0 0 1 0.1
Hampala bimoculata 0 5 10 5 I 21 65 6.6
Lobocheilus bo 0 4 7 5 0 16 58 5.9
Nemotabramis steindachneri 5 6 9 5 26 153 15.5
Osteochilus spilurus 0 0 I 3 7 0.7
Paracrossochilus acerus 1 4 7 5 0 17 102 10.4
Paracrossochilus vittatus 0 0 1 0 0 1 4 0.4
Puntius binotatus 0 0 0 1 2 5 0.5
Bardodes collingwoodi 0 I 5 3 0 9 71 7.2
Rasbora argyrotaenia 3 6 10 5 25 148 15.0
Rasbora sumotrana 3 3 9 5 1 21 32 3.2
Rasbora sp. 0 I 0 4 0 5 8 0.8
Tor douronensis 3 2 10 5 21 51 5.1
Parawaous sp. 4 3 10 15 1.5
Pseudogobiopsis sp. 1 0 0 3 3 0.3
Gastromyzon bomeensis 1 1 5 0 8 33 3.3
Gastromyzon /asciatus 1 0 0 0 2 3 0.3
Gastromyzon monticola 0 2 3 3 0 8 22 2.2
Gastromyzon lepidogaster 5 3 3 3 0 14 18 1.8
Gastromyzon punctulatus 0 1 2 0 4 7 0.7
Gastromyzon ridens 0 0 0 2 0 2 3 0.3
Glaniopsis hanitschi 3 0 0 0 4 7 0.7
Glaniopsis gossei 1 0 0 0 0 0.1
Neogastromyzon nieuwenhuisi 8 4 6 5 1 26 54 5.4
Neogastromyzon sp. I 0 0 0 0 1 2 0.2
Protomyzon whiteheadi 2 3 2 4 1 12 16 1.6
Protomyzon griswoldi 0 0 1 0 0 1 0.1
Protomyzon sp. 1 0 0 0 0 2 0.2
Macrognathus moculatus 1 1 5 5 1 13 18 1.8
Mastacembelus unicolor 0 0 0 1 2 4 0.4
Krytopterus sp. 0 0 0 0 0.1
Ompak sp. 0 0 I 0 0 1 0.1
Silurichthys hasselti 0 1 0 2 0 3 7 0.7
Silurus sp. 0 0 1 0 0 1 I 0.1
Glyptothorax mojor 0 1 I 2 5 9 0.9
Total no. species in each stream order 20 22 27 30 15 44
Max. no. species in each stream order 7 12 19 25 13
382 SATISH C. CHOY
The most widespread species, collected from over twenty sites, were Nematabramis
steindachneri, Rasbora argyrotaenia, R. sumatrana, Tor douronensis, Neogastromyzon
nieuwenhiusi and Hampala bimaculata. Of these, only N. nieuwenhuisi is a rheophilic
species. Overall, Nematabramis steindachneri, Rasbora argyrotaenia and Paracrossochilus
acerus were most numerous, with over a hundred specimens of each being caught. Eight
species (Clarias teijsmanni, Glaniopsis gossei, Protomyzon griswoldi, Chela sp., Mystus sp.,
Krytopterus sp., Ompak sp. and Silurus sp.) seem rare (Table 3).
Water quality of the streams was good and unpolluted, with low nutrient and high
dissolved oxygen content (Table 4). Smaller streams generally had lower conductivity and
total suspended solids (TSS). Following rain, anion, cation and total suspended solids
increased markedly. There was no correlation between species richness or diversity and
water quality (P > 0.10).
Only seven species were collected at any site above 400 m. They were Gastromyzon
fasciatus, G. lepidogaster, Glaniopsis hanitschi, G. gossei, Neogastromyzon nieuwenhiusi,
Neogastromyzon sp. and Sundareonectes sabanus, all belonging to the torrent loach family
Balitoridae. With a single exception, no species was collected at a site with a gradient
exceeding 20 %. The exception (Station I, gradient 41 %) is an artifact of the method of
calculating gradient: all species were collected in an area just before a series of steep slopes
with waterfalls; the slopes were included in the calculation of the gradient, according to the
standardised methodology. No fish was found beyond these slopes, even though the gradient
eventually decreased.
The mean number of fish species per site increased from 2.4 (+ 0.53 S.E.) for first order
streams through 8.1 (+ 1.08 S.E.) for second order streams, 10.4 (+ 1.17 S.E.) for third
order streams and 19.0 (+ 2.46 S.E.) for fourth order streams, but decreased to 13 (+ 0
S.E.) for the fifth order stream.
Table 4. Physico-chemical characteristics of some of the streams in the study area. For stations 19 & 25 the upper
limits were obtained following rain. S, sungai (river).
S. Sitam Trib. S. Sitam S. Belalong S. Temburong

(Stn 7) (above Stn 9) (Stn 19) (Stn 25)
Width (m) 5 0.4 15 20
Depth (m) 0.3 0.04 1.5 1.1
Current (ms'!) 0.75 0.17 0.65-3.7 2.0
Temperature ("C) 23-25 24-25 24-25 24-25
Conductivity (jLScm'!) 36 15 24-35 26
pH 6.8 6.8 5.9-6.9 5.4
Diss. Oxygen (mgl'!) 8 7.6 7.6 7.7
CO2 (ppm) 3 3 3 3
Total Hardness (mg!"'CaCO,) 14 14 8 8
Ca Hardness (mgl") 4 4 4 4
NO,-N (mg!"!) <0.2 0.83 0.12-0.42 0.07-0.21
PO,-P (mgl") <0.2 0.04 0.04 0.04
Calcium (mgl'!) 0.49 0.78-1.50 0.27-0.31
Magnesium (mgl") 0.79 0.91-1.30 0.02-0.88
Sodium (mg!"') 0.99 1.32-1.90 1.49
Potassium (mgl'!) 0.21 0.20-0.72 0.15-0.22
Chloride (mgl") 0.23 0.23-0.85 0.15-0.31
SO, (mgl'!) 1.85 3.30-5.54 4.31
TSS (mgl'!) 7.1 10-250 10-320
Table 5. Regression equations for log transformed data.
Log (no. species)= 5.137 - 0.672 log (elevation), n=35, r=0.65

Log (no. species)= 2.580 - 0.464 log (gradient), n=35, r=0.70
Log (no. species) = 1.067 + 1.199 log (stream order), n=26, r=0.82
Log (no. species)= 1.990 + 0.297 log (drainage area),n= 12, r=0.95
Table 6. a) Regression equations and correlation, r for log transformed data. All are significant at P<O.OI (2-tailed).
b) Kendall's rank correlation coefficient (Kendall's tau) between number of species and physical attributes of sites.
All Kendall's tau are significant at P<O.OI (2-tailed). Highest number of species, largest drainage area, lowest
stream order, lowest elevation and lowest gradient were ranked as I. c) Partial Kendall's tau with elevation effect
controlled (Daniel 1990). All are significant at P<O.OI (2-tailed).
Elevation Gradient Drainage area Stream order
a) No. species 0.65 0.70 0.95 0.82

n=35 n=34 n=12 n=35
b) No. species 0.45 0.42 0.83 0.74
n=40 n=40 n=12 n=40
c) No. species n.a. 0.31 0.61 0.52
The number of fish species at a site was generally higher at high drainage area, low
elevation, high stream order and low gradient in decreasing order of relationship (P<O.Ol,
2-tailed; Tables 5 & 6). Partial Kendall's tau (Daniel 1990) was also calculated to control
the influence of elevation on Kendall's tau values for gradient, drainage area and stream
order with number of species (Table 6). Again, the relationships were significant.
A cladogram of ecological zones based on Jaccard's coefficient of similarity of species is
given in Figure 2. Two distinct ichthyological groupings were identified in the study area;
those species which are confined almost exclusively to areas below waterfalls (which act as
dispersal barriers) and those which are found mainly above. The latter is confined to twelve
species, belonging to the torrent loach family Balitoridae. These species are Neogastromyzon
nieuwenhuisi, Gastromyzon borneensis, G. fasciatus, G. lepidogaster, G. punctulatus,
Protomyzon whiteheadi, Protmyzon sp., Nemacheilus sp., Neogastromyzon sp., Glaniopsis
hanitschi, G. gossei and Sundareonectes sabanus; the last six being found exclusively above
waterfalls.
0.000 1.000
Above, riffle
Above, run & pool
Below, fast
Below, slow
Below, demersal
Below, pelagic
Below, benthic
Below, neustonic
Figure 2. Cladogram of ecological zones based on species similarity. Scale is euclidean distance obtained using single
linkage method. Above, streams above first encountered major waterfall (> 3m high), acting as barrier; Below,
streams below any first major waterfall; fast (velocity > 0.5 ms") and slow (velocity < 0.5 ms") refer to flowing
condition; neustonic (near surface), pelagic (mid-water), demersal (near stream bed) and benthic (on stream bed)
refer to the zones in the water column.
384 SATISH C. CHOY
Table 7. Habitat classification of the freshwater fish in the Belalong-Temburong watershed.
I. BELOW WATERFALLS:
a) Pools & slow-flowing channels
i) Neustonic: Nematabramis steindachneri & Rasbora argyrotaenia.
ii)Pelagic: Hampala bimaculata, Lobocheilus bo, Osteochilus spilurus, Barbodes collingwoodi, Puntius
binotatus, Rasbora sumatrana, Rasbora sp., Tor douronensis, Krytopterus sp. & Chela sp.
iii) Demersal: Anguilla marmorata, Clarias leiacanthus, Clarias teijsmanni, Glyptothorax major,
Macrognathus maculatus, Mastacembelus unicolor, Mystus baramensis, Silurichthys hasselti, Silurus sp.,
Leiocassis micropogon, Pangio mariarum, Lobocheilus bo, Hampala bimaculata, Tor douronensis,
Osteochilus spi/urus, Barbodes collingwoodi, Puntius binotatus, Ompak sp., Krytopterus sp. &
Mystus sp.
iv) Benthic: Pangio mariarum, Parawaous sp. & Pseudogobiopsis sp.
b) Runs & riffles (fast flowing channels)
i) Demersal: Lobocheilus bo, Hampala bimaculata, Glyptothorax major, Silurichthys hasselti, Pangio
mariarum, Mystus baramensis, Mystus sp., Leiocassis micropogon, Clarias leiacanthus, C. teijsmanni,
Anguilla marmorata, Tor douronensis, Macrognathus maculatus, Mastascembelus unicolor, Silurus sp.,
Ompak sp. & Krytopterus sp.
ii) Benthic: Gastromyzon bomeensis, G. lepidogaster, G. monticola, G. punctulatus, G. ridens, G. fasciatus,
Neogastromyzon nieuwenhuisi, Paracrossochilusacerus, P. vittatus, Protomyzon whiteheadi, P. griswoldi,
Parawaous sp. & Pseudogobiopsis sp.
2. ABOVE WATERFALLS: all benthic species

a) Pools: Neogastromyzon nieuwenhuisi, Neogastromyzon sp., Nemacheilus sp. , Glaniopsis gossei, Sundareonectes
sabanus, Gastromyzon bomeensis & G. lepidogaster.
b) Runs & riffles: Glaniopsis hanitschi, Gl. gossei, Nemacheilus sp., Neogastromyzon nieuwenhuisi, Gastromyzon
bomeensis, G. fasciatus, G. punctulatus, Protomyzon whiteheadi & Protomyzon sp.
In pools and slow flowing areas below waterfalls, Nematabramis steindachneri and Rasbora
argyrotaenia were most abundant numerically while in faster flowing zones, Paracrossochilus
acerus and Gastromyzon spp. were predominant (Table 7). In the deep, slow-flowing
channels, neustonic or surface species were Nematabramis steindachneri and Rasbora
argyrotaenia, pelagic species were other cyprinids such as Hampala bimaculata, Lobocheilus
bo, Osteochilus spilurus, Tor douronensis, Puntius binotatus and Barbodes collingwoodi,
demersal species were the catfishes and the benthic species were the gobies and loaches.
There was little vertical stratification in the shallow fast-flowing zones, due to the
compression of this stratum.
DISCUSSION
The water quality of the streams in the Belalong-Temburong watershed, above the confluence
(Kuala Belalong) is good and unpolluted. Lee et ai. (1990) suggested that the most suitable
aquatic environment for fish life is water which has pH of between 6.5-9,0, dissolved oxygen
content of greater than 5 mgl· 1 and total suspended solids (TSS) of less than 80 mgl· l • All
these conditions are being met in the Belalong-Temburong river system although they may
not be fulfilled temporarily after heavy rain, when TSS increases significantly.
Increase in species richness with increasing stream order and drainage area has been
demonstrated by Mahon (1984) for the rivers in the temperate region. Except for Bishop
(1973), Livingstone et al. (1982) and Watson & Balon (1984a, b), such analysis has not been
done for tropical streams. Livingstone et al. (1982) and Watson & Balon (1984a, b) found
species richness to be positively correlated to drainage area.
Fausch et al. (1984) used relationships between stream order and number of fish species
to develop indices of biotic integrity for use in comparing streams within a biogeographic
region. Stream order is an indicator of stream size (width, depth and discharge). Gradient
and elevation and water velocity decrease with increase in stream order and generally
substrates chanre from rocky to silty (Kuehne 1970). The present study suggests that species
richness increases with increase in stream order. Reduced species richness in low order
streams result<. from steep gradient, rapid changes in flow, unstable substratum, small habitat
volume, limited habitat complexity and low autochthonous production (Vannote et at. 1980).
Because stream order is generally easy to determine from topographic maps, and because it
reflects other physical attributes which could influence absence or presence of aquatic
species, it could form a basis for comparing streams in icthyological studies (Kuehne 1970).
If these relationships are valid, then stream order could be useful in predicting species
numbers or presence of particular species in different watersheds within a given region.
In this study, however, drainage area was more strongly correlated than stream order.
Analysis of published data from Southeast Asia (Bishop 1973, Furtado & Mori 1982, Inger
1955, Lee et at. 1990, Roberts 1989, Samat 1990a, b, Taylor 1989) also indicated this,
suggesting that drainage area is a better physical attribute to use than stream order, at least
for Southeast Asia. Estimating drainage area from topographic maps is more accurate than
determining stream order. In tropical rainforest areas, where rainfall is high and intermitten!
many streams are ephemeral and attributing stream order to a particular stream may not be
accurate. Furthermore. most of the smaller streams are not shown on topographic maps,
which makes assessment difficult and unreliable.
The lower correlation of stream order. elevation and gradient to species richness is likely
to be the result of geographic heterogeneity of the areas under study. Streams having low
and high topographic relief within a small area result in extremes of gradients and low
species richness.
The segregation of fish above and below any major waterfall is expected. Waterfalls and
cascades form effective barriers to migration and dispersal (Schlosser 1982). All the species
found above waterfalls belong to the family Balitoridae and exhibit morphological and
behavioural specialisations to live in torrent streams. In above-waterfall streams, several of
these species occur in slow flowing pools whereas in below-waterfall streams none are found
such areas. Above waterfalls, there is an absence of other fish but in pools below any major
waterfall there is an abundance of competitors and predators such as the cyprinids. bagrids,
clarids and silurids. Thus, predation by and/or competition from other species prevent the
balitorids from colonising the slow-flowing areas of below-waterfall streams.
Many published studies have shown increased community diversity in stream fishes from
upstream to downstream areas (Bishop 1973, Gorman & Karr 1978, Sheldon 1968) with most
emphasizing species addition. Bishop (1973) also noted replacement of one species by
another. In the present study, increased species richness from upstream to downstream was
initially by species replacement, particularly when a major waterfall was encountered, and
then for below-waterfall streams. by species addition. Increase in species diversity has been
attributed to increase in available habitat and decrease in environmental fluctuations (Lowe-
McConnell 1975). Downstream, there is an increase in the availability of deep channels.
Increased species diversity in these channels was a result of different species occupying the
different zones of the vertical stratum, suggesting that fish are prone to using the stream
environment in a 3-dimensional way (Gorman & Karr 1978, Sheldon 1968).
ACKNOWLEDGEMENTS
I wish to express my gratitude to the guides and RGS employees from the Batang Duri
long house who helped with logistics and sampling. Special thanks to Dr. D.S. Edwards,
other Biology Dept. colleagues and the RGS personnel for support and encouragement and
386 SATISH C. CHOY
to Mr. Samhan Nyawa for his company on many of the rather difficult sampling trips. This
research was funded by grant number UBD/T/RG.12 from Universiti Brunei Darussalam.
LITERATURE CITED
BISHOP, J.E. 1973. Limnology of a small Malayan river, Sungei Gombak. Monographiae Biologicae 22. Junk,
The Hague. 485 pp.
BURRIDGE, M.G. 1992. Systematics of the Acanthophthalmus kuhlii complex (Teleostei: Cobitidae), with the new
description of a new species from Sarawak and Brunei. Copeia 1: 172-186.
CRANBROOK, EARL OF. 1981. The vertebrate faunas. In Whitmore, T.C. (ed.). Wallace's Line and plate
tectonics. Clarendon Press, Oxford.
DANIEL, W.W. 1990. Applied non-parametric statistics. PWS Kent. Boston, USA. 635 pp.
FAUSCH, K.D., KARR, l.R. & YANK, P.R. 1984. Regional application of an index of biotic integrity based on
stream fish communities. Transactions Of the American Fisheries Society 113: 39-55.
FURTADO, J.I & MORI, S. 1982. Tasek Bera. Junk, The Hague. 413 pp.
GORMAN, O.T. & KARR, J.R. 1978. Habitat structure and stream fish communities. Ecology 59: 507-515.
INGER, R.F. 1955. Ecological notes on the freshwater fishes of a coastal drainage of North Borneo. Fieldiana:
Zoology 37:47-90.
INGER, R.E & CHIN, P.K. 1962. The freshwater fishes of Northern Borneo. Fieldiana: Zoology 45:1-268.
KOTTELAT, M. 1989. Zoogeography of the fishes from Indo-Chinese inland waters with an annotated checklist.
Bulletin of the Zoological Museum, University of Amsterdam 12(1): I-54.
KUEHNE, R.A. 1970. Application of the Horton stream classification to evaluate faunal studies. Pp. 367-370 in
Weist, W.G. Jr. & Greeson, P.E. (eds). Hydrobiology - bioresources of shallow water environments. American
Water Resources Association. Urbana, Illinois.
LEE, Y.H., LEE, N., GHAFFER, M.A. & SAMAT, A. 1990. Water quality and fish fauna distribution in the
rivers of the Tawau Hills Park, Sabah, East Malaysia. pp. 132-133, in Extended abstracts of papers presented at
the International Conference on Forest Biology and Conservation in Borneo. Kota Kinabalu, Sabah, 30 July-3
August, 1990.
LIVINGSTONE, D.A., ROWLAND, M. & BAILEY, P.E. 1982. On the size of African riverine fish faunas.
American Zoologist 22:361-369.
LOWE-McCONNELL, R.H. 1975. Fish communities in tropical freshwater: their distribution, ecology and
evolution. Longman, London. 337 pp.
MAHON, R. 1984. Divergent substructure in fish taxocenes of north temperate streams. Canadian Journal of
Fisheries and Aquatic Sciences 41 :330-350.
RAINBOTH, W.J. 1991. Cyprinids of Southeast Asia. In Winfield, I.J & Nelson 1.S. (eds). Cyprinid fishes -
systematics, biology and exploitation. Chapman & Hall, London.
ROBERTS, T.R. 1989. The freshwater fishes of western Borneo (Kalimantan Barat, Indonesia). Memoirs of the
California Academy of Science 14: 1-210.
SAMAT, A. 1990a. Taburan dan populasi ikan air tawar di beperapa altitud di Taman Kinabalu, Sabah, Malaysia.
Pertanika (Sabah) 13(3):341-348.
SAMAT, A. 1990b. A study offreshwater fish ecology of streams in logged and un logged areas of Danum Valley.
Technical Report, Danum Valley Committee/Royal Society. 5 pp.
SCHEFLER, W.e. 1980. Statistics for biological sciences. 2nd Edition. Addison-Wesley, California, USA. 230 pp.
SCHLOSSER, I.J. 1982. Fish community structure and function along two habitat gradients in a headwater stream.
Ecological Monographs 52:395-414.
SHELDON, A.L. 1968. Species diversity and longitudinal succession in stream fishes. Ecology 49: 193-198.
SOUTHWOOD, T.R.E. 1978. Ecological methods with particular reference to the study of insect populations.
Chapman & Hall, London. 524 pp.
TAYLOR, R.S. 1989. A survey of the fish fauna of the Tabin Wildlife Reserve, Sabah. Sabah Museum Monographs
3:87-99.
VANNOTE, R.L., MINSHALL, G.W., CUMMINS, K.W., SEDELL, l.R. & CUSHING, C.E. 1980. The river
continuum concept. Canadian Journal of Fisheries and Aquatic Sciences 37:130-138.
WATSON D.J. & BALON, E.K. 1984a. Structure and production of fish communities in tropical rainforest streams
of northern Borneo. Canadian Journal of Zoology 62:927-940.
WATSON D.l. & BALON, E.K. I 984b. Ecomorphological analysis of fish taxocenes in rainforest streams of
northern Borneo. Journal of Fish Biology 25:371-384.
ZAKARIA-ISMAIL, M. 1991. Zoogeography and biodiversity of the freshwater fishes of South-east Asia. Paper
presented at the Regional Seminar on Ecology and Conservation of Southeast Asian Marine and Freshwater
Environments including Wetlands. Institute of Advanced Studies, University of Malaya, Malaysia, November 1991.
Analysis of factors contributing to the stability of steep hillslopes in the
tropical rainforest of Temburong, Brunei Darussalam
A.P. DYKES
Department of Geography, King's College, London, United Kingdom
ABSTRACT. Tectonic uplift and associated channel incision in southern Temburong District, Brunei
Darussalam, and adjacent parts of Sarawak, have given rise to a deeply dissected topography which is
characterised by steep slopes and sharp ridges. These steep hillslopes are prone to both shallow and deep-
seated failure although all recent examples (not yet revegetated) fall into the former category. The Infinite
Slope Model for shallow planar failures is used to examine the stability of characteristic slopes for the
range of hydrological conditions observed in the field. Matric suction in unsaturated soil significantly
enhances the ability of these slopes to resist failure, but typical wet-season rain events can cause failures
due to positive pore water pressures beneath perched water tables in the soil. Seasonality of the climate
is shown to be an important factor in the stability regime of the hillslopes.
KEY WORDS: hydrology, hillslope stability, climate, topography, soils, tectonics, Borneo.
Reality is a dynamic system, variable in time and space, integrating all pathways, and
ranging between the extremes. (C.A. Troendle 1985, p.350)
INTRODUCTION
This paper is intended to give a broad introduction to the physical characteristics and
processes of the tropical rainforest of Brunei Darussalam's Temburong District. It will
describe the main features of the climate, hydrology and geomorphology of the region,
before focussing on the apparent contemporary dominance of shallow mass movement over
other hillslope processes in order to illustrate how seasonal climatic patterns can directly
affect geomorphic processes beneath a primary rainforest vegetation cover. This region of
rainforest was the focus of the Brunei Rainforest Project 1991-92, an international research
effort organised jointly by the Universiti Brunei Darussalam and the Royal Geographical
Society (U.K.).
The research described in this paper was conducted within a designated 50 km 2 area based
on the new Field Studies Centre (FSC) south of Kuala Belalong, in the Temburong District
of Brunei. Brunei Darussalam is an oil-rich microstate of 5,765 km 2 situated on the
northwestern coast of Borneo. 59% of the country is covered with primary rainforest, with
a further 22 % also under forest cover (Collins et al. 1991). Its population of 268,000 (1991
census) is largely concentrated along the coastal strip. Temburong District is physically
separated from the remainder of Brunei by a narrow region of Sarawak. Most of
Temburong's 1,288 km 2 is covered with primary rainforest, the southernmost 488 km 2
belonging to the Batu Apoi Forest Reserve. Kuala Belalong lies on the northwestern
boundary of this reserve, and the designated research area extended southwestwards into the
reserve. All field data was collected in the lower Belalong valley and in the valley of the
Engkiang river, a major tributary of the Belalong river (Figure 1).
387
388 A.P.DYKES
BRUNEI
TEMBURONG DISTRICI'
......
......
,,
I
I
.....
I .....
1
,
I
• 5elllamanl
= Road
Mounlalnunmits
10kln
I
Figure 1. Map of Temburong District, Brunei Darussalam, showing locations referred to throughout the text.
THE RESEARCH ENVIRONMENT
Climate
Brunei experiences a tropical climate which is influenced by the monsoon seasons of South
East Asia. These are caused by the seasonal movement of the pressure systems in tropical
latitudes. The Inter Tropical Convergence Zone (ITCZ), which defines the Equatorial low
pressure belt, moves south across the Equator during November and, from December to
March, Brunei and northern Borneo experience northeasterly winds (the 'North East
Monsoon'), which curve into the ITCZ over Indonesia. The ITCZ moves north across the
Equator during May, and the 'South West Monsoon' sets in over Brunei from June to
September as the southeast trade winds curve across the Equator to become southwesterly
winds over northern Borneo (Figure 2). Since the dominant direction of the monsoon winds
is roughly parallel with the northwest coast of Borneo during both seasons, the effects of the
monsoons in terms of rainfall seasonality in Brunei are greatly ameliorated. Long-term
means of rainfall data (e.g. Figure 3), show peaks in May and November which coincide
with the overhead passage of the ITCZ, and in some areas the start of the North East
Monsoon appears to bring slightly enhanced rainfall.
Hillslope stability in tropical rainforest 389
1100E 11S0E 1200E 1100E 11SQ E 1200E
SON
0°
50 S
North east monsoon (Dec.- Mar.) South west monsoon (Jun.- Sep.)
Figure 2. Monsoon wind patterns over Brunei and the rest of Borneo.
Mean rainfall 1973-90 {mm}

60~-
500
400
300
200
100
o
JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
Figure 3. Mean monthly rainfall at Semabat Agricultural Station (see Figure 1), 1973-90 inclusive.
Much of Brunei's rain falls in short duration, high intensity and localised events, mostly
due to small-scale convection cells which are accentuated in areas of high relief, although
longer periods of steadier, less intense rainfall can occur in the interior due to the rising and
cooling of moist monsoon air masses over the mountains. The combination of monsoon,
convection and orographic effects results in a rainfall regime noted for its inconsistency and
variability, both spatially and temporally, although broad seasonal trends are apparent in the
long term. The mean annual rainfall exceeds 2300 mm throughout Brunei (Grant 1984); in
Temburong, the annual mean of around 3600 mm at Bangar increases inland to over 4000
mm at Semabat, and the area around Kuala Belalong should expect 4000-4500 mm of rain
annually. Table 1 shows the monthly and annual variability of the rainfall recorded at
Semabat, which is just 9 km west-northwest of Kuala Belalong, whilst Table 2 illustrates the
orographic effect on rainfall totals using data recorded by automatic weather stations (A WS)
located at Kuala Belalong (800 metres north of the FSC) and on the summit of Bukit
Belalong, just 7.6 km apart horizontally but at altitudes 853 metres different, between July
1991 and March 1992 inclusive.
390 A.P.DYKES
Dry periods lasting three or four months are irregular but not infrequent in this part of
Borneo. The total of 377 mm of rain recorded between 1 January 1992 and 31 March 1992
at Kuala Belalong is comparable with dry spells recorded at Semabat and other nearby
stations in Sarawak at other times. At Nanga Medamit, 32 km south-west of Kuala Belalong,
1962.5 mm of rain fell during the 12 months June 1984-May 1985 inclusive (c.f. annual
mean 4040.7 mm; data supplied by the Malaysian Meteorological Service). Such dry periods
usually include the months of February and March, or July and August, but appear to be
local rather than regional in extent.
The incidence of extremely wet periods is equally variable. At Semabat, 1559.1 mm fell
in Apr/May 1984 and 1602.1 mm in Nov/Dec 1987, whilst at Ukong (35 km west of Kuala
Belalong), 1063.2 mm fell in January 1963 (data supplied by the Malaysian Meteorological
Service). It is likely that most of this monthly total resulted from a very small number of
extreme rain events, or even a single exceptional storm. At Kuala Belalong, an event of
116.5 mm of rain falling in 75 minutes was recorded at the Field Studies Centre on 19
September 1991. Five days later, 139.5 mm fell throughout an 11-hour period. This latter
storm also produced 181.5 mm of rain in 10 hours at Bukit Belalong, of which 67 mm fell
in one hour, and 100.5 mm within two hours. Figure 4 illustrates the daily rainfall pattern
recorded by the weather station at Kuala Belalong throughout September 1991, and shows
how one or two large rainfall events can dominate the monthly total. The hourly rainfall
intensity profile of the large storm of 24 September 1991 is characteristic of tropical
thunderstorms generally, although five-minute intensities of over 168 mm/hr have been
recorded during smaller storms. The storm of 24 September 1991 is shown in Figure 5
along with those of the smaller events on the preceding and following days, as recorded at
Bukit Belalong. On the basis of the examples given here, analysis of magnitude-frequency-
intensity relationships for rain events in the region might perhaps provide a better indication
of the overall rainfall regime (Day 1979).
Daily mean temperatures at sea level vary little from 25-26°C in any month, falling to 20-
23°C at 915 metres altitude, and generally slightly lower during the North East Monsoon due
to increased cloud cover and reduced sunshine hours. The diurnal range is around 10°C
Table 1: Rainfall summary for Semabat, Temburong; 1973-90. Coefficient of variation = (St.Dev.lMean) x 100.
Source: Kilanas Agriculture Department, Brunei Darussalam.
Month Mean Standard Minimum Maximum Variation Coefficient

(mm) Deviation (mm) (mm) (%)
JAN 354.8 169.6 29.7 637.5 47.8
FEB 275.0 179.6 89.8 655.7 65.3
MAR 305.4 146.5 135.9 724.0 48.0
APR 384.7 158.5 101.3 734.3 41.2
MAY 409.4 184.8 216.9 824.8 45.1
JUN 289.3 196.5 104.9 890.8 67.9
1UL 269.2 113.9 117.2 482.6 42.3
AUG 273.5 200.9 21.3 721.2 73.5
SEP 351.1 198.8 73.1 783.8 56.6
OCT 338.7 109.7 231.0 691.2 32.4
NOV 492.1 151.3 289.9 757.2 30.7
DEC 376.1 185.5 69.9 752.2 49.3
ANNUAL 4119 779 2845 5328 18.9
Hillslope stability in tropical rairiforest 391
Table 2: Monthly rainfall and rainless days in the Kuala Belalong research area, 1991-92.
Kuala Belalong Bukit Belalong

Altitude: 60 metres Altitude: 913 metres
Month Rainfall (mm) Rlless days Rainfall (mm) R/less days

JUL '91 148.0 17 190.0 14
AUG '91 131.5 14 180.5 15
SEP '91 492.5 4 651.5 3
OCT '91 210.5 9 278.5 5
NOV'91 564.0 3 717.0
DEC '91 292.0 6 330.0 5
JAN '92 153.0 14 277.5 12
FEB '92 121.5 17 91.5 13
MAR '92 102.5 18 235.5 15
TOTAL 2215.5 102 2952.0 83
throughout the year. Mean onshore wind speeds tend to be low, but instantaneous speeds
associated with vigorous convection cells can reach 25-30 mls for short periods during storms
in Brunei generally (Grant 1984). An hourly average wind speed of 4 mls was attained only
once between May 1991 and April 1992, on the exposed summit of Bukit Belalong.
Relative humidity in the open varies daily between 95-100% during the night to around
75 % in the early afternoon (Grant 1984); relative humidities at Kuala Belalong exceeded 99%
every night, and fell as low as 57% between Ipm and 2pm on 19 June 1991, although the
low-70s are the usual lower values. On Bukit Belalong, relative humidities tended to stay
lower during the night, but rarely dropped below the low-80s during the day, due to the
higher vapour pressures in the cooler air at the higher altitude. Grant (1984) also reports a
mean annual evaporation of 1831 mm at Brunei Airport for the period 1978-81, measured
using an evaporation pan. This appears significantly greater than estimates for the
mountainous interior: the mean annual evaporation measured at Ukong was 1382 mm for the
period 1963-77, the monthly mean ranging from 104.1 mm in February to 128.5 mm in May
(data from the Malaysian Department of Irrigation & Drainage). A water balance calculation
of mean annual evapotranspiration over a 5-year period for mature dipterocarp forest at
Mendolong, Sabah, came to 1540 mm (MaImer 1992), with a mean annual rainfall of 3352
mm for the same period, whilst Bruijnzeel (1983) reported a mean annual evaporation of
1345 mm for a forested drainage basin in central Java (mean annual rainfall 4770 mm). On
the basis of these figures, southern Temburong might expect 1350-1450 mm of
evapotranspiration losses annually.
Geology
The underlying geology of the research area is generally very simple, although the
regional geological history is extremely complex (see James 1984, Wilford 1961). The
whole of the Batu Apoi Forest Reserve is underlain by the Temburong Formation,
sedimentary rocks of Middle Oligocene to Early Miocene age (around 16-30 million years
old) of shale lithology, described by Wilford (1961) as a sandy, dark grey to black rock with
abundant plant remains, occasionally encompassing thin beds of hard, fine to medium grained
quartzitic sandstone, and thin limestone beds (the latter extremely rare). The formation is
strongly folded, though the intensity and amplitude of folding decreases northwestwards. In
the lower Belalong valley the shale beds dip eastwards at 40-6<f and occasionally steeper
(Wilford 1961, Map Sheet 2).
392 A.P.DYKES
Rainfall (mm)
140,-------------------------------------------------
120 -+ ....................................... .
Total rainfall = 492.5mm
100 + ................................................
801··················································· ..•.•..........•...•..••...........................................................................••...................................•.................................... ,
60+··················································· ....................................
40
20~· ..................................•
5 10 15 20 25 30
September 1991
Figure 4. Daily rainfall recorded in September 1991 by tbe automatic weatber station at Kuala Belalong.
Hourly rainfall (mm)

70
60
23/9/91 24/9/91 25/9/91
50
I I
40
30
20
10
o 1.11 II III•. III

i5 18 21 24 03 06 09 12 15 18 21 24 03 06 09 12 15 18 21 24
Time (hours): 23-25 September 1991
Figure 5. Hourly rainfall intensity profiles of tbe storms of 23-25 September 1991 at Bukit Belalong.
The area has experienced considerable tectonic uplift in geologically recent times, although
estimates of uplift rates vary. Wilford provides a detailed denudation chronology of the
entire region of Sarawak within which Brunei is enclosed, though it lacks measured dates for
any of the geomorphological features he uses to justify his account. He suggests a rate of
uplift of 1.1 mm/year in southeast Temburong, decreasing towards the northwest (giving an
estimate of 0.75 mm/year at Kuala Belalong). Recent palaeomagnetic dating of cave
sediments in the nearby Gunung Mulu National Park, approx. 40 km southwest of
Temburong's southern border, suggests a rate of uplift there of 0.19 (±0.02) mm/year for
at least the last 2 million years (A. Farrant & P. L. Smart, personal communication), which
is probably the best available upper estimate for southern Temburong since the Mulu
anticline has uplifted to a much greater extent than regions surrounding it.
Soils
Intense chemical weathering of the shale bedrock has given the area a relatively thin cover
"f residual clay soil. The depth of undisturbed soil cover is thought to only rarely exceed
a couple of metres or so, although considerable depths of soil and unsorted weathered shale
fragments are likely to occur in some restricted localities marked by the deposition zones of
major landslides. The soils of the research area were first described by Ashton (1964), who
summarised his profile observations by concluding: "The. catena at Belalong from high ridge
to river side showed ... a sequence from friable relatively dry soils, with roots continuing
deep below the surface, to wet sticky impervious soils with shallow roots" (p.26). Soil
profiles examined in detail during 1991192 have been characterised by Ross & Dykes (this
volume), and physical properties of the soils which affect the hillslope hydrology will be
discussed later in this paper.
Vegetation
Ashton (1964) gave the first detailed account of the vegetation of this region, describing
it as 'Lowland Mixed Dipterocarp Forest'. The results of detailed studies and surveys from
the 1991-92 expedition (this volume) will provide a more extensive coverage of species
assemblages and physiographic characteristics to complement Ashton's earlier work. The
implications of the vegetation cover for any analysis of slope stability may be considerable,
given especially the dominance of the giant dipterocarp trees over other species in terms of
physical size and species assemblages; these implications and assumptions made in respect
of them will be discussed later.
Topography and Drainage

The topography of the lower Belalong Valley and its tributaries is characteristic of the
whole of southern Temburong. Channel gradients are steep: the mean slope of the Belalong
river for the first few kilometres upriver from Kuala Belalong exceeds 0.5 %, and the
Engkiang river gradient increases to over 2.4 % within 3 km upstream from its confluence
with the Belalong river. The channels are cut into the bedrock of the valley floors, and in
many places the larger channels contain large quantities of coarse bedload material. Steep
slopes rise from the sides of the channels to form deep, V-shaped valleys. There are no
floodplains. Minor tributaries with channel gradients often exceeding 400 (84%) descend the
main valley sides within their own sub-valleys.
Main valley hillslopes occasionally exceed 45° or even 60°, whilst slope angles in the
tributary valleys are usually in excess of 45°, although the slopes are much shorter. The
major catchments and the minor sub-valleys are separated by narrow, often knife-edged
ridges, and almost all hills lopes bear the scars of major landslide events. Permanent drainage
density in the region appears to be fairly high, but cannot be reliably determined from maps
since the maps are often produced from air photographs which do not show ground surface
features such as drainage channels because of the dense vegetation cover (Morgan 1976).
There is a much higher density of storm-activated channels, which should be included in any
estimate of drainage density because "... most land forming processes are more active ...
when water is flowing in the intermittent first-order channels than during base flow" (Morgan
1976, p.323). Morgan (1976) estimated the "effective" drainage density (i.e. including
storm-activated channels) of the dissected uplands of West Malaysia to be around 8-12
km/km2, subject to the assumptions he applied to his map data.
394 A.P.DYKES
The total relief of the research area is about 863 metres, from Kuala Belalong itself (the
river junction) at 50 metres above sea level, rising via the main ridge east of the Belalong
river (c.500 m) to the summit of Bukit Belalong at 913 m a.s.l.; the ridge on the west side
of the Belalong river averages 300 m. Beyond the research area, the Belalong catchment
rises to 1192 m at Bukit Lesong, at the southern edge of Temburong, and the Temburong
river itself rises on the slopes of Bukit Pagon (1850 m).
HYDROLOGICAL PROCESSES
The rainfall regime experienced in this region of northern Borneo is highly variable in space,
time and magnitude. The frequency distribution of rainfall events determines the
hydrological regime of the hillslope soil profiles, and this in turn controls the development
of the slope forms by denudation processes and the flood response of the channels. The
nature of channel discharge regimes can, depending on the fluvial characteristics of the
channel, directly affect slope form by undercutting of the lower parts of the hillslopes, and
the overall slope form greatly influences the hillslope hydrology. The problem of analysing
any single component of the drainage basin system stems from the interrelationships
identified above, and the integration of hillslope and channel processes within a single model
to simulate drainage basin evolution is now being attempted by a number of workers (e.g.
Willgoose et al. 1991). This paper will describe the general hydrological characteristics of
the Temburong rainforest, before analysing the linkages between hillslope hydrology and
slope form with respect to shallow mass movement.
Water enters the catchment as rainfall, and that water which is not transferred back to the
atmosphere by evaporation or transpiration, leaves the catchment in the form of channel
flow. Some of the intermediate components of the tropical rainforest hydrological cycle can
not yet be confidently quantified (Bruijnzeel 1990), especially the forest canopy factors of
interception, throughfall and stemflow, which determine the amount of rain which reaches
the ground. Bruijnzeel's (1990) review of published interception and throughfall data
concludes that whilst stemflow usually accounts for no more than 0.5-2 % of rainfall, average
interception for lowland rainforest is 13 % (range 4.5-22). At Kuala Belalong, measured
canopy interception was exactly 18% of rainfall if 1% stemflow is assumed.
The rainfall patterns experienced around Kuala Belalong have been described already, but
as yet the likely evaporation and evapotranspiration rates can only be estimated from data
collected elsewhere. River discharges fluctuate widely and rapidly in response to rainfall,
as the stage hydrograph of the Belalong river (catchment 150 km2) shows (Figure 6). The
lowest observed flow was measured near Kuala Belalong towards the end of the extremely
dry period in early 1992. The discharge of 1.17 cumecs probably represents the groundwater
base flow from the entire catchment (i.e. a catchment yield of approx. 6700 litres/hectarel
day). Flash floods with estimated peak discharges in excess of 120 cumecs occurred three
times within a 14-month period ending 14 June 1992 (equivalent to a catchment-wide yield
of 29,000 litres/hectare/hour). The stage hydrographs shown in Figure 6 illustrate the typical
wet season condition, with discharge rarely falling below about 10 cumecs (calculated for a
stage of 2 metres). Neither the location, spatial extent or the intensity distributions of the
storms which produced these flood events are known, but rainfall on Bukit Belalong which
contributed to these flood peaks is shown for illustration.
What is clear from the discharge pattern in Figure 6 is that rain falling on any part of the
catchment is producing extremely rapid flood responses. Overland flow is not a rare
occurrence in tropical forests (e.g. Bonell et al. 1983), but it is usually of limited spatial
extent due to the high infiltration capacities of undisturbed forest soils (Bonell & Gilmour
60
Rainfall (mm) at Bukit Be,61ong
40
20
3.5
~ 3.0
;;
E
2.5
2.0 -h..,...,...,...,.-,--r-,--r-,--r,..,..:>lh-""""''''''''''''''''''''''''''''''''''''''''''''''''''''''''"T""T""T""T''''''''''''''''''''''''''''''''
Date and time

Figure 6. Stage hydrograph for the Belalong river at the FSC for a IO-day period in November 1991.
1978) and results from the build-up of saturated wedges in the soil due to rapid subsurface
flow from upslope, which reach the surface in hollows or adjacent to channels (e.g. Moore
et al. 1986). Overland flow was observed on undisturbed ground in the Kuala Belalong area
in the form of storm-activated channels originating from hollows where saturated wedges
would be expected to develop. Measured saturated infiltration rates (subject to allowance for
problems as discussed by Youngs 1991) were extremely high at the soil surface (average
656.2 mm/hr, n=6, range 503-878) and at surfaces cut into the B-horizon (wet season
average 69 mm/hr, n=9, range 6-267), thus under the range of 60-minute rainfall intensities
recorded at Kuala Belalong between June 1991 and April 1992 (up to 67 mm/hr), almost all
the rainwater which reaches the ground should enter the soil. Subsurface throughflow
therefore appears to be the dominant contributor to the storm hydrographs. Considering the
hillslope profile in a schematic manner, and leaving aside the mathematical treatments of soil-
water relationships (the interested reader is referred to Hillel 1971, for these; or to Kirkby
1985, for their incorporation within hillslope hydrological analyses), it is assumed here that
all of the infiltrating rainwater moves vertically down through the soil profile as far as the
bedrock interface, then downslope along the base of the soil profile parallel to the surface
with negligible losses to the bedrock. Unsaturated flow between storms will supply enough
water to the saturated lower parts of the hillslopes to maintain discharge of water into the
channels between rainfall events (Weyman 1970), whilst downslope flow during storms is
likely to be saturated, creating perched water tables above the bedrock within the soil profile
if the average subsurface flow velocity is not too great (Pearce & O'Loughlin 1985). The
extent of the saturated zone at the lower extent of the soil profile varies within and between
storms, hence the area of slope which contributes to the discharge hydrograph of the channel,
and the channel stormflow itself, similarly vary (see review by Troendle 1985).
That saturated wedges rapidly develop in hollows as well as adjacent to channels during
heavy rain is demonstrated by soil tension measurements from mercury manometer
tensiometers near Kuala Belalong. Figure 7 illustrates the rapid build-up of saturated
conditions at a depth of 90 cm in the soil at tensiometer nest D, near the bottom of a small
396 A.P.DYKES
Hydraulic head (em water)

60
50
40
30
20
~
10
o ~
----- ~
-10
~
o 2 3 4 5 6
Days (1st record =
0945 on 2/5/92)
Figure 7. Changing soil water conditions over six days in early May 1992. measured at tensiometer D3 (90 cm deep
in the soil). A positive hydraulic head shows the height of a water table above the tensiometer cup; negative values
are tensions.
, ~ Soil pit
Tensiometer nest
Main footpath
Contours at 2m intervals.
heights above river level
oI 10
I
me/res
Figure 8. Map showing arrangement of tensiometers and soil pits on the research hillslope, with tensiometer nest
E nearest the bottom of the small topographic hollow.
topographic hollow (see Figure 8), during two storms in May 1992 (86.8 mm and 47.0 mm
of rain respectively), and the very fast saturated subsurface drainage of the perched
groundwater once the input of rainwater has eased. During both of these storms, a perched
water table had developed within 20 minutes of the onset of rain, but most of this water had
drained within about 90 minutes at this site, and within 5-6 hours at the deepest tensiometer
nearest the slope foot. In neither of these events did the saturated wedge reach the deepest
tensiometer furthest up the slope. (Similarly large moisture fluxes are experienced during
HUlslope stability in tropical rainforest 397
rainfall under dry season conditions, although there is rarely sufficient rainfall at these times
of year to cause saturation.) 'Translatory flow', i.e. displacement of downslope soil water
into the channels by new rainwater from above (e.g. Bruijnzeel 1983), may account for the
initial rapid channel responses to rainfall, but the rapid drainage of saturated zones as
illustrated above suggests that direct throughflow from a given storm dominates the
hydrograph, with significant contributions from the numerous storm-activated channels which
originate in saturated hollows on the hillslopes.
The rate and nature of water movement through the hillslope profile is normally dependent
on the moisture status of the soils, which is controlled jointly by the physical characteristics
of the soils and by the climatic regime. The soils of southern Temburong have very high
clay contents (Ross & Dykes, this volume), which enables them to retain more moisture at
a given tension than coarser-textured soils due to adsorption of water onto the clay minerals
as well as capillary resistance to unsaturated flow in the micropores (Hillel 1971). The
relationship between moisture content and soil tension is usually unique to a particular soil:
the moisture retention curve for the soils at Kuala Belalong shows an initially rapid decline
of soil moisture with low applied tensions, unusual for clay soils generally but indicative of
the occurrence of macropores within these soils, and is fully described by the empirical
equation:
8 = 83. 7 - 3.87 In(~) (1)
for ~>O, where ~ is the tension (bars) and 8 is the percentage saturation of the soil: the
mean saturated moisture content for the soil (i.e. ~=O) was 105.10% (n=12, s.d. =5.99)
by volume pore space, based on a mean porosity of 49.53% (n=12, s.d. =6.08). Most of
the measured change in the soil moisture content occurred over the first bar (100 kPa) of
applied tension, which is slightly more than the highest tensions measured in the field. The
seasonal variation of soil tensions encountered here therefore covers most of the range of
moisture contents possible in these soils; this seasonal variation not only affects the channel
response to rainfall at different times, but also the probability of slope failures occurring at
any given time.
HILLS LOPE PROCESSES AND SLOPE FORM
Introduction
Hillslopes are initially produced by channel incision into an uplifted land mass, their
length and slope angle being largely determined by the available relief and drainage density
(Strahler 1950). Thus an area with a high drainage density and high relief will be
characterised by short but steep slopes, and since steep slopes supply large amounts of
material to the channels, then the channels must be steep in order to remove the eroded
material. The actual removal of material from a hillslope will either be weathering-limited
or transport-limited; in humid tropical regions the latter is dominant, i.e. rates of weathering
and soil production are greater than (or, on a geological timescale, equal to) the available
long-term transport capacity provided by water or gravity. Rapid mass movement is the most
significant slope process in steep, forested tropical landscapes (Chorley et al. 1985), with
water and gravity being the causative agents. Small-scale erosion of surface soil by overland
flow, splash erosion and creep is largely inhibited by the vegetation, especially the dense
networks of shallow and surface roots (Bonell et at. 1983); overland flow generally only
occurs in undisturbed forests on saturated channel margins and stream head hollows (Chorley
et al. 1985). Preliminary data from a hillslope plot experiment near Kuala Belalong clearly
illustrates the degree of protection from splash and runoff erosion afforded to the soil by the
398 A.P.DYKES
surface litter layer (Ross & Dykes, this volume).

HilIslopes evolve by erosion, either retreating parallel to themselves or declining from the
top downwards, with the dominant erosion mechanism being largely determined by the
climate. The concept of threshold slopes may enable the process history of a given landscape
to be interpreted using the observable range of characteristic slope angles (i.e. the inclination
of the steep, straight part of any given profile: Chorley et al. 1985). Hillslopes which
initially exceed their calculated maximum stable angles will decline by mass movement
processes, until they are lowered sufficiently to be stable with respect to further movement
and so enable small scale denudation processes of creep and wash, etc., to control the
subsequent evolution of the hillslopes. Parallel retreat by mass movement may also occur
if the threshold slope angle for stability is exceeded and shallow failures dominate. This
situation arises where channel incision is rapid, and hillslope erosion occurs at a similar rate:
a dynamic equilibrium situation in which shallow failures maintain steep slopes. In this
situation, steep and straight slope forms will tend to dominate the observed slope features.
In cases where the hillslope response is not rapid enough to maintain an equilibrium form
with the channel, convex slope forms may occur, indicating that the hillslopes have not yet
adjusted to very rapid incision. "Differences in hillslope form should not be the result of the
disparity between uplift and denudation, but probably are the result of the difference between
the two components of denudation, rates of channel incision and hillslope erosion" (Schumm
1963, p.H8).
Shallow Mass Failure: the Infinite Slope Model

The simplest form of failure is that of a layer of soil or weathered material sliding
downslope along a surface which is parallel to the surface profile of the slope. The analysis
of hillslope stability with respect to this type of failure is also simple, involving a single
linear equation, but is direct in its application. Although this infinite slope model was
developed to analyse shallow landslides in the London Clay (Skempton 1964, Skempton &
Delory 1957), it has been widely used for interpreting observed limiting or threshold
inclination of slopes where landslides are thought to be responsible for slope decline
(Chandler 1982), and has been adapted and used in analyses of stream head extension
through shallow failure (e.g. Dietrich et al. 1992, Montgomery & Dietrich 1989). However,
"the question of whether conventional methods of stability analyses are equally applicable to
very steep slopes as to relatively gentle ones is not one which has received much attention
in the soil mechanics literature" (Wesley 1977, p.127, in reference to massive, shallow
surface failures on steep slopes in Indonesia).
p
II
Figure 9. The Infinite Slope Model. See text for explanation.

HilZslope stability in tropical rainforest 399
The infinite slope analysis is illustrated in Figure 9, and is used in this paper to calculate
the 'Factor of Safety' of a hillslope under given input conditions. This is generally defined
as the ratio of the forces which promote downslope movement to those which resist
movement, i.e. the tangential (T) and normal (P) components of the vertical force due to
gravity acting on the mass W, which act on the base of a given slice of the hillslope profile.
A factor of safety of 1.0 indicates the critical condition at which the slope reaches the point
of instability; below this value, the slope will fail since the downslope force exceeds the
strength of the slope material. Since the single slice is assumed to represent conditions at
every slice and therefore the hillslope as a whole, lateral forces acting on it from the slices
on either side are assumed to be equal and opposite in effect, and can therefore be removed
from the equation. The factor of safety is thus:
F = c· + (yzcos 2 1} - u)tan0' (2)

yzsin I}cos I}
here the geometric terms are B, the slope angle, and z, the depth of the slip plane; the soil
shear strength terms are c', the cohesion and 0', the angle of internal resistance (friction
angle) of the soil; and the hydrologic terms are y, the bulk unit weight of the soil (dependent
on the moisture content) and u, the pore water pressure, where:
(3)
for positive pressures, with Yw = unit weight of water, and m = height of water table above
the slip plane. Negative pore pressures (soil tensions) can be incorporated directly in the "u"
term of the equation for tensions of up to about 20 kPa (Professor E. N. Bromhead, personal
communication), although Fredlund (1987) has analysed the effect of suction on the effective
shear strength of a soil and has detailed more rigorous methods for incorporating suction
effects into stability analyses.
One of the major problems with stability analysis such as this is the extent to which soil
strength parameters obtained in laboratory tests reflect the true conditions in the field (e.g.
Bromhead 1978). Strength tests on samples from Temburong have not yet been completed,
and so this problem will not be addressed here.
THE SOUTHERN TEMBURONG LANDSCAPE
Analysis of HilZslope Stability

Given that soil shear strength data are not yet available from Kuala Belalong, a modified
version of the infinite slope model has been run for assumed soil strength parameters, based
on Wesley's (1977) Indonesian results and on a simple back-analysis of a recent landslide.
That the landscape in this region is inherently unstable has already been indicated; the
purpose of this paper is to evaluate the relative importance of environmental factors on
stability conditions in this region.
The analyses described below assume a friction angle of 42.5°, slightly higher than
Wesley's (1977) results, and a soil cohesion of 10.5 kPa. The latter was obtained by
conducting a back-analysis of a hypothetical slope of 45° which had failed (i.e. F= 1.0), a
soil depth of 2.0 metres to the shear plane (as observed at the upper edge of a shallow failure
in the Engkiang valley thought to have occurred on 24 September 1991), the friction angle
400 A.P.DYKES
42.5° and worst-case groundwater conditions with the water table at the ground surface and
therefore maximum possible pore water pressures within the slope. This is likely to have
been the case during the rainfall of 24 September 1991. The parameter values assigned
above are those which were used as the basis for the analyses which follow, with the soil
moisture characteristic for these soils (equation (1» being explicitly incorporated within the
stability model, and saturation (~=O) assumed when 8=105%.
A theoretical evaluation of the relative importance of each of the variables incorporated
in the stability model can be made by means of a sensitivity analysis of the model (Pearce
& O'Loughlin 1985). This procedure involves the calculation of a base factor of safety using
the medians of realistic ranges of values for all variables, and its subsequent recalculation
as the value of each variable in turn is changed whilst the others are held constant. The
chosen base values are those given in the previous paragraph, and the results have been
plotted as relative percentages (Figure 10). The analysis shows that the factor of safety is
most sensitive to changes in Band 0' and is also quite sensitive to the cohesion. It is
relatively insensitive to higher unit weights or greater depths than those used as the basis for
this analysis, although they do become progressively more influential with respect to
increasing stability as they become unrealistically low in value. The influence of pore water
pressures is presented by the height of the water table above the shear plane, m, which
increases the factor of safety as the water table, and pore water pressures, fall from their
maximum at the point of failure. The importance of the soil shear strength parameters, and
the caution with which this and the following discussion based on assumed soil parameters
should be treated, is therefore demonstrated.
In attempting to account for ultimate stable slope angles under natural conditions,
Chowdhury (1978) raises the potential problem for evolutionary studies of the possible
decrease of soil shear strength parameters due to weathering or geological factors, but
demonstrates that "... long-term stability is controlled by slope inclination, pore water
pressure and effective shear strength parameters" (p. 91). In this study, the emphasis is on
% change in Factor of Safety

+10r-~_.
+80
+60
+40
+20
-20
-40
-60
-80
_10L-~
-100 -80 -60 -40 -20 o +20 +40 +60 +80 +100
% change in stability variables
Figure 10. Sensitivity analysis of variation in factor of safety for infinite slope stability analysis.
accounting for the occurrence of landslides under the present climatic and tectonic regime,
and the initial model results in the sensitivity analysis presented above agree with
Chowdhury's comment on the dominant controls on stability. In the Temburong setting, soil
shear strength variation is assumed to be negligible in the fully weathered residual soil
profile, and the inclination of any given slope is assumed to be constant over time (due to
parallel retreat) as long as the equilibrium status of the slope/channel unit under analysis does
not change. The soil profiles on the steep slopes under discussion may not, however, attain
the fully weathered state: this may result in the Factor of Safety being underestimated
slightly, which is not serious as it errs on the side of caution (the slope may not fail even if
predicted to do so). Shear strength and slope angles provide static controls on the stability
analysis of a hillslope in the context of this study, whereas the hydrological regime
determines the actual stability at any given moment in time.
The dynamic hydrological processes illustrated earlier can now be incorporated into the
discussion of the stability regime. The theoretical effects of soil tension and of water table
depth (which produces positive pore water pressure according to equation (3), above) on the
calculated factor of safety are plotted for different soil profile depths in Figure 11 and for
different slope angles in Figure 12. The large increase of stability for very small rises in
tension (-1.0 m head = 0.1 bars (10 kPa) tension), and the rapid decrease of stability as pore
water pressures increase show that the hydrological inputs are of significant importance. The
seasonal pattern of soil moistureltension changes (Figure 13) shows that any given slope
would be extremely unlikely to fail during dry season conditions due to the high tensions
experienced within the soil profile, and that even during a very wet season rapid drainage
of the soils result in dominantly unsaturated conditions except during and immediately
following heavy rain (Figure 7). As Figure 11 shows, even very low tensions can
significantly raise the factor of safety of a hillslope, especially on steeper slopes, thus
maintaining stable conditions between potentially critical rainfall events.
Factor of Safety
6 ,--------------------------------------------------------,
~ z=O.5m ---£- z=lm --*- z=2m --B-- z=3m --* z=4m

5
O~-,_r.
-1 -O.B -0.6 -0.4 -0.2 o 0.2 0.4 0.6 O.B 1

Effective hydraulic head (m water)
Figure 11. Variation in calculated slope stability with changing soil hydrological conditions, showing how these are
affected by the soil profile depth.
402 A.P.DYKES
Factor of Safety
2.5 -~
~ 35 deg .•lope -t3-- 45 deg.slope --*- 55 deg .•lope

2.25
-1 -O.B -0.6
-0.4 -0.2 o 0.2 0.4 0.6 O.B 1
Effective hydraulic head (m water)
Figure 12. Variation in calculated slope stability with changing hydrological conditions on slopes of different
inclinations.
Soil suction (kPa)
10,-~ ________~
- Dl (30em) -*- D3 (DOem)
225 245 265 285 305 325 345 363 20 40 60 82 100 120 135
Julian day 1991-92 (13/6/91-14/5/92)

Figure 13. Seasonal soil suction changes measured at tensiometers Dl (30 cm deep) and D3 (90 cm deep) over a
nine month period ending 14 May 1992. Julian day 1 = 1st January; 32 = 1st February, etc.
The rapid destruction of tensions and associated water table build-up illustrated in Figure
7 shows how unstable conditions could be reached in minutes under sufficiently heavy
rainfall. Analysing for a given slope, under known antecedent soil moisture conditions, there
will be a certain range of rainfall intensity profiles for individual storms which would
generate instability. Calculating these for 'typical' wet and dry season conditions, and
comparing them with previous rainfall records in terms of magnitude-frequency relationships,
will ultimately enable the probability of failure of a given slope to be estimated for both wet
and dry season conditions, subject to a satisfactory means of defining these seasons. This
part of the research is still under development.
There are problems with the analysis of the steep lower valley slopes which have not yet
been addressed. The first is specific to this geomorphic situation, and is the model
requirement of an infinitely long, planar hillslope with a shallow slip plane parallel to the
surface, and the assumption that the depth of water table above the slip plane is also constant
along the entire slope length. The development of saturated wedges in the lower parts of a
hillslope, adjacent to a channel or hollow, will result in significantly higher positive pore
water pressures in the lower slope region, but there may be no saturated conditions upslope.
In addition, active channel erosion and undercutting of slope foot locations may remove the
lowest part of the soil profile, leaving an unsupported bottom 'slice' and reducing the overall
stability of the slope by unloading of the toe. Under these conditions a factor of safety can
be calculated using the method of analysis developed by Morgenstern & Price (1965) for slip
surfaces, ground surfaces and water table surfaces of any shape (Bromhead 1988).
The second problem is that of quantifying the effect of vegetation on the stability of the
steepest slopes with respect to shallow failure. "Vegetation that may be growing on a slope
has traditionally been considered to have an indirect or minor effect on stability, and it is
usually neglected in stability analyses. This assumption is not always correct, and for certain
forested slopes with relatively thin soil mantles, it has been shown to be significantly in
error" (Greenway 1987, p. 187).
Greenway's (1987) review considers vegetation effects as being either hydrological or
mechanical, and discusses in particular the impact of root strength, distribution and
morphology on the overall shear strength of the soil mass. There does not yet appear to be
any published data concerning the physical or spatial properties of root distributions for
rainforest tree species, nor on the actual spatial size distribution of trees within natural forest.
In Temburong, it appears that few trees of significant size occurred on the steepest slopes,
and that throughout the forest few roots penetrated more than a few centimetres into the soil,
except perhaps directly beneath some of the largest trees. There was no visible evidence of
roots penetrating the bedrock beneath a soil mantle, although Ashton (1964) reported that this
did occur on ridgetop sites. In terms of the slopes under analysis, it is assumed that no trees
of significant size occur, so the surcharge due to the weight of the trees can be ignored, and
it is also assumed that the reinforcing effect of those roots which do reach the depth of the
bedrock interface is negligible.
The hydrological effects of vegetation are less difficult to estimate, but accurate
quantification remains a considerable problem for hydrological and stability research
(Bruijnzeel 1990). The hydrological factors are either adverse to stability, in that vegetation
promotes soil infiltration capacities by increasing the soil surface roughness and organic
content and by causing dessication cracks to appear in the soil following moisture uptake by
roots, or beneficial to stability, by reducing the amount of rainfall which reaches the soil by
as much as 24% (Vis 1986) and then by removing the water which does enter the soil and
transpiring it to the atmosphere. On the steep hillslopes in Temburong, the role of slope
inclination in reducing infiltration is an unknown but possibly significant factor which may
actually increase the effective stability of the steepest slopes under wet conditions. Soil
moisture losses through evapotranspiration are of great importance in maintaining stable
conditions; in this environment plants are likely to transpire at their maximum potential rate
almost all the time, since soil moisture deficits are not expected to occur in any month on
the basis of rainfall and mean evaporation data discussed earlier.
The relationship between slope stability and the depth of the soil profile above the shear
plane is worthy of further comment. Figure 10 implies that if all other parameters are held
constant, there is a threshold profile depth with respect to stability, in that any increase in
depth beyond that threshold has relatively little influence on the stability of the hillslope. In
terms of the varying hydrological conditions within the profile, it is obvious that a given
water table depth affects a far smaller proportion of a deep profile than of a shallow one.
404 A.P.DYKES
The major implication is in analysing the evolution of the hillslope in terms of material
production by weathering, and material loss by erosion or solution. For a soil profile of any
depth to occur, weathering rates must exceed rates of removal of the weathered products.
As the threshold depth for stability is approached, the probability of failure of the slope
increases, until failure occurs and erosion by mass movement instantaneously exceeds long-
term production. Thus one might expect to find a long-term equilibrium between weathering
and erosion, an equilibrium which cannot be tested for because the timescale over which it
applies is a function of the prevalent weathering rates, but which is concordant with the
earlier discussion of slope development by parallel retreat.
Given that the initial parameter values used in the calculations above have been assumed,
largely on the basis of a back-analysis of one small (approx. 500 m3) landslide, the high
sensitivity of the model to the shear strength of the soil means that any application of results
obtained so far on the basis of assumed parameter values must be regarded as somewhat
tenuous. However, the preceding paragraph provides an approach for analysing the actual
contemporary status of the Temburong landscape in evolutionary terms. Retaining the
assumed parameters used above, for a 45° slope there appears to be a threshold soil profile
depth of about two metres, above which the slope may fail if no other parameter value
changes. The threshold profile depth with respect to failure will be smaller for steeper slopes
and greater for less steep slopes. The return period of rainfall events required to cause a
hillslope of given profile depth to fail, will also need to be calculated and accounted for in
probabilistic terms. If the topographic lowering of the landscape by shallow mass movements
is in equilibrium with channel incision, most soil profiles at any given slope angle would be
expected to be at, or close to, the calculated threshold depth for that slope. Stability analyses
using measured values of soil shear strength will ultimately be performed to predict the
expected hills lope inclination-profile depth relationship for the range of slopes encountered
in the field. These results will be taken to the landscape to see how, or even if, they account
for the observed hillslope forms.
Hills/ope Characteristics in southern Temburong

Hillslope profiles measured in the Belalong and Engkiang valleys in Temburong display
many distinctive features. Twenty-two profiles were measured, and four of these are shown
in Figure 14. The profiles were measured by taking slope angles over lengths of 5 to 15
metres, along a compass bearing at right angles to the line of the valley floor (whenever
possible). The lines of the profiles were chosen in the field in an attempt to avoid distortions
due to tributary valleys. The main effect of the generally high drainage density is to limit
the areal extent of any given valley side hillslope, so that the sides of the main valleys
comprise short steep footslopes with long, undulating, knife-edge ridges, or relatively steep
tributary valleys leading to very steep stream-head slopes.
The profiles in Figure 14 show many of the features encountered in the field. The lower
slopes of the Engkiang valley are the steepest slopes generally, with the mean slope angle
exceeding 48° over a 100+ metre long segment on profile 4E. The hillslopes on the west
side of the valley are less steep, sloping with the dip of the bedrock as in the main Belalong
valley. These lower slopes are either straight or slightly convex, although the lower part of
profile 4W displays a markedly concave section due to an old large, deep-seated landslide.
Many of the lower slopes display more recent signs of shallow failures. The upper slopes
in the Engkiang valley are generally less steep, with segments averaging around 300 but
always with small sections considerably greater than 45°, usually marking the headwalls of
large, deep-seated movements. The slopes of the tributary valleys appeared to be generally
steeper than many of the main valley slopes. These slopes often exceeded 50° as measured
over several tens of metres.
In the main Belalong valley, the eastern slopes are again the steepest, with mean
All slope angles in degrees.

Examples of slopes > 40 degrees are shown.
Large numbers indicate mean inclination of slope segments.
rm Sg. Belalong
47
I
35.6
50
29.2 44
43 33.5
~ 36.8
Sg. Belalong 50
rid,e
42
.Jl00
;etxes
43
41
27.7 54 1 51 37.6
0 100
metres
Figure l4a. Valley side profiles of the Belalong valley just south of the FSC, looking downstream.
44
48
Sg. Engkiang
27.0
44
54
45 50
37.3 45 47 48.7
47 Sg. Engkiang
33.3 43 28.8
45
44 58
47
33.2 47 43 43.8
Figure 14b. Valley side profiles of the lower Engkiang valley, within 2.5 kIn of the Belalong river. Diagram scale
and details as for Figure 14a.
inclinations exceeding 35° in some places. The western slopes rarely average 300, but, as
in the Engkiang valley, all the slopes contain very steep sections (up to 60°) which are either
adjacent to the river channel or at the top of depressions caused by deep-seated mass
movements. Profile 2E includes the slope directly above a stream-head position and clearly
shows the steep character of stream-head locations in this environment.
The observed spatial distribution of major deep-seated landslides on the measured profiles
did not follow any significant pattern, although they were relatively uncommon on the steep,
406 A.P.DYKES
straight slopes adjacent to the main rivers, where shallow failures predominated. At least
31 major deep-seated landslides (e.g. 20-400 metres wide, lengths usually unknown) were
directly encountered whilst measuring the profiles, together with a further 10 smaller failures
and a number of slopes in the Engkiang valley where shallow slips had occurred relatively
recently (e.g. within the previous couple of years).
Shallow failures seem to be confined to the steep, straight slopes adjacent to the channels,
with several recent (1991) examples being noted along the Engkiang, Belalong and
Temburong rivers. They can also be found on the steep upper slopes adjacent to the ridges,
as a late 1992 example near the Field Studies Centre shows. The scar of an enormous
hillside collapse above the confluence of the Engkiang and Belalong rivers was also found
to be retreating by shallow failure, part of it having done so in 1990. In addition, it is likely
that such failures occur in stream-head hollows (Deitrich et al. 1992, Montgomery &
Dietrich 1989), although no obvious examples of this have been encountered so far.
A digital elevation model (OEM) produced from the Geographical Information System
(GIS) database for the research area (Wills, this volume), was analysed to produce a map of
slope angles for the entire area. This approach to landscape stability evaluation is well
established, but is dependent on a good digitised topographic dataset being available (e.g.
Carrara et al. 1991). Despite inaccuracies inherited from the original aerial survey of the
region, from which the OEM is derived, the obtained map of slope angles does show that
the steepest slopes are generally of limited spatial extent, and are concentrated either near
the valley floors, which is not unexpected given the regional context of rapid uplift and
incision which prevails here, or adjacent to the crests of the interfluve ridges. Francis (1987)
notes that "... in areas apparently dominated by steep slopes, low-angles slopes are always
present and often cover a large proportion of catchment areas" (p. 608). The most frequently
occurring slope angle class within the OEM is 25-2<)0, with almost as many slopes of 30-34°.
In other words, the research area is actually dominated by steep slopes.
Provisional Interpretation of Observed Slope Forms

A very crude application of the stability analysis results to the suite of measured slope
profiles would imply that soil profiles on the lower slopes of the Engkiang valley will be
thinner than on the upper slopes of the main valley or in the Belalong valley, and that even
less soil could remain stable on the steepest slopes of the minor tributary valleys, if the
valley slopes are in equilibrium with the contemporary channel positions with respect to
shallow mass movement. If soil profile depths generally differ from the critical predicted
depths over a wide range of hillslopes, then equilibrium between the hillslopes and channels
cannot be assumed, and other processes such as deep-seated mass movement would need to
be investigated. The absence of any quantitative information about soil depth distributions
on the hillslopes in southern Temburong, and the current unavailability of measured soil
shear strength parameters, precludes a more detailed application of the analysis at this stage.
The observed steep and straight or slightly convex lower valley slopes with generally thin
soil profiles, and the concentration of shallow failures along these slopes, suggests that in the
Engkiang valley at least, the lower slopes are probably in equilibrium with the rapidly
incising channel, and that shallow slope failures are encouraged by progressive steepening
of the slope by channel incision. The apparent convexity of the Engkiang valley sides as a
whole suggests that incision by the Engkiang river is currently proceeding more rapidly than
the valley as a whole can adjust to. In the Belalong valley, where the main valley slope
forms are generally straight but where soil profiles are thought to be considerably less deep
than their predicted thresholds using the infinite slope analysis, the slopes may be in
equilibrium with the channel due to deep-seated failures, with shallow landslides occurring
where lateral erosion and undercutting has produced localised steepening of slopes at the
channel edges.
When the work currently in progress reaches the stage of allowing more concrete theories
to be formulated concerning the equilibrium status of the hillslopes in southern Temburong,
a wider view of the evolution of the entire drainage basin will be possible for this regional
setting of tectonic uplift and subsequent landscape denudation beneath a primary rainforest
cover.
CONCLUSIONS
The physical characteristics of the landscape of southern Temburong have been described
and, to some extent, explained in terms of the interrelated hydrological and geomorphic
processes active within the drainage basin system. The tropical climate allows a complete
rainforest cover to flourish, but its seasonality of rainfall can directly influence the temporal
occurrence of major geomorphic events such as landslides.
Analysis of the stability of a hypothetically characteristic 45° slope, and further stability
calculations using the modified infinite slope model, demonstrate the critical importance of
hydrological conditions on the instantaneous stability of the hillslopes, and show that matric
suction in unsaturated soil (soil tension) is probably the most important factor in accounting
for the maintained stability of these hills lopes where the soil profile is close to a critical
threshold depth for the angle of slope. High tensions experienced by the soils during the dry
seasons will prevent failure conditions from being attained during many storms whose
magnitudes could cause failure under wet season conditions.
No firm conclusions can yet be drawn concerning the equilibrium status of any hillslope-
channel units in the research area, but it appears that parallel slope retreat by means of
shallow failure is maintaining steep, straight slope forms in the lower slopes of the Engkiang
valley, and that these slopes are in equilibrium with the Engkiang river channel. This
situation is thought to prevail along most of the tributary valleys in the Belalong and
Temburong catchments.
Many of the geomorphic processes acting on this region are not yet known in any
quantitative sense, and even some of the basic hydrological parameters can not be utilised
with any great degree of confidence. The slope stability work introduced in this paper would
be greatly enhanced by data on spatial variations in soil profile depth, spatial distribution of
tree sizes with respect to topographic and geomorphological features, and some quantitative
data on root size and depth distributions through the soil profiles.
Finally, more general research questions raised by this work include the effect of slope
angle on soil surface processes and the spatial variability of infiltration generally, rates of
weathering and denudation by solution, and the links between mean rates of sediment
production and transportation in the rivers and how they relate to seasonal hydrological
conditions.
ACKNOWLEDGEMENTS
This research was undertaken as part of the Brunei Rainforest Project 1991-92 organised by
Universiti Brunei Darussalam and the Royal Geographical Society. The author would like
to acknowledge the support of the Project's corporate patrons (Royal Brunei Airlines, The
Baring Foundation, Diawa-DICAM, Greencard Trust, Hongkong Bank, Morgan Grenfell and
Nomura-NIMCO) and Brunei Shell Petroleum Sdn Bhd, for their commitment to the Kuala
Belalong Field Studies Centre. In particular, RGS sponsorship of this research is gratefully
acknowledged, together with the logistical support provided in Brunei by C. Prebble, A.
Simpson and LT. Wills (RGS), and Dr D.S. Edwards, Dr Kamariah Abu Salim and Samhan
bin Nyawa (Biology Dept., UBD).
408 A.P.DYKES
The advice and encouragement given throughout by Professor John Thomes (King's
College London) has been invaluable, both in the field and subsequently.
At Kuala Belalong, assistance was provided by many people throughout the 17 months of
fieldwork. In particular, special thanks go to Kuni Anggat, whose dedication and effort
towards the fieldwork for this research during the latter 10 months was of immense value.
Thanks also go to Gawing Anggat, Jali Amba, Simpurai Anggat, Inggul Timbang and
Samsudin Panya, who gave considerable help at various times, with further assistance from
Impin Lasah, Nyua Anggat, Ramlah Sapong, Isah Laju, Mujah Ngambat, Osman Linggi, and
others. Additional help in the field came from Prof J.B. Thomes, J.T. Wills, Dr S.M. Ross,
M. Mulligan, Y. Willcox, D. Buck, A. Simpson, C.A. Pendry, M. Welsh, and members of
2nd Battalion, King Edward's Own Gurkha Rifles. The helicopter was provided by Royal
Brunei Armed Forces.
Additional practical and logistical support in Brunei was provided at UBD by all members
of the Biology Dept., and by Dr. Goh Kim Chuan and Dulima Jali (Geography Dept.). The
Dept. of Civil Aviation provided the Semabat rainfall data.
In the U.K., the practical and logistical help of all the technical/support staff in the
Geography Dept. at King's College London, and Ms C. Prebble and Mrs V. Simonds at the
RGS are acknowledged, as are NERC (Scientific Services) and Mr. M. Walker (IoH) for the
preparation and loan of two automatic weather stations to the Project. Thanks also go to the
many people with whom relevant discussions took place; also to Ha1crow Engineering Ltd.
for providing the Malaysian meteorological data, and to Mr. I.D. Bloomfield for assistance
with the computer programming.
LITERATURE CITED
ASHTON, P.S. 1964. Ecological studies in the Mixed Dipterocarp forests of Brunei state. Oxford Forestry Memoirs
No. 25. Clarendon Press, Oxford.
BONELL, M. & GILMOUR, D.A. 1978. The development of overland flow in a tropical rainforest catchment.
Journal of Hydrology 39:365-382.
BONELL, M., GILMOUR, D.A. & CASSELLS, D.S. 1983. Runoff generation in tropical rainforests of northeast
Queensland, Australia, and the implications for land use management. In Hydrology of Humid Tropical Regions
with Particular Reference to the Hydrological Effects of Agriculture and Forestry Practice. (proceedings of the
Hamburg Symposium, August 1983). lARS Publ. No. 140.
BROMHEAD, E.N. 1978. Large landslides in London Clay at Herne Bay, Kent. Quarterly Journal of Engineering
Geology 11:291-304.
BROMHEAD, E.N. 1988. Slope Stability Analysis Package: User's Manual. Unpublished handbook.
BRUIJNZEEL, L.A. 1983. Evaluation of runoff sources in a forested basin in a wet monsoonal environment: a
combined hydrological and hydrochemical approach. In Hydrology of Humid Tropical Regions with Particular
Reference to the Hydrological Effects of Agriculture and Forestry Practice. (proceedings of the Hamburg
Symposium, August 1983). IAHS Publ. No. 140.
BRUIJNZEEL, L.A. 1990. Hydrology of moist tropical forests and effects of conversion: a state of knowledge
review. International Hydrological Programme (UNESCO), Paris. 224 pp.
CARRARA, A., CARDINALI, M., DETTI, R., GUZZETTI, F., PASQUI, V. & REICHENBACH, P. 1991. GIS
techniques and statistical models in evaluating landslide hazard. Earth SUlface Processes and Landforms 16:427-
445.
CHANDLER, RJ., 1982. Lias Clay slope sections and their implication for the prediction of limiting or threshold
slope angles. Earth SUlface Processes and LandfomlS 7:427-438.
CHORLEY, R.J., SCHUMM, S.A. & SUGDEN, D.E. 1985. Geomorphology. Methuen, London. 605 pp.
CHOWDHURY. R.N. 1978. Slope Analysis (Developments in geotechnical engineering; 22). Elsevier, Amsterdam.
423 pp.
COLLINS, N.M., SAYER, J.A., & WHITMORE, T.C. 1991. The Conservation Atlas of Tropical Forests: Asia
and the Pacific. Macmillan, London.
CROZIER, MJ. 1986. Landslides: causes, consequences and environment. Croom Helm, London. 252 pp.
DAY, M.J. 1979. Landslides in the Gunong Mulu National Park, Sarawak, E. Malaysia. Unpublished report to the
Royal Geographical Society (Mulu Expedition Library).
DIETRICH, W.E., WILSON, CL, MONTGOMERY, D.R., McKEAN, 1. & BAUER, R. 1992. Erosion thresholds
and land surface morphology. Geology 20:675-679.
FRANCIS, S.C. 1987. Slope Development Through the Threshold Slope Concept. Pp. 601-624 in Anderson, M.G.
& Richards, K.S. (eds). Slope Stability. Wiley, Chichester.
FREDLUND, D.G. 1987. Slope Stability Analysis Incorporating the Effect of Soil Suction. Pp. 113-144 in
Anderson, M.G. & Richards, K.S. (eds). Slope Stability. Wiley, Chichester.
GRANT, CK. 1984. Geographical Overview. Pp. 11-33 in James, D.M.D. (ed.). The Geology and Hydrocarbon
Resources of Negara Brunei Darussalam. Muzium Brunei, Bandar Seri Begawan.
GREENWAY, D.R. 1987. Vegetation and Slope Stability. Pp. 187-230 in Anderson, M.G. & Richards, K.S.
(eds). Slope Stability. Wiley, Chichester.
HILLEL, D. 1971. Soil and Water: Physical Principles and Processes. Academic Press, Inc., New York. 288 pp.
JAMES, D.M.D. 1984. Regional Geological Setting. Pp. 34-42 in James, D.M.D. (ed.). The Geology and
Hydrocarbon Resources of Negara Brunei Darussalam. Muzium Brunei, Bandar Seri Begawan.
KIRKBY, M.L 1985. Hillslope hydrology. Pp. 37-75 in Anderson, M.G. & Burt, T.P. (eds). Hydrological
Forecasting. Wiley, Chichester.
MALMER, A. 1992. Water-yield changes after clear-felling tropical rainforest and establishment of forest plantation
in Sabah, Malaysia. Journal of Hydrology 134:77-94.
MONTGOMERY, D.R. & DIETRICH, W.E. 1989. Source areas, drainage density, and channel initiation. Water
Resources Research 25:1907-1918.
MOORE, I.D., MACKAY, S.M., WALLBRINK, P.J., BURCH, G.J. &O'LOUGHLIN, E.M.1986. Hydrological
characteristics and modelling of a small forested catchment in southeastern New South Wales. Pre-logging
conditions. Journal Of Hydrology 83:307-335.
MORGAN, R.P.C. 1976. The Role of Climate in the Denudation System: a Case Study from West Malaysia. Pp.
317-343 in Derbyshire, E. (ed.). Geomorphology and Climate. Wiley, London.
MORGENSTERN, N .R. & PRICE, V.E. 1965. The analysis of the stability of general slip surfaces. Geotechnique
15:79-93.
PEARCE, A.J. & O'LOUGHLIN, CL 1985. Slope Stability and Land Use. Water Resources Monograph 11.
American Geophysical Union.
ROSS, S.M. & DYKES, A.P. (this volume). Soil conditions, erosion and nutrient loss on steep slopes under mixed
dipterocarp forest in Brunei Darussalam.
SCHUMM, S.A. 1963. The disparity between present rates of denudation and orogeny. US Geological Survey
Professional Paper 454-H.
SKEMPTON, A.W. 1964. Long-term stability of clay slopes. Geotechnique 14:77-102.
SKEMPTON, A.W. & DELORY, F.A. 1957. Stability of natural slopes in London Clay. Pp. 378-381 in Proc.
4th Int. Con! of Soil Mechanics & Foundation Engineering VoL 2. London.
STRAHLER, A.M. 1950. Equilibrium theory of erosional slopes approached by frequency distribution analysis.
American Journal of Science 248:673-696, 800-814.
TROENDLE, CA. 1985. Variable source area models. Pp. 347-403 in Anderson, M.G. & Burt, T.P. (eds).
Hydrological Forecasting. Wiley, Chichester.
VIS, M. 1986. Interception, drop size distributions and rainfall kinetic energy in four Colombian forest ecosystems.
Earth Surface Processes and Landforms 11:591-603.
WESLEY, LD. 1977. Shear strength properties of halloysite and allophane clays in Java, Indonesia. Geotechnique
27:125-136.
WEYMAN, D.R. 1970. Throughflow on hillslopes and its relation to stream hydrographs. International Association
Of Scientific Hydrology. XV Annee, No.3, September:25-33.
WILFORD, G.E. 1961. The Geology and Mineral Resources of Brunei and adjacent parts of Sarawak with
descriptions of Seria and Miri Oilfields. Memoir 10, Geological Survey Department, British Territories in Borneo.
319 pp.
WILLGOOSE, G., BRAS, R.L & RODRIGUEZ-ITURBE, L 1991. Results from a new model of river basin
evolution. Earth Surface Processes and Landforms 16:237-254.
WILLS, J. T. (this volume). The creation of a Geographical Information System for the Kuala Belalong Field Studies
Centre.
YOUNGS, E.G. 1991. Infiltration measurements - a review. Hydrological Processes 5:309-320.
The effects of timber harvesting and forest conversion on peat swamp
forest dynamics and environment in West Kalimantan
HERUJONO HADISUPARTO
Department of Forestry, Faculty of Agriculture, Tanjungpura University, Pontianak, West

Kalimantan, Indonesia
ABSTRACT. The peat swamp forests of Kalimantan, formerly covering 6.3 million hectares, are
found primarily along the western and southern parts of Indonesian Borneo and are separated by
the Sunda shelf from those of Sumatra's east coast. Studies conducted on production forests in
West Kalimantan indicate that the presence of natural regeneration stages favour continued
sustainability. Silvicultural practices ensure sustainability when the residual trees left after logging
are adequate and there is protection of the resource from unplanned relogging and forest
conversion.
KEY WORDS: peat swamp forest, sustainability, silviculture, forest population dynamics.
INTRODUCTION
The remaining tropical rain forests in the world are largely concentrated in three regions,
namely America (Amazon basin), Africa (Zaire and Congo) and Indo-Malaya (primarily in
South East Asia). Among the three regions, the Indo-Malayan rain forest contains the greatest
richness in species (Van Steenis 1971, Whitmore 1985). The Indo-Malayan rain forest region
includes different forest formations such as dry-land rain forest, wet-land rain forest and
monsoon forest.
Peat swamp forest is a very distinct forest formation of wet-land rain forest, which is
generally found on flat terrain, along with fresh water swamp forest and mangrove forest.
According to Jacobs (1988) peat swamp forest is a successional stage of fresh water swamp
forest which is characterised by having soils with a low pH, lack of minerals, and very slow
natural decomposition of the litter.
In Southeast Asia, extensive peat swamp forests are found in western and southern parts
of Borneo, primarily in West Kalimantan (West Indonesian Borneo) in the down-stream area
of large drainage basins. Such forests are also found in eastern Sumatra and Malaya.
Anderson (1963) indicated that peat swamp forests in Sarawak and Brunei occupy nearly the
entire coastline and the largest, Maludam Peninsula peat swamp, stretches inland up to 64
km. Two of the most important timber species in these forests are Shorea albida, a Bornean
endemic with a range extending from north-west Kalimantan to Brunei and Gonystylus
bancanus, one of the most valuable endemic timber species of West Kalimantan. This latter
species is endangered as a result of logging and forest conversion.
This paper describes the effects of timber harvesting and conversion on peat swamp forest
dynamics and environment with special reference to West Kalimantan. Field data included
here was obtained from ecological studies conducted primarily by the staff and students of
the Department of Forestry - Faculty of Agriculture at the Tanjungpura University in
Pontianak.
411
D. S. Edwards et al. (eds.), Tropical Rairiforest Research - Current Issues, 411-415.
412 HERUJONO HADISUPARTO
EFFECTS OF LOGGING ON PEAT SWAMP FOREST DYNAMICS
Total forest land area in West Kalimantan according to the Recognised Forest Land Use
Survey (Menteri Pertanian 1982) comprises 9,204,425 hectares (ha) or 62.5 % of the total
provincial area. It consists of protection forest (2,047,125 ha) parks and forest reserve
(1,339,880 ha) and production forest in all types of formation totalling 5,817,420 ha. This
includes an assigned convertible production forest area of 1,505,670 ha. Peat swamp forest
area in the province makes up 1,878,200 ha including a total of 857,460 ha which has been
designated as protected (Tim RSTRP-Kalbar 1991).
Timber harvesting of tropical rainforest in Indonesia following Foreign Investment Law
was probably initiated in the peat swamp forest. In West Kalimantan, semi-mechanical
logging using small railroads was introduced by several foreign loggers coming from
Sarawak (Malaysia) in the late sixties. The timber trade boom of the early eighties has caused
significant changes in forest structure and composition. The application of the Indonesian
selective Cutting System (TPIITPTI) has not been successfully practiced. In order to attain
forest sustainability damage to residual growing trees and regeneration stages should be
avoided by preventing unplanned re-logging and uncontrolled forest conversion.
Ecological studies on peat swamp forests were conducted in several selected locations in
West Kalimantan. Study locations included: Mandor nature reserve (unlogged forest);
logged-over forest of Sukadana; logged-over forest of Air Hitam near Pontianak; production
forest of Nanga Tayap (Ketapang). These four sites represent different facets with respect
to forest population dynamics, as summarized in the following paragraphs.
Unlogged peat swamp forest (Mandor Nature Reserve)

This peat swamp forest is located on the fringe of kerangas (heath) forest within the nature
reserve. The peat swamp forest was found to be more diverse than the kerangas forest (Table
1). These two adjacent habitats had a relatively high species similarity of poles (SS index =
0.64); on the other hand both habitats showed a lower species similarity of trees (SS index
= 0.30). This indicates that disturbance had occurred in the nature reserve and that the
vegetation is still in the process of succession towards a climax forest (Hadisuparto &
Syafruddin 1990). There were 86 species of poles and trees in the peat swamp forest,
whereas in the kerangas 75 species were found. species with the highest importance value
indexes in the peat swamp forest were Dryobalanops beccarii, Xylopia malayana, Calopyllum
sp. and Gonystylus bancanus, whereas in the kerangas forest the highest values were for
Tristania sp., Dryobalanops beccarii, Shorea pachyphylla and Calophyllum sp. Forest gaps,
in terms of canopy openings, was found to be 10.5 % lower in peat swamp forest compared
to Kerangas. The natural forest community of peat swamp forest in Mandor nature reserve
consisted of 14.7% seedlings, 54.7% saplings, 20.0% poles, and 10.7% trees. This suggests
that a seasonal natural regeneration has occurred which may result in a normal forest stand
in the future.
Table 1. Species diversity and similarity indexes between peat swamp and kerangas forests at Mandor Nature
Reserve.
Diversity Index Similarity Index

Forest type Pole Tree Pole Tree
Kerangas 1.33 0.67 0.64 0.30
Peat swamp 1.42 0.73
Peat swamp forest dynamics in Kalimantan 413
Table 2. Species number and similarity indexes between unlogged and logged-over forest of Sukadana. From
Hadisuparto et at. 1990.
Regeneration stage Number of species Similarity Index (%)

Unlogged Logged-over
Seedling 61 40 49.9
Sapling 59 52 46.8
Pole 53 35 33.7
Tree 73 32 31.3
Logged-over peat swamp forest of Sukadana

This forest is growing on an alluvial plain and overlies white sand deposits (Leighton,
1990). It is located between a low-land mixed dipterocarp forest and brackish water swamp
and mangrove forests. This peat swamp forest was designated as production forest and was
logged more than 10 years ago. The study compares the residual growing stand with the
unlogged forest in the area.
In the unlogged forest seedlings and saplings were dominated by Diospyros sp. while trees
were dominated by Dipterocarpus sp., whereas in the logged-over forest Gluta renghas,
Diospyros sp. and Litsea sp. were dominant. Ramin (Gonystylus bancanus) was not found
at any stage of its growth in the logged-over forest. However, the species similarity indexes
between unlogged and logged-over forests indicate that naturally regenerating seedling and
saplings of many species grew commonly in both unlogged and logged-over forest. In other
words, the forest stand could be restored gradually if no further disturbance or interference
of the successional stages occurred.
Logged-over forest of Air Hitam

The study conducted here included the analysis of vegetation following timber harvesting.
Sample plots were located in 5-year and 1O-year logged-over forests. The results indicate
that 5 and 10 years after logging, recovery is occurring, although some changes in forest
structure and composition might have occurred. Table 3 suggests that the residual stand was
sufficient to provide enough stock of each regeneration stage from seedling to tree. Dominant
species present in both logged-over forests (with the highest importance value indexes) were:
Gonystylus bancanus, Dryobalanops sp., Litsea sp., Gluta renghas and Cratoxylon
arborescens. Non-dipterocarp species dominate the forest community in this peat swamp
formation.
Nanga Tayap production forest

The effects of timber harvesting in this peat swamp forest formation was studied by means
of vegetation analyses before and after logging. This 'Ramin Forest' was assessed through
Table 3. Forest dynamics on logged-over peat swamp forest of Air Hitam.
Regeneration stage 5-year logged- over to-year logged-over

No. species Stems/ha No. species Stems/ha
Seedling 24 28375 26 38938
Sapling 24 2790 23 3600
Pole 24 830 26 915
Tree 25 182 25 188
414 HERUJONO HADISUPARTO
Table 4. Population dynamics ofRamin (Gonystylus bancanus) on the peat swamp production forest of Nanga Tayap.
Regeneration stage Stems/ha Stems/ha

Before logging After logging
Seedling 6450 5600

Sapling 656 488
Pole 26 18
consideration of its dominant species (Table 4). Ramin (Gonystylus bancanus) is the most
valuable timber in the peat swamp forest which may have been seriously affected by logging
practices in West Kalimantan. Natural regeneration of this species has been considered more
successful than artificial planting. There are still adequate seedlings and saplings in the forest
following the logging, although poles are relatively rare, indicating heavy damage during
logging. Silvicultural practices aimed at maintaining natural regeneration are needed to ensure
future standing stock of this valuable species.
EFFECTS OF FOREST CONVERSION ON ENVIRONMENT
Peat swamp forest in Kalimantan (Indonesian Borneo) formerly covered 6.3 million ha. Like
fresh water swamp forest, the peat swamp forest has been rapidly cleared for agriculture and
other land uses. The acceleration of peat swamp forest conversion in Kalimantan is primarily
linked to the geographical distribution of the forests which are mostly along rivers and on
coastal plains, hence they are more readily exploited for timber. Increased transportation
facilities in such areas has promoted settlement and other land uses and has exacerbated the
situation.
In West Kalimantan, the peat swamp forest comprises 1,878,200 ha or about 12.8% of
the province. The forest extends from the north to south to follow the coast behind the
brackish waters. The ecological role of peat lands, especially in conserving water and
controlling flooding in down-stream areas, was given due attention and through a Presidential
Decree, peat swamp forest with a peat depth of 3 or more meters was assigned as protected
areas. In accordance with the Presidential Decree, 45.7 % of the peat land area in West
Kalimantan was designated as protected forest, which can only be cultivated with a special
release by the government. However, assignment to protected status is not easy and is not
instantly applicable because it requires a comprehensive study on peat depth all over the
forests.
The remaining peat swamp forests have been continuously threatened by local migrants
who wrongly consider that peat lands in their forested state are less useful (economically)
than cleared land. After the forests are cleared however, it is found that the peat cannot be
cultivated directly and as a result, people would rather continue exploiting the residual trees
of the remaining forests. A very degraded forest is finalIy cleared by burning (in the dry
season) and drainage channels made to allow cultivation. Deeper peats have recently become
a commercial target as energy sources and such use will eventually destroy the environment
completely if there is no integrated central planning followed by proper implementation.
Population growth and the need for land in Indonesia, both for agriculture and settlement,
in some parts through transmigration projects, has to be carefully planned. Conversion of
peat swamp forests should be based on firm ecological principles. Peat swamp forest is
known to be a renewable resource but once it is cleared, the peat will eventually degrade and
the site cannot return to the original state. It becomes non-renewable as a natural forest
habitat.
Peat swamp forest dynamics in Kalimantan 415
CONCLUSION
The peat swamp forest of West Kalimantan is one of the largest natural ecosystems in the
Indonesian rainforest. The forests extend along the coastal plain behind the brackish water
and along the rivers inland within the upstream section of Kapuas river basin.
The effects of timber harvesting in this distinctive forest formation have changed
significantly the forest structure and composition. However, studies show that forest
sustainability may be attained if sufficient residual growing trees are left standing and if
enough natural regeneration occurs, provided that unplanned re-logging and uncontrolled
conversion of the forest can be avoided. Ramin (Gonystylus bancanus), the dominant species
of the peat swamp forest in this area, has possibly become an endangered species, with
natural regeneration considered more successful than artificial planting for this species. The
remaining peat swamp forests have been threatened by increasing population pressure, mainly
by migrants.
Forest conversion has been accelerated since most people consider that peat lands in their
forested state are less useful than if they are cleared. Conversion of peat swamp forest to
other uses should be based on ecological as well as economic principles.
ACKNOWLEDGEMENTS
I would like acknowledge the work of our staff and students for the preparation and field data
collection; in particular:- Mr. Abdurrani Muin, Mr. Syafruddin Said, Ms. Augustin
Lumangkun, Mr. Dirhamsyah, Mr. Kambodja, Ms. Ina Kartini and Mr. Agus Purnomo.
LITERATURE CITED
ANDERSON, J .A.R. 1963. The flora of peat swamp forests of Sarawak and Brunei including a Catalogue of all
recorded species of flowering plants, ferns and fern allies. Gardens' Bulletin. Singapore 20:131-228.
HADISUP ARTO, H. & SY AFRUDDIN S. 1990. Species diversity of two adjacent minor habitats in Mandor nature
reserve of West Kali11lllntan. Paper presented at the International Conference on Forest Biology and Conservation
in Borneo. Kota Kinabalu, Sabah, Malaysia. July 30 - August 3, 1990.
HADISUPARTO, H., ABDURRANI M., LUMANAGKUN, A. & DIRHAMSYAH. 1990. Evaluasi beberapa
plas11lll nut/ah pada hutan produksi di Kalimantan Bara(. (Evaluation of several genetic resources on selected
production forests in West Kalimantan). Kerjasama Kantor Menteri Negara KLH & PSL Univ Tanjungpura,
Jakarta.
JACOBS, M. 1988. The tropical rainforest. Springer Verlag, London. 295 pp.
LEIGHTON, M. 1990. Ecological interactions between plants and vertebrates: Comparison among rainforest
habitats. A report on several research topics of Gunung Palung Projects.
MENTERI PERTANIAN. 1982. Penunjukan areal hutan di Wilayah Propinsi dati I Kali11lllntan Bara( sebagai
kawasan hutan (TGHK). (Assigned forest areas for forest land use in West Kalimantan Province). SK. No.
757/Kpts/Um1l01l982, Jakarta.
STEENIS, C. G. G.J. van. 1971. Plant conservation in Malaysia. Bulletin du Jardin Botanique National de Belgique.
41:189-202.
TIM RSTRP-KALBAR. 1991. Rencana Struktur Tata Ruang Propinsi Kali11lllntan Barat Tahun 2005. (Land use
planning of West Kalimantan province towards 2005). Pemda Tkt. I. Kal. Barat.
WHITMORE, T.C. 1985. Tropical rainforests of the Far East. 2nd ed. Clarendon Press, Oxford. 352 pp.
Socio-economic values and tropical rain forest conservation
PETER EATON
Universiti Brunei Darussalam, Bandar Seri Begawan 2028, Brunei Darussalam
ABSTRACT. Many threats to tropical forests and their biodiversity result from social and economic
factors. Policies to conserve the forests depend on being able to demonstrate their economic value and
on social measures involving local community participation. Research to calculate the total economic
value of the forests has involved estimating the direct use value of forest products, the indirect values
from environmental benefits and services, and option and existence values. Threats to the forest have
endangered the social and economic welfare of many indigenous groups; their survival and the
conservation of the forests can often be achieved by reinforcing traditional systems of land tenure and
environmental management.
KEY WORDS: Borneo, conservation, national parks, New Guinea, tropical forest, valuation.
INTRODUCTION
Global concern over issues such as the depletion of tropical rain forests and the preservation
of their biodiversity formed a major part of the agenda of the United Nations Conference on
Environment and Development held in Brazil in 1992. Products of the conference resulted
in a Convention on Biological Diversity and a statement of principles for the management
of forests. These and other significant documents such as Agenda 21 and the Declaration on
Environment and Development have emphasised the need for the sustainable development,
conservation and rational uses of these forests. Whether the good intentions, rhetoric and
compromises of the conference will provide the foundations for effective policies and actions
still remains to be seen, but one point of general agreement in the discussions there and
elsewhere is that most of the problems associated with the tropical rain forests are social or
economic in their origins. If, for example, one analyses the reasons for deforestation in
Southeast Asia, immediate causes which are usually identified are commercial logging,
shifting cultivation and conversion of forest to other types of land use. However, these in
turn are largely a result of other problems such as excessive population growth, unequal
patterns and terms of trade between industrial and developing countries, and the financial
operations of political and economic elites. In addition, there are often inappropriate systems
of land tenure which may both fail to protect the customary rights of indigenous forest
dwellers and also force the landless and hungry into clearing forest on marginal land
basically unsuited to long-term agriculture.
In the analysis of such problems and the formulation of policies to solve them, the social
sciences necessarily have a key role to play. However, it should be emphasised that in issues
of tropical rain forest management and conservation there is no place for "intellectual
marking of territory" (Machi is 1992, p.2) by anyone discipline, but instead there is a need
for cooperation and sharing of knowledge, not only between social and natural scientists, but
also involving foresters, protected area managers and the indigenous people who are directly
dependent on the resources of the forest.
Policies to conserve tropical rain forests depend on being able to demonstrate their
economic value to governments, their communities and rural populations. In the long term,
to be successful they also depend on the support and participation of these local communities.
417
D. S. Edwards et at. (eds.), Tropical Rainforest Research - Current Issues, 417-423.
418 PETER EATON
In this paper I shall discuss some of the methods used by economists in attempting to
measure economic values of the tropical rain forests, and then go on to consider measures
for local involvement in their conservation.
THE TOTAL ECONOMIC VALUE OF TROPICAL RAIN FORESTS
Attempts to assess the total worth of tropical forests have generally been based on both
marketable and non-marketable values. The most comprehensive calculations include direct
and indirect use values, and in addition what are termed option, bequest and existence values.
Direct uses involve the consumption of forest products, including use for recreation, research
and education. Indirect uses include vital environmental functions such as watershed
protection and nutrient cycling. In addition, the assessment of total economic value should
encompass option values, the amount individuals and society are willing to pay for future
uses of the forests, and bequest values which relate to the amount the present generation is
willing to pay to preserve use for their descendants and future generations. Finally,
allowance should be made for existence values that represent how much individuals (or
collectively society) are prepared to pay for an area to be preserved, even though they are
not going to use it themselves.
Methods of calculating these different environmental values vary and some are necessarily
more subjective and imprecise than others. In the case of commercial timber and forest
products it is possible to calculate a market value. For subsistence production and
consumption it may be possible to estimate substitute values from these products or similar
goods bought commercially. Changes in productivity, the cost of defensive expenditure,
replacement costs, and the use of shadow projects that provide substitute environmental
services, can all be used as means of evaluating environmental changes and resource values.
Surrogate markets can be used to estimate the value people place on forest conservation
areas; entry fees, foregone earnings and travel costs providing an indication of how much
people are prepared to pay to visit these areas. Option, bequest and existence values are
more difficult to reliably assess and often rely on contingent valuation techniques by which
people indicate how much they are willing to pay to prevent degradation to an environmental
asset.
Examples of the economic valuation of environmental assets provide persuasive arguments
for conservation. In the case of the Amboseli National Park of Kenya it as been estimated
that each lion was worth US$27,OOO a year in tourist earnings, and that the earnings of the
park averaged about US$40 per hectare, far more than could be produced from agriculture
(Western 1984). For a tropical rain forest reserve in Costa Rica research using travel cost
Table 1. Classification of the total economic value of tropical forests. (Sources: Munasinghe 1992, Pearce 1990).
DIRECT USE INDIRECT USE OPTION VALUE EXISTENCE VALUE

Sustainable timber Nutrient cycling Future Uses Value from knowledge
products of continued existence
Non-timber products Watershed protection Biodiversity Heritage
Medicine Air pollution reduction Conserved habitats Endangered species
Plant genetics Micro-climatic
functions
Recreation Carbon store
Education
Research
Tropical rainforest conservation 419
Table 2. Valuation Methods Relevant to Tropical Rain Forests.
Market value of products and services

Market value of surrogate products
Changes in productivity
Defensive expenditure
Rep lacement costs
Shadow projects
Travel costs
Contingent valuation - willingness to pay
Table 3. Financial returns from different forest uses, US$lha (Source: Peters et al. 1989)
Non. timber harvesting - fruit and latex - 6330

Clear-cutting timber harvesting - 1001
Plantation timber harvesting and pulpwood - 3184
Cattle ranching - 2960
methodology suggested a value for the reserve of US$I,250 per hectare (Tobias &
Mendelsohn 1991). The change-in-productivity technique was used in a study of Amazonian
forest in Peru (Peters et at. 1989); the results of this study can be seen in Table 3 and are
interesting in that they show that the financial returns from harvesting non-timber products,
mainly wild fruit and latex, are greater than alternative land and resource uses such as clear
felling, forest plantations, and conversion to cattle ranching.
Valuation methods have direct relevance to benefit-cost analysis of conservation and
development projects. A study was made of the Khao Yai national park in Thailand (Dixon
& Sherman 1990) that calculated benefits from tourism, wildlife preservation, and watershed
protection against management expenses and foregone opportunity costs from other types of
land use such as logging and agriculture. The conclusion of this study was that although the
park was 'socially beneficial', government intervention was needed for resource protection
as the directly appropriable benefits from exploitation would outweigh the diffused benefits
and returns from protection if the area was left to private ownership and market forces.
Here. in Brunei Darussalam, a national park is being developed in the Ulu Temburong
area where the Kuala Belalong Field Studies Centre is located. It is possible to assess the
potential benefits from this park under a number of different headings which include
biodiversity and species conservation, watershed protection, research and education,
recreation and tourism, the use of forest products on a sustainable basis, and the provision
of local employment and income.
Not all the types of resource use listed in Table 4 are compatible, and this poses problems
for park planning and management. This can be partly solved by a system of zoning which
restricts certain uses to particular areas, separating for example strictly conservation and
research uses from recreation and tourism. A zoning plan which has been proposed for the
park provides for an intensive use administrative area near the confluence of the Temburong
and Belalong rivers, a medium intensity use recreation area for outdoor activities along the
Temburong river, a low intensity wilderness zone, a scientific research and education area
in the catchment area of the Belalong river, and security and restricted zones near the
international boundaries.
Brunei is fortunate in that it has been able to conserve a large proportion of its forests,
and at present there are few economic pressures to exploit its timber resources. Elsewhere
in Southeast Asia the arguments for commercial logging have often proved more persuasive
than those for conservation areas. The economically quantifiable benefits listed in support
420 PETER EATON
Table 4. Potential benefits from the Ulu Temburong National Park.
I. Preservation of biodiversity, species and ecosystems.

2. Preservation of ecological processes.
3. Watershed protection.
4. Scientific research.
5. Recreation and tourism.
6. Environmental education.
7. Local employment and income.
8. Regional/district development.
9. Sustained local use of forest products.
10. Aesthetic, spiritual and cultural.
of logging have included the timber industry's contributions to the gross domestic product,
foreign earnings and government revenue. If, for example, we look at Brunei's neighbouring
East Malaysian states, we find that during the past twenty years royalties and taxes from the
timber industry in Sarawak and Sabah have formed a large proportion of these states'
government revenues. It can, however, be argued that much of the income from the
exploitation of timber resources has not been recovered by the states, and has not been
adequately redistributed among the rural indigenous populations, whose forest resources have
often been destroyed or degraded. There is, for instance, considerable illegal logging in both
states which does not come into the national accounts or contribute to revenue. Even using
the official statistics, the proportion of the profits or resource rent which goes to the state
is much lower than we have been led to believe. One study of resource rent recovery
(Vincent 1990) gave a figure of only 18.4 per cent for Sarawak and 46.2 per cent for
Sabah. This and other analyses (Repetto & Gillis 1988) would seem to indicate that much
of the economic benefit from logging in Southeast Asian countries has been enjoyed by a
very small section of the population rather than spread among the community as a whole.
Conventional national accounting, the compilation of national income and gross domestic
product, has not in the past made any allowance for the destruction or depletion of natural
resources, or for the defensive expenditures necessary as a result of environmental
degradation. These have been pioneered by the governments of France, Norway and Canada
(Pearce et al. 1989), but have also been applied to developing countries and the loss of
tropical forests. Estimates of the depletion of forest, soil and petroleum resources between
1971 and 1984 have shown that, although the gross domestic product was calculated to have
grown at an average annual rate of 7.1 per cent during this period, if allowance had been
made for the loss of natural resources then the average annual growth rate would have been
only 4 per cent (Repetto et al. 1989). This type of national accounting which measures
losses of natural capital is necessary if governments are to plan for the sustainable use of
tropical rain forests.
A form of direct use of the forests which is frequently underestimated by economists is
that of non-timber forest products. In Borneo these have always been of great significance
to the indigenous people, both as a vital part of the subsistence economy and as the basis for
early trade. The great range of plant and animal products has been an indication of both the
rich biodiversity of the forests and the traditional knowledge and skills which have been
applied to their utilisation.
The variety of forest products was recorded in many of the early descriptions of Borneo
and its inhabitants (Hose & McDougall 1912, Low 1948). The role of the forest in the
subsistence economy has also featured in the work of more recent anthropologists, for
example the studies of the Iban by Freeman (1955) and Sandin (1967).
Ethnobotanical studies have spanned the boundaries between the natural and social
sciences, showing that rural communities still use a large number of forest products for food,
handicrafts, construction, medicines and ceremonies. A survey of four Iban longhouses in
the Pantu area of Sarawak (Pearce et al. 1987) identified 142 plant species as being collected
for food and other purposes. Research into traditional medicines has provided further
evidence of the utilisation and potential value of the forests. A study of three Kenyah
villages on the Apoh Kayan plateau of Kalimantan found 213 different plant species that the
local people used for medical purposes (Leaman et al. 1992). Another study of one village
in Central Kalimantan identified 107 medicinal plants as being in general use (Riswan et al.
1992). In the Gunung Mulu area of Sarawak 38 species of plants were identified as being
used for medicinal purposes by the Penans (Ked it 1982). The tropical rain forests also serve
as the habitat for a variety of wildlife species, some of which are important as a source of
food and significant to the local economy. Their value is again difficult to quantify, but one
study in Sarawak (Caldecott 1986) has estimated the total worth of the wildlife harvest as
being between M$83 million and M$115 million each year.
LOCAL INVOLVEMENT IN TROPICAL RAIN FOREST CONSERVATION
The importance of non-timber products to rural indigenous communities living in or near the
forests provides them with an important motive to attempt to conserve forest resources, and
also to maintain and protect their rights to use them. Experience has shown that efforts to
preserve the forest by establishing national parks, reserves or other types of protected area
are unlikely to gain support if these rights are ignored and excluded. Conservation often
proves more successful if local people are involved, if they feel it is to their advantage and
that their traditional rights are being protected. It has also been found that traditional forms
of land tenure and resource management may prove more effective in protecting the
environment than introduced laws and institutions. The World Commission on Environment
and Development recognised these principles when it advised: "The recognition of traditional
rights must go hand in hand with measures to protect the local institutions that enforce
responsibility in resource use" (WCED 1987, pp.115-6).
In Southeast Asia local involvement and the protection of traditional rights have not
always been a feature of conservation policy. In Thailand and Indonesia the rapid expansion
of national parks and reserves has often resulted in conflict with resident populations and
competing land use demands. The Gunung Leuser Park in Sumatra is an example of
continuous encroachment through illegal logging and farming (Wells et al. 1990). The
forests of the Khao Yai Park were reduced by 10 per cent during the first twenty-five years
of the park's existence (Dixon & Sherman 1991, p.l00); further north in Thailand it has
been estimated that 43 per cent of the Doi Inthanon Park has been affected by clearing for
agriculture and fuelwood collecting (Dearden et al. 1992).
More success has been achieved where parks have been shown to be economically
beneficial to local people, especially through providing employment and income from
tourism. This seems to have been a factor in creating support among Kadazan villagers for
the Kinabalu Park in Sabah (Jacobson 1991), and has helped to create more favourable local
attitudes to the Khao Yai Park (McNeely 1988, pp.153-5). The creation of linkages between
protected areas and local economies has often helped to promote both rural development and
conservation, as happened in the case of the Dumoga-Bone Park in Sulawesi when it was
perceived that the protection of the forests was ensuring a continuous supply of water for
irrigation projects in neighbouring areas (Wells et al. 1990).
422 PETER EATON
Although certain traditional rights to hunt and collect forest products may be allowed in
parks, there has generally been a reluctance in Southeast Asia to encourage conservation
based on customary tenure. Where indigenous groups have attempted to protect forest
resources this has often led to conflict with the government, as in the case of the Penans,
Kayans and Kelabits who blockaded the logging roads in Sarawak (Sahabat Alam Malaysia
1989, Hong 1987). In order to find a network of protected areas based on traditional tenure,
it is necessary to examine conservation in Papua New Guinea, a country on the periphery of
the Southeast Asia region.
In Papua New Guinea 97 per cent of all land is held under customary systems of tenure,
under which rights to use land and forest resources depend on membership of the kinship
group or clan. Traditional rules and communal controls have in the past helped to protect
resources by excluding outsiders, imposing seasonal bans to prevent overhunting, and
protecting certain species and areas by taboos (Morauta et al. 1982). Over three-quarters of
the country is still covered by forest, and its products and wildlife remain important to the
subsistence economy. Social and economic changes have, however, placed increased
pressure on these resources as have technological innovations, such as the replacement of the
stone axe by the chainsaw, and the bow and arrow by the shotgun.
National parks on state-owned land have not proved very successful in Papua New
Guinea. Their development and expansion have often been hindered by land disputes and
compensation claims (Eaton 1982, 1985, Gare 1986). Only a relatively small area, 127
square kilometres, has been reserved for parks, and even these are subject to encroachment.
Customary tenure has proved a constraint to the national parks system, but it has provided
the basis for other types of protected area, in particular those known as wildlife management
areas. These are established by rural communities on their own customary land; village
representatives form management committees, designate the boundaries, and draw up
regulations designed to solve local problems of wildlife depletion. These rules vary
according to the needs of the particular area, but generally include prohibitions on hunting
with firearms, and destruction of the forest. Complete or seasonal protection is provided for
certain valued but endangered wildlife species; hunting by outsiders is either not allowed, or
subject to licences, royalties and quotas. Wildlife management areas have varied in their
effectiveness, but have proved popular with the customary landowners who see them as a
means of formal ising their traditional rights and ensuring that they retain the power to retain
the power to manage their own local resources. During the period from 1975 to 1992,
eighteen wildlife management areas were established, plus five other conservation areas on
customary land. Their total area is 10,529 square kilometres, over sixty times that protected
under the country's national parks legislation.
Similar community participation has proved effective in the western or Indonesian part of
New Guinea (Craven et al. 1992), but has not yet been a feature of conservation programmes
in other parts of Southeast Asia, where there often remains a need to reinforce rather than
destroy local systems of resource management. If we wish to conserve biological diversity,
we may find that one of the most effective means is by supporting local institutions that also
preserve cultural diversity.
LITERATURE CITED
CALDECOTT, J. O. 1986. Hunting and Wildlife Management in Malaysia. World Wildlife Fund of Malaysia,
Kuala Lumpur.
CRAVEN, I., PURBA, P., SETIONO, D., WOMISWOR & D. & THAMRIN, S. 1992. Community Involvement
in Protected Area Establishment and Management in Irian Jaya. Paper presented at IVth World Congress of
National Parks and Protected Areas, Caracas.
DIXON, K.A. & SHERMAN, P.B. 1990. Economics of Protected Areas. Earthscan, London.
EATON, P. 1982. Land Tenure and National Parks. Pp. 223-226 in Morauta, L., Pernetta, J. & Heaney, W. (eds).
Traditional Conservation in PapUfl New Guinea. Institute of Applied Social and Economic Research, Port
Moresby.
EATON, P. 1985. Land Tenure and Conservation: Protected Areas in the South Pacific. South Pacific Commission,
Noumea.
EATON, P. 1986. Grassroots Conservation: Wildlife Management Areas in PapUfl New Guinea. University of
Papua New Guinea, Port Moresby.
FREEMAN, 1. Report on the !ban of Sarawak. Government Printing Office, Kuching.
GARE, N. 1986. The Making of a National Park. Parks 11:13-18.
HONG, E. 1987. Natives of Sarawak. Institut Masyarakat, Penang.
HOSE, C. & MCDOUGALL, W. 1912. The Pagan Tribes of Borneo. Frank Cass & Co., London.
JACOBSON, S.K. 1991. Resident Attitudes about a National Park in Sabah, Malaysia. Pp.250-262 in West, P.C.
& Brechin, S.R. (eds). Resident Peoples and National Parks. University of Arizona Press, Tucson.
KEDIT, P. 1982. An Ecological Study of the Penan. Sarawak Museum Journal XXX, 51:225-279.
LEAMAN, D.L, YUSOF, R. & ARNASON, LT. 1992. Kenyah Medicinal Plants: Beyond the Inventory. Pp.
456-459 in Ismail, G., Mohamed, M. & Omar, S. (eds). Forest Biology and Conservation in Borneo. Yayasan
Sabah, Kota Kinabalu.
LOW, H. 1848. Sarawak; its inhabitants and productions. Richard Bentley, London.
MCNEELY, LA. 1988. Economics and Biological Diversity. International Union for the Conservation of Nature
and Natural Resources, Gland.
MORAUTA, L., PERNETTA, J. & HEANEY, W. 1982. Traditional Conservation in PapUfl New Guinea. Institute
of Applied Social and Economic Research. Port Moresby.
MUNASINGHE, M. 1992. Biodiversity Protection Policy: Environmental Valuation and Distribution Issues. Ambio
12:3.
PEARCE, D. 1990. An Economic Approach to saving the Tropical Forests. London Environmental Economics
Centre, London.
PEARCE, D., MAKANDYA, A. & BARBIER, E.B. 1989. Blueprintfor a Green Economy. Earthscan, London.
PEARCE, K., AMAN, V. & 10K, S. 1987. An Ethnobotanical Study of the Iban Community of the Pantu
Sub-district, Sri Aman, Division 2, Sarawak. Sarawak Museum Journal XXXVII, 58:223-270.
PETERS, C., GENTRY, A. & MENDELSOHN, R. 1989. Valuation of an Amazonian Rainforest. Nature
339:665-6, 227-236.
REPETTO, R., MAGRATH, W., WELLS, M., BEER, C. & ROSSINI, F. 1989. Wasting assets: natural resources
in the national income accounting. World Resources Institute, Washington.
REPETTO, R. & GILLIS, M. 1988. Public Policies and the Misuse of Forest Resources. Cambridge University
Press, Cambridge.
RISWAN, R., MAHYAR, U.W., SANGAT-ROEMANTYO, H. & BURLEY, 1.S. 1992. Ethnobotany of several
medicinal plants in Harowu village, Central Kalimantan. Pp. 441-455 in Ismail, G., Mohamed M. & Omar, S.
(eds). Forest Biology and Conservation in Borneo. Yayasan Sabah, Kota Kinabalu.
SAHABAT ALAM MALAYSIA. 1989. The Battle for Sarawak's Forests. Penang.
SANDIN, B. 1967. The Sea Dayaks of Borneo before the White Rajah Rule. Macmillan, London.
TOBIAS, D. & MENDELSOHN, R. 1991. Valuing Ecotourism in a Tropical Rainforest Reserve. Ambio 20:91-93.
VINCENT, J.R. 1990. Rent Capture and the Feasibility of Tropical Forest Management. Land Economics
66:212-223.
WELLS, M., BRANDON, K. & HANNAH, L. 1990. People and Parks: Linking Protected Area Management with
Local Communities. World Bank, Washington.
WESTERN, D. 1984. Amboseli National Park: Human Values and the Conservation of a Savanna Ecosystem. Pp.
93-100 in McNeely, 1. & Miller, K. (eds). National Parks: Conservation and Development. Smithsonian
Institution, Washington.
WORLD COMMISSION ON ENVIRONMENT AND DEVELOPMENT (WCED). 1987. Our Common Future.
Oxford University Press, Oxford.
Impact of rainforest destruction - the Siberut Island case, Sumatra,
Indonesia
F. GOELTENBOTH and K.H.TIMOTIUS
Universitas Kristen Satya Wacana, Fakultas Sains dan Matematika, fl. Diponegoro 52-60,
Salatiga 50711, Indonesia
ABSTRACT. Indonesia is covered with about 143,969,000 ha of tropical forest, stretching from
Aceh in Northern Sumatra to Irian Jaya on the island of New Guinea. About 13,000 islands, 6,000
inhabited, form the archipelago with its very distinct biogeographical regions. One group of
islands in the Wallacea region, the four Mentawai Islands off the shore of Western Sumatra have
a very unique ecosystem with an array of endemic species of flora and fauna rarely found
anywhere else. This is due to the isolation of these islands from mainland Sumatra and the rest
of the Sunda shelf for the past 500,000 years, and the relative stability of the islands ecosystem
until recent days. Over 65 % of the animal species and 15 % of the plant species are thought to be
endemic. The major island, Siberut, has an area of 448,000 ha, 90% of which is covered by
Lowland Dipterocarp Rainforest, 3% by Riverine Forest, and 5% by Sago Palm Swamp Forest.
The high endemism rate and the unique culture of the indigenous population of about 24,000
people (18,000 still living their traditional lives with a sedentary form of permanent garden
agriculture) were reasons for the incorporation of Siberut Island into the Man and Biosphere
Program of the United Nations in 1981. Since then a continuous struggle has been going on to
find the optimal way to reconcile biosphere necessities and development. Main impacts come
from logging companies who hold 256,000 ha of concession area; plans for the establishment of
palm oil plantations; transmigration plans for about 10,000 families from Sumatra and Java;
government resettlement schemes for the indigenous population; and steadily increasing tourism.
KEY WORDS: Mentawai Islands, Wallacea, endemic flora and fauna, conservation, exploitation.
INTRODUCTION
The island of Siberut, off the shore of Western Sumatra is the largest of the four Mentawai
Islands (Figure 1). It is well recognized that the island belongs to one of the most outstanding
environments of the globe (Ave et at. 1990, McNeely 1978, Mitchell 1982, Shefold 1980,
Whitten et at. 1979, Whitten et al. 1980, Whitten et at. 1981).
About 65 % of the recorded animals (Tables 1 & 2) and 15 % of the recorded plant species
are endemic. Since the mid-Pleistocene, Siberut has been essentially an oceanic island, and
its fauna and flora has evolved in isolation from the dynamic evolutionary events on the
Sunda Shelf. The evolutionary processes on this island have led to a dynamic equilibrium
resulting in a very high number of endemics. Long isolation has allowed the survival of
relicts of an early Indo-Malayan fauna and flora and these forms are of great importance to
studies of the evolution of the modern fauna and flora of South East Asia (Mitchell 1982).
Since 1981 a part of Siberut Island has been identified as a National Biosphere Reserve
under the Man and Biosphere Program of UNESCO. Despite this, the island is under
increasing pressure due to various logging operations and other ongoing development
processes.
General Facts
Siberut lies within 00 55' to 2° 00' South and 98° 31' to 9~ 30' East. It is the largest of
the four Mentawai Islands, being 112 km in length and 40 km in width, and with an overall
425
426 F. GOELTENBOTH AND K.H. TIMOTIUS
1
N
"",.
Figure 1. Location map of Siberut Island (after Whitten et al. 1980).
Figure 2. Drainage map of Siberut Island with linguistic regions occupying mostly an entire
river catchment basin (after Whitten et al. 1980).
Impact of rairiforest destruction on Siberut, Indonesia 427
Table 1. Endemism of Mentawai mammals (after Whitten et al. \980).
Level of endemism
Species Genus Species Subspecies None
Insectivores
Tupaia glis siberu +
Ptilocercus lowii continentis +
Bats
Pteropus hypomelanus enganus +
Pteropus vampyrus mnlaccensis +
Cynopterus shinx angulatus +
Macroglossus sobrinus fratemus +
Emballonura monticola +
Hipposideros galeritus +
Myotis muricela abbotti +
Kerivoula harwickei engana +
Megadermn spasmn ssp. ?
Pangolins
Manis javanica +
Squirrels
Callosciurus melanogaster +
Sundascirus fraterculus +
Laricus obscurus +
Iomys sipora +
Petinomys hageni lugens +
Hylopetes sipora +
Rats
Chiropodomys karlkoopmnni +
Rattus tiomnnicus mentawi +
Rattus sabanus siporanus +
Rattus surifer pagensis +
Civets
Paradoxurus hermnphroditus siberu +
Hemigalus derbyanus sipora +
Otters
Aonyx cinera +
Deer
Cervus unicolor oceanus +
Primates
Macaca pagensis +
Presbytis potenziani +
Hylobates klossii +
Simias concolor +
area of about 448,000 ha. It is a hilly island with the highest peak at 384 m above sea level.
Siberut was uplifted as a relative flat surface with some sedimentary beds having been
overturned and dipping at up to 900 • Exposures are very scarce, consisting of tuffs,
sandstones, and claystones of the Sagulubek Formation. The rather regular sedimentary
formation is of a young geological age, and erosion rates have been high, resulting in the
Table 2. Endemic subspecies of birds (after Whitten et al. 1980).
Species Character difference from

mainland or other island forms
Smaller Larger Colour
Honey Buzzard Pemis apivorus ptilorhynchus + +
Crested Serpent Eagle Spilomis cheela sipora + +
Red Cuckoo-Dove Macropygia phasianella elassa +
Blue-Rumped Parrot Psittinus cyanurus pontius +
Chestnut-breasted Malkoha Phaenicophaeus curvirostris oenicaudus +
Green Broadbill CalyptomelUl viridis siiJeru +
Ashy Drongo Dicrurus leaucophaeus siiJeru + +
Spangled Drongo Dicrurus hottentotus viridinitens + +
Black-naped Oriole Oriolus chinensis mentawi +
Dark-throated Oriole Oriolus xantonotus mentawi +
Magpie Robin Copsychos saularis pagiensis +
Black-naped Monarch Hypothymis azurea leucophila +
development of a rugged landscape with many relatively short rivers and streams (Figure 2).
The hills rise steeply almost without interruption from river terraces. As a result, a very
complex drainage pattern can be found on Siberut, with no clear island divide. The dominant
soil type is red-yellow podsolic of tropudults with dystropepts and troporthents on eroded
sites. The larger rivers are usually 30-40 m wide up to one kilometer inland from the coast,
and are navigable throughout the year. Eight catchment areas are identified, varying in size
from 110-550 km 2 with a calculated mean runoff of 7-35 m3 sec· l • Siberut has extensive areas
of hillrocky tuffaceous sedimentary plains, extensive coalescent estuarine or riverine plains,
and some minor valley floors within the hill region.
The river system is fed by a wet equatorial climate with no distinct dry season. Annual
rainfall is 3364 mm, temperature oscillates between 20 0 and 32 0 C, humidity is 82-86%,
mean wind velocity is 24-34 km, and mean annual sunshine is 1290 hours. According to the
Koppen System the climate can be described as tropical, rainy, isothermal with a hot
summer, no dry season, little seasonal temperature changes and a bimodal rainfall
distribution. In every month, more than 50% of the days have rain. The high rainfall on
non-resistant soil results in an ongoing formation of silt terraces in the lowlands of the east
coast in particular. Subsidence has also occurred along the east coast and is reflected in a
very irregular coastline with bays, capes, islets, and coral reefs. The west coast is rather
straight and lined with broad sandy beaches.
Natural vegetation and fauna

The natural vegetation of Siberut is tropical rainforest with some distinctive features
(Figure 3).
1. Primary Dipterocarp Forest. This type of forest is confined to the high ridges and hills.
In comparison to other Dipterocarp forests the genus Dipterocarpus is represented in
greater numbers than the genus Shorea. Palaquium and Hydnocarpus species emerge from
the 30-35 m high canopy. Large woody climbers and epiphytes are not abundant and the
ground layer is relatively sparse with many saplings and a low herb layer. Fan and fishtail
palms are absent and the rattans usually low and stemless.
2. Primary Mixed Forest. This type is found on the slopes and low hills with many tree
species of the families Myristicaceae, Euphorbiaceae, Dilleniaceae and Dipterocarpaceae.
Legumes, in contrast to Sumatra, are rare. The commonest emergent tree genera are
Impact of rainforest destruction on Siberut. Indonesia 429
0-""",
Dc- .... .........
Figure 3. Different habitats of Siberut Island and their use by four endemic primates (after Whitten et al. 1980) .
Bilou: Kloss Gibbon - Hylobates klossii; Joja: Mentawa Langur - Presbytis potenziani; Simakobu: Pig-tailed Langur-
Simias concolor; Bokkoi: Mentawai Macaque - Macaca pagensis.
Shorea, Dipterocarpus, Dialium, Pentan and Durio. The forest is rich in climbers and
epiphytes, the ground vegetation dense with palms dominating. The canopy height is 15-30
m with fewer emergents than in the Dipterocarp forest.
3. Freshwater Swamp Forest. This has a specialized and limited tree flora dominated by
Terminalia pheilocarpa. The ground layer is composed of feather palms, rattans, pandans,
and aroids. The soil is persistently wet and studded with pneumatophore roots.
4. Mangrove Forest. This type is only found on the shoreline of the Eastern coast. Usually
Rhizophora species are replaced by taller Bruguiera trees where the substrate is exposed
at low tide. The habitat is very specialized and the stands are composed of very few
species. Estuaries of large rivers are fringed with Nypafruticans.
5. Barringtonia Forest. This forest type is found on the west coast, with pure stands of
Casuarina equisetifolia and tree-shrub associations of Barringtonia, Hibiscus, and
Pandanus. Eugenia grandis and CalophyLLum inophyLLoides can be found on headlands.
The major botanical collections on Siberut were made by Kloos in 1924 and described by
Ridley in 1926 (Ridley 1926). Besides the fact that 15% of the recorded 330 plant species
are endemic, it is remarkable that plant species and genera which are also found elsewhere
in South East Asia often have quite peculiar characteristics on Siberut. Xanthophytum villarii
(Rubiaceae), otherwise known only as a small tree from the Philippines, is a tall tree on
Siberut, while Phaeomeria minor (Zingiberaceae) is the smallest in the genus. In Siberut,
Metroxylon sagu is extraordinarily large with a trunk 18 m in length and 1.5-2.0 m in girth,
whereas on average they reach only 9.5 m in length and 0.75-1.20 m in girth in Malaysia
(Whitmore 1975).
The importance of the Siberut forests as a plant genetic resource is highlighted by the fact
that on the smaller islands each plant species tends to fill a broader niche than on the
mainland or larger islands. Species are usually adapted to a wider range of environmental
conditions and are likely to be more suitable for silviculture or cross-breeding than the more
narrowly specialized forms. One very important fruit tree originating from forests in Siberut
is the durian (Durio spp.).
Table 3. Forest plants utilised by man (after Whitten et al. 1980).
Category Number of species

Fruit and root crops 31
Seed grains 3
Spices 2
Animal forage 7
Medicine 10
Fibre 6
Resin 5
Gums 5
Insecticides 4
Total 73
Table 4. Endemic primates on Mentawi Islands compared with other islands with
endemic primate species (after Whitten et al. 1980).
Island(s) Area (lan2) Number Ratio Area/Primates

Mentawai 7,000 4 1,750
Sumatra 473,370 473,370
Sulawesi 189,037 7 27,005
Kalimantan 725,478 4 181,369
Madagascar 587,042 19 30,897
Taiwan 35,962 35,962
Of the 330 plant species Kloos collected in 1924, 73 (22 %) represent genera of economic
importance to man (Table 3). The forest of Siberut is clearly essential for the continuing
existence of its rich endemic wildlife. Of 31 terrestrial species of mammals, 28 are
dependent on forest (Whitten et al. 1980). All four of the endemic primates (Table 4) are
dependent on forest to some extent. Rattan fruits are a significant food item for 3 of the 4
species and one such rattan (Calamus manan) is only found in mature forest. A similar
pattern of dependence exists for the 107 species of birds.
This interrelationship is also true for many of the rainforest plants which are dependent
on animals for seed dispersal, pollination and insect control. Thus the forest of Siberut, as
all other natural forests, comprise a complex interacting system of plants and animals which
are dependent on each other for their continued existence.
The main force reducing forest diversity on Siberut is modern man.
Human effects on the biosphere of Siberut

Five rather clear categories can be identified:
1. Undisturbed areas composed of forest showing negligible effects. About 70% of the
Western part of the island belongs to this category, but virtually all these areas have been
given out as timber concessions since the 1970's.
2. Moderately disturbed areas where human effects have noticeably altered the appearance
of the forest. For example: areas around settlements, areas of commercial rattan
collection, and mangrove areas.
3. Very disturbed areas where traditional systems of agriculture have been practiced. These
are found mostly near navigable rivers and flat swampy areas dominated by Sago and
Nipa palms.
4. Totally disturbed areas which include planned government and missionary villages, and
Impact of rainforest destruction on Siberut, Indonesia 431
Concessions not under operation:

A P.T. Sumber laya (ndah ( ince 1972)
8 Cagar AJam (since 1977);
C P.T.Swnber Surya SemesUi (since 1977).
D Gazeued reserve area Teitei Balti
Concessions under operation:
§ P.T. Cirebon Agung\ P.T. Tridaru (since
1970, 70,000 hal
lIlJ P.T. Carya Phannin Pulau Siberut ( ince
1969,36,000 hal
P.T. Kayu Siberut\ P.T. Inkappa (since
1973 . 110.000 hal.
Present base camp for logging operations
Present logging sites
Boundary of concession areas
Logging roads
Boundary of local administration area
Figure 4. Timber concessions and concessionaires (status as at January 1993).
..
Figure 5. Heavily disturbed areas due to logging activities (Status as at January 1993).
their surroundings. For administrative reasons the traditionally scattered populations have
been relocated in newly created, mostly coastal, resettlements. This practice is prone to
create ecological wastelands through which the exchange of plants and animals is greatly
hampered.
Table 5. Major impacts of logging operations on the environment and the population.
Shallow sea Coral reef destruction due to siltation

Mangroves Destruction of fish nurseries; erosion by sea
Estuaries Loss of habitat; loss of species
Primary Mixed forest Changes of habitats; grassland areas split forest into
Primary Dipterocarp forest 'islands' too small for maintenance of some species;
increased erosion; depletion of various resources
Secondary forest Increased erosion; reduced fertility
Brackish lake areas
Nipa belt ) Loss of habitat and associated species
Barringtonia beach forest )
West coast
5. Areas disturbed by logging, including all those where commercial logging takes place
(Figures 4 & 5). At present 256,000 ha, more than 50% of the total area of Siberut
Island, is logged by 3 major companies. This occurs despite the fact that major parts of
Siberut have been declared or are needed as Biosphere reserve areas.
Due to inappropriate practices by contractors and sub-contractors, the extraction of 10% of
the timber results in about 65 % destruction of the forest. The felling of one major tree and
its extraction may result in the damage of about 0.2 ha of forest. The logging roads on soft
clay soil, up to 60 m wide, result in extremely high erosion rates. The logging does not
benefit the Siberut people (Table 5); less than 3 % of the work force is from Siberut itself.
'---'--' <:>
10km
Figure 6. Proposed area for the planned National Park (198,000 ha) t;:-= =3; the two timber concession areas under
operation until the end of 19941:'i::S!3; the two recent resettlement areas for the local people .~ After Momberg
1994, unpublished. ~
Impact of rainforest destruction on Siberut, Indonesia 433
CONCLUSIONS
The long isolation of Siberut has allowed a relict fauna to survive and new forms to evolve.
This gives the island a very high rate of endemism, and offers a natural laboratory for the
study of island biogeography. If logging is allowed to continue, these endemic forms of
animals and plants will be gone forever.
Traditionally, conservation was assured by cultural sanctions and a complex system of
taboos. Any method of conservation has to take into consideration the traditional systems of
balanced exploitation by recognizing the rights of the indigenous people. Present logging
practices must be seen as a threat to both the livelihood and ancient culture of the
Proto-Malayan people and to the widely varied endemic fauna and flora.
At the very least, an extended totally protected area of about 190,000 ha is needed to
guarantee the continuation of some of the unique fauna and flora (Mitchell 1982) (Figure 6).
At present only 90,000 ha lies within gazetted reserve areas. Transmigration settlements,
palm oil plantation schemes, uncontrolled tourism, and the ongoing resettlement programme
are major threats to the uniqueness of this island. Use of the island for the transmigration of
up to 10,000 families from Java and Sumatra would be disastrous. In addition, the lack of
agreement between timber producers and forest conservation authorities on the safe
exploitation of the forest in the buffer zones around the reserves will result in a loss for
world heritage, the indigenous people, and the scientific community.
LITERATURE CITED
AVE, A. & SUNITO, S. 1990. Medicinal plants of Siberut. World Wildlife Fund Report, Gland, Switzerland.
186 pp.
MCNEELY,I.A. 1978. Siberut: conservation oflndonesia's Island Paradise. Tiger Paper 5(2):16-19.
MITCHELL, A.H. 1982. Siberut Nature Conservation Area West Sumatra Management Plan 1983-1988. World
Wildlife Fund Report, Bogor, Indonesia. 139 pp.
RIDLEY, H.N. 1926. Spolia Mentawiensia - The flora of the Mentawai Islands. Kew Bulletin 1926:56-94.
SHEFOLD, R. 1980. The Siberut Project. Survival International Reviews: 4-12.
WHITMORE, T.C. 1975. Tropical Rainforests of the Far East. Oxford University Press, Oxford.
WHITTEN, A.J., WHITTEN, I.E.A. & HOUSE, A.P.N. 1979. Solution for Siberut. Oryx 15:166-169.
WHITTEN, A.J. & WHITTEN, J.E.A. 1980. Saving Siberut: A conservation masterplan. World Wildlife Fund
Report, Bogor, Indonesia. 134 pp.
WHITTEN, A.J. & SARDER, Z. 1981. Masterplan for a tropical paradise. New Scientist 91:230-235.
WHITTEN, A.I., DAMANlK, S.J., ANWAR, I. & HISYAM, N. 1984. The ecology of Sumatra. Gadja Mada
University Press, Yogyakarta, Indonesia. 583 pp.
Higher-order categories in Brunei Dusun Ethnobotany: the Folk-
Classification of Rainforest Plants
JA Y H. BERNSTEIN
Department of Sociology & Social Anthropology, Eliot College, University of Kent at

Canterbury, Canterbury, Kent CT2 7NS, United Kingdom
KEY WORDS: Brunei Darussalam, Borneo, ethnobotany, Dusun, human-rainforest interactions.
INTRODUCTION
For more than 100 years botanists, agricultural scientists, and foresters working in island
Southeast Asia have collected indigenous names for plants (e.g. Burkill 1935, Corner 1940,
Ochse 1931, Quisumbing 1951). The reasons for this are plain. There needs to be some
provisional way of identifying species new to science, scientists need to communicate with
local people about plants, and various government departments require a convenient set of
names for efficient administration. Unfortunately, it is also hazardous to use names and
other isolated fragments of local knowledge about plants outside of the ethnobotanical context
from which they are drawn (Ellen, this volume). Terms and associated data have not always
been collected carefully, sometimes displaying linguistic and cultural ignorance. In the case
of Brunei, Ashton's (Ashton 1964, also Pukul & Ashton n.d.) use of indigenous terms for
trees keyed to scientific terms is open to confusion. Native terms are used as if they were
vernacular synonyms for species names, but there is an unexplained overlap of names. In
some cases where there are no known vernacular names new ones are suggested.
This paper presents a preliminary report on the overall structure of Brunei Dusun folk-
botanical classification system, and may go some way towards illustrating some issues and
problems of which rainforest biologists should be aware.
Brunei is an excellent location for an ethnobotanical study seeking to lay bare general
principles of classification. The tropical rainforests of Southeast Asia have one of the highest
biodiversities on earth, and in Brunei, through good fortune and effective planning, they are
well-preserved. The cultures of the Austronesian peoples who have lived in Borneo for over
two thousand years have co-evolved with the rainforest environment, and their survival has
depended upon knowledge of the forest (Ellen, in press). Bornean groups have always
relied on forest products, not only for food but to obtain materials needed for medicine,
clothing, and other aspects of technology. The extraction of forest products, both for
consumption and for trade, has long been an essential element of the economies of traditional
Bornean populations, even those who define themselves first and foremost as 'farmers' (King
1993, pp. 77-102). With rapid socio-economic change occurring throughout Borneo (Cleary
& Eaton 1992) traditional patterns of rainforest use have undergone massive change, and are
now in decline.
The Area and its Inhabitants

The field site for the project reported here is the Dusun village of Merimbun, located in
the Rambai mukim (administrative sub-district) in Tutong district. Tutong, one of Brunei's
four districts, is the traditional homeland of the Dusun, who are among seven ethnic and
linguistic groups constitutionally recognized as native Bruneians. Hence they are classified
435
D. S. Edwards et aJ. (eds.), Tropical Rainforest Research - Current Issues, 435-450.
436 JA Y H. BERNSTEIN
for government purposes as 'Malay', and for this reason there are no contemporary statistics
on the number of Dusun. A well considered estimate (Antaran 1993, p. 19) is that there are
5,000 non-Muslim Dusun in Brunei. It is clear, however, that the Dusun population is
decreasing as the result of marriages to Malays and Chinese. Furthermore, traditional
Dusun language and culture is not being effectively reproduced (Kershaw 1992). In large
part, this is due to social and economic mobility: with a large public sector economy,
Bruneians are abandoning traditional economic pursuits to join labour forces away from their
villages. As this occurs, traditional social formations, and the cultural knowledge which
sustains them, decline.! Another crucial factor in the transformation of Brunei's traditional
indigenous peoples is conversion to Islam.
Traditional Dusun culture revolves around the cultivation of dry rice and fruit orchards.
Values relating to these activities are expressed in ritual and belief, for example in the
shamanic temarok ceremony and in bird omenism. Other major economic pursuits include
hunting, fishing, and the collection of various forest products. Until recent years rubber
cultivation and taPping also contributed to the Dusun economy. 2
Merimbun village is situated at a large lake (tasek), and this is a source of fish for all
villagers. It is also a bathing spot and a source of potable water when other supplies are
depleted. Like most of Brunei, it is also surrounded by forests. The forests in this region
are of mixed dipterocarp type, and are the source of many of the economic products that
provided the subsistence of the traditional Dusun.
The Merimbun area has been made accessible by metalled roads leading into the village,
and has been developed as a recreation site by the provision of bridges, pavilions, and toilets.
On Sundays and public holidays it is a favoured spot for picnickers and nature-lovers. It is
considered that the area has been degraded ecologically by this development by increased
water pollution due to the use of motor boats, among other reasons. Heavy use of walking
paths and the construction of roads has also disturbed and even destroyed local habitats
(Wong 1992). Villagers note that fish are less plentiful than in bygone days, and that they
now have to walk increasingly further to reach game-hunting grounds. But it is the rapid
social, cultural, and economic change, rather than environmental degradation, that is leading
to what seems the inexorable decline of forest use and ethnobotanical knowledge. Much of
the folk botanical lexicon may no longer be in use in 20 years time as Malay (and English)
supplant the Dusun language. 3
A Framework for Ethnobiological Classification

Folk-classification differs in several important respects from the scientific, biotaxonomic
classification of plants which is primarily a reflection of the position of a plant in evolution.
In scientific taxonomies, reproductive and genetic criteria are highly significant:
classifications are based to a great extent on the structure of reproductive parts (flowers,
fruits, or spores). Folk classification, on the other hand, is relatively unconcerned with
these matters, though it does reflect empirical criteria - particularly morphology, ecology,
and (in the case of animals) behaviour - that are used in scientific classification. However,
these are never entirely independent of other cultural criteria, as the present analysis will
demonstrate.
Berlin (1992) suggests that plant classifications everywhere display evidence of certain
universal regularies. He claims that the reason non-literate peoples have elaborated extensive
domains of ethnobiological knowledge - and why these are strikingly similar in structure and
content - is "human beings' inescapable and largely unconscious appreciation of the inherent
structure of biological reality" (Berlin 1992, p. 8). The identity of a plant is immediately
apprehended, due not to cultural logic but to the perception of 'discontinuities in nature'.
For Berlin, folk classification relies upon the "natural system" of classification, in which
plants and animals are categorized in accordance with "their overall morphological structure,
Brunei Dusun Ethnobotany 437
or morphological plan" (ibid., p. 9).

Berlin claims, additionally, that living things are classified in terms of the following ranks,
listed below from most inclusive to most specific.
(1) kingdom: a taxon with one member including all taxa of lesser rank.
(2) life form: a taxon containing a small number of highly distinctive morphotypes based
on the recognition of gross morphological structure and ecological adaptation.
(3) intermediate: a taxon grouping together a small number of generic taxa, most
commonly members of life-form taxa, showing marked perceptual similarities.
(4) generic: a taxon naming the smallest fundamental biological discontinuities easily
recognized in any particular habitat.
(5) specific: a taxon partitioning the generic taxa into two or more members, differing in
a few ways.
(6) varietal: a taxon further dividing specific taxa.
In this model, generally, taxa at all ranks except for those of kingdom and intermediate
are lexically labelled; taxa at those ranks are usually unlabelled (covert), and some life-form
taxa may also be unlabelled. Another axiom of Berlin's model is that terms at the generic
rank are primary, meaning that they "are precisely the names of common speech" (Berlin
1992, p. 53). Psychologically, they are more salient than names at other ranks.4
The significance of Berlin's theory is that it challenges the idea that folk classifications
of plants bear no resemblance to biotaxonomic classifications, but rather are determined to
a great extent, if not completely, by cultural values, culture-specific logic, or the structure
of language.
In this paper I use Berlin's model as a framework to discuss my data on one aspect of
Dusun ethnobotany: the grouping of generically-named plants into more inclusive, 'higher-
order' categories. I use the terms "life-form", "intermediate" and "generic" accepting the
model of a ranked order they presuppose. I agree with Berlin that categories are
predominantly constructed on the basis of their biological distinctiveness, and not because
of independent cultural valuation and classification systems. At the same time I share the
view of some of his critics, such as Ellen (1986, 1994) and Morris (1993), that categories
derived from criteria that are neither morphologic nor phenotypic may intrude into the
'natural system of classification'. even though such categories are indistinct. For this reason
I include in my analysis higher-order categories elicited in the course of fieldwork that do
not fit comfortably in the Berlin scheme. I conclude not only that higher-order categories
occur at different ranks, but that they are weighted differently, and that it is possible to
differentiate the most salient categories from less salient ones.
THE PROJECT AND ITS FIELD METHODS
The University of Kent - Brunei Museum project on human-rainforest interaction has among
its aims the documentation of the current significance of rainforest resources and the
ecological parameters and opportunities presented by rainforest in the lives of rural Brunei
Dusun. By extension, it seeks to further our understanding of other comparable parts of
Southeast Asia. The project also aims to contribute to knowledge about human impacts and
dependency on rainforest (indigenous and commercial), and the realistic valuation of
rainforest as an economic resource. As part of the research it was necessary to obtain data
on people's knowledge and uses of specific rainforest plants. This was accomplished
primarily through the collection of voucher specimens.
Plant collection was undertaken in the Merimbun area for four months in 1992 and three
months in 1993. Four principal informants were engaged in this work. All were in their late
fifties or older, with an extensive knowledge of forests, having spent most of their lives
engaged in traditional subsistence activities, including hunting. Normally, only one

informant collected plants on anyone trip; however, on certain occasions two informants
were used simultaneously. At the time of collection the informant would provide the Dusun
name of the plant, the name of the location and the kind of habitat in which it was found,
information about its uses, and any other information about the plant. Optionally at first, but
later more regularly, the informant was asked to categorize the plant. Each evening,
following collection, plants were preserved for further processing. Specimens were
sometimes shown to other informants to obtain their identifications, before they were
removed for processing.
Five hundred and thirty-five herbarium specimens of various kinds were collected, for
which 426 different names (exclusive of synonymous or alternative names for identical
plants) were provided. Collections were made in uncultivated areas, and almost no cultivated
plants were included. 5 These specimens are currently being identified by the Royal Botanic
Gardens at Kew and the Forestry Centre in Brunei. At the time of writing, about 300 of the
specimens have been identified. A further 132 additional plant names were collected where
no specimens were taken, through inventories enumerating each plant found within
circumscribed areas of forest. For this exercise two informants named every plant they could
recognize. Most plants were identified at the generic rank, while some were identified at a
sub-generic rank; those plants for which generic names were unknown were identified at a
more inclusive rank. In the course of the exercise, involving the identification of many
thousands of plants, only one plant could not be identified to any category.
Other plant names were elicited in the course of interviews in which informants were
asked to name plants similar to those given in prompts. The names of plants were put to
several informants, consisting mainly of people not involved in plant collection work. They
were asked to state what kind of plant it was and give any other information they could about
it.
Another technique used was to present informants with sets of cards containing the names
of plants which they were then asked to sort into groups that were similar in some way. For
illiterate informants the names of plants were read aloud. They were then asked what the
plants so grouped had in common, and how they contrasted with those in other groupings.
NOMENCLATURE AS EVIDENCE FOR CATEGORIZATION
The names of plants provide a major key to the ordering of an ethnobotanical domain, though
not an entirely reliable one (Conklin 1962). There is no overall name for all plant life which
would correspond with the kingdom rank. From the names of plants themselves the
following list of category names may be drawn up: akau (vine or liana), bulu' (bamboo),
kayu (tree), kulat (mushroom), usak (alternatviely bunga) (flower), urnbus (vegetable), and
uwai (rattan). These terms are used regularly in Dusun as the first element in a polynomial
plant name (e.g., akau gamuri, kayu baru, kulat kamanci). These are all higher-order
categories, groupings of generically named plants. They are the most inclusive groupings
within the un-named domain 'plant'. These names, among others, were elicited when
informants were asked, given a certain plant, what sort of a plant it was. How is it
possible to determine which of these terms represent culturally salient categories
corresponding with Berlin's "life-form" rank?
The specification of this term is difficult. It indicates, as Wierzbicka (1985) states, "a
category of categories" (p. 189), "a kind including many different kinds" (p. 192). Usak
(flower) and urnbus (vegetable) are excluded on grounds that they are not completely
distinctive: a plant can belong simultaneously in one of these categories and in some other
group. bulu' (bamboo) is associated with other, non-life-form categories, though it does not
overlap with any of them. For this reason it does not seem distinctive enough to be a life-
form, despite the fact that it is, in a way, a category of categories. The remaining
categories, akau, kayu, kulat, and uwai are treated here as life-forms.
The distinction between trees and vines is general though not universal in folk-
classification (Brown 1977), and these are extremely common features of the forest
environment. The focal perceptual difference between a tree and a vine is immediately
apprehendable, and it is doubtful that there is much cultural variation in the construction of
these categories in clear cases. Berlin (1976, p. 385) considers them (along with herbs) to
represent "such distinct perceptual discontinuities that their recognition may constitute a
substantive near-universal in prescientific man's view of the plant world."
The distinction between a tree and a vine is based on morphology (stem growth pattern),
but is not recognized in scientific biological taxonomy. Plants whose stem growth pattern
is intermediary between that of a "tree" and a "vine" must be included in one of these
categories, and the basis of such a plant's categorization seems arbitrary in terms of
morphology.
Rattans (climbing palms) are a distinctive category in Dusun terms, but they meet at least
some of the specifications for 'vines' as well. While recognizing the similarity between akau
and uwai, the Dusun do not consider uwai to be a kind of akau. In conducting my plot
inventories I found that informants would initally identify rattans as vines at first, but then
correct themselves when I repeated the name to them. It appears that the focal definition of
akau excludes rattans. One non-palm was identified at the time of collection by a highly
reliable informant as uwai asu-asu (JHB 280),6 though I questioned him as to whether it
really was uwai, and in my field notes for the specimen I wrote: "Sort of a lake grass, said
to be a rattan. Not found on dry land. Useless, except for binding." The specimen was
subsequently identified as Flagellaria indica. 7 I have also collected, with another informant,
uwai asu-asu bukid (JHB 377, not yet identified), which like uwai asu-asu was thornless,
though I did not question that it was a rattan.
Of 210 kinds of plants collected that were categorized as kayu (some ambiguously so; see
below), the names of only ten were prefixed by the life-form name; one name included the
term alternatively; one name ended with the term.
Turning to the akau category, of 71 kinds of plants collected, the names of 32 were
habitually prefixed by the word akau, two alternatively used the prefix, and one included the
life form name as the second term.
In the uwai category all but one of 16 kinds contained the life-form name uwai as a first
term. In the kulat category all 13 kinds contained the word kulat as the first term.
It will be seen that in the case of kayu, the category term is rarely obligatory, for akau
it is commonly obligatory, for uwai it is nearly always obligatory, and for kulat it is always
obligatory. This fact underlies the high salience of kayu and the decreasing salience of the
other categories, and supports Brown's (1977) theory that 'tree' is the first life-form term to
emerge in an evolutionary sequence in any language. 8
In other categories, as well, the category name appears as part of the name: usak
(alternatively bunga) ('flower') and bulu' ('bamboo').9 There are several other kinds of
plant, but their category-labels are not similarly used. While these are higher-order
groupings, they do not correspond to the 'life-form rank' as defined. The usak category
seems roughly equivalent to 'herbaceous plant', which is often a life-form (Brown 1988), but
the Dusun term is not used consistently to denote this. The criteria for the category seem
not to be purely morphological but to be influenced by what Wierzbicka (1985) points out
is not exactly "function" or "utility" but more generally a relation to people. Some woody
plants are called usak, and this is probably due in part to their size. Some, but by no means
all, plants in the usak category are "decorative plants." Finally, the usak category is not
culturally salient. Only 13 specimens were placed primarily in this category, though some
marginal plants, difficult to place in any life-form category, were placed here, as well as in
one or more other categories. Furthermore, in an enumeration of plants in forest plots, two
informants were more frequently unable to provide names at the generic rank for these than
for other life-forms; in many cases usak were said not to have a name. bulu' is represented
by no more than five different kinds, and is not considered to be completely distinct from
some other kinds of plants.
It must be noted that viewing kulat, uwai, and possibly akau as life-forms violates
Berlin's third principle of ethnobiological nomenclature, which holds that life-form and
generic taxa should have primary names, while subgeneric taxa should (with exceptions) have
secondary names (Berlin 1992, p. 29).
Berlin (1992, p. 27) distinguishes between primary and secondary plant (and animal)
names. A primary plant name may be simple (e.g. oak) or complex (e.g. skunk cabbage,
forget-me-not). To give an example of a secondary name, and explain the difference
between such a name and a complex primary name, Berlin (1992, p. 28) writes:
"Sugar maple can be shown unambiguously to be a secondary plant name in that (a)
one of the name's constituents, maple, is also the name of the taxon that immediately
dominates the category labelled by the expression sugar maple, and (b) sugar maple
occurs in a contrast set whose members are also laeled by names that include that same
constituent (red maple, Norway maple, Oregon maple, vine maple, ... n maple). In
contrast, complex primary names such as tulip tree or catfish do no show this
distributional pattern. These expressions occur as the names for taxa in contrast sets
some of whose members are labelled by simple primary names such as hickory, ash,
and poplar (all kinds of tree), or bass, crappie, and carp (all kinds of fish)"
According to the third principle of ethnobiological nomenclature mentioned above, any
category name that habitually appears in a taxon, as is the case in uwai and kulat, should
not be a life-form but a genus. My difficulty in applying this principle to Dusun ethnobotany
is that it is at odds with the definition of a life-form as a kind including 'many different
kinds' (though it must be admitted that such a vague criterion begs the question).
Furthermore, the category prefix akau appears in about a half of generically-labelled vines.
There seems to be no difference between the categorizing function of the term akau and that
of uwai or kulat.
No ethnobiologist appears to have recognized rattans as a life-form (Brown 1977);
Bartlett (1940) even presents his data on rattans as exemplifying the concept of folk-genus.
But to see uwai as a generic we would have to accept the premise that the primary names
of all 19 kinds of uwai are not generic but specific. 1O The generic level taxon would be
empty, for there is no focal, generic, unmarked uwai. This contradicts Berlin's axiom that
generic taxa are primary in any folk-biological system of nomenclature (Berlin 1992, pp. 52-
101). The main evidence that uwai is not a folk-generic in Dusun is that it contrasts with
kayu and akau, at the life-form rank. If uwai were a folk-generic it would presumably be
included in some other life-form, probably as akau.
The same scepticism about the life-form status of rattans holds for kulat (fungi). Fungi
are completely distinct from all other life-forms; they are infrequently encountered in forests.
Some informants said that mushrooms growing out of the ground could be put into the more
inclusive category of sakot, roughly translated as 'weed' (see below). However, kulat meets
a fundamental criterion of a life-form: it is a kind that includes many different kinds. Unlike
folk-specifics fungi differ not in a few distict ways but in many general ways.
Life-Forms
I will argue that kayu is the most salient plant life-form. As I have indicated, far more
forest plants in that category were collected than in any other, meaning that they are
encountered and recognized more often than other kinds of plants.
That trees are the most prominent and frequently-occurring feature of the forest
environment is obvious enough, and this is reflected at the level of folk terminology in our
analysis of two randomly selected patches of land, both on hilly slopes, one categorized as
entalun ('forest', never previously cleared in known memory) and the other categorized as
gapu' (secondary growth, cleared about 20 years ago). In enumerations made by one
informant, in both the entalun and gapu' plots, 13 of the 20 most frequently mentioned plant-
names were in the kayu category, followed, again in both cases, by 4 in the akau category.
Considering the entalun survey only, this informant provided the names for 103 kinds of
kayu, 26 kinds of akau, 8 kinds of uwai, and 21 names for all other kinds of plants, not
counting plants he could not identify or remember the names for. A second informant (who
surveyed an adjacent area half the size the one surveyed by the first informant) named 114
kinds of kayu, 19 kinds of akau, 10 kinds of uwai, and eight kinds of other plant.
As mentioned above, the difference between kayu and akau is not so much phylogenetic
as morphological. Plants in both categories have many of the same uses: edible fruit and/or
leaves; medicines, poisons, and antidotes; and numerous specialized uses, especially for
leaves. The fruit of vines is rarely important in the trade economy compared with those of
trees. But it is the role of trees in providing wood for building things that gives trees their
unique cultural, social, and economic precedence in the plant world. The stereotypical
purpose of vines, on the other hand, is as binding material. This purpose is not, however,
unique to the akau category, but is shared with rattans (uwai). While this is a reason to
reject a purely utilitarian explanation of life-form categories (such as that offered by Randall
& Hunn 1984), it does seem plausible that utilitarian factors play some role in their relative
salience.
The categories kayu and akau are semantically oppositional and dialectical, with kayu
being primary. While about half of the kinds of akau are referred to habitually as akau X,
very few trees are referred to habitually as kayu X rather than simply X.11 Some terms
occurred in both akau and kayu categories:
Akau Kayu
1. Akau melia' (JHB 155) Dichapetalum? Melia' (JHB 275) Diospyros
2. Akau Sengkarai ORB 321) Agelaea borneensis Sengkarai Polyathia sp. (Brunei checklist)
(Hook. f.) Merr.
3. Akau tis (uncollected) Tis (JHB 149) Galeariafulva (Tu!.) Miq.
4. Akau terakang ORB 277) Oxyceros (kayu) terakang (uncollected)
5. Bagu akau (JHB 137) Gnetum Bagu kayu (JRB 147) Gnetum
The fact that the life-form akau is marked in three cases seems to substantiate the claim that
constituents of the kayu category are psychologically and culturally primary vis-a.-vis those
of the akau category (Brown 1984, cf. Berlin 1992, pp. 27-28). The use of the category
term following the generic term in the case of bagu is explained by the premise that while
the other plants are different things, bagu akau and bagu kayu are two kinds of one thing,
and hence the category names are used as adjectives. Another example of this structure is
found in the names of two kinds of cassava (Manihot utilissima), e.g. ubi kayu ('woody
cassava') and ubi bogo ('creeping cassava').
An exception to the generalization that the kayu form is primary because it is unmarked
vis-a.-vis genera in other life-forms is kayu tungod (JHB 212, Saurauia), which contrasts
with tungod (JHB 174, Pentaphragma), a 'grass' in the telasai category, the salience of
which is low.
There is another instance of a term being identified nomenclaturally as both kayu and
akau:
Akau piantok (JHB 264) (unidentified at time of writing)
Kayu piantok (JHB 126) Rutaceae?
Piantok (JHB 165) Celastraceae?
All of these plants are usually categorized as akau rather than kayu. Informants did not
group it together with other kayu. As one informant said, kayu piantok "is considered
kayu, but in reality it is akau. It is sliced and made into chopsticks (supit)". It seems to
be a marginal type. In field notes both akau piantok and kayu piantok are described as
'vines', while piantok is described as a small plant. In this case plants that are different
both phylogenetically and morphologically but are used in the same way occupy a single
category.
A similar anomaly is kayu sangga (JHB 247, Schejflera), categorized by all informants
as akau.
The categories kayu and akau are sufficiently complex and rich that informants
knowledgable about forests are able to categorize them according to various dimensions such
as hardness and size. Many trees and vines are associated with specific, distinctive
purposes, such as the manufacture of a certain object or a role in a certain ritual. Uwai
(rattans), on the other hand, is a much smaller category. With the exception of uwai pios
(JHB 13, Calamus comptus J. Dransf.), which is used in a rice planting ceremony, rattans
have only five contemporary uses: binding, weaving, frame-building, edible fruit, and edible
shoots. 12
Another phylogenetically focused group, that I consider to meet the criteria for a life-
form, is kulat. Unlike kayu, akau, and uwai, they are not ubiquitous in the forest, and they
are relatively insignificant culturally and economically.
Fungi are unmistakably different from all other kinds of living things, but are regarded
by Dusun as plants. A certain number are edible, and one or two varieties may fetch high
prices, but they are not an important part of either Dusun diet or economy, or of the forest
landscape. An indicator of the very specific utility of kulat (as food) is that my informants
were unable to provide generic names, for most inedible mushrooms lack generic names,
even though it was recognized that they vary one from the other. Inedible mushrooms may
be called kulat andi kanon (inedible mushroom) or kulat raat (bad mushroom). \3
However, kulatjelundong (JHB 289) (literally 'shade mushroom') is generically named, and
the category so identified is even subdivided further: kulat jelundong purak (white shade
mushroom) and kulat jelundong gapu' (secondary growth shade mushroom). None of the
edible mushrooms collected were identified at a sub-generic rank.
An alternative possibility is that kulat is not a life-form but a folk-generic unaffiliated
with any life-form (see discussion of 'unaffilated generics' below). By this analysis edible
mushrooms as well as inedible kulat jelundong are marked off from generic kulat at the
rank of folk-species. I resist this conclusion because it seems to me that the differences
between the various kinds of inedible fungi are perceptually recognized as different genera
without being coded. 14
Intermediate categories
I have identified life-forms through category names that appear as proper parts of generic
rank taxa. Some generically-named plants fall outside all life-forms, but are grouped in other
categories. Berlin (1992) properly calls these categories 'intermediate', since they occupy
a rank between generic and life-form.
Among Dusun ethnobotanical categories are those that may be glossed as 'grasses',
'gingers', 'bananas', and 'ferns', as well as various kinds of palms. These do not constitute
life-forms as they are not labelled by any single over-arching term. For example, several
kinds of ginger (including those called encalongon [JHB 41, Plagiostachys crocydocalyx],
kunyit [Curcuma], sagang [JHB 43, Etlingera ?punicea] , and sumbang [JHB 39,
Homstedtia reticulata)) were said to be sorts of tumid, while others (those called terabak)
Table I. Higher-order Dusun plant categories (excluding cultivated plants).
Life-forms
kayu (tree)
akau (vine)
uwai (rattan)
kulat (fungus/mushroom)
Intermediates
Licuala plams (silad-benjiru-ukang)
Grasses (kumpau-telasai-rumput)
Ferns (gerajai-paku-Iimputong-engkubuk-kuban)
Gingers (tumid-Iingkuas-layoh)
Bananas (punti-rutai-binci-encarawan-powow)
Problematic higher-order categories

usak (flower)
sakot/sakot tanah ('weed')
raun (leal)
umbus/sancam (vegetable)
were categorized as kinds of lingkuas. It is possible that other gingers, called layoh (Iayoh
lamatai JHB 372, Globba brachyanthera K. Schum. var rubra R.M. Smith; layoh entaloo
JHB 467, not yet identified), are not included in either of these groupings. While there is
a clustering around focal members of these groups, the categories are also reciprocally
related to one another. Although there is no term for a single category uniting all gingers,
such a grouping is recognized by Dusun. The idea of covert categories in ethnobotanical
classification, as proposed by Berlin, Breedlove & Raven (1968) is now generally accepted,
but identification has proved to be a controversial matter (e.g. Berlin 1974, 1992, Brown
1974, Ellen 1993, Hunn 1977, Taylor 1984).
Another set of plants that do not fit into any other higher-order grouping but are related
to each other are Musa spp. ('banana') plants. True bananas (Musa sapientum) are called
pooti, but several other genera are found in the wild: rutai, (JHB225, Musa gracilis) binci,
encarawan, and pow ow (all uncollected).
'Grasses' seem a plausible candidate for life-form status, as Dusun kumpau, can be
glossed by rumput, the Malay word for 'grass' .15 However, in the course of research I
found that several plants categorised by Dusun informants as rumput, and which I
considered 'grasses', were said not to be kumpau. Kumpau occurs as a folk-generic name
for Paspalum conjugatum Berg. (Graminae) (JHB 316), but also refers to an overarching
category of grasses each of which have separate folk-generic names. Most grasses are
included under this overarching cover-term, although tall grasses in both the Gramineae and
Cyperaceae families are excluded. Telincim (JHB 230, Paramapania, and JHB 318,
Mapania sumatrana (Miq.) Benth. ssp. sumatrana) and teIasai (e.g. telasai biasa, JHB 219,
Seleria), both Cyperaceae, were said not to be kumpau. A Gramineae, kumpai (JHB 298,
no generic identification), as well as two Cyperaceae grasses, purun, (JHB 287, Lepironia
articulata) and tarupuk (JHB 347, Phragmites karka (Retz.) Steud.) were classified as
telasai. Kumpau and telasai, therefore, seem to suggest two prototypes around which
similar folk genera may be clustered conceptually.
Similarly, ferns are classified as engkubuk, gerajai (alt. derajai, rajai), gerintik,
limputong, paku, and kuban, without there being an overall Dusun term for 'ferns'. They
are not grouped together as a life-form, though there is some reciprocal labeling, e.g. gerajai
classified as a kind of limputong. Informants classify them as rumput or sancam

('vegetable'). One informant, asked to classify engkubuk (JHB 223), was able to reply only:
"not kayu, not akau."
The final covert intermediate cluster includes palms of the Licuala genus and possibly an
Areca. The three Licuala are benjiru (L. spinosa,16 cf. JHB 80), silad (JHB 60, 61, L.
orbicularus), and ukang (JHB 168). These were categorised reciprocally: silad was said to
be a kind of benjiru and vice versa; there was no overall category name. Another palm,
landai tiwow (JHB 181, Areca minuta) , was also said by two informants at the time of
collection to be a kind ofukang; another informant subsequently classified it as kayu. Other
informants who had heard of this plant could not classify it at all.
Two other categories of palm are clustered around salient, focal members. The first
grouping consists of palms similar to pinang, (Areca catechu, the betel-nut palm). As with
the grass kumpau and telasai (see above), pinang is a folk generic that is the focal member
of a larger grouping. Five palms of this kind have been collected. One is pinang derato 17
(JHB 465, not yet identified), the binomial directly indicating that it is folk-species of the
folk-genus pinang. Another is landai tiwow (JHB 181, Areca minuta), the name of which
means "peacock's tail-feathers". However, four other palms in this group have uninomials,
including barang (JHB 216, Pinang salicifolia) and raring (JHB 278, Cyrtostachys
renda).18 Such names do not subdivide a recognized folk-genus. Sub-generic taxa should,
according to Berlin, be labelled with secondary names: "linguistically complex expressions,
one of whose constituents indicates a category superordinate to the form in question" (Berlin
1992, p. 28) - unless one taxon is the prototype of the genus or is ·of "major cultural
significance" (ibid., p. 30). The betel-nut palm (Areca catechu) is without doubt highly
significant culturally throughout Southeast Asia, but barang and raring are decidedly
insignificant.
Another palm category has as its focal member luba (uncollected), and includes tiad (JHB
182, Eugeissonia minor), bidang (Borassodendron borneense 19), ramok (JHB 161, Arenga
undulatifolia) and iyul (uncollected).
Problematic higher-order categories

Four or five supra-generic categories used by the Dusun in classifying plants have some
of the characteristics of life-forms, but seem very different in other ways. They appear to
be able to contrast with life-forms, and are treated like life-forms by Dusun informants who,
when given a plant name and asked to classify it, provided the category terms to be
discussed. They differ, however, in two important respects: they are not based on
morphological or phenotypic stereotypes or indeed on any kind of visual cognitive prototype,
and they are used idiosyncratically. Some members of these categories may belong in other
categories. Others are placed in these categories by default, as they do not belong in any of
the main groups. The categories are labelled (1) usak (Malay: bunga) (2) sakot or sakot
tanah, (3) sancam (Malay: sayur) or umbus20, and (4) raun (Malay: daun). These
groupings occupy much of the same classificatory space as a comprehensive 'herbaceous
plant' life-form, which does not exist in Dusun. However, none of these categories can
clearly be described as life-forms since they do not specify rank.
The term usak, besides meaning 'flower', is used in Dusun in the same way as other life-
form terms. Thus, if a Dusun informant is asked what kind of plant X is, he or she is able
to respond: 'usak'. This term seems to mean 'decorative plant', 'shrub', or 'small
herbaceous plant'. However, it overlaps with other categories, including kayu: a woody
plant such as usak bila pinggan, (JHB 8, 192, Ixora, and 91, Pavetta?, both Rubiaceae, plus
378, 385, and 416, unidentified) is classified primarily as usak, though it is also considered
kayu. Usually a low shrub, it can grow up to four meters high. An attribute of usak is that
it is close to the ground. A plant classified primarily as usak may be optionally included in
other categories. The reverse is not the case. Usak as a plant category seems in some ways
to imitate a life-form, but it may also include growth stages of plants which when mature are
obviously members of other, more clearly-defined, life-forms.
The word sakot refers to any detritus lying around that cannot be identified. It may refer
to rubbish. In the case of plants the term sakot tanah (or sakot bwni) is often used,
suggesting a literal translation of 'ground detritus' (or 'earth detritus'). More idiomatically,
it seems similar to English 'weed': a useless or even undesirable, relatively small plant. In
Malay the term rumput generally has this connotation. And indeed the boundary between
the 'grasses' and sakot is permeable. Miau-miau (JHB 37, Lophatherium gracile Brogn.)
is categorized by some as sakot, and by others as kumpau or rumput. But plants lexically
classified elsewhere may also be placed into this group, e.g. bunga rampai (JHB 152,
Dianella) and akau belan (JHB 48, Merremia borneensis). Grasses, such as those in the
telincim group are categorized as sakot. Galugut (JHB 116, moss) is put in this category,
as are those mushrooms that grow out of the ground. A few informants put basulok (JHB
345, Senna alata (L.) Roxb.) in this category. There is considerable variation between
informants as to which plants they classify as sakot. But the category is not haphazard, nor
is it a "special purpose" category (Berlin et af. 1966, p. 274), as it does not have any single,
'special' purpose. Nor, on the other hand, can all plants described in this way be said to be
useless, annoying, or disvalued, as the term 'weed' implies, since a number have medicinal
or other uses. Sakot seems to be less a category with a clearly defined focus than a residue
including culturally less important species that tend to be small and not noticed as individual
plants but rather as clusters or clumps.21 Usak, in contrast, groups plants in ways closely
resembling those of a life-form, and it may perhaps be considered a 'pseudo-life-form'.
Unlike a true life-form it lacks an ethnobiological focus and does not specify rank.
The categories sancam ('vegetable') and raun ('leaf') are different. They are also
residual, used for plants that do not belong in the main groupings. The terms are most often
used for lamba (JHB 72, Curculigo vi/osa (Kurz) Merrill). This is a big leafy plant with no
stalk. Informants varied in how they classified it. The generic category lamba is quite
complex, and not all its consituents are edible. Lamba gabuk ('macaque lamba') (JHB 19,
Curcu/igo racemosa Ridley), like lamba (alternatively called lamba biasa, 'ordinary lamba'),
is an edible member of the family Hypoxidaceae. Lamba-lamba22 (JHB 146, 160,
Corymborkis) and lamba entalun ('forest lamba') (JHB 198, 392, Niewiedia) are both
orchids, and are not eaten.
Eight other plants were classified as sancam or sayur.23 For example, one informant
categorized the shrub kelupak manuk ('chicken kelupak') (JHB 33, not yet identified) as
sayur, saying that it was neither kayu nor akau.
Besides the lamba group, Araceae plants such as latu (including raun latu, JHB 199,
Homalomena cordata (Houtt.) Schott) are classified as raun. Allocating a plant to the
category of raun indicates that an attempt to put it into any other overarching category would
be unsuccessful, but that it is possible to single out its outstanding characteristic. Like usak,
raun is the term for a part of a plant. As for sancam (vegetable), it appears to be a
functional category, like English 'toy' (Wierzbicka 1984), defined in terms of the use of its
fleshy parts. Wierzbicka argues that such functional categories should not be treated as
collective groupings. By this argument sancam, umbus, or raun may not be plant
categories at all, and plants such as lamba and kelupak manuk may be 'unaffiliated
generics' (see below).
Unlike the categories of folk genera and life-forms, the objects grouped together under
the heading of usak, sancam, raun, sakot, or umbus lack any "inherent physical nature or
'essence'" (Atran 1993, p. 198) uniting them. However, the categories are natural in that
they appear in natural discourse; sakot in particular is used contrastively with kumpau,
kayu, akau, and the others as a category grouping together various genera. These
problematical categogies are not exclusively contrastive: a plant could belong primarily to
a category such as sakot and secondarily to a life-form. Plants occupying the fuzzy zone
between kayu and akau, such as kayu piantok (JHB 126 Rutaceae?, also piantok, JHB 165,
Celastraceae?), or nunuk ('strangling figs', e.g. JHB 445, not yet identified) cannot similarly
be included in both categories.
Unaffiliated generics?
Three kinds of plant are unaffiliated to any of the life-forms, intermediate-rank categories,
or the problematical categories just described. Like life-forms they are 'kind of kinds',
contrasting with other such groupings. Unlike life-forms, there are only a few members of
each grouping, and these are closely related biologically, rather than "cross-cut[ting]
biologically natural groupings of organisms" (Berlin 1992, p. 167). In the case of the first
two categories to be considered, it is awkward to consider them as folk genera. Berlin
(1992, pp. 171-176) has addressed the problem that unaffiliated generics "cannot easily be
removed from the contrast set they share with full-blown life-form taxa" (ibid., p. 174). If
one accepts Berlin's ethnobiological rank system, one must agree with him that one cannot
have one's cake and eat it too (Berlin 1992, p. 175) by placing a category in both life-form
and generic ranks. If Berlin's schema is correct, then a taxon may occupy only one
ethnobiological rank. Rank, however, is distinguished from 'taxonomic level' (ibid., p. 64),
which refers to structural position within a hierarchy as well as cognitive aspects. 24
Bamboos are uncommon in the Merimbun area. Only two kinds of bamboo were
collected, bulu' baluy (JHB 21, Graminae, JHB 430, not yet identified) and bulu' embiling
(JHB 431, not yet identified). I also collected Dusun terms for four other bamboo: bulu'
aur, bulu' batong, bulu' gana, and bulu' ranap.25 Linguistically, bulu' suggests a folk-
genus. But unlike other folk-genera sub-divided and labelled at the folk-specific rank, there
is no unmarked, or 'regular' (biasa) , 'genuine' (bonor) bulu'. Also, informants did not
classify bulu' into any life-form or other higher-order category. Rather, they classified
members of this group simply as bulu'. Thus it seems that bulu' could be described as a
life-form. Some informants considered bulu' to be affiliated with certain palms, such as tiad
(JHB 182, Eugeissonia minor) and raring (JHB 278, Cyrtostachys renda).
A clearer example of an unaffiliated generic is bamban, members of the Marantaceae
family, including bamban (JHB 289, Donax) , bamban aig (JHB 249, Schumannianthus
prob. dichotomus (Roxb.) Gagnep.). and bamban batu (JHB 517, not yet identified). They
are affilated with, and may be compared to, bulu' in that they have a similar general
structure and may be used in the same way (constructing fish-traps and similar object), but
in terms of their basic categorization they stand alone in the bamban category.
CONCLUSION
The goal of this paper has been to provide a provisional outline of Dusun ethnobotanical
classification in terms of its most inclusive categories. From this study we can see how the
Dusun order knowledge of the forest environment with which they have interacted. Through
the collection of forest plants of all kinds, along with sociocultural information about those
plants, a series of general, inclusive, higher-order categories emerge. By analyzing these it
can be seen that they do not have the same cognitive or logical status. I have interpreted my
data using as a starting point the concepts and terminology developed by Berlin (1992). The
results lead me to conclude that Dusun general ethnobotanical categories are not symmetrical
but lopsided, and may shade into each other, and that non-taxonomic categories of the sort
Berlin would exclude are an intrinsic feature of Dusun ethnobotanical classification, and
provide only ambiguous evidence for completely separate cross-cutting 'special purpose'
classifications.
In the social and psychological process of classification there is a yearning to find
symmetry, but if we agree with Berlin's premise that ethnobiological classification follows
nature's plan, the appearance of asymmetry should not be surprising - natural discontinuities
are not evenly distributed. My data show that the general rules set forth by Berlin for
recognizing categories at various ranks hold in the main, but become unworkable in certain
cases.
The Dusun evidence shows that higher-order ethnobotanical categories do not have equal
weight. Life-forms are more salient than other categories, but even the life-forms show
differential salience. On the basis of nomenclatural evidence, prominence in the
environment, and knowledge about different kinds, kayu is the most salient life-form,
followed by akau, uwai, and kulat. As life-form salience decreases, categories are more
easily treated as folk-genera. Likewise, some polytypic unaffiliated folk-genera have some
characteristics of life-forms. In marginal cases, deciding whether a category is a highly
polytypic unaffiliated generic or a minor life-form is almost a matter for subjective judgment,
taking into account number of members and degree of biological focus. Bulmer (1974, p.
23) calls attention to this problem, and ultimately dispenses with Berlin's concept of life-form
and indeed the whole rank system. And Ellen's (1993, p. 116) finding for Nuaulu
ethnozoology holds true Dusun ethnobotany: "there is a gradation of inclusive categories,
some highly salient culturally and widely shared, others - decreasingly - less so." Ellen, too,
is moved to question the existence of life-forms and the reality of ethnobiological rank. But
I find that these concepts ring true in the main: they provide a helpful framework in which
to make sense of most of Dusun ethnobotanical classification. I agree with Hays (1983, p.
593) that while Berlin's system requires "slight modifications ... critics' insistence on the
need for a radically different approach in folk biology is unwarranted. "
Categories of intermediate rank, constructed around prototypes, include various plants not
included in life-forms. Other categories based on selected characteristics rather than overall
morphology may be used as life-forms in classifying plants. These heterogeneous, non-
taxonomic categories intrude in the 'natural' system of classification which Berlin says is
paramount.
Anomalous plants may be categorized marginally in several groups, though never in more
than one major life-form. Patagar anak (JHB 100, Schizaea dichotoma Swartz), for
example, is ambiguous, being woody but tiny, and growing straight out of the ground. It
has been classified as kayu, usak, and sakot. Similarly, badoi (JHB 256, Costus globosus
BI.) is classified variously as usak, ukit, uwai, and tumid. It may be ascribed to one of
these categories on the basis of its use, for example, as a minor food, or as binding material;
or because of the way it grows low to the ground. Where perceived features do not result
in a conclusive, unambiguous category, informants seeking to categorize a plant must
consider use. Thus, a plant may be categorized with rattans (uwai) on the grounds that it
is used to manufacture the carrying basket, saging. Certain plants may not be fully
accommodated in any larger groupings, even though they may have affinities with plants in
other categories.
ACKNOWLEGMENTS
The research described in this report was made possible by U.K. Economic and Social
Research Council award ROOO 23 3028 to Roy Ellen for a study of "The Ecology and
Ethnobiology of Human-Rainforest Interaction in Brunei". Travel funds to attend the
RGS/UBD conference were provided by the British Academy. I thank the Brunei Museum
(particularly the Ethnography Section) for its collaboration, and The Royal Botanic Gardens,
Kew, and the Department of Forestry in Brunei for their assistance. Bantong Antaran of the
448 JAY H. BERNSTEIN
Brunei Museum collaborated in the collection of data and provided helpful suggestions on this
paper. Above all I thank Roy Ellen not only for critically reading several earlier drafts of
the paper but also for his guidance and support through all stages of the study.
LITERATURE CITED
ANTARAN, B. 1993. The Brunei Dusun: An Ethnographic Study. M.Phil. thesis, University of Hull.
ASHTON, P.S. 1964. A ManUflI of the Dipterocarp Trees of Brunei State. Oxford University Press, London.
ATRAN, S. 1993. Folk biological cognition. Current Anthropology 34:195-198.
BARTLETT, H.H. 1940. The concept of genus. 1. History of the generic concept in botany. Bulletin of the Torrey
Botanical Club 67:349-362.
BERLIN, B. 1972. Speculations on the growth of ethnobotanical nomenclature. LangUflge in Society 1:51-86.
BERLIN, B. 1974. Further notes on covert categories and folk taxonomies: a reply to Brown. American
Anthropologist 76:327-331.
BERLIN, B. 1976. The concept of rank in ethnobiological classification: some evidence from Aguaruna folk botany.
American Ethnologist 3:381-399.
BERLIN, B. 1992. Ethnobiological Classification: Principles of Categorization of Plants and Animals in Traditional
Societies. Princeton University Press, Princeton, NJ.
BERLIN, B., BREEDLOVE, D.E. & RAVEN, P.H. 1966. Folk taxonomies and biological classification. Science
154:273-275.
BERLIN, B., BREEDLOVE, D.E. & RAVEN, P.H. 1968. Covert categories and folk taxonomies. American
Anthropologist 70:290-299.
BROWN, C.H. 1974. Unique beginners and covert categories in folk biological taxonomies. American Anthopologist
76:325-327.
BROWN, C. H. 1977. Folk-botanical life-forms: their universality and growth. American Anthropologist 79:317-
342.
BROWN, C.H. 1984. LangUflge and Living Things: Unifonnities in Folk Classification and Naming. Rutgers
University Press, New Brunswick, NJ.
BULMER, R.N.H. 1974. Folk biology in the New Guinea highlands. Social Science Information 13:9-28.
BURKILL, I.H. 1935. A Dictionary of Economic Products of the Malay Peninsula. Crown Agents for the Colonies
on behalf of the Governments of the Straits Settlements and Federated Malay States, London.
CONKLIN, H.C. 1962. Lexicographical treatment of folk taxonomies. In Householder F. W. & Saporta S. (eds).
Problems in Lexicography. Indiana University Research Center in Anthropology, Folklore, and Linguistics.
CLEARY, M. & EATON, P. 1992. Borneo: Change and Development. Oxford University Press, Singapore.
CORNER, E.J.H. 1940. Wayside Trees of Malaya. 2 vols. Government Printing Office, Singapore.
DRANSFIELD, J. 1992. Rattans of Sarawak. Royal Botanic Gardens, Kew, Surrey.
EDER, J. 1988. On the Road to Tribal Extinction: Depopulation, Deculturation, and Adaptive Well-Being among
the Batak of the Philippines. University of California Press, Berkeley.
ELLEN, R. 1986. Ethnobiology, cognition and the structure of prehension: some general theoretical notes. Journal
of Ethnobiology 6:83-98.
ELLEN, R. 1991. Grass, grerb or weed: a BUlmerian meditation on the category monote in Nuaulu plant
classification. In Pawley A (ed.). Man and a Half: Essays in Honor of Ralph Bulmer. Uniprint, Aukland.
ELLEN, R. 1993. The Cultural Relations Of Classification: An Analysis Of NUfluiu Animal Categories from Central
Seram. Cambridge University Press, Cambridge.
ELLEN, R., in press. Modes of subsistence and ethnobiological knowledge: between extraction and cultivation in
Southeast Asia. In Warren & Slikkerveer (eds). Pithecanthropus Centennial, Vol. C. Pithecanthropus
Foundation, Leiden.
ELLEN, R. 1994. Review of Ethnobiological Classification, by B. Berlin. Man.
ELLEN, R. (this volume). Putting plants in their place: Anthropological approaches to understanding the
ethnobotanical knowledge of rainforest populations.
ELLEN, R. & BERNSTEIN, J., in press. Urbs in rure: cultural transformation of the rainforest in Brunei.
Anthropology Today.
HAYS, T.E. 1983. Ndumba folk biology and general principles of ethnobotanical classification and nomenclature.
American Anthropologist 85:592-611.
HUNN, E.S. 1977. TzeJtal Folk Zoology: The Classification of Discontinuities in Nature. Academic Press, New
York.
KERSHAW, E.M. 1992. Final shifts: some why's and hows of Brunei Dusun convergence on Malay. Paper
presented at Borneo Research Council 2nd Biennial International Conference, Kota Kinabalu.
KING, V.T. 1993. The Peoples of Borneo. Basil Blackwell, Oxford.
MORRIS, B. 1993. Review of Ethnobiological Classification, by B. Berlin. American Anthropologist 95:733-734.
OCHSE, J.J. 1931. Vegetables of the Dutch East Indies, English edition. Department Landbow, Buitenzorg.
PERANIO, R.D. 1972. The Bisaya. In LeBar F. M. (ed.). Ethnic Groups of Insular Southeast Asia, Volume I:
Indonesia, Andaman Islands and Madagascar. Human Relations Area Files, New Haven.
PERANIO, R.D. 1977. The Structure of Bisaya Society: a Ranked Cognatic Social System. Ph.D. thesis, Columbia
University.
PUKUL, H. & ASHTON, P. n.d. A Check List of Brunei Trees. Brunei Forestry Department, Bandar Seri
Begawan.
QUISUMBING, E. 1951. Medicinal Plants of the Philippines. Technical Bulletin no. 16, Department of
Agriculture and Natural Resources, Manila.
RANDALL, R. & HUNN, E.S. 1984. Do life forms evolve or do uses for life? American Ethnologist 11:329-349.
RIDLEY, H.N. 1925. The Flora of the Malay Peninsula. L. Reeve, Ashford, Kent.
TAYLOR, P. 1984. "Covert categories" reconsidered: identifying unlabeled classes in Tobelo folk biological
classification. Journal of Ethnobiology 4: 105-122.
TAYLOR, P. 1990. The Folk Biology of the Tobelo People: A Study of Folk ClaSSification. Smithsonian
Contributions to Anthropology, no. 34. Smithsonian Institution Press, Washington.
WIERZBICKA, A. 1984, Apples are not a "kind of fruit": the semantics of human categorization. American
Ethnologist 11 :313-328.
WIERZBICKA, A. 1985. Lexicgraphy and Conceptual Analysis. Karoma Publishers, Ann Arbor.
WONG, M. 1992. The Future of Tasek Merimbun. Unpublished Position paper, Brunei Museum, Bandar Seri
Begawan.
ENDNOTES
1. Compare Eder (1987) on 'deculturation' - the loss of cultural and social identity. See also Ellen &
Bernstein, in press.
2. The only complete ethnographic description of the Brunei Dusun is Bantong Antaran's thesis (Antaran
1993). However, these people are very closely related - culturally, socially, and linguistically - to the
Bisaya of Sarawak (see Peranio 1972, 1977). The main difference between the two groups is a political
one: they occupy territories in different states.
3. Comparative studies have shown that specialised terms such as those for plants are among the last to
disappear from a lexicon. However, in the Dusun case that knowledge of the plant world needed to
preserve this part of the lexicon is itself disappearing in the younger generation, not only because traditional
subsistence is giving way to wage employment, but also because parents now tend to speak to their children
in Malay rather than Dusun (Kershaw 1992).
4. Wierzbicka (1985) challenges the universality of this claim, offering as counter-evidence the primacy of
higher-level (viz. life-form) terms in English - especially bird, tree, and f'lSh. The case, she says, is the
same in other European languages. She suggests this may be due to the fact that native English-speakers
do not live close enough to nature for various kinds of some plants and animals to be important. See also
Berlin (1972).
5. On cultigens and cultivars see Antaran 1993, pp. 71-79. The author provides Dusun terms for 17 types of
common vegetable, 32 types of seasonal fruit tree, and 12 types of non-seasonal fruit. He also provides
names for 15 varieties of rice, and discusses the cultivation of tobacco and sago.
6. ruB followed by a number refers to a specimen in the herbarium of the University of Kent - Brunei
Museum Tasek Merimbun project. Sets are held at the Brunei Forestry Centre, the Royal Botanic Gardens,
Kew, and the University of Kent. At the time of writing only 55 percent of specimens have been identified.
I indicate genus and species name if available, providing family names only in instances where I wish to
show a contrast.
7. F. indica is also categorized as a rattan in Malaysia, Java, and Sumatra (Burkill 1935, pp. 1024-1025; see
also Bartlett 1940, p. 353).
8. The Dusun ethnobotanical data do not support the rest of Brown's evolutionary theory, in which 'grerbs'
(grasses plus herbs) emerge as life-forms before 'vines', and in which herbaceous plants and bushes but
not rattans are life-forms.
9. Because of its lack of complexity, buluh is considered an unaffiliated generic rather than a life-form, and
is described under the section heading, 'Unaffiliated generics?'.
10. In rattans as in other plant classes, my herbarium specimens and other data are not by any means complete.
Dransfield (1992) has documented, for Sarawak (Brunei's neighbour), 102 species of rattans in 13 genera.
11. This points to a distinction between uninomial (one-word) and binomial (two-word) nomenclature. When
the life-form term in not included in a binomial plant name, the first term is as a rule the generic name and
the second name is a sub-generic modifier. Dusun ethnobotanical classification below the genus level will
be discussed in another paper.
12. Before the introduction of metal scrapers rattan was also used to scrape cassava tubers.
13. It is unclear from my study that other Dusun communities also lack generic names for inedible mushrooms.
It may also be the case that women's ethnomycological knowledge exceeds that of men.
14. Cf. Taylor's (1990, p. 64) discussion of 'residue' and 'residual taxa'.
15. For interesting discussions of the ethnobiological rank or status of 'grass' see Brown (1984, pp. 20-21) and
Wierzbicka (1985, p. 190).
16. The identification of a specimen in the Brunei Checklist Project, Royal Botanic Gardens, Kew.
17. Derato is a god in Dusun cosmology. The nuts of this palm are believed to be used and preferred by
Derato. Hence this palm is 'Derato's pinang'.
18. Two other palms, uncollected, were also included in this grouping:- dudor and uwat endali'.
19. Specimen in the Brunei Checklist Project Database.
20. I provisionally consider these terms to be synonymous, but it is possible that some distinction exists between
plants grouped as sancam and those grouped as umbus.
21. Ellen (1991) has considered the classificatory status of a very large but heterogeneous collection of plants,
called 'monote', among the Nuaulu. Like Dusun sakot, they are a residual grouping and lack focal
members. Ellen considers and ultimately rejects the term 'weed', with its underlying criterion of dis-utility,
as a gloss for monote.
22. The use of reduplication in this case seems to mean 'like lamba, or lamba-ish.'
23. Additionally, two plants are identified in their names as umbus, which also means 'vegetable': umbus
kenawai (JHB 353) and umbus tunjong (JHB 132), neither identified as yet.
24. Berlin considers level to be "trivially easy to define," while the definition of 'rank' is "quite slippery"
(1992, p. 65). However, he never explains exactly what he means by level.
25. The staff of the Royal Botanic Gardens, Kew, have provided me with a listing from their Brunei Checklist
Project Database, still in progress at the time of writing. The Kew database includes the following Dusun
terms for bamboos.
Buloh baden (alt. badan) Dinochloa
Buloh gana Gigantochloa
Buloh betung (alt. abang) Gigantochloa aff. levis
Buloh nanap Schizostachyum latifolium
The Checklist also includes two specimens identified, though not specifically in the Dusun language, as
buluh balui (Gigantochloa and Gigantochloa sp. nov.).
The Socio-Economic Benefits of Indonesian Forest
MUBYARTO
Faculty of Economics, Gadjah Mada University, lalan Kaliurang, Bulak Sumur Complex,
Yogyakarta 55200, Indonesia
KEY WORDS: forest products, timber concessions, exploitation, rural development, legislation.
INTRODUCTION
Since the beginning of the New Order Government, forest products, especially timber, has
become the most important export product of Indonesia after oil. Some 60 million ha of
Indonesian production forest (out of a total of 144 million ha), has been contracted out to
579 concessionaires, most of them in East and Central Kalimantan (255 concessionaires or
Hak Pengusahaan Hutan (HPH), 24 million ha), Maluku and Irian Jaya (64 HPH, 10 million
ha), and Riau-Jambi (96 HPH, 9 Million ha). The revenue from forest exploitation has been
so large that these regions have experienced a timber bonanza, giving a degree of prosperity
to the majority of their population. In East Kalimantan, the economic impact has been so
dramatic that people from other regions have flooded the province to take part in enjoying
the bonanza.
According to Manning (1971, p.56) "The boom has brought a tremendous influx of people
into the region - Bandjarese from South Kalimantan, Bugis fishermen, and farmers and
townsfolk from Makassar, under-paid and under employer East Javanese, and Malaysians
from Sabah and Sarawak. Within the region too there has been a substantial movement of
labour into the timber and associated trades. Dajak Farmers and hunters have left their
traditional pursuits to join bandjir kap teams and earn anything from Rp 15,000-25,000 to
as high as Rp 100,00-150,000 per month. The wide variation in wages results from the fact
that extraction of timber is almost totally dependent on a heavy rainfall (bandjir) which swells
rivulets so that logs can float downstream and government officials have left prestigious
positions to take up jobs in timber companies where they may earn anything from 15,000 to
45,000 rupiahs per month. The great variety of imported goods in the night markets, the
flash of Japanese motor scooters down the narrow streets of Samarinda, the provincial
capital, and the constant traffic of speedboats or tugs towing rafts of logs down the Mahakam
and returning with supplies for the isolated timber camps all attest to the new economic
prosperity brought about by the boom. "
Legislation issued by the Indonesian Government to facilitate forest exploitation has
included the Foreign Investment Law (1967), the Domestic Investment Law (1968), the Basic
Forestry Law No. 5/1967, and Government Regulations No. 22/1967 and No. 21/1970. The
influence of all this legislation has been very effective in facilitating large scale investment.
So much so that after 20 years of forest exploitation (since 1970) there is now growing
concern about the sustainability of our tropical forest. The free hand and generous economic
stimuli given to HPH holders have resulted in these concessionaires becoming uncontrollable
in maximising their profits.
In response to this behaviour, the government is currently working on a variety of
arrangements to appropriate a greater share of the forest economic rent and to find ways to
451
D. S. Edwards et al. (eds.), Tropical Rairiforest Research - Current Issues, 451-456.
452 MUBYARTO
force the HPH holders to help rural development in the concession areas. These are the so-
called social forestry programmes which have become one of the more ambitious aspects of
the next five year plan (1994-1999).
SOCIO-ECONOMIC BENEFIT: THEORY AND PRACTICE
Theoretically, forest exploitation should provide multiple benefits to many people of the
region where the forest is located, and to the national economy as a whole. The benefits
should be in the form of increasing supply of forest products, increasing income and foreign
exchange, and of course increasing employment.
In practice, however, forest exploitation has rarely resulted in such benefits. Initially,
foreign investors found difficulties in getting skilled labour for forest exploitation, forcing
them to bring manpower from their own countries. By 1971 there were 4800 foreign
labourers, mostly from Malaysia and the Philippines, working especially in East Kalimantan.
Early in the boom period, the Government became aware of the possible conflict between
the short term objective of income and profit maximisation and the long term objective of
forest sustainability, and entrusted the State forest corporation Perhutani to hold monopoly
power. This monopoly, accompanied by government control of foreign investment did not
last long, due perhaps to inefficiency and weak management. In 1970 the joint venture
between Perhutani and a Japanese company was dissolved and Perhutani suffered a loss of
US$ 10 million. Eventually a Government Regulation (PP No. 2111970) was applied. Thus
began a new phase in forest exploitation, where the Government gave great freedom to
private companies, national as well as foreign, to operate, even going so far as to freeze the
customary right of forest utilization to avoid possible problems for the concessionaires.
The stipulation to freeze the local rights to collect forest products clearly indicated that
the Government 'sided' with the investors, allowing the companies to operate without
difficulties, and with the understanding that the government should collect large royalties
which could be used for national development. It was also expected that the forestry
companies would create employment for local people.
Government regulations did anticipate possible dangers of deforestation as a result of such
encouragement. Thus article 8 stipulated that in principle HPH holders (the forest
companies) are allowed to do only selective logging and that they are required to practice
forest conservation and reforestation.
This stipulation, however, is seldom implemented, mainly because there are not enough
foresters to enforce the practice in the field. In theory, the companies should submit an
annual working plan and a five year working plan to which they should adhere. But again
because of the lack of supervision, local forestry representatives are usually forced to
approve all the plans without question. It is therefore understandable that there is a great
danger of corruption at the various levels of the forest bureaucracy.
THE FOREST AND THE PEOPLE'S ECONOMY
Article 33 of 1945 Constitution formulated by the then economist Vice-President of the

Republic of Indonesia emphasizes the people's aspect of the economy. This aspect has
always been the basis of state guidelines, law, and other legislation. The present National
Development Plan (Garis Besar Haluan Negara) of 1993 gives even stronger emphasis on
developing the people's economy. In recent years, it has become clear that local people in
some provinces which are rich in forest resources have not received direct benefits. Equity
in wealth distribution, which is supposed to be the country's official policy (within the
Socio-economic benefits of Indonesian forest 453
framework of the trilogy of development), does not seem to be easily implemented in the
regions.
The New Order Government which in 1966 inherited an almost bankrupt economy where
poverty was widespread, had few alternatives when designing a recovery programme.
Because economic growth was a prerequisite for further development, the Government chose
a growth-oriented policy in all sectors, and for this it turned to forest and petroleum products
(oil) to finance the economic recovery programme. The Government first introduced the
Foreign Investment Act, followed by the Basic Act on Forestry, both in 1968, and in the
following year the Government issued the Domestic Investment Act.
In order to facilitate forest exploitation, the Basic Forestry Act (No. 5/1967) anticipated
possible problems of the forestry companies vis a vis local people possessing customary
rights. In one of the articles (No. 17), the customary right is recognized but exercising such
rights must not "disturb" companies holding concessionary rights granted by the Minister of
Agriculture/Forestry in Jakarta.
This 'neutral' stipulation, however, was changed drastically in Government Regulation
No. 5/1970, which explicitly froze the customary rights of the local community, in order for
forest exploitation to operate without impediment. Local people who had been dependent for
their livelihood on forest resources for generations, were forced to sacrifice their sovereignty
to the incoming investors in order for the latter to help the Government earn much-needed
foreign exchange. This resulted in perpetuating local people's poverty, and of course
exacerbated the income and social gaps, even during the time when the country as a whole
was achieving impressive economic growth.
Now in the face of the second long term National Development Plan which gives priority
to equity programmes, there must be radical change in the development strategy, emphasizing
the alleviation of poverty. In other words the new strategy must not be solely growth
oriented but more equity oriented to achieve social justice for all Indonesian people.
The Government (i.e. the Ministry of Forestry) has indeed started equity programmes, the
most recent of which has been the HPH Bina Desa in which the HPH holder must help local
rural development. The programme is new and faces many problems, especially because the
profit orientation of the HPH holders makes it difficult for them to operate useful rural
development programmes. It is also apparent that the programmes rely too much on the
'ingenuity' of HPH holders. In reality the HPH holders in turn rely too much on the
regional representatives of the Ministry of Forestry.
Socio-economic and cultural problems, and also that of shifting cultivation, can be solved
only through correcting the many different acts (laws) concerning forests and forest
exploitation. Above all, the two acts (No. 5/1967 and Government Regulation No. 21/1970)
which are not helping the development of the people's economy must be modified.
The people's economy is a self-reliant economy. Sustainable development must be based
on the development of the people's economy, especially in the rural areas. It is this rural
people's economy which should become the basis of national economic resilience. And in
the forestry sector, shifting cultivators are the principle actors. The shifting cultivators must
be protected and must be encouraged to become permanent farmers. Such considerations are
included in the social forestry programme already formulated in the 1993 State Guidelines.
THE SHIFTING CULTIVATORS AND FOREST SQUATTERS
In its overall effort to sustain the forest, the Government faces difficulties in pinpointing who
is responsible for damaging the forest. The easiest thing is to accuse the shifting cultivators
because, based on their name, they are destroying the forest to grow food crops. Thus the
only way to control them is to force them to stop the practice and to settle them as farmers.
454 MUBYARTO
The Ministry of Forestry has made an effort to do this through training cadres of settled
farmers (KANIT AP).
In 1989, it was estimated that over one million families practiced shifting cultivation,
covering an area of 7.3 million ha, so that each family of five would cultivate 6.9 ha. Even
though the Government has attempted to reduce shifting cultivation, there is strong evidence
that the area affected keeps increasing. An intensive study has shown that forest damage
(estimated to be expanding at 1.3 million ha per year) is not caused solely by shifting
cultivators but is also the result of deliberate logging for short term gain. It is suspected that
the HPH holders, through close cooperation and under the protection of powerful individuals
may be involved in such illegal logging because the 'stolen' timber is much cheaper. It is
reported that application of strict controls on timber cutting in Samarinda for example,
quickly increased the timber price in Balikpapan to a 'normal' level.
Native local people involved in shifting cultivation have indeed carefully practiced
sustainable forest management, and are likely to become productive settled farmers if seeds,
fertilizer, and farm implements can be purchased cheaply or obtained on credit. On the other
hand Government action is needed to discourage the illegal forest squatter, rather than those
exercising traditional rights. At the same time the illegal logging must be stopped.
The Ministry of Forestry is the most suitable agency to take responsibility for ensuring
the sustainability of the forest. In the 1993 State Guideline, it is stipulated that:
1. Forest exploitation and forest product utilization is to be organized, jointly with the local
Government, in such a way as to guarantee the highest revenue for the State, in line with
the interests of the people living in and around the forest area.
2. Forest exploitation must result in minimal damage and utilization of the forest must be
conducted within the framework of the law.
3. Forestry extension must be aimed at upgrading forestry personnel and increasing the
participation of the local community in the form of social forestry.
FOREST ECONOMIC RENT
The economic aspect which has attracted much debate among writers is the 'inequitable'
sharing of forest economic rent among the HPH holders, the Government, and the local
community. Following a national seminar on the economic aspects of forest exploitation held
in October 1992, the journal Prisma (No.6, 1992) focussed on this problem.
The problem was the subject of a Doctoral Dissertation submitted to London University,
and a summary of this research was presented in BIES in July 1979. It was shown that
during 1973/1974, the average economic rent collected by the Government was as low as
30 % of the fob price, while the portion received by the timber company was more than two
times greater (63% of the fob price) (Ruzicka 1979).
It is astonishing that 13 years later, there had been no improvement in the government
share, which even dropped to 8% in 1989 and 17% in 1990 (Mubariq 1992). In absolute
terms the 1989 economic rent in terms of stumpage value should have been US$ 99.24 per
m3 of log, so that for the total timber production of 31,215,000 m3 the total economic rent
would have been US$ 3,098,000,000. However, in the case of logs which were not exported
(export was banned completely in 1985) the economic rent, based on the stumpage value was
US$ 29.66 per m3 , so that the uncollected economic rent amounted to US$ 13.66 or 46%.
Thus the domestic price of timber, which at that time was only 55 % of the fob price,
decreased the stumpage value of logs to only one third (31 %).
It is now clear that the failure of the Government to collect the economic rent represented
a 'hidden subsidy' to the logging industry, resulting in the domestic price of timber being
lower than it should have been. This in turn created an excessive demand for logs. The great
Socio-economic benefits of Indonesian forest 455
demand for logs on the domestic market pushed up the price and this resulted in excessive
logging. Based on this economic analysis, if the Government desires to encourage sustainable
tropical forest logging in Indonesia, there is a need to collect a greater share of the total
economic rent which should be used by the Government to finance reforestation.
In recent years the Government has encouraged the development of forest plantations
(Hutan Tanaman Industri, HTI) using the funds collected from the Reforestation Fund (DR).
The development of HTI will of course reduce the pressure of demand on the natural forest,
especially for the supply of pulp and paper factories. In 1991/92 it was reported that HTI
areas reached 516,038 ha, consisting of 207,058 ha in Sumatra, 165.331 ha in Kalimantan,
117,417 ha in Java, and 26,232 ha in other regions (Soeharto 1992). That HTI developed
faster in Sumatra than in Kalimantan is an indication that the remaining natural forest area
is still much larger in Kalimantan. In addition, most of the pulp and paper factories are
located in Sumatra, primarily in Riau and North Sumatra.
There is a danger that encouraging the HTI policy may accelerate deforestation. This is
likely because there is a provision that allows transfer of HPH concessions to HTI. This
means that the requirement for selective logging can be changed to clear felling in order for
concession holders to become eligible to apply for HTI concessions. The HTI holder is then
eligible to receive government capital support of up to 46% of the total investment; the
remainder can be financed by credit from the reforestation fund at zero interest rate. Under
this scheme, there is a danger that these incentives will encourage the creation of fictitious
areas of HTI to obtain subsidies, which may be used for other investment projects. At the
same time the HTI becomes a 100% government project owned by private companies.
RECOMMENDATION TO THE HPH HOLDERS
The HPH holder is a business entity and has a 'mission' to earn profit. Since sustainable
harvesting requires an allocation of funds, this means reducing possible profits. The
Government must work out how to allow the company to earn sufficiently attractive profits
but at the same time set aside sufficient funds to sustain the forest.
In addition, since the Government desires each company to absorb as many labourers as
possible, there is a need for other incentives to achieve this latter objective, for example by
offering an incentive for every 100 or 1000 labourers employed. It is in this context that the
social forestry approach requires the HPH holder to involve more and more local
participation in forest exploitation, or, looking at the problem from the other side, that the
company must pay more attention to rural development. If cooperation can be improved
between the company and the local people, not only can government supervision be reduced,
or made easier, but community participation will also be made more effective.
At the present time, relationships between the HPH company and the local people are not
cordial, the causes of which can be traced back to government regulations. In China,
reforestation is the responsibility of the whole village community, so that the individuals who
actually carry out the reforestation are controlled directly by the village community. If this
system can be applied in Indonesia, it would increase community participation in forest
sustainability. However, in order to do this the Government and the Parliament must take
the necessary steps to change some forest legislation. Government regulations which
separated the HPH company and the local people must be abolished and replaced with
regulations emphasising the social forestry approach stipulated in the 1993 State Guidelines.
Of course this must be implemented in the context of increasing regional autonomy and
decentralization related to the implementation of Law No. 5/1974 on government in the
regions.
On the other hand, the HPH holder and the forest industries must realize the need to apply
456 MUBYARTO
this social forestry approach. The HPH company needs to establish a special section and to
appoint special personnel responsible for applying social forestry approach. The need for
a sustainable forestry management programme is urgent since the target date to achieve
sustainability, set down by the International Tropical Timber Organisation (lTTO), of which
Indonesia is a member, is the year 2000. Some people consider that Indonesia will be able
to satisfy the ITTO criteria in 1995. However, if we examine present forest conditions in
some regions, one cannot be so optimistic.
Companies will need to set aside some portion of their profit to contribute to the
achievement of sustainable forest management. They will have to pay more attention to rural
development. It must be noted that many of the HPH holders do not really know what to
do because of the multitude of regulations issued by the Ministry. which change too quickly.
It is recommended that the HPH companies should be more active in proposing policies to
the Ministry.
CONCLUSIONS
The socio-economic benefits of the Indonesian forest are now being questioned because of
the demand from the people to improve living conditions. Sustainable forest management
must be achieved in order to guarantee a decent standard of living for the next generation.
ITIO set the target of sustainable forest management by the year 2000 and Indonesia is
optimistic that this target can be met. However, not all of our working methods are
sufficiently good. In fact much of the legislation must be modified to accommodate the
needs of a wider sector of society.
The 1993 State Guidelines explicitly mentioned that forest destruction must be stopped and
that forestry companies should pay higher royalties to the Government. If these funds can
be paid to the Government early every year, then the Government is in a much better
position to implement sustainable forest management.
LITERATURE CITED
MANNING, C. 1971. The Timber Boom. Bulletin of Indonesian Economic Studies Vll(3).
MUBARIQ, A. 1992. Rente Ekonomi da1am Eksp10itasi Hutan Tropis. Prisma 6.
RUZICKA, I. 1979. Rent Appropriation in Indonesia Logging: East Kalimantan. Bulletin of Indonesian Economic
Studies XV(2).
SOEHARTO. 1992. Pidato Kenegaraan (presidential State Address 1992). Pp. 242-243.
Putting plants in their place: Anthropological approaches to
understanding the ethnobotanical knowledge of rainforest populations
ROY ELLEN
Department of Sociology and Social Anthropology, Elliot College, University of Kent at

Canterbury, Canterbury, Kent CTl 7NS, United Kingdom
INTRODUCTION
Over the last decade or so ethnobotany has assumed a scientific prominence previously
denied it. It is endorsed by institutions with a high international profile (Kew, the Royal
Geographical Society, WWF, UNDP, UNESCO), has a market value placed upon it by
foresters, agronomists, development advisers and pharmacologists, and has become pivotal
in preserving the cultural identity and knowledge of indigenous peoples whose traditional way
of life is under threat (Posey 1990). Ethnobotanical knowledge has, therefore, become both
economic commodity and political slogan. This is particularly true with respect to the plant
knowledge of rainforest peoples, as these peoples are often those with the highest media
profile. However, in our eagerness to exploit a product and to demonstrate its usefulness,
there has been a tendency to oversimplify what ethnobotany entails and just how it can be
useful. I argue in this paper that we must not be narrow-minded or simplistic in our
conception of ethnobotanical knowledge, and that to take anything less than a broad
culturally-contextualised approach may miss the point of the relevance of indigenous
knowledge altogether.
In an historical review of the subject published a little over a decade ago, Richard Ford
(Ford 1978, p. 29) says of ethnobotany that it represents a common discourse but lacks a
unifying theory. That this is still the case is in part due to the historical differences between
what we might call biological ethnobotany and anthropological ethnobotany. I The first
operates within a bioeconomic paradigm. At its narrowest this has been simply the study of
plants used by indigenous people (whoever they might be). Data thus provided have often
been no more than species identifications against vernacular names, and lists of uses. In its
most recent version this approach has become heavily identified with the discovery of new
applied botanical and phytochemical possibilities, and with the economic valuation of
rainforest (Peters et al. 1989); and so becomes virtually synonymous with what is widely
understood as the remit of economic botany. Although the global significance of such work
should not be underestimated, bare lists are of limited application (see e.g. Alcorn 1981a,
p. 228).
The second kind of ethnobotany operates primarily within a cultural-linguistic paradigm,
is concerned with indigenous rules and categories and is particularly evident in the work of
Harold Conklin, Brent Berlin and those who have been influenced by them in the US; and
in the work of those within a Francophone tradition closely associated with what is now the
Laboratoire d'Ethnobiologie-Biogeographie in Paris, and which finds much of its intellectual
legitimation in the writings of Claude Levi-Strauss. Although there is some overlap between
IThis is similar to (though by no means identical with) other distinctions which are sometimes drawn, such as
between different 'levels' of ethnobiological enquiry (Bulmer 1974).
457
D. S. Edwards et al. (eds.), Tropical Rainforest Research - Current Issues, 457--465.
458 ROY ELLEN
the data and findings of the two kinds of ethnobotany, it is the second which - especially
when situated in systematic botanical investigation - comes closest to what Ford (ibid., p. 44)
believes best exemplifies what ethnobotanists should be doing: namely, the direct
investigation of the relations between plants and humans, placing plants in their total cultural
context.
I wish to explain here why it is vital to adopt a 'progressive contextualist' (Vayda 1983)
approach, beginning at the species level and working outwards; an approach which locks
specific local knowledge within increasingly more general but denser culturally-relative
paradigms, and which links indigenous ecological know-how to general subsistence and social
behaviour. This is not - nor should it be - privileged or esoteric anthropological
methodology, but a strategy of practical relevance to botanists, foresters and all those
working in the development field.
The main weakness of the first kind of ethnobotany, then, is one of definition: that for
many it has become simply that empirical knowledge embodied in fixed folk traditions which
can be of value to global science. A pharmacologist looking at ethnobotany in this sense thus
sees potential drugs, a botanist scientifically-unrecorded species, food scientists new foods,
materials scientists new materials (for example, milled coconut fibre as a substitute growing
medium for peat), and the Body Shop some new politically-correct cosmetic. Such an
approach tends to reduce ethnobotany to partial unconnected bits: or to put it another way,
knowledge is transformed into information. In the process much of potential value is lost.
Rather than generating selected bits of information in a framework determined by the quite
specialised requirements of conventional biological science and taxonomy, we should be
focussing on connected systems of local knowledge, informed by an understanding that such
knowledge is intrinsically highly variable between individuals and populations,
fundamentally situational, and dynamic. My remaining remarks - which put more flesh on
these bare assertions - fall under four headings: (a) species-focussed empirical knowledge,
(b) knowledge of the rainforest as an entity, (c) the interconnectedness of social and mundane
knowledge, and (d) domains of ethnobotanical knowledge defined functionally. I shall
conclude by returning to the question of variability and change.
SPECIES-FOCUSSED EMPIRICAL KNOWLEDGE
By species-focussed knowledge I mean those data which an investigator will primarily elicit
with a single voucher specimen as the unit of analysis, or multiple specimens of the same
'natural kind'. This will include all names which informants give to a particular specimen,
the terminology used for its parts, knowledge of variation in form, its reproductive cycle,
seasonal and geographic distribution, growing conditions; in fact, all information that a field
botanist would ordinarily seek concerning a particular species being investigated. The
secondary data, secondary only in some putative logical order of elicitation, refer to cultural
uses, and practices with respect to the specimen type. These include how it is located,
collected, prepared for use, protected and cultivated, including of course its recognised
effects: material, nutritional, medicinal or symbolic; and evaluations of the plant type and
particular specimens of the plant. As Hays has stressed (Hays 1974, 1982), it is really only
when we examine our data in this way that we can see what constitutes a 'use', and that one
use is by no means functionally equivalent with another. A third dataset at this level consists
of associated ecological knowledge. This includes knowledge of the identity and activity of
predators and diseases, the effects which the plant has on the habitat in which it is found and
on associated flora, and the effects they have on the plant. Rich data in this area has
emerged over the years with respect to specific cultigens, for example the important work
of Paul Richards on folk understandings of Zonocerus infestations of cassava in southern
Ethnobotanical knowledge of rainforest populations 459
Nigeria (Richards 1979, 1985). More recently, and of more concern to rainforest specialists,
Darrell Posey (Posey 1988) has shown how Amazonian Kayap6 maintain buffer zones
between gardens and forest which contain plants with nectar-producing glands on their
foliage which have the effect of drawing away aggressive ants and parasitic wasps from
crops. Then there are the effects of different patterns of rainfall, predation and human
extraction; the role of other organisms in the dispersal of seeds, and the use made of the
plant by other non-human organisms as food. Nuaulu, for example, in the lowland
rainforests of Seram in eastern Indonesia, are well-informed on many species not because
they are directly of use to humans, but because they represent the food of animals which they
hunt, particularly cassowary, pig and deer. In other words, plants have to be understood as
part of the web of forest life, not simply in isolation.
Although such knowledge as discussed here can be presented in terms of a checklist for
the species as a whole, it is important to realise that most folk-botanical knowledge differs
fundamentally from conventional scientific knowledge in not being organised abstractly
within some convenient general-purpose classification, but rather with respect to particular
contexts, defined perhaps in terms of different subsistence activities (hunting, planting trees,
collecting, artefact manufacture, burning swidden trash, or whatever). Moreover, how that
knowledge is apprehended by people will be determined by culturally relative coordinates of
sense perception which sometimes deviate sharply from the expectations of
scientifically-trained personnel; for example, the significance of olfactory and textual stimuli
compared with the purely visual. Thus, much knowledge is inaccessible except via a
research strategy which allows a multi-focal approach; and if investigators additionally wish
to appreciate how local people make key subsistence decisions, they must attend to the
categories of knowledge locally applicable (Ellen 1982). To do all this requires a general
understanding of the underlying principles of ethno-botanical knowledge (e.g. Berlin et al.
1974, Conklin 1954), but goes beyond the more formal systematical aspect of folk
classification, to what we might call folk-autoecology (rather, that is, than folk-synecology).
But, then, that ethnobotany and the study of indigenous knowledge of particular plants needs
to be ecological has been recognised for over half a century (Carter 1950, Jones 1941), even
if the implications have seldom been followed through, or its relevance fully understood.
Species-focussed knowledge is, additionally, variable between individuals, and so may not
be fully documented until a number of people have been queried regarding the same species,
or indeed the same individual has been interviewed on the same subject on a variety of
occasions. Moreover, since such knowledge is situational, the optimal conditions for finding
out are with respect to growing specimens, at different stages in the life-cycle, in contexts
where there is a predetermined subsistence goal, such as cutting bush for swiddens or
collecting fungi for food. Just as knowledge is not fixed in its contemporary distribution,
so it also changes through time: the result of generations of trial-and-error testing, extensive
experimental evidence, enormous individual specialist and collective experience; new plants
moving in and out. This does not mean that mistakes are never made, but simply that daily
routine pragmatic considerations act as a major incentive for identifying effectiveness. As
in science as a whole, there may be bad folk science; but there is also a wealth of good
science (Richards 1993). It is this which pharmaceutical firms are now taking advantage of,
but they and us also need to appreciate how that experience has been gained.
KNOWLEDGE OF THE RAINFOREST AS AN ENTITY
Ethnobotanical studies of rainforest peoples have uniformly demonstrated an impressive

breadth of knowledge of plant species, in addition to a depth of knowledge of particular
significant species. A recent attempt to collate data on the total inventories of plant
3
460 ROY ELLEN
categories for different subsistence populations (Brown 1985, p. 44) shows strikingly how
rainforest populations have repeatedly been found to yield much longer lists than populations
living in other environments, lists which may consist of between 800 and 2000 items at the
upper end. Clearly, to some extent this reflects the subsistence necessity of those who
extract from such environments, and also the strong tendency for non-significant and
non-useful species to go unnamed or to be lumped within larger generic categories. Though
it is still debated whether or not human populations could ever have entirely survived on
rainforest plant matter, other than when they can rely on palm starch (Headland 1987),
breadth of knowledge is undoubtedly a key part of any adaptive strategy.
In addition to this, it has become clear that systematic encyclopaedic knowledge is situated
within folk-models which reflect an ability to connect observations at the species level with
informed perceptions about forest structure and dynamics. What constitutes 'forest' is, of
course, something we might expect to vary cross-culturally, but even if we restrict ourselves
to focal shared meanings it is clearly a complex categorical construction. Thus, the Nuaulu
generic term wesie (Ellen 1993) is anything but uniform or empty in the way they perceive,
understand and respond to it. It is more like a mosaic of resources, and a dense network of
particular places each having different material values. In this sense it is much like the
modem scientific modelling of rainforest as a continuous aggregation of different biotopes
and patches, varying according to stages in growth cycles, degree of regeneration, underlying
geology, altitude, geography and natural contingency. In such a view, simple distinctions
between 'secondary' and 'primary' begin to look pretty academic. Seventy-eight percent of
the 272 forest trees identifiable by Nuaulu have particular human uses, and it is through
their uses that they are apprehended (Ellen 1985), wherever they are found. The picture is
similar elsewhere. Carneiro (Carneiro 1988, p.79) reports that, on average, Panare, Tembe,
Urobu and Chacobo peoples of the Amazon basin use at least two-thirds of the tree species
growing in the forests (see also Carneiro 1978). Such peoples, like the Indonesian Nuaulu,
being forest-fallow swidden cultivators, also have sensitive understandings of how forest
changes as a consequence of different soils, selective extraction, cutting and burning; and of
the regrowth stages following abandonment. It is this knowledge which permits such
strategies - despite persistent rumours - to be self-sustaining (Dove 1983a, b), and which
underlies its deliberate application in a way which assists the recovery of degraded areas
(Shanley 1991). The classic description of this knowledge is still that by Conklin (Conklin
1957) for the Hanunoo of Southeastern Mindoro.
THE INTERCONNECTEDNESS OF SOCIAL AND TECHNICAL KNOWLEDGE
Empirical knowledge of plants, as I have referred to so far, does not exist apart from a
broader socially-informed understanding of the world, in some kind if hermetically sealed
vacuum from which other aspects of culture are excluded. Detailed knowledge of plant
reproduction or symbiosis may, for example, comfortably co-exist with beliefs about the
world which have not been empirically tested in a conventional scientific sense. Everything
is seen as connected through chains of mutual causation to give rise to a complex notion of
nature. Indeed, it is often plausibly argued that rainforest peoples have cosmologies which
in certain respects anticipated the systems view of the world which underlies modem ecology;
as argued, for example, by Reichel-Dolmatoff (1976) for the Tukano of the northwest
Amazon. Such a view is also reconstructable for the Nuaulu, and we can infer their
conceptualisation of forest and attitudes towards it from systematic ethnobiological data, from
human subsistence practices with respect to it, from the rules of land tenure, from general
statements about forest, from rules relating to the extraction of resources and sanctions
consequent upon their infringement, indirectly from myths, stories, taboos, and so on. From
these we can distill four general characteristics which, taken together, summarise the most
important aspects of Nuaulu conceptual engagement with 'forest'. The first characteristic has
already been discussed, namely that forest is never experienced as homogeneous. The
remaining three are that forest is a complex categorical construction, that there is an inner
connection between history, identity and forest, and that forest is not morally neutral.
Although uncut forest is recognised by the Nuaulu as a single entity (wesie), it contrasts
in different ways with other land types depending on context. It may contrast with wasi
(owned land, which may sometimes display very mature forest growth), emphasising a jural
distinction; with nisi (garden land), emphasising human physical interference; or with niane
(village), emphasising landforms: empty as opposed to well-timbered space, inhabited (dwelt)
as opposed to uninhabited space, untamed as opposed to tamed space, all with various
symbolic associations and practical consequences for Nuaulu consumers. Although there are
no Nuaulu words for either 'nature' or 'culture', it is in the various and aggregated senses
of wesie that the Nuaulu come closest to having such a term, and from which the existence
of an abstract covert notion of 'nature' can reasonably be inferred.
The values which Nuaulu attach to forest are thus multi-faceted and differential,
simultaneously materially useful and culturally meaningful. And in the same way that the
material uses to which forest is put must be understood in specific and local terms, so too
the social implications. While the Western conception of environment is something which
is 'opposed' to people, or some kind of medium in which we dwell, and which is therefore
bounded; the Nuaulu conception of environment is not as a space in which they hang, but
much more like a series of fixed points to which particular clans and individuals are
connected. These points are objects in an unbounded landscape linked to their appearance
in myths; use of land is at every turn inseparable from specific sacred knowledge, sometimes
mutually contradictory and obscure, though never absent.
The undeniable effect of merging practical usefulness, mythic knowledge and identity in
the construction of the category wesie is to give it a moral dimension. That is, there are
right and wrong ways in which to engage with forest which arise in part from the specific
social histories of parts of it, but also from its intrinsic mystical properties. Forest is
unpredictable, dangerous and untamed, and various attempts are made to control it. This is
reflected in the inferential symbolic opposition between 'nature' and 'culture' evident in most
ritual, in the specific rituals conducted prior to cultivating forest, in the charms which are
used to protect travellers in the forest, in the prohibitions on certain behaviours and
utterances while in the forest, in the correct ritual disposal of its products.
The practical implications of the interconnections between the social and the environmental
can be very important, and it is often the case that subsistence practices triggered by cultural
beliefs (for example, linked to prohibitions) appear to regulate resources. It is in the context
of all this that we must understand Nuaulu ritual restrictions (sasz) on harvesting certain
forest products at particular times. Certainly, the effect of all of these things may well be
to conserve resources and maintain biodiversity, and in particular cases people may
consciously do so. But we should not confuse effect with purpose. Much of what appears
to be 'ecological balance' amongst rainforest peoples is either illusory or simply a beneficial
function of low population densities and benign subsistence practices. However, when
governments and other agencies interfere and seek to introduce 'rational' measures to
conserve resources, ignorant of local cultural representations of the forest, their purposes
may be meaningless to local peoples, as Richards (Richards 1992) has shown for Mende
living on the edges of the Gola in Sierra Leone. Similarly, governments acting with the best
of intentions may interfere with cultural regulators (purposeful or inadvertent) which are
often more sensitive, and in the long term, more effective (e.g. Morauta et al. 1982).
462 ROY ELLEN
FUNCTIONALLY-DEFINED DOMAINS OF ETHNOBOT ANI CAL KNOWLEDGE
All human knowledge is represented and articulated not only through the systematic lens of
general-purpose classifications, such as those offered by science, but also through the
domains constructed around specialised applied knowledge. I have already noted how
specifically 'botanical' knowledge is embedded in the particular contexts in which plants are
encountered, but knowledge can also be presented in terms of functionally-defined
abstractions, such as in the context of medical or veterinary knowledge, specific craft
specialisations (such as house building), and so on. Indeed, many peoples who might
otherwise deny formal knowledge of ethnobotany, may in fact have a considerable practical
working knowledge known to them under other headings (see e.g. Brown 1992, Telban
1988). Thus, the unit of analysis (in this case most usually the voucher specimen) can
always be organised into arrays which demonstrate significant discontinuities other than those
found in classifications based predominantly on morphology.
VARIATION AND CHANGE
Ethnobotanical knowledge and practice must, therefore, be understood contextually, but it

must also be recognised as being intrinsically variable and subject to change. One of the
problems with the kind of narrowly-conceived ethnobotany which I have been criticising is
that it conceals much of this. We need to take into account intra-cultural variation, often
hidden in over-generalised accounts which imply that one respondent is as good as any other
with respect to a particular subject. We now have convincing demonstrations that knowledge
may vary qualitatively and quantitatively according to crucial social variables, such as
gender, and even within groups defined in terms of a single key social indicator (e.g. Hays
1974). We also need to take into account variation between different rainforest populations,
reflecting for example different modes of subsistence (Brown 1985), though it may be
artificial to separate out populations on the basis of their apparent degree of interaction with
forest, degree of acculturation or integration into the market. I take a broad view of what
constitutes a rainforest population, including all those (such as peasants) interacting with the
rainforest (Denslow & Padoch 1988). Knowledge may pass between superficially different
groups which are in contact or who extract from similar biotopes. We must be wary of
inventing a category of 'traditional' peoples whose knowledge is regarded as somehow
pristine and superior, however much the temptation.
Narrow conceptions of the ethnobotanical enterprise also sustain a non-dynamic view of
particular human populations. In the popular imagination, peoples of the tropical rainforest
have hitherto been remote, isolated, living in more or less the same place, unchanging, living
'in harmony' with their surroundings. In fact, we now have plenty of evidence to the
contrary. Ironically perhaps, it is ethnobotany which has provided much of this, in terms
of the movement of plant species between isolated points on all sides of the globe. The
migration of the main plant domesticates is well-documented, though we know less of the
impact this has had on the forests themselves. Certainly, human settlement has introduced
many varieties of cultivated aws into the lowland forest of Seram in the Nuaulu area, both
deliberately and inadvertently. For example, Tectona grandis and Toona sureni are now
established, though they were probably introduced during the seventeenth century; and the
ornamental Delonix regia cannot have been planted before the nineteenth (Ellen 1985,
p. 563). And the picture is no different as far as foraging populations are concerned (e.g.
Fox 1952), where a nomadic lifestyle has served well to distribute certain humanly-valued
trees widely. An excellent example of this is the role of the Mbuti of the Ituri forest of Zaire
(Ichikawa 1992) in the propagation of genera such as Canarium and Landolphia. Many
rainforest peoples selectively fell trees, protect valued species, replant the tops of wild
Dioscorea tubers, transplant palm suckers, on a sufficiently systematic basis for us to speak
of ' rainforest management'. Alcorn (Alcorn 1981b, p. 410) reports for the Huastec of
Mexico's Sienna Madre that 63 percent of 800 wild species documented had uses and 25
percent were actively manipulated. The selective extraction of wild species, strategic
burning, and swiddening at optimal conditions may combine to give rise to distinctive patches
and new opportunities for colonisation; and there are numerous examples - such as those
provided by the Kayap6 (Posey 1988, p. 89) - of the deliberate preservation of corridors of
mature forest between plots as some kind of biological reserve. The cumulative effect of this
in some areas must be considerable, and persistent human interventions may be presumed to
have co-evolutionary consequences. Such disturbances not only improve the rainforest as a
human resource base but contribute in significant ways to its structural patchiness and
biodiversity. They also signal, along with the gratuitous destruction of vegetation for
short-term gain, such as the felling of entire trees to catch a single arboreal mammal (Ellen
1986, Rambo 1985), that the practical impact (if not the cosmological representation) of even
the demographically smallest human group need not always be benign. Many tropical forest
peoples have complex histories of movement, no better illustrated than from the intricate
ethnic mosaic of northern Borneo (see e.g. Freeman 1970). Trade and exchange has existed
for centuries between upriver peoples, including remote foraging populations, linking them
to peoples of the forest fringes, the estuaries, and ultimately the global economy (Dunn 1975,
Hoffman 1984).
We now have a considerable body of ethnobotanical knowledge and a growing volume of
literature pertaining to rainforest peoples (Conklin 1972); for southeast Asia alone there is
a critical mass (Barrau (n.d.». But the data are of variable quality, ranging from the
fragmentary notes of field botanists, the systematic inventories of an active generation of
colonial officials (e.g. BurkillI935), and detailed long-term studies within an anthropological
context (e.g. Friedberg 1990, Revel 1990, see also bibliography in Brown 1985). The
priority now, as I see it, is to organise such information into accessible databases and
recognise the differential quality of our data (some of which is shoddy, much of which is
inadequate). But perhaps even more importantly we need to shift our focus from abstracted
empirical data on particular plants to contextual accounts. Indigenous peoples have
perceived, interacted with and made use of tropical rainforest in historically diverse ways,
and it is clearly important for those making recommendations in the fields of conservation
and sustainable forest management to take this into account.
ACKNOWLEDGEMENTS
This paper is a by-product of an ESRC funded project (ROoo 23 3088) on 'The ecology and
ethnobiology of human-rainforest interaction in Brunei'. I would like to thank Jay Bernstein
for his comments, and am additionally grateful to the British Council for the award of a
travel grant to make attendance at the UBD-RGS conference possible. Perhaps I should
explain - at the risk of seeming an imposter - that my own empirical work has primarily
focussed on ethnozoology. This paper might have had such a focus, and still dealt with
essentially similar issues, though I judge that current debates concerning the role of
indigenous rainforest knowledge assume that this is largely a matter of ethnobotany.
464 ROY ELLEN
LITERATURE CITED
ALCORN, J.B. 1981a. Factors influencing botanical resource perception among the Huastec: suggestions for future
ethnobotanical enquiry. Joul7Ull of Ethnobiology 1:221-230.
ALCORN, J. B. 1981 b. Huastec noncrop resource management: implications for prehistoric rainforest management.
Human Ecology 9:395-417.
BARRAU, J. (n.d.). An ethnobotanical guide for anthropological research in Malayo-Oceania. UNESCO Science
Cooperation Office for Southeast Asia, (paris).
BERLIN, B., BREEDLOVE, D. & RAVEN P. 1974. Principles of Tze/tal plant classification: an introduction to
the botanical ethnography of a Mayan-speaking people of the highland Chiapas. Academic Press, New York.
BROWN, C. H. 1985. Mode of subsistence and folk biological taxonomy. Current Anthropology 26:43-53.
BROWN, K. 1992. Medicinal plants, indigenous medicine and conservation of biodiversity in Ghana (GEC 92-36).
Norwich: CSERGE Working Papers, UEA.
BULMER, R. N. H. 1974. Folk biology in the New Guinea Highlands. Social Science Information 13:9-28.
BURKILL, T. H. 1935. A dictionary of the economic products of the Malay peninsula (2 volumes). Crown Agents
for the Colonies, London.
CARNEIRO, R. 1978. The knowledge and use of rainforest trees by the Kuikuru Indians of central Brazil. In
Ford, R.I. (ed.). The nature and status of ethnobotany. Anthropological Papers 67, Museum of Anthropology,
University of Michigan, Ann Arbor, Michigan.
CARNEIRO, R.L. 1988. Indians of the Amazonian forest. In Denslow, 1. S. and Padoch, C. (eds). People of the
tropical rainforest. University of California Press, Berkeley.
CARTER, G.P. 1950. Ecology-Geography-Ethnobotany. Scientific Monthly 70:73-80.
CONKLIN, H.C. 1954. The relation of Hanunoo culture to the plant world. Unpublished Ph.D. dissertation. Yale
University, New Haven, Connecticut.
CONKLIN, H.C. 1957. Hanunoo agriculture, a report on an integral system of shifting cultivation in the Philippines.
Food and Agricultural Organization of the United Nations, Rome.
CONKLIN, H.C. 1972. Folk classification: a topically arranged bibliography of contemporary and background
reference through 1971. Department of Anthropology, Yale University, New Haven, Connecticut.
DENSLOW, 1. and PADOCH, C. (eds) 1988. People of the tropical rainforest. University of California Press,
Berkeley.
DOVE, M.R. 1983a. Forest preference in swidden agriculture. Tropical Ecology 24:122-142.
DOVE, M.R. 1983b. Theories of swidden agriculture and the political economy of ignorance. Agroforestry Systems
1:85-99.
DUNN, F.L. 1975. Rainforest collectors and traders: a study of resource utilization in modern and ancient Malaya.
Monographs of the Malaysian Branch of the Royal Asiatic Society 5.
ELLEN, R.F. 1982. Environment, subsistence and system: The ecology of small-scale social informations.
Cambridge University Press, Cambridge.
ELLEN, R.F. 1985. Patterns of indigenous timber extraction from Moluccan rain forest fringes. Journal of
Biogeography 12:559-587.
ELLEN, R.F. 1986. What Black Elk left unsaid: on the illusory images of Green primitivism. Anthropology Today
2:8-12.
ELLEN, R.F. 1993. Rhetoric, practice and incentive in the face of the changing times: a case study of Nuaulu
attitudes to conservation and deforestation. In Milton, K. (ed.). Environmentalism: the view from anthropology.
ASA Monograph 32. Routledge, London.
FORD, R.I. 1978. Ethnobotany: historical diversity and synthesis. In Ford, R.I. (ed.). The.nature and status of
ethnobotany. Anthropological Papers 67. Museum of Anthropology, University of Michigan, Ann Arbor,
Michigan.
FOX, R.B. 1952. The Pinatubo Negritos: their useful plants and material culture. Philippine Journal of Science
81:173-414.
FREEMAN, J.D. 1970. Report on the /ban. LSE Monographs 41. Athlone, London.
FRIEDBERG, C. 1990. Le Savoir Botanique des Bunaq. Percevoir et Classer dans Ie Haut Lamaknen (Timor,
Indonesie). Memoires du Museum National d'Histoire Naturelle, Botanique 32. Museum National d'Histoire
Naturelle, Paris.
HAYS, T. E. 1974. Mauna: Explorations in Ndumba ethnobotany. Ph-D. dissertation, University of Washington,
Seattle.
HAYS, T. E. 1982. Utilitarianladaptationist explanations of folk biological classification: some cautionary notes.
Journal of Ethnobiology 2: 89-94.
HEADLAND, T. N. 1987. The wild yam question: how well could independent hunter-gatherers live in a tropical
rainforest ecosystem. Human Ecology 15: 463-492.
HOFFMAN, C. 1984. Punan foragers in the trading networks of southeast Asia. In Schrire C. (ed.). Past and
present in hunter-gatherer studies. Academic, London.
ICHIKAWA, M. 1992. Traditional use of tropical rainforest by the Mbuti hunter-gatherers in Africa. pp.
305-317 in Itoigawa, N., Sugiyama, Y., Sackett, G. P. and Thompson, R. K. R. (eds). Topics in primatology:
behaviour, ecology and conservation. University of Tokyo Press, Tokyo.
JONES, V. H. 1941. The nature and scope ofethnobotany. Chronica Botanica 6:219-21.
MORAUTA, L., PERNETRA, J. and HEANEY, W. (eds) 1982. Traditional conservation in Papua New Guinea:
implications for today (Monograph 16). Institute for Applied Social and Economic Research, Papua New Guinea.
PETERS, C. M., GENTRY, A. H. and MENDELSOHN, R. O. 1989. Valuation of an Amazonian rainforest.
Nature 339: 665-6.
POSEY, D. 1988. Kayap6 Indian natural-resource management. In Denslow, J. S. and Padoch, C. (eds). People
of the tropical rainforest. University of California Press, Berkeley.
POSEY, D. A. 1990. Intellectual property rights: what is the position of ethnobiology? Journal of Ethnobiology 10:
93-98.
RAMBO, A. T. 1985. Primitive polluters: Semang impact on the Malaysian tropical rainforest ecosystem.
Anthropological Papers 76, Museum of Anthropology. University of Michigan, Ann Arbor, Michigan.
REICHEL-DOLMATOFF, G. 1976. Cosmology as ecological analysis: a view from the rain forest. Man (N.S.)
11: 307-318.
REVEL, N. 1990. Fleurs de Paroles: Histoire Naturelle Palawan. Volumes I and 2. Peeters/SELAF, Paris.
RICHARDS, P. 1979. Community environmental knowledge and African rural development. IDS Bulletin 10:
28-36.
RICHARDS, P. 1985. Indigenous agricultural revolution: ecology and food production in west Africa. Hutchinson,
London.
RICHARDS, P. 1992. Saving the rain forest? Contested futures in conservation. In Wallman, S. (ed.).
Contemporary futures: perspectives from social anthropology. ASA Monographs 30. Routledge, London.
RICHARDS, P. 1993. Natural symbols and natural history. In Milton, K. (ed.). Environmentalism: the view from
anthropology. ASA Monograph 32. Routledge, London.
SHIRLEY, P. 1991. Recovery and utilisation of ethnobotanical know/edge: the Rio Capim, Eastern Amazonia.
Final Report: Ethnobotanical exchange between Asia and Amazonia. UNDP, New York.
TELBAN, B. 1988. The role of medical ethnobotany in ethnomedicine: a New Guinea example. Journal of
Ethnobiology 8: 149-169.
VAYDA, A. P. 1983. Progressive contextualization: methods for research in human ecology. Human Ecology 11:
265-281.
Valuing a tree: the ethics of environmental evaluation
DAVID H. BENNETTl
Aboriginal and Torres Strait Islander Commission
ABSTRACT. The paper looks at four sets of argwnents for valuing trees: economic, instrumental,
deeper ethical, and cultural. On the economic level, I consider material and functional values of trees.
On the instrumental level I look at the argwnents of John Passmore for the preservation of trees (or any
part of the environment) on a use basis. I also look at an additional argwnent that rainforest lock in
deadly viruses. On the deeper ethical level, I give the example of A1do Leopold's extension argwnent
and Deep Ecology. On the cultural level, I consider an idealised traditional relationship between
Australian Aborigines and their environment.
KEYWORDS: economic, instrumental, land ethic, deep ecology, Australian aborigines, value.
There are a number of ways to value a tree or a forest. Four sets of arguments for valuing
trees that I should like to discuss briefly are economic, instrumental, deeper ethical, and
cultural.
ECONOMIC VALUE
The first set of arguments are economic arguments. Perhaps one of the simplest ways to
value a tree or forest is economically. A usual way of valuing a tree economically is for its
material value, such as woodchips or timber. In the case of woodchips the value can be
determined by some formula like:
V=H+P+T+R+X+M
where Value (price per tonne) = Harvesting + Processing + Transporting + Royalties +
Taxes + Profit Margin or some similar set of variables, where value, expressed in dollars
(or yen or marks, etc.) per tonne, equates to the costs of obtaining a tonne of woodchips plus
a margin for profit. Considerations such as labour, are built into the variables of harvesting,
processing, transporting, and so on. What should be noted is that there is no variable in this
equation for value of the tree or forest. The value of a tree or forest is an externality to the
calculation. The tree is not an object of economic consideration, let alone of ethical
consideration.
This is not so in the case of trees used for timber. A timber equation would include a
factor for the type of tree or type of forest
V=H+P+T+R+X+M+L
where Value (price per tonne) = Harvesting + Processing + Transporting + Royalties +
Taxes + Profit Margin + Timber Type.
Here the tree enters the equation, because factors such as the quality, origin, colour,
hardness, scarceness, and end purpose for the timber, among other things, impact on the
price per unit. So a plank of teak (Tectona grandis) will cost more than the same quantity
lpresent address: Native Title Section, Land, Heritage and Environment Branch, Aboriginal and Torres Strait
Islander Commission, P.O. Box 17, Woden, ACT 2606, Australia.
467
468 DAVID H. BENNET
of Monterey pine (Pinus radiata). But considerations such as these do not take into account
such factors as durable wood products being carbon prisons or carbon sinks, which remove
and keep out of the atmosphere at least some of the 5.5 billion tonnes per year of
anthropogenic carbon emissions. That requires another way of determining the economic
value of a tree, besides its material value.
Another and more interesting economic analysis is the functional value of a tree. One
such economic functional analysis estimates the value of a single mature tree at $196,250 (in
Singapore dollars).
According to Professor T.M. Das of the University of Calcutta "a tree living for 50 years
will generate $31,250 worth of oxygen, provide $62,000 worth of air pollution control,
control soil erosion and increase soil fertility to the tune of $31,250, recycle $37,500 worth
of water and provide a home for animals worth $31,250. This figure does not include the
value of the fruits, lumber or beauty derived from trees". 2
Here it is the services that a tree provides that determine its value. This method is crude,
of course. It indicates only an average functional value for a hypothetical tree. The
functional value of a given tree of a given species might be more or less than this amount.
The point remains that the economic value of a tree cannot be reduced to its materials. A
living tree has values that it does not retain when converted to chips or lumber. It will no
longer remove carbon from the atmosphere even if it may continue to imprison carbon it
trapped while alive. Even a standing dead tree will have functional values that it will cease
to have when converted to chips or lumber. It will no longer provide shelter, food, and
protection to non-human species, independent of any value it may have for humans. And this
point can be taken in the opposite direction. Just as the total value of a single tree is not the
value of one aspect, so the value of many trees is not merely the sum total of each of those
trees. A forest is more than a collection of trees. It is an ecosystem. A system whose
material and functional values exceed the value of its component parts, including not just
trees, but undergrowth, epiphytes, animal species, the relationships among these and other
elements, and so on. A forest is not just a stand of trees. Its value is not just a stand of
trees. A mere stand of trees is a tree plantation.
But there other ways, besides economic, of valuing the services that a tree provides. Also
there are other services, instrumental services, that a tree provides.
INSTRUMENTAL VALUE
The second set of arguments are instrumental arguments. Instrumental value construes nature
and the environment simply as a means to ends. Instrumental value is attributed to an item
that is in some way of value to or useful to another, usually to human ends and values. But
the ends are not necessarily economic ends. The ends may be ethical. Instrumental value
arguments are usually arguments for the ethical value of a thing.
There are a variety of instrumental arguments that may be used to argue for the value of
a tree. The Australian philosopher, John Passmore, while he does not argue specifically for
trees, lists economic, scientific, recreational, renewal, aesthetic, and posterity varieties. To
take the last argument first, the 'Posterity Argument' holds that the present generation has
a duty to hand over the world "to our successors in a better condition" (Passmore 1974,
p.91). Part of handing over the world in a better condition (from the ecologic, as opposed
to the economic, perspective) is not to destroy species or let them perish through contributory
2Poster from the Singapore Zoological Gardens, The Value of a Tree, based on Update Forestry, Michigan State
University.
Valuing a tree 469
negligence and thereby deprive posterity of the opportunity to observe, study, and otherwise
have the experience of the diversity enjoyed by the present generation. For example, this
is the last generation with a chance to preserve pristine rainforests. If this generation does
not do it, no other can. With a 'Posterity Argument' Passmore is not calling for passive
restraint on the part of humans in their dealings with the environment. For him, handing
over the world in a better condition means "transforming the world into a civilised state"
(Passmore 1974, p.178). Betterment through transformation may mean the preservation of
some rainforest, for example, but with the major responsibility to other humans, it could
mean the destruction of parts of the environment considered harmful or 'un-civilised'.
Civilisation has too often been on a collision course with environmentalism. But in the
words of Rachel Carson, "The question is whether any civilization can wage relentless war
on life without destroying itself, and without losing the right to be called civilized?" (Carson
1962, p.99).
Yet, Passmore attempts to soften the harsh impact of this line of reasoning by a second
argument: the 'Gene Pool Argument', or as another Australian philosopher, Bill Grey
(formerly William Godfrey-Smith) labels it, the 'Silo Argument' (1979). As humans utilise
the plant and animal resources of the world they modify them through cultivation and
domestication, but a modification for one purpose, say breeding cattle to produce more beef
or greater drought resistance, can cause undesirable side effects, say less hybrid-vigour.
Therefore, it is in humanity's own interest to keep wild stocks as a reservoir or gene pool.
Passmore holds "a purely economic argument will suffice to establish at least a prima
facie case against the clearing of wilderness [and] the destruction of species" (1974, p.102).
But obviously the 'Gene Pool' or 'Silo Argument' is reasonable in the case of trees only for
preserving sample wild counterparts of cultivated species that produce fruit, blossoms,
timber, resins, or other economically valuable products. Also it could be objected that a
prima facie case does not hold if it is more economic to destroy a wilderness or species for
immediate gain. "Uncontrolled timber extraction, slash and burn agriculture and large scale
aggressive plantation development are destroying genetic materials" (Tinggal 1992, p.143).
These same practices have placed some rainforest species on the edge of extinction and some
rainforests on the edge of annihilation.
To protect species that are not currently held to be valuable, Passmore appeals to the
'Rare Herb Argument'. "A species often turns out to be unexpectedly useful, a tropical plant
to contain pharmacologically valuable substances" (Passmore 1974, p.102). Grey claims of
this argument, "The rare herb argument is a sound argument in favour of preservation on the
grounds of prudence, but it is not, on the face of it, a moral argument" (Godfrey-Smith
1980, p.31). If the argument is effective, it need not be a moral argument. But prudence
has not prevented the destruction or near destruction of rainforests and the myriad of species
that inhabited them. In Australia alone "up to 75 per cent of rainforest has been lost" by
direct or indirect human intervention since 1788, the establishment of the first white colony
(World Commission on Environment & Development 1990, p.44). Hopefully the 'Rare Herb
Argument' could be applied to argue against the obliteration of additional species; but it says
nothing of the numbers, conditions, or manner under which these reservoirs are kept and
could permit the decimation of these species to plantation populations, botanic gardens, or
token specimens.
Related to the 'Rare Herb Argument' is the 'Lock Argument'. The 'Lock Argument' is
not one suggested by Passmore. If the 'Rare Herb Argument' holds that there are potentially
beneficial substances locked in rainforests that may one day prove valuable, as an example,
for their ability to prevent or cure diseases then the 'Lock Argument' holds that there are
potentially dangerous substances locked in rainforests. Fifty percent of the world's biological
diversity is in rainforests.
470 DAVID H. BENNET
"The rain forest, being by far the earth's largest reservoir of both plant and animal
species, is also its largest reservoir of viruses, since all living things carry viruses.
Just how large the tropical reservoir of viruses is no one knows, but here is one way
to consider the question. The earth is estimated to contain between three million and
thirty million species of plants and animals. Most of the species are fungi, insects, and
non-insect arthropods, such as ticks and mites, and the bulk of them live in tropical
forests. Viruses often adapt to one or two species. For example, human beings carry
more than a hundred different cold viruses that are adapted almost exclusively to the
human host. If we suppose that every species carries one virus exclusively adapted to
it, then there may be from three to thirty million strains of viruses. Possibly the
number of virus strains is much larger than that - perhaps a hundred million - but
nobody has ever tried to count them.
When an ecosystem suffers degradation, many species die out and a few
survivor-species have population explosions. Viruses in a damaged ecosystem can
come under extreme selective pressure. Viruses are adaptable: they react to change
and can mutate fast, and they can jump among species of hosts. As people enter the
forest and clear it, viruses come out, carried in their survivor-hosts - rodents, insects,
soft ticks - and the viruses meet Homo sapiens" (Preston 1992, p.62).
Exactly how many viruses and diseases are lurking in rainforests may not be the most
important question, when the virulence of some that have already emerged is considered.
Not how many there are, but how lethal they are, may be the question. Among those viruses
suspected to have come from rainforest is H.L V. "The human immunodeficiency virus looks
like an example rather than a culminating disaster. As lethal viruses go, H.LV. is by no
means nature's preeminent display of power" (Preston 1992, p.62). Apart from H.LV.,
there is Ebola Zaire, "a slate wiper in humans" (Preston 1992, p.59). It kills 88 per cent
of the people it infects. At least 16 other lethal viruses have also emerged (Preston 1992,
p.62).3
Thus to destroy a rainforest is to break the 'lock'. If the viruses remained with their hosts
and if rainforests were of no other value, then the current rapid rate of their destruction (3
hectares per minute) could be a preventative. But the viruses do not remain with their hosts
and rainforests do have other values besides 'locking away' lethal viruses. The emergence
of the viruses appears to be a natural consequence of the ruin of the tropical biosphere. This
says nothing of the threat these viruses pose to other species.
Prudence would demand that humanity not knowingly endanger itself or other species.
Prudential arguments are pleas for prudent behaviour and are typically grounded in human
interests.' Prudential arguments are arguments encouraging humans (whose Latin binomial,
Homo sapiens, even carries the idea of wisdom in it) to exercise wisdom, although mainly
the wisdom of protecting human interests. These arguments have a powerful appeal because
they pander to human interests, but they do not acknowledge the value of the environment
or parts of it, such as a tree, except so far as they promote human well-being or survival.
Another complementary argument to the 'Rare Herb Argument' is the 'Laboratory
Argument'. The 'Laboratory Argument' states the obvious thought that unless possible
sources of pharmacologically valuable substances are preserved for experimentation then the
diversity of humanity'S environmental laboratory is diminished. As with the 'Silo
3 Some of the emerging viruses are Lassa, Rift Valley, Oropouche, Rocio, Q fever, V .E.E., Guarnarito, Ross River,
Monkeypox, Dengue, Chikungunya, Hantaan, Machupo, Junin, Mokola, Duvenhage, Le Dantec, Kyasanur Forest
virus, Semliki virus, Crimean-Congo, Sindbis, O'nyong-nynong, Marburg and Ebola.
, These interests may be quite oblique, amounting for instance to merely thinking that something (valuable) is there.
Valuing a tree 471
Argument', the thrust of the 'Laboratory Argument' is that biological modification and
experimentation by humans requires natural species as norms, as base lines and as fresh
resources. Grey states: "This is the argument that wilderness areas provide vital subject
matter for scientific inquiry which provides us with an understanding of the intricate
interdependencies of biological systems, their modes of change and development, their energy
cycles, and the sources of their stabilities. If we are to understand our own biological
dependencies, we require natural systems as a norm, to inform us of the biological laws
which we transgress at our peril" (Godfrey-Smith 1979, p.311).
In addition to the environment providing material resources, Passmore also holds "that
there is refreshment as well as enjoyment to be found in wandering through wild country.
(Not only recreation but re-creation; it renews one's sense of proportion)" (Passmore 1974,
p.107). This can be split into two arguments: the 'Recreation Argument', where humans use
wilderness, such as a rainforest, as their gymnasium and the 'Cathedral Argument' where
humans draw spiritual revitalisation from wilderness.
But the instrumental argument upon which Passmore builds much of his overall case for
a stronger adherence to existing ethics and ethical principles rather than the development of
a new ethics is the 'Aesthetic Argument'. Passmore holds that, "to justify action against the
beauty-destroying polluter; at most what is needed is a strengthening of existing moral
principles" (Passmore 1974, p.56). The 'Aesthetic Argument' is not limited to protecting
the environment against polluters. As another Australian philosopher, H.J. McCloskey
observes, "we hold the beautiful and mysterious in nature in awe and wonder, see much as
precious, and view with moral shock and even moral outrage the wilful destruction of
beautiful species" (McCloskey 1983, p.34). The 'Aesthetic Argument' conveys the concept
of 'beauty-destroying', which Passmore finds is a "more explicitly Western tradition that it
is wrong unnecessarily to destroy - a principle embodied in the concept of 'vandalism'"
(Passmore 1974, p.124). This argument places the moral onus on the destroyer. It is
intended to protect the environment from '''wanton' destruction where no defence of the
destruction can be offered" (Passmore 1974, p.124).
Passmore states that these arguments should be enough to preserve the environment and
give value to a tree. But they have certainly failed on the first count. As is apparent from
just one statistic, Edmund O. Wilson graphically captures the American attitude towards
tropical rainforests, "More money is spent in the bars of New York City in two weeks than
is spent annually on tropical research" (Gordon & Suzuki 1990, p.52). If human interests
change and humans, at least in the dominant Western and Western styled societies, placed
even less value on the environment than they already do, then these arguments fail as well.
These arguments still do not place value on the tree itself, for itself. For that a different
value needs to be recognised, an intrinsic value.
DEEPER ETHICAL VALUES
The third set of arguments are deeper ethical arguments. These are arguments that recognise
that instrumental values are not enough to protect the environment or rainforest or a tree.
These deeper ethical arguments recognise the value of natural items for their own sake.
While instrumental arguments can point people in the correct direction as far as saving
the environment, they are only a first step. Further steps are needed. One further step is
Aldo Leopold's land ethic. He recognised that items in the natural environment have
value-in-themselves as well as or despite any value they may have for humans. He summed
this up in the maxim, "A thing is right when it tends to preserve the integrity, stability, and
beauty of the biotic community. It is wrong when it tends otherwise" (Leopold 1966,
p.262). He calls for an extension of ethical consideration to the land, or more broadly and
472 DAVID H. BENNET
accurately, extending the ethical community to include the ecosphere. Leopold sought to
make the ethical and biotic communities co-extensive. His extension principle is summed
up as, "The land ethic simply enlarges the boundaries of the community to include soils,
waters, plants, and animals, or collectively, the land" (Leopold 1966, p.219).
The value of a tree on Leopold's land ethic is the value of one being among many in the
ethical community. Thus moral consideration is extended to trees. This does not necessarily
mean that a tree is extended equal moral consideration to say a human in every case, nor
does it necessarily mean that a tree looses out to a human in every case.
But further steps are possible. It is often maintained that what makes the difference
between something that is valuable merely because it is of use and something that is valuable
in-itself is that the latter has a telos or goal of its own. This is not simply that a tree passes
from a seed to a sapling to maturity to death. This is closer to the idea that Aristotle put
forward about the virtue of an item is its telos or that for which it is intended. This is the
difference between instrumental value and intrinsic value. A thing may have both
instrumental value and intrinsic value. This is generally held to be the case with human
beings. This means recognising that a tree's functions of 'carbon sink' and 'oxygen pump'
are more than simply instrumental values but a 'virtue' of the tree.
A still further step is to take a deeper environmental perspective such as Deep Ecology.
"Deep Ecology is a normative and policy- and lifestyle-oriented theory" (Sylvan 1985, p.43).
Like the science of ecology, which deals with the relationships between organisms and their
environments, Deep Ecology is concerned with the place of organisms in their environment,
but it is bound up with value judgments about the organisms, their environments, and the
relationships between them, and with the wisdom of living by the principles revealed by
ecology, and not, as with the science of ecology, with doing experiments that reveal the
principles. In a sense, Deep Ecology uses ecology to overcome the embarrassment of
science at not being able to convert knowledge into wisdom and state what is most needed -
an appreciation of how to value the environment and ecological relationships.
The argument from Deep Ecology is that while humans are entitled to protect their own
vital needs, they are not entitled to pollute the habitats of other species for non-vital reasons.
In seeking to protect their short-term vital needs humans often damage systems that will meet
their long-term needs. Long-term viability is part of determining a vital need. The human
pollution of the environment is imprudent because it affects humans, but it is also wrong of
humans to ignore that they share the environment with other species that have a claim to its
use and an interest in its quality. Pollution modifies habitats and directly kills other species.
Once again humans need to remember that they need other species more than other species
need them. Humans are more likely to miss rainforests than rainforests are likely to miss
humans.
Resource depletion is another form of human interference. In an interview, Arne Naess,
founder of Deep Ecology, observed, "The shallow ecology movement talks only about
resources of mankind, whereas in deep ecology we talk about resources for each species"
(Bodian 1982, p.lO). While the most common concern is with the depletion of
non-renewable resources -i.e., there are finite quantities of fossil fuels such as coal and
petroleum available and once they are depleted they cannot be replaced (not, at least, in
human time spans) - there is also a significant and growing concern about what have been
considered renewable resources. The depletion of the systems by which renewable resources
are produced would be an eco-catastrophe for all species. Supposedly renewable resources
are being depleted faster than they can be replaced. A case in point is the circular
connection exists on Christmas Island with Abbott's Booby, nesting trees, and guano mining.
"The conservation of Abbott's Booby provides a classic example of a conflict between

the preservation of a scientifically interesting species - which occurs nowhere else and
Valuing a tree 473
whose value is difficult to access - and the utilisation of valuable resources. There are
surface phosphate deposits derived from bird dropping over countless years. The main
breeding areas of the Abbott's Booby coincide with areas rich in phosphate" (Ovington
1978, p.106).
The mining of the phosphate requires removing the trees in which the Boobies nest and
thus destroys the breeding habitat of the Boobies, which are the source of the phosphate.
The collection of the resources destroys the source. The source is depleted along with the
resource. Here the trees are a vital resource for the Boobies and the guano is a vital
resource for the trees. The value of the trees 'is difficult to assess', but the value of the loss
of the trees means the loss of the Boobies and the guano industry.
Another more tragic example is the depletion of rainforests. The 'lungs of the world' are
burnt out of existence. Vast resources are wasted by the burning of the rainforests to create
new grazing lands.
But if the value of things in themselves is recognised then while a tree may still be used,
it must be used to meet vital needs rather than mere convenience. This leads to the fourth
form of argument, cultural values and the way, in this case, that Australian Aborigines both
use and respect trees.
CULTURAL VALUE
A fourth set of arguments are cultural arguments. A tree can form part of a culture's
identity. As Bill Neidjie, an Aborigine from Kakadu in the far north of Australia, says:
"When you cut tree,

it pump life away,
all the same as blood in my arm
If you feel sore ....
headache, sore body,
that mean somebody killing tree or grass
You feel because your body in that tree or earth.
Nobody can tell you,
you got to feel it yourself." (Neidjie 1985, p.82)
A valuable lesson that can be learned from Aborigines is respect for the environment.
Aboriginal traditions can help Western societies realise a role in the environment compatible
with ecological principles, or as Leopold said, compatible with the continuing "integrity,
stability and beauty of the biotic community". An Aboriginal viewpoint is very much like
Leopold's, in that, it is concerned with the maintenance of the whole. No part can exist
without the whole. Human beings are like any other species in this regard - they are part
of the whole. As part of the whole, they are interrelated with the other parts. An appeal
can be made for the maintenance of the other parts in order to maintain humans. What
Aborigines have accepted is the concept that everything in the environment is connected to
everything else. The connections may not be direct and they may be very weak, but they
are nonetheless connected. It is almost too obvious to remark that if Western societies could
continue to manipulate the world without any serious environmental repercussions to
themselves, they would. Yet if it is accepted that the part cannot exist without the whole,
then preserving the whole is in the interest of each part. Insofar as humans are a part of the
whole it is in their best interests to preserve, protect or nurture the whole. If humans
recognise that they are part of the environmental/ecological whole and not separable from
474 DAVID H. BENNET
the whole nor outside the whole, then the other parts can be seen as valuable for maintaining
the whole and thus maintaining human interests. If human preservation is subordinated to
environmental preservation on the grounds that human preservation is not possible without
environmental preservation, then human preservation becomes instrumental to environmental
preservation. This is what Aborigines practiced. This is stepping into the web of life as a
participant rather than standing outside of it as a voyeur.
In summary, Aboriginal individuals feel a unity with and responsibility for specific objects
or phenomena or species, but this bond with these specific elements does not preclude use
of those objects or phenomena or species. The relationship is one of spiritual unity and
mundane utility. Thus a tree is more than a resource, for those with a special relationship
to a given species of tree, it is part of their very being.
CONCLUSION
We profit more from trees than just their dollar value. They perform vital functions from
which we profit. They are part of the same ecosystems of which we are a part and they and
we are interconnected functionally and ecologically. To realise the value of a tree we must
look beyond its dollar value, we must take lessons from older cultures who realised the
interconnectedness of the environment.
While traditional Aboriginal practices and attitudes cannot contribute so much to the
specific content of changes, they can provide inspiration for general changes in context for
Western societies. For example, Aborigines have a generalised attitude towards the
environment that can be specifically applied. This attitude can be likened to the maxim of
'Think globally, act locally'. Aborigines practiced a form ofbioregionalism. Ways in which
the context of Western practices and attitudes can be changed include but are not limited to
acknowledging that the distinctions between humans and other species (us/them) are less
sharp than it is usually made out to be and that therefore humans are subject to the same
ecological pressures and limitations as other species in their dealings with the environment;
that a life simple in means can be rich in ends and that such a life can lower human impact
on the environment; that Western societies should be encouraged to 'dwell in situations of
inherent value'; that the relationships between humans and other species is not solely a matter
of the other species functional utility to humans; that Western societies can better live in
accord with their environment by accepting ecological principles and living according to the
character of their particular bioregion. "We might observe what we have been doing badly
to perpetuate our own welfare at the expense of diminishing the welfare of other species.
Then we might consider Aboriginal approaches to grasp what, more elementally, we might
better have pursued - not in imitation of Aborigines, but symbolically of their attitudes
towards the maintenance of their environment and their participation in it and tempo of life.
We might consider Aboriginal approaches that would permit us to be happier creatures and
give us and other species a greater hope of ecological survival" (Bennett 1991).
LITERATURE CITED
BENNETT, D.H. 1991. Animal Rights and Aboriginal Concepts. Pp.53-69 in Croft, D.B. (ed.). Australian
People and Animals in Today's Dreamtime: The role of comparative psychology in the management of natural
resources. Praeger Press, New York.
BERNDT, R.M. 1974. Australian Aboriginal Religion. E.l. Brill, Leiden.
BERNDT, R.M. 1979. A Profile of Good and Bad In Australian Aboriginal Religion. Australian and New Zealand
Theological Review 12:17-32.
BODIAN, S. 1982. Simple in Means, Rich in Ends: a conversation with Arne Naess. Ten Directions 7:10-12.
Valuing a tree 475
CARSON, R. 1962. Silent Spring. Houghton Mifflin, Boston.

GODFREY-SMITH, W. 1979. The Value of Wilderness. Environmental Ethics 1:309-19.
GORDON, A. & SUZUKI, D. 1990. It's a Matter of Survival. Allen and Unwin, Sydney.
LEOPOLD, A. 1966. A Sand County Almanac. Oxford University Press, New York (First published 1949).
McCLOSKEY, H.J. 1983. Ecological Ethics and Politics. Rowman and Littlefield, Totowa, N.J.
NEIDJIE, B., DAVIS, S. & FOX, A. 1985. Kakadu Man ... BilI Neidjie. Mybrood, Sydney.
OVINGTON, D. 1978. Australian Endangered Species: mammals, birds and reptiles. Cassell Australia,
Stanmore, NSW.
PASSMORE, J. 1974. Man's Responsibility for Nature. Duckworth, London.
PRESTON, R. 1992. Crisis in the hot zone: a reporter at large. New Yorker, 26 October 1992.
ROSE, D. B. 1988. Exploring an Aboriginal Land Ethic. Meanjin 3:378-387.
STANNER, W.E.H. 1960. On Aboriginal Religion. Oceania 30:245-278.
STANNER, W.E.H. 1979. White Man Got No Dreaming: essays 1938-1973. Australian National University
Press, Canbera.
STREHLOW, T.G.H. 1970. Geography and the Totemic Landscape in Central Australia: a functional study. Pp.
92-140 in Berndt, R.M. (ed.). Australian Aboriginal Anthropology: rrwdem studies in the social anthropology
of the Australian Aborigines. University of Western Australia Press for the Australian Institute of Aboriginal
Studies, Nedlands, W.A.
SYLVAN, R. 1985. A Critique of Deep Ecology. Australian National University, Department of Philosophy,
Canberra (Discussion Papers in Environmental Philosophy, 12).
TINGGAL, SERUDIN. 1992. Brunei Darussalam Fruits in Colour. Universiti Brunei Darussalam, Brunei
Darussalam.
WORLD COMMISSION ON ENVIRONMENT AND DEVELOPMENT 1990. Our Common Future. Australian
edition. Oxford University Press, Melbourne.
Extraction and Tropical Rain Forest Conservation in Eastern Brazil
ROBERT A. VOEKS t
Department of Geography, California State University, Fullerton, CA 92634, USA
ABSTRACT. Management and extraction of non-timber products represents an alternative to destructive

exploitation of Latin America's tropical forests. Well adapted to subsistence level economies, extraction
has nevertheless yet to prove itself ecologically sustainable when exposed to outside market forces and
economically competitive when confronted with more destructive land uses. By means ofa case study,
this paper examines two objectives of the extraction option: (1) Can extraction remain economically
competitive over the long term? and (2) Can management for extractive products contribute to the
conservation of tropical forest biodiversity? Piassava palm (Attalea funifera Mart.) is endemic to the
tropical rain forests of Bahia, Brazil. Its fibre has represented a commercially important product since
the 16th century, first in the fashioning of anchor ropes and later in the manufacture of brooms and
brushes. Destructive harvest methods during Brazil's Imperial period resulted in near extirpation of the
species. Privatization of the royal forests led to more sustainable harvest and management techniques.
Fibre extraction now represents an economically sustainable land use. Piassava-rich lands are managed
as long-fallow, second growth forests. An inventory 0.3 ha of managed piassava land revealed the
presence of 52 tree and 19 shrub species, most of which also inhabit primary forests. By serving as a
temporary repository of primary forest species, long-fallow forest managed for extractive products
subsidizes the conservation of tropical biodiversity. These results suggest that (1) extraction of native
forest products can compete economically with more destructive land uses, particularly if the land is of
marginal value; (2) provided that sustainability incentives exist, private ownership is not necessarily
inconsistent with the successful management of extractive resources; (3) relatively high species richness
can be maintained as part of commercial extractive systems; and (4) secondary forests may represent the
best candidates for the development of extractive enterprises.
KEY WORDS: non-timber products, Piassava palms, Attaleafunifera, fibres, conservation, sustainable
management.
INTRODUCTION
Sustainable management and extraction of non-timber products offers a promising alternative

to destructive exploitation of Latin America's tropical forests. Under ideal conditions,
extractive enterprises provide meaningful employment for native and other forest-dwelling
people, protect a large measure of the biodiversity and contribute to the local and regional
economy. Indigenous land use practices, employing a combination of short and long fallow
swiddens, sparing and transplanting of valuable species, and simple extraction, represent
time-tested procedures for sustainable forest exploitation (Bennett 1992, Denevan et ai. 1985,
Dufour 1990). Aside from protecting biological values (Chapin 1992), subsistence
management and extraction methods provide the framework within which sustainable
agrosystems can be developed for non-native peoples (Alcorn 1984, Altieri & Merrick 1987,
Clay 1988). The establishment of reserves around the collection of fruit, nuts, fibre, and
latex seems a logical extension of the sustainable extraction philosophy. If not a panacea for
the problem of rain forest degradation (Browder 1990), extractive reserves at least represent
a notable improvement over current practices (Allegretti 1990, Fearnside 1989).
'Present address: Department of Geography, Universiti Brunei Darussalam, Bandar Seri Begawan 2028, Brunei
Darussalam.
477
478 ROBERT A. VOEKS
As a model for regional development and conservation, petty resource extraction exhibits
several serious shortcomings. First, extractive activities retain a dubious reputation in Latin
American history (Anderson 1990). Brazil's forested regions witnessed a succession of
boom-and-bust extractive cycles during the colonial period, resulting in the near extinction
of valuable plant and animal species. Policymakers, as a result, continue to view extractive
enterprises as both agronomically regressive and environmentally destructive (Alvim 1989).
Second, as a serious development alternative, extraction must compete favorably with other
land uses. Small-scale studies on the collection of latex, fruits, and nuts (Peters et at. 1989)
and forest medicinals (Balick & Mendelsohn 1992) suggest that petty extractive values exceed
those of forestry and cattle ranching activities when considered over the long term.
Secondary forests appear particularly desirable, as they are noted to be hot spots for useful
forest products (Grenand 1992, Kohn 1993). Coupled with the potential benefits of
protecting future pharmaceuticals, crop genetic resources, as well as less tangible
environmental values (Myers 1984, Randall 1991), extraction would seem, at least in
principle, a reasonable alternative to present practices. In reality, extractive systems seldom
persist when confronted with the short-term economic benefits of more destructive forms of
land use. Third, considerable biodiversity is maintained within extractive systems that are
subsistence based, such as those practiced by Amazonian Indians. It is unclear, however,
whether the social mechanisms that encourage environmental conservation in subsistence
economies will persevere as extractive products become increasingly responsive to outside
market forces.
This paper considers the viability of non-timber forest extraction by means of a case study
from the Atlantic rain forests of eastern Brazil. With a history reaching back over four
centuries, the collection of piassava palm fibre (Attaleafunifera Mart.) represents one of the
longest running extractive activities in the New World tropics. The exploitation of this palm
is examined in terms of meeting two goals of tropical forest conservation and development:
(1) Is it economically sustainable over the long term?, and (2) Does it conserve a relatively
high level of biological richness?
PlASS AVA BIOLOGY AND GEOGRAPHY
Piassava palm is endemic to the Atlantic coast of Bahia, Brazil (Bondar 1942, Silva and
Vinha 1982), where it is distributed from roughly 130-17ooS latitude (Figure 1). Reaching
15 to 20 m in height, the species is easily recognized in the field due to the presence of erect
10 m fronds covered with droopy leaflets. The valuable sheath fibre is 3-5 m long and is
available for harvest soon after the fronds have sloughed off the terminal meristem.
Flowering and fruiting occurs throughout the year. Outcrossing is effected by sap beetles
(Mystrops sp.) and weevils (Phyllotrox tatianae) (Voeks 1985). Individual palms change
gender during their life cycle, producing first staminate and later pistillate flowers (Voeks
1988b). Fruits average 300-400 gm, and contain edible kernels.
The regional climate is a Koppen Af tropical rain forest type. Mean annual rainfall ranges
from 1800-2100 mm, with no pronounced dry season. Mean monthly temperatures reach a
low of 21°C in August and a high of 25°C in February (Milde 1983). Soils grade from
oxisols in the interior zones to spodosols nearer the coast. Responsive to this edaphic
transition, the vegetation changes from tall, broadleaf evergreen formations, mata higr6fila,
to shorter, heath-like forests, restinga. Piassava is facultatively restricted to sandy spodosols,
although it readily colonizes oxisolic areas that have been deforested (Voeks 1990).
The Atlantic coastal forests of Brazil may represent the most speciose tropical forests in
the world (Anon. 1993). The vegetation is characterized by largely buttress-free, broadleaf
Rain forest conservation in Brazil 479
6.._
1;-:-:.:-:-:-:.:-:-:-:
6.._
17·~-'OCJ
Figure 1. Distribution ofpiassava palm (Attaleafunifera). North of the city ofValemra, A.funifera is replaced by
its shrub form, A. acaulis, from which sheath fibre is also extracted.
evergreen trees with relatively heavy loads of lianas and epiphytes (Mori et al. 1983). The
high percentage of endemic tree species in the Bahian forests suggests lengthy geographical
isolation from Amazonian formations (Mori et al. 1981). With over 95% of the biome
converted to agricultural and urban landscapes, these forests represent one of the most
threatened tropical biotas on Earth (Fonesca 1985, Mori 1989, Myers 1988).
FIBRE EXTRACTION: 1587-1987
The presence of piassava palm and its probable indigenous use was recorded within a few
days of the Portuguese discovery of Brazil. Ship scribe Pero Vaz de Caminha reported that
"the entire beach is covered with palms" and, later, that the huts of the natives were covered
with fibre (Braga 1968, p. 33). Since coconut palms were yet to be introduced to eastern
South America, Caminha was most likely describing piassava, which is the most numerous
native palm at the site (Andrade-Lima 1984).
The first commercial use of piassava was recorded in 1587 by Gabriel Soares de Sousa,
a sugar plantation owner. Able to withstand long periods of immersion in sea water,
piassava was employed on local ships as a cable material (Sousa 1938). A century later, in
his 1699 passage through Bahia, Captain William Dampier similarly reported that piassava
480 ROBERT A. VOEKS
cables were "strong and long lasting" and were "let to hire to European ships" (Masefield
1906, p. 388). During the 18th and 19th centuries, piassava became the cable material of
choice for Atlantic shipping (Lapa 1968, p. 92). The Count of Atouguia noted their value
to the local shipping industry in 1751 (Atougia 1925, p. 186), and by 1757 piassava cables
were being exported to foreign countries (Caldas 1951, pp. 221-2). Prince Maximilian von
Wied-Neuwied, in his 1816 passage through Bahia, described the fashioning of cables by
African slaves (Wied-Neuwied 1940, p. 224). In 1820 Spix and Martius (1928, p. 122) noted
that piassava rope making represented the principal occupation of the pacified coastal Indians.
But while some touted cable making as a "lucrative branch of commerce" for the region
(Henderson 1821, p. 314), the economic contribution of this cottage industry to colonial Bahia
was in fact negligible (Bahia 1771-1777, Bahia 1789-1796, Camara 1789, Vilhena 1969).
The utility of piassava fibre as a broom and brush material was discovered in the
mid-1800's. Employed as dunnage to pack Brazilian sugar, piassava fibre often accumulated
as rubbish on foreign docks. An enterprising dock worker brought some to a British brush
maker who, it is reported, fashioned the first piassava fibre broom (Kiddier 1927). By the
1840s, piassava brooms were well known in Europe and the Americas. In Cuba, they were
referred to as "cleaning angels" (Amaral 1925, pp. 89-90). Sir William Hooker (1849, p.
122) commented that "few have walked the streets of London without remarking that of late
years these streets are . . . kept peculiarly neat and clean, by the stiff fibres of a new
material for making brushes and brooms . . .". As the value of piassava fibre to the broom
and brush industry grew, its utility as a cable material apparently fell out of favour (Dodge
1897).
Responsive to the burgeoning international market, piassava fibre extraction became the
principal industry in 19th century southern Bahia, surpassing sugar, cotton, and coffee in
value (Silva Campos 1981). The British vice-council to Bahia noted that piassava was " ...
much procured and exported to Europe" (Wetherell n.d., p. 115). By the 1880s, nearly
nineteen thousand tons were reaching the foreign markets, most arriving in Engilmd, France,
Germany, and the United States (Relatorios 1867-1940). "In the future", predicted one
contemporary observer, piassava "promises to be ... one of the great sources of receipts
for the province" (Carige 1882, p. 39).
Bahia's 19th century fibre boom was short lived. Sustained harvest and export of ten to
twenty thousand tons per year during the 1880s was followed by a precipitous decline in the
1890s (Figure 1). For the next six decades, annual collection figures stabilized at one to
three thousand tons, and exports fell to a fraction of their former importance. Faced with
dwindling supplies and increasing costs, broom manufacturers in Europe and North America
switched to less costly African fibres (Kiddier 1927, Moraes 1911).
The piassava fibre crash was associated with over-exploitation of the resource. Until
1888, much of Bahia's forested landscape belonged to the Imperial government.
Administration of the royal woods and their resources was delegated to the various relevant
municipalities. In order to attain access rights to piassava and other non-timber products,
local entrepreneurs were required to apply for a license and pay a user fee to the municipal
government. In their letters of application, fibre merchants carefully outlined the precautions
that would be taken to protect the forests and, in particular, how the fibre would be collected
in a non-destructive manner. Such assurances were largely unnecessary, however, as
municipal authorities lacked the resources to properly oversee their vast holdings. Moreover,
although forest regulators were well aware of ongoing forest abuses, user fees represented
an important source of revenue for the municipalities (Letters 1864-1889).
Incentives to harvest piassava on a renewable basis existed at neither the harvest nor the
regulatory level. Fibre merchants, as a result, collected the fibre by the most destructive
means possible. A municipal judge complained that exporters were "causing total destruction
because ... they cut the whole tree in order to collect the fibre" (Letter 1881). Another
observer noted that "the present mode of obtaining [fibre] is to cut the tree down ... a very
foolish proceeding, considering the time that the fallen nuts take to germinate and grow into
bearing trees" (Booth 1889, p. 239). This "destructive method of harvest", according to a
provincial employee, was carried out "because those with licenses have nothing to lose, since
the forests belong to the state" (Letter 1885).
Given these destructive extraction methods, 5 % of the entire piassava population was
destroyed on an annual basis (Webering 1937). By the 1890s, piassava was commercially
extinct as an export commodity. As a poorly regulated extractive product subject to intense
international demand and inappropriate modes and levels of exploitation, piassava supplies
declined rapidly. Its demise may have been exacerbated by labour shortages associated with
the abolition of slavery and the inability of the region to attract immigrant workers (Leao
1982, p. 315). There is, however, little evidence linking piassava harvest to dependence on
slave labour. Extraction was and continues to be carried out by caboclos, mixed descendants
of African, Amerindian, and Portuguese ancestry. In any case, by the end of the century,
another boom-and-bust extractive cycle had run its course.
After decades of quiescence, piassava production began to rebound in the mid-twentieth
century. Total collection doubled during the 1950s and by 1967 had exceeded the highest
figures recorded in the 1880s. The 1987 harvest exceeded 66 thousand tons, 3.5 times
higher than the fibre boom years (Figure 1). Wholesale values climbed from US$300,OOO
in the 1950s to over US$16,OOO,000 in the 1980s. Unlike the earlier peak production years,
however, the recent upsurge in production has been accompanied by decreasing dependence
on export markets. With over 99% of the raw material destined for broom and brush
manufacturers within Brazil, piassava exports have slumped to 400 tons per year, barely 2 %
of their 1880s total. Declining interest in piassava, according to European importers, is due
to the relatively high cost of Bahian fibre and the popularity of monofilament fibres (Voeks
1987, pp. 208-210).
Improvements in fibre management and harvest appeared as fibre-rich lands passed from
public to private control. Formation of the Brazilian Republic in the latter part of the
century led to the opening of the royal woods for settlement. Fibre-bearing forests that were
not already in private hands were settled by locals and incoming colonists. Some were
incorporated into cacao-based latifundia. Most, however, appear to have been occupied by
peasant farmers.
Occupation of piassava-rich forests represented anything but a land rush. Rather, with
nutrient-poor, sandy soils, heath forests were viewed as being of marginal agronomic value.
Piassava was associated with neither quick wealth nor social status. It was and continues to
be perceived as a 'poor man's' pursuit. With a strong incentive to protect the value of their
properties, peasant smallholders initiated the non-destructive collection methods that are in
place today.
Under current practices, piassava fibre is collected on a yearly cycle. Rather than killing
the palms, as was done previously, fibre collectors climb the trees and cut away from the
nearly senescent outer fronds with a machete. After being stripped from the fronds, the fibre
is hauled to a common cleaning area and manually separated. The borra fibre, a papery
material, is sun dried, woven into thatch, and sold as roofing material for fashionable homes
and beach cabanas. The more valuable strand fibre is transported to fibre merchants in the
nearest city. After being cleaned, sorted into grades, and cut, the fibre bales are forwarded
to broom and brush manufacturers throughout Brazil (Voeks 1988a).
Piassava is frequently managed as part of a long-fallow system. Every thirty to sixty
years the vegetation is cut and burned. Only piassava individuals are spared. Manioc is
planted during the first and second year after the burn so that some economic return is
realized prior to the palms achieving harvestable status. Following the fire, previously
established piassava seedlings vigorously resprout through the charred remains. In five to
482 ROBERT A. VOEKS
eight years, the palm fronds have grown large enough to produce harvestable fibre. Within
20 years, the forest is numerically dominated by adult piassava (Voeks 1988c). Given the
present mode of harvest and management, piassava palm fibre has made the transition from
destructively exploited export product to sustainably managed regional resource.
BIOLOGICAL RICHNESS
Piassava's transition from destructive to sustainable management during this century means
that two of the objectives of extractive development are being met, that is, creating
employment for the local population and contributing to the regional economy. Nevertheless,
piassava palm is often managed by fire. Every half century or so, the heath forest
biodiversity is converted to ash and atmospheric gas. Given this level of disturbance, is the
management of this extractive resource consistent with the goals of protecting biological
diversity?
This question was examined by enumerating the tree and shrub species in a 0.3 ha (25 m
x 120 m) plot. Slashed and burned in the 1930s, the plot is managed as long fallow piassava
forest. All trees and palms with a dbh of 5.0 cm or greater were marked with aluminum
tags. Voucher specimens were collected over an 18 month period (1984-86). Shrubs were
also collected, although the large number of individuals was not recorded. Over 90% of the
collections were made during flowering or fruiting. Thus, although many specimens have
yet to be identified to species, I am confident that most collections identified as distinct
species will withstand further scrutiny. Vouchers are permanently housed at the herbarium,
Centro de Pesquisas do Cacau, ltabuna, Bahia.
Enumeration results are listed in Table 1. A total of 52 tree and palm species (or
morphospecies) distributed among 22 families were recorded. Nineteen shrub species were
distributed among 11 families. Among the trees, the most speciose families included the
Myrtaceae (6), Rubiaceae (4) and Myrsinaceae (4). The most species-rich shrub families
included Melastomataceae (5), Rubiaceae (3), and Euphorbiaceae (2).
Numerically dominated by piassava (85 of 185 individuals), the plot nevertheless retains
several characteristics in common with old growth tropical forest. For example, fully 60 %
of the tree species are represented by a single individual. Excluding piassava, the 51 tree
species were represented on average by less than two individuals. Further similarities
between this second growth forest and old growth are evidenced by calculating Simpson Ds
and Shannon-Wiener H' diversities. Including piassava in the calculations, Ds = 0.78 and
H' = 3.90, figures that are diverse by temperate standards but not by primary tropical forest
standards. Excluding piassava from the calculation, however, yields figures of Ds = 0.98
and H' = 5.35. Comparing these latter figures with those reported from old growth rain
forests (cf. Bongers et al. 1988, Knight 1975, Uhl & Murphy 1981) indicates that this plot
is diverse even by primary forest standards.
The high species diversity exhibited by this second growth plot is not an isolated
occurrence. Comparable results have been reported for a plot near this study area (Passos
1984), as well as others in northern Amazonia (Saldarriaga et al. 1986) and Malaysia
(Kochummen & Ng 1977). In their summary of secondary tropical forests, Brown & Lugo
(1990) report that species richness often reaches that of old growth forest within 80 years of
disturbance. In some cases, secondary forest accumulates more woody species than the
forest it replaced.
Table 1. Species Enumeration of 0.3 ha Second Growth Plot. T=tree, P=palm, S=shrub, NC=no collection,
VC=various collections.
Species Abund- Life Coli. Species Abund- Life Coli.

ance form No. ance form No.
Anacardiaceae Melastomataceae
Tapirira guianensis Aubl. T 108 Clidemia capitellata S 27
Annonaceae var. dependens
Annona salzmannii Alph. DC. T 65 Henriettea succosa DC. S 8
Annona sp. T 45 Miconia albicans Steud. S 60
Aquifoliaceae Miconia amoena Triana 2 T 82
/lex sp. 4 T 94 Miconia pileata DC. S 17
flex sp. 5 T 54 Ossaea retropila Triana S 46
/lex sp. T 63 Mimosaceae
Arecaceae Abarema filamentosa (Benth.) S 71
Attalea funifera Mart. 85 TP VC Pittier
Attalea humilis Mart. 1 TP NC Albizia pedicellaris (DC.) T 68
Bactris sp. SP NC ined.
Geonoma blanchetiana TP 29 Moraceae
Wendland ex Drude Ficus sp. T 9
Boraginaceae Myrsinaceae
Cordia nodoso Lam. 1 S 61 Conomorpha oblongifolia A. T 21
Cordia superba Cham. 1 T 66 DC.
Cordia sp. 3 T 34 Cybianthus gardneri (A. DC.) T 18
Chrysobalanaceae G. Agostini
Couepia impressa ssp. S 53 Myrsine umbellata Mart. 2 T 41
cabraliae Prance Rapanea sp. 1 T 5
Hirtella bahiensis Prance T 72 Myrtaceae
Licania lamentanda T 28 Myrcia sp. 4 T 96
Clusiaceae Myrcia sp. 1 T 51
Clusia sp. 2 T 10 Myrcia sp. 2 T 109
Clusia sp. 2 T 43 ? 1 T 69
Symphonia globulifera L. T 44 4 T 73
Erytbroxylaceae 6 T 59
Erythroxylum mattos-silvii T. S 58 Nyctaginaceae
Plowman Pisonia sp. 2 T 85
Euphorbiaceae ? T 47
Chaetocarpus sp. 2 T 30 ? S 121
Croton selowii Baill. S 2 Piperaceae
Pera sp. 5 T 89 Ottoma sp. S 40
Sebastiania sp. S 31 Polygonaceae
T 35 Coccoloba sp. 1 T 66
Fabaceae Coccoloba sp. 1 T 84
Andira sp. T 102 Coccoloba sp. 3 T 97
Platymiscium blanchetii T 19 Rubiaceae
Benth. Psychotria erecta (Aubl.) S 111
Swartzia apetala Raddi. T 12 StandI. et Steym.
F1acourtiaceae Psychotria jambosioides S 52
Casearia commersoniana T 78 Schlecht
Cambess. Psychotria mapourioides DC. 1 T 7
Casearia sp. T 83 Psychotria sp. 6 T 77
Icacinaceae Salzmannia nitida DC. S 3
Emmotum affine Miers. T 80 ? T 57
Lauraceae T NC
Persea sp. 2 T 55 Sapindaceae
? 2 T 62 Cupania oblongifolia Mart. 2 T 11
Malpighiaceae Sapotaceae
Byrsonima sericea DC. 9 T 79 ? T 88
Marantaceae Solanaceae
Stromanthe pOrteana A.Gris. S 20 Solanum polytrichum Moric. S
484 ROBERT A. VOEKS
Table 2. Species Diversity: Old and Second Growth Forest. Simpson Index: Ds = l-(Eni(ni-l)/N(N-l).
Shannon-Wiener Index: H' = (N log N - Eni log ni)/N.
Simpson Ds Shannon-Weiner H'

0.93 4.80 Knight 1975 (panama)
0.97 5.86
0.95 5.38
0.97 5.71
0.95 5.27
0.94 3.63 Bongers et al. 1988 (Mexico)
0.92 4.80 Uhl & Murphy 1981 (Venezuela)

0.96 5.40
0.96 5.30
0.93 5.00
0.78 3.90 This study - study site + piassava

0.98 5.35 This study - study site - piassava
The biological value of long-fallow forest management is further evidenced by comparing

the species list (Table 1) with a list of the trees and shrubs known to inhabit primary forests
in the region (Mori et al. 1983). The majority of trees and shrubs enumerated in the second
growth plot are seen to represent components of old growth tropical forest. Examples
include the trees Tapirira guianensis, Annona salzmannii, Cordia nodosa, and Swanzia
apetala. Several locally endemic old growth species were also recorded in the plot, including
Couepia impressa ssp. cabraliae, Hirtella bahiensis, and Erythroxylum mattos-silvii. It is
also noteworthy that nearly all the species in the plot were found to be reproductively active.
Thus, rather than representing a destructive form of land use, long-fallow forest management
appears to create fluctuating second-growth repositories for the conservation of primary forest
biodiversity.
Piassava fibre extraction appears to be compatible with the maintenance of biodiversity
values. By discouraging less profitable and more destructive land uses, fibre extraction
subsidizes the existence of the remaining members of the forest community. Heath forest
owners, with a vested interest in maintaining the biological integrity of their property,
represent ideal land stewards. As an adjunct to biological preservation in parks and reserves,
extraction should be considered a serious conservation alternative.
SUMMARY
Extraction of non-timber forest products represents one of the mix of development options
for tropical rain forests. In the case of piassava fibre extraction, local employment of forest
dwellers is realized, species-rich second growth forests are conserved, the region produces
an economically valuable product, and the country achieves self-sufficiency in a commodity
that otherwise might need to be imported.
Although it is difficult to generalize from a single example, these results at least suggest
the following. First, extraction of native forest products can, under some circumstances,
compete economically with more destructive land uses. In the case of Bahia's fibre-rich
heath forests, the marginal agronomic value of sandy spodosolic soils has allowed piassava
extraction to dominate the economy for over four centuries. Given the relatively low
production of extractive systems, marginality may prove to be a prerequisite to long term
sustainability. Second, private ownership of native forests is not necessarily inconsistent with
the successful management of extractive resources. In the case of piassava, transfer of
property from the state to the private sector was associated with the transition from
destructive to sustainable practices. This does not mean that collective ownership or
customary rights management are necessarily less desirable than private ownership. Rather,
serious incentives for resource protection at the grass roots level appear to be essential to
long term management of extractive resources. Third, species richness figures from
disturbed forests are high, sometimes as high as old growth forests. Periodic disturbance of
old growth forests, in this case to enhance the population status of a particularly valuable
species, may be consistent with the goals of biodiversity preservation. Finally, secondary
forests are increasing in Latin America at the expense of old growth. Learning to sustainably
exploit these already disturbed tropical ecosystems might relieve some of the pressure to
remove the increasingly rare primary forest.
ACKNOWLEDGEMENTS
Field work was funded in part by a Fulbright Exchange Fellowship. The logistical support
supplied by the Centro de Pesquisas do Cacau, Itabuna, Bahia is gratefully acknowledged.
The project could not have been accomplished without the assistance and advice of A. M.
de Carvalho, L. A. Mattos e Silva, T. S. Santos and S. G. da Vinha. I also thank Glyn
Davies for constructive criticism of the manuscript.
LITERATURE CITED
ALCORN, J. B. 1984. Development policy, forests and peasant farms: reflections on Huastec-managed forests'
contribution to commercial production and resource conservation. Economic Botany 38:389-406.
ALLEGRETTI, M. H. 1990. Extractive reserves: an alternative for reconciling development and enviromnental
conservation in Amazonia. pp. 252-264 in Anderson, A. (ed.). Alternatives to deforestation. Columbia
University Press, New York.
ALTIERI, M. A. & MERRICK, L. C. 1987. In situ conservation of crop genetic resources through maintenance
of traditional farming systems. Economic Botany 41:86-96.
ALVIM, P. de T. 1989. Tecnologias apropriadas para a agricultura nos tr6picos umidos. Agrotropica 1:5-26.
ANON. 1993. New in the botanical book of records: highest tree diversity in the world. Field Notes from the New
York Botanical Garden 2:1-2.
AMARAL, L. 1925. A explora~ de fibras na Bahia (1843). Anais do Arquivo PUblico da Bahia 14:89-90.
ANDERSON, A. B. 1990. Extraction and forest management by rural inhabitants in the Amazon estuary. pp. 65-85
in Anderson, A. (ed.). Alternatives to deforestation. Columbia University Press, New York.
ANDRADE-LIMA, D. de 1984. A botanica da carta de Pero Vaz de Caminha. Rodriguesia (Rio) 36:5-8.
ATOUGIA [to the Royal Court, 1751] 1925. Letter to the Royal Court. Anais do Arquivo Publico da Bahia
14:185-86.
BAHIA [to Portuguese Arsenal] 1771-1777. Notices of shipments of piassava anchor ropes. Inventtirios dos
documentos relativos ao Brasil II: 261, 269, 318, 345, 369. Centre of Bahian Studies, Salvador.
BAHIA [to Portuguese Arsenal] 1789-1796. Notices of shipments of piassava anchor ropes. Inventtirios dos
documentos relativos ao Brasil ID:112, 115, \17, 249, 263, 282, 283, 290, 291, 309, 311, 312, 316, 317, 319, 322,
348,350,357,383,390. Centre of Bahian Studies, Salvador.
BALICK, M. J. & MENDELSOHN, R. 1992. Assessing the economic value of traditional medicines from tropical
rain forests. Conservation Biology 6: 128-130.
BENNETT, B. C. 1992. Plants and people of the Amazonian rainforests. BioScience 42:599-607.
BONDAR, G. 1942. A piassaveira e outras palmeiras Attaleianeas na Bahia, Instituto central de fomento economico
da Bahia, boletim no. 13. 73pp.
BONGERS, F., POPMA, J., MEAVE DEL CASTILLO, J. & CARABIAS, J. 1988. Structure and floristic
composition of the lowland rain forest of Los Tuxtlas, Mexico. Vegetatio 74:55-80.
BOOTH, W.S. 1889. Bahia piassava. Kew Bulletin of Miscellaneous Information 34:327-42.
BRAGA, R. 1968. Pero Vaz de Caminha: carta a el Rey Dom Manuel. Editora Sabia, Rio de Janeiro.
486 ROBERT A. VOEKS
BROWDER, J. O. 1990. Extractive reserves will not save the tropics. BioScience 40:626.
BROWN, S. A. & LUGO, A. E. 1990. Tropical secondary forests. Journal of Tropical Ecology 6: 1-32.
CALDAS, J. A. 1951. Noticia geral de toda esta Capitania da Bahia desde 0 seu descobrimento ate 0 presente ano
1759. Reprint edition, Tipografia Benitiana Ltd.
CAMARA, M. F. 1789. Ensaio da descriyao fizica, e econornica da Comarca dos I1heus na America. MemOrias
Economicas da Academia Real das Ciencias de Lisboa, pp. 304-350.
CARIGE, E. 1882. GeograjUl, physica e politica da provincia da Bahia. Irnpresa Economica, Bahia.
CHAPIN, M. 1992. Coexistence of indigenous peoples and the natural environment in Central America. Research
and Exploration Spring (special map supplement), 16 pp.
CLAY, J. W. 1988. Indigenous peoples and tropical forests. Cultural Survival, Inc. Cambridge, MA.
DENEVAN, W. M., TREACY, J. M., ALCORN, J. B., PADOCH, C., DENSLOW, J. & PAITON, S. F. 1985.
Indigenous agroforestry in the Peruvian Amazon: Bora Indian management of swidden fallows. pp. 137-155 in
Hemming, J. (ed.). Oumge in the Amazon Basin. Vol. I., man's impact on forests and rivers. Manchester
University Press, Manchester.
DODGE, C. R. 1897. A descriptive catalogue of useful fiber plants. U.S. Department of Agriculture Report No.9.
DUFOUR, D. L. 1990. Use of tropical rainforests by native Amazonians. BioScience 40:652-659.
FEARNSIDE, P. M. 1989. Extractive reserves in Brazilian Amazonia. BioScience 39:387-393.
FONESCA, G. A. B. 1985. The vanishing Brazilian Atlantic forest. Biological Conservation 34:17-34.
GRENAND, P. 1992. The use and cultural significance of the secondary forest among the Wayapi Indians. pp. 27-
40 in Plotkin, M. & Famolare, L. (eds). Sustainable harvest and marketing of rain forest products. Island
Press, Washington D.C.
HENDERSON, J. 1821. A history of the Brazil, comprising its geography, commerce, colonization, aboriginal
inhabitants, etc. Longman, Horst, Reese, Orme & Brown, London.
HOOKER, Sir Wm. 1849. Piayaba; fibre and fruit of the coquilla nut, Attaleafunifera, Mart. Hooker's Journal of
Botany and Kew Garden Miscellany 1:121-23.
IIBGE (lnstituto Brasileiro de Geografia e Estatistica) 1946-1982. Anuario estatistico do Brasil. Conselho Nacional
de Estatistica, Rio de Janeiro.
KIDDIER, W. 1927(?). The brushmaker and the secrets of his craft: his romance. Jonathan Cape, London ..
KNIGHT, D. H. 1975. A phytosociological analysis of species-rich tropical forest on Barro Colorado Island,
Panama. Ecological Monographs 45:259-284.
KOCHUMMEN, K. M. & NG, F. S. P. 1977. Cited in Whitmore, T. C. 1984. Tropical Rain Forests of the Far
East. Clarendon, Oxford.
KOHN, E. 0., in press. Some observations on the use of medicinal plants from primary and secondary growth by
the Runa of eastern lowland Ecuador. Journal of Ethnobiology.
LAPA, J. R. 1968. A Bahia e a corrida da India. Nacional, Sao Paulo.
LEAO, S. 1982. The evolution 0/ agricultural land use patterns in the state 0/ Bahia, Brazil. Ph. D. dissertation,
University of Western Ontario.
LETTER [Judge Jose Augusto Beirbosa Coelho to Senator Conselheiro Joao Lustosa? da Cunha Paranagua] 13
October 1881. Colonial Section, Agriculture-Piassava. Public Archive of Bahia, Salvador.
LETTER [Comrnissao de Mediyao de Terras to Jesse Martins Sampaio] 2 September 1885. Colonial Section,
Agriculture-Piassava. Public Archive of Bahia, Salvador.
LETTERS 1864-1889. Colonial Section, Agriculture-Piassava. Public Archive of Bahia, Salvador.
MASEFIELD, J. 1906. Dampier's voyages, by Captain William Dampier. Dutton, New York.
MILDE, L.C.E. 1983. Estudo de precipitQf,:ao diana: regimes pluvio-metricos e modelos de distribuirao para a
regiao cacaueira do sudeste da Bahia. Unpublished Master's Thesis, Federal University of Paraiba.
MORAES, P. de, 1911. A piassabeira: a extinyao dos piassabaes naturais e a sua necessaria cultura. Cluicaras e
Quintais 3:15-16.
MORl, S. A., BOOM, B. M. & PRANCE, G. T. 1981. Distribution patterns and conservation of eastern Brazilian
coastal forest tree species. Brittonia 33:234-45.
MORl, S. A., BOOM, B. M., CARVALHO, A. M. de & SANTOS, T. S. 1983. Southern Bahian moist forests.
The Botanical Review 49:155-232.
MORl, S. A. 1989. Eastern extra-Amazonian Brazil. pp. 427-454 in Campbell, D.G. & Hammond, H.D. (eds).
Floristic inventory o/tropical countries: the current status o/plant systematics, collections, and vegetation. New
York Botanical Garden, New York.
MYERS, N. 1984. The primary source: tropical forests and our future. Norton, New York.
MYERS, N. 1988. Threatened biotas: "hot spots" in tropical forests. The Environmentalist 8:187-208.
PASSOS, J. R. 1984. Estudo comparativo da vegetafQo em tres estagios da sucessao e a determinarao da area
minima de amostragem para inventario jloresa na EstafQo Ecologica do Pau-brasil. Centro de pesquisas do
cacau. Unpublished manuscript.
PETERS, C. M., GENTRY, A. H. & MENDELSOHN, R. O. 1989. Valuation of an Amazonian rainforest. Nature
339:655-656.
RANDALL, A. 1991. The value of biodiversity. Ambio 20:64-68.

RELATORIOS. Associa9ao Comercial da Bahia. 1867-1940. Wilcke, Picarde Co.
SALDARRIAGA, J., WEST, D. C. & THORP, M. L. 1986. Cited in Brown, S.A. & Lugo, A.E. (1990).
Tropical secondary forests. Journal of Tropical Ecology 6: 1-32.
SILVA, L. A. & VINHA, S. G. da 1982. A piassaveira (Attalea funifera Mart.) e a vegeta9ao associada no
Municipio de IlMus, Bahia, Centro de pesquisa do cacau, boletim tecnico no. WI, IlMus, BA, Brasil.12p.
SILVA CAMPOS, J. 1981. Cr6nica tin Capitania do Sao Jorge de Ilheus. Nacional, Rio de Janeiro.
SOUSA, G. S. de. 1938. Tratado descritivo do Brasil em 1587. n.p.
SPIX, VON & VON MARTIUS, C. F. P. 1928. Atraves tin Bahia: excertos tin "obra reise in Brasilien ", translated
by da Silva, P. & Wolf, P. Imprensa Official, Bahia.
UHL, C. & MURPHY, P. G. 1981. Composition, structure and regeneration of a tierra firme forest in the Amazon
Basin of Venezuela. Tropical Ecology 22:219-237.
VILHENA, L. dos S. 1969. A Bahia no seculo XVIII. Volwne I. Editora ltapoa, Salvador.
VOEKS, R. A. 1985. Preliminary observations on the reproductive ecology of the piassava palm (Attaleafunifera).
Anais, Academia Brasileira de Ciencias 57:524-525.
VOEKS, R. A. 1987. A biogeography of the piassava fiber palm (Attalea funifera Mart.) of Bahia, Brazil. Ph.D.
dissertation, University of California, Berkeley.
VOEKS, R. A. 1988a. The Brazilian fiber belt: harvest and management of the piassava fiber palm (Attalea funifera
Mart.). Advances in Economic Botany 6:262-275.
VOEKS, R. A. 1988b. Changing sexual expression of a Brazilian rain forest palm (Attalea funifera Mart.).
VOEKS, R. A. & da VINHA, S. 1988c. Fire management of the piassava fiber palm (Attalea funifera) in eastern
Brazil. Yearbook, Conference of Latin Americanist Geographers 14:7-13.
VOEKS, R. A. 1990. Edaphic limitation of a Brazilian rainforest palm: the role of energy allocation and competition
for sunlight. Physical Geography 11: 154-171.
WEBERING, V. A. 1937. Notas sobre a piassaveira. Bahia Rural 5(49/50): 1828-30.
WETHERELL, J. n.d. Brasil apontamentos sobre a Bahia: 1842-57, trans. M.P. do Rio Branco, S.A. Artes
Graficas, Salvador.
WIED-NEUWIED, Prince Maximilian. 1940. Viagem ao Brasil. Editora Nacional, Sao Paula.
Geographic information systems in rainforest ecosystem research: the
Sarawak pilot model
EBERHARD F. BRUENIG and SONJA CSOMOS
University of Hamburg, Leuschnerstr. 91, D-21031, Hamburg, Germany, and Forest

Department Sarawak, Wisma Sumber Alam, Petra Jaya, 93660 Kuching, Sarawak, Malaysia
THE DEVELOPMENT OF CONCEPT
In 1963, a 20 ha permanent ecological research plot, now Forest Department Research Plot
(RP) 146, was established in an area of Kerangas and Mixed Dipterocarp forest in Sabal
Forest Reserve, West Sarawak. All trees over 10 cm dbh were enumerated and the
topography sketched in 20 x 20 m recording units (exactly: chain-squares). The trees were
not mapped because we did not realize at the time the great heterogeneity of soils and species
associations. The conventional traditional textbook analyses of species richness, tree species
diversity (evenness of mixture), and general growing stock patterns and densities in the
sixties were easily determined (Weiscke, 1982). However, we encountered problems with
determining associations between tree species, between species and soil features, and between
patterns and assumed dynamic trends. The chain-square grid was too coarse to catch the fine
fabric of soil and vegetation heterogeneity which varies at a scale of metres and even less.
Another disadvantage was that plant locations within the 20 x 20 m square were not
recorded. This made it impossible to analyse species associations and cohort dynamics
among the smaller trees. Unless dynamic cohorts and neighbourhood conditions can be
determined and observed in time series of remeasurements, any measurements of growth,
nutrient stocks and cycles and hydrological conditions will be difficult to interpret. The
shifting-mosaic dynamics have proved to be so strong and the time-series of patterns of the
shifting mosaics among the smaller trees so turbulent (Poker 1992) that research into
processes which are influenced by the autecological properties and physiological state of the
trees at the point of measurement is a shot in the dark unless the fine-grained spatial
information is available.
Consequently, in our next ecosystem study in the seventies, the International Amazon
Ecosystem MAB-pilot project at San Carlos de Rio Negro, we established and demarcated
a coordinate grid covering the 20 ha research area. This was intended (Bruenig 1977, 1978)
to provide a permanent frame of reference for the coordination and linking of research
activities by different scientists and different disciplines. This purpose was achieved but a
further deficiency emerged. Research into the interactions between stationary plants and
mobile animals and humans required larger areas and observations outside the 20 or 50 ha
vegetation research plots. The coordinate grid had to be extended accordingly far beyond
the boundaries of the natural-ecological ecosystem research area.
This was done in the eighties in the two tropical natural forest and restoration ecosystem
studies in Guangdong (Xiaoliang barren coastal ecosystem) and Hainan (Bawangling
evergreen tropical forest ecosystem) which were part of the Chinese-German Coordinated
Ecological Research Programme (CERP). The purpose was to:
establish a coordinate grid (1 dm resolution) for the natural ecosystem study as in San
Carlos de Rio Negro
extend the grid into the landscape beyond to include agricultural and degraded lands
489
490 EBERHARD F. BRUENIG AND SONJA CSOMOS
areas under native customary rights of the indigenous Li people, and village ecosystems
to coordinate and link the data sources of research with those of practical forestry,
agriculture, infrastructural and general economic and social development
to secure safety and integrity of the research area by means of integrating the research
area into the landscape ecosystem which not only widened the scope for research, but
also made local authorities and planning agencies better aware of the existence of
the area and also of research as a source of information of value to themselves.
OVERALL OBJECTIVES
The ultimate objectives for establishing the spatial information system for ecosystem research
are to:
improve the efficiency of collecting and utilizing research data
improve inter-disciplinary analyses of associations and cause-and-effect relationships
provide a stable spatial reference for longterm time series
improve the security and integrity of the research area against abuse by scientists and
encroachment from outside
provide a means to prevent data and projects to be lost and forgotten
stop the notorious habit of scientists to re-invent the wheel and to repeat work by
making it impossible for newcomers not to recognize existing data and knowledge
from past and ongoing projects.
At the same time it was hoped that spatial linkages and easier identification and access
would lead to better integration of forestry and general land-use practice and would
eventually reduce:
misallocation of sites to unsuitable uses
failures to match species with natural and economic site conditions
application of obsolete or unsuitable techniques
wastage of time, money, natural resources and opportunities as a result of ignorance
of existing data sets and knowledge
REMAINING PROBLEMS
While scores of diverse proprietary brands of geographic information systems have recently
flooded the market, none has yet met our needs adequately. Also, none has been or is
sufficiently user-friendly that it could be efficiently used for research and application in a
tropical rainforest country. Improvements of the quality of ecosystem research by the use
of efficient spatial (geographic) information systems are urgently needed. This applies at all
levels of the ecosystem hierarchy (Gross mass in Bruenig et al. 1986):
natural ecosystem level (forestry, landscape)
technological ecosystem level (forest management and conservation)
economic ecosystem level (local, regional and national economy, tactical policies)
socio-political ecosystem level (formulation and decision of strategic policies).
A dramatic increase in demand for spatial information systems which provide links between
research and practice and between the various sectors involved in land-use analysis, planning
and monitoring has arisen from the decision and pledge by the member states of the
International Tropical Timber Organisation (lITO) to achieve sustainability of forest
management by the year 2000 (lITO 1991), to which Sarawak has particularly committed
itself. This target will not be met unless an efficient information system is in place on time,
that means immediately. But information systems which adequately answer our specific
Geographic Information Systems - The Sarawak model 491
needs of forest ecosystem research, forest management and forest and environmental
conservation at all levels are not yet available.
METHODOLOGICAL APPROACHES
The problem has been attacked by our team (E.F. Bruenig, Jackson Lau, Steven Chan, Sonja
Csomos, Stephen Price) at three levels (Figure 1):
(1) Research
advancement and improvement of the Sabal - San Carlos - Bawangling concept for
GIS ecosystem research
establishing operational data linkages of the ecosystem GIS for basic scientific
research in natural rainforest ecosystems with existing or evolving special research
GIS (e.g. genetics, autecology silviculture, growth and yield, protection utilization,
biodiversity) and higher-level ecosystem research GIS (socio-economic,
anthropological, socio-political).
(2) Practice of Management, Economics, Sociology
Developing operational linkages between existing sectoral GIS (e.g. between
agriculture, forestry, infrastructure, industrial development).
(3) Integrated Multi-Sectoral Information System (IMSIS)
This is a new development of a comprehensive GIS with a holistic structure which
integrates the sectoral GIS of research and practice which we developed over the past
three years in the form of an integrated multi-sectoral information system (Figure 1).
The primary objective is to provide a solid, sound and reliable data and information
base for technical planning and monitoring of forestry and agricultural land use and
for general economic and social development. A further objective is to link scientific
research, application-oriented experimentation and practice within an efficient,
comprehensive information system. IMSIS is particularly suited to prevent or solve
conflicts arising out of ignorance or disregard of conflicting interests and intentions,
of conservation needs, native customary rights, local attitudes, local spiritual and
cultural requirements. It was a condition that it should be flexible and applicable to
the different levels of local and regional forest management planning and monitoring
and of regional, state-wide and national land-use and economic planning and
monitoring. IMSIS is designed to incorporate any proprietary brand of GIS which
a user may have already installed or may want to use.
IMSIS in combination with data from remote-sensing provides graphic and non-graphic
data for:
information of past and present states (statistics, areas, etc.)
information on existing sectoral plans and ongoing planning projects in the area
covered
analysis of relationships and interactions between various sectors (mutual supports,
conflicts, synchronization of operations)
analysis of sensitivities and dynamics of the sectoral systems involved.
Combination with dynamic system analysis and simulation permits the assessment of
possible future states of ecosystems (Bossel & Bruenig 1992, Bruenig et al. 1986). IMSIS
can, if required and wished by government, also be used as an instrument for exchange of
data with international regional data stores of regional networks or with global data stores
(Figure 1).
STATE OR
NATIONAL
LEVEL
IMSIS
I I
Specialized Agricu~e
Sectoral Fisheries
Research Village Trade
:; I :7:l:J-
-, I:
Research :;:
"1
5 : . I~S
~
I
and Conservation
Eco-Tourism
'~d Micro & Macro

Research Economics
socio--economics Trade, Markets
.a ~T '~
~-rlGIS.+_
Global·linked
Research
and Networ1<
Monitoring
I 18=
Il GIS rf--____
Population
Education
Hea~h
---j
INTERNATIONAL-LEVEL GIS
ego GLOBAL CHANGE
MONITORING, CITES,
GEMS, GRIT etc.
Figure 1. The structure of the harmonized and integrated geographical information system complex serving research
and practice of sustainable forest management, land·use and socio-economic development and monitoring. The
various GIS of local and specialised research projects (left) are linked directly among themselves. They may also
be linked directly to corresponding GIS in practice (right) if desired. The more efficient analysis and synthesis
function is supplied by the comprehensive Integrated Multi-Sectoral Spatial Information System (IMSIS). IMSIS
supplies landuse , socio-econornic and socio-cultural information for the identification and assessment of capacities
and opportunities, the analysis of potential conflicts and sources of instability and for long-term monitoring and
evaluation of research and practice.
ACTIVITIES TO SOLVE THE PROBLEM
The development of the GIS technology for individual projects of scientific ecosystem
research is an open-ended, ongoing process. The standards for data collection and processing
in rainforest ecosystem research are well established since the inception of the IBP and MAB
programmes (Bruenig 1977). Adherence to these standards assures compatibility of data, but
leaves the user freedom to decide which GIS he wants or develop to suit particular needs.
IMSIS, as a new concept and technique, has been developed to the state of an operational
prototype which can be used on a PC or work station. The next phase is field testing in a
suitably heterogeneous forested and agricultural region of Sarawak. This will then lead to
Geographic Information Systems - The Sarawak model 493
full operationality for use in r~seach and forestry practice in 1994. The ecosystem research
GIS for individual ecosystem research projects will simultaneously be further advanced,
based on the experiences from recent field work in Sabal Forest Reserve, RP 146. The plot
is now being re-established and extended southwards to cover a research area of
approximately 600 ha Kerangas and Mixed Dipterocarp forest, including the remaining 12
ha of the original 20 ha plot.
CONCLUSION
Rainforest ecosystem research and tropical forest research has, until now, been rather
inefficient and its results have been largely ineffective in influencing forestry practice and
the general public. Forestry practice has notoriously and generally fallen short of targets and
is specifically yet far from the lITO Target 2000. Among the major reasons for this are the
lack of knowledge of research done and its results, lack of communication and coordination
between forest research and practice, and between forestry and the other sectors of the socio-
economic, social and political systems. Underlying these shortcomings, which are crucial
obstacles to holistic approaches to sustainable development, is the notorious failure of
governments to support the development and installation of comprehensive, integrated and
multisectoral user-friendly spatial information systems which woUld benefit themselves, the
private sector and the public. An efficient, user-friendly GIS for rainforest ecosystem
research is essential to improve research quality and the application of results to management
and policy practice. IMSIS is therefore essential for advancement and improvement as an
instrument of integration and a vehicle of communication between ecological science, the
practice of forest management and conservation, and the practice and politics of landuse.
The still existing wide gap between information needs and supply has most recently again
been confirmed by the 14th Commonwealth Forestry Conference in Kuala Lumpur. The
final recommendations of the conference state "that new developments in fields such as
biotechnology, geographical information technology and valuation of non-market benefits (of
forests) be promoted" by Commonwealth Governments in their very own interest.
ACKNOWLEDGEMENTS
The authors are thankful to the German Research Foundation (DFG), the Federal Ministry
for Research and Technology and the Forest Department, Sarawak, for supporting the
ecosystem research projects which provided the opportunities for developing the ecosystem
GIS and the higher-level IMSIS. Special thanks are due to Miss Julia Philip for processing
the manuscript, to Dr. T. Schmaltschinski and Mr. William Then for computer-drafting.
LITERATURE CITED
BOSSEL, H. & BRUENIG, E.F. 1992 Natural Resource Systems Analysis. Ecological and Socio-economic Systems
and Sensitivity Analysis for the Conservation and Management of Forest Ecosystems and Natural Resources in
South-East Asia. Feldafing, Food and Agriculture Development Center (ZEL), DSE, Dok.Nr. 1655 A/a-TK 78-
300-89.
BRUENIG, E.F. (ed.) 1977. Transactions of the International MAB-IUFRO Workshop on tropical rainforest
ecosystem research. Chair of World Forestry, Hamburg University. Special Report No.1. pp. IV + 363.
BRUENIG, E.F. 1979. The International Amazon Rainforest Ecosystem Pilot Project at San Carlos de Rio Negro:
Vegetation Classification and Structure. pp. 67-100 in Bruenig, E.F. & S. Adisoemarto, S. (eds). Transactions
of the 2nd International MAB-JUFRO Workshop on Tropical Rainforest Ecosystem Research, Jaknrta, 21.-
25.10.1978. Chair of World Forestry, Hamburg University, Special Report No.2.
BRUENIG, E.F., BOSSEL, H, ELPEL, K.-P., GROSSMANN, K.-D., SHCNEIDER T.W., WANG, ZHU-HAO
& YU, ZUO-YUE 1986. Ecological-socioeconomic system analysis and simulation: A guide for application of
system analysis to the conservation, utilization and development of tropical and subtropical land resources in
China. Hamburg, Inst. World Forestry. Feldafing, ZEL-DSE; and Bonn, German MAB Nat. Conun. Mitt. 24.
pp. XV + 388.
BRUENlG, E.F. & SYNNOTT, T.J. 1977. An outline for a basic interdisciplinary research progranune with
emphasis on a conunon approach to research methodology and measurement techniques. pp. 265-339 in Bruenig,
E.F. (ed.). Transactions of the international MAB-lUFRO Workshop on tropical rainforest ecosystems research.
Chair of World Forestry, Hamburg-Reinbek, Special Report No.1. Revised Edition, 1978.
ITTO 1991. Beyond the guidelines - an action program for sustainable management of tropical forests. ITTO
Technical Sevices No.7. pp. VI + 187.
WEISCKE, A.S. 1982. Structure and functions offorest ecosystem: structural comparison between Kerangas and
Coatinga. Unpublished thesis, Chair of World Forestry, Faculty of Biology, University of Hamburg. pp. ill
+ 83 (in German).
POKER, 1. 1992. Mosaikdynamik in tropischen Regenwald, (Mosaic dynamics in the tropical rainforest).
Unpublished thesis, University of Hamburg, cyclostyled. pp. 220.
The creation of a geographical information system for the Kuala Belalong
Field Studies Centre
JOHN T. WILLS
Department of Geographical and Environmental Sciences, University of Huddersfield,

Queensgate, Huddersfield HDI 3DH, United Kingdom
ABSTRACT. Whilst the applications for Geographical Information Systems (GIS) are expanding, their
use in tropical rainforest research is relatively undeveloped. At the conclusion of many scientific
expeditions and projects, the data produced becomes dispersed between a large number of individuals and
institutions, resulting in some information being under-utilised or even lost. Apart from the traditional
monograph, many projects involving large numbers of scientists have also lacked the means to illustrate
their research findings visually. GIS offers a means to display, store and manipulate geographically
referenced data. Recording data in this way ensures its long term survival and use, allowing for frequent
update, the revisiting of sample sites and an expansion in the uses of the information for scientists and
non-scientists alike. This paper details the aims, methods and problems encountered in the creation of a
GIS for The Kuala Belalong Field Studies Centre in Temburong District, Brunei Darussalam, as part of
the UBD/RGS Brunei Rainforest Project 1991-92. Topographical data relating to the project area was
digitised and results of the scientific research added as they became available in the form of database files
linked to geographical points or areas. Problems were encountered with the accuracy of locational data
and in converting some data into a suitable form for input into the GIS. A number of suggestions are
given to aid the creation of future GIS's.
KEY WORDS: tropical rainforest, GIS, Brunei Darussalam, Borneo.
INTRODUCTION
"A Geographical Information System is an organised collection of computer hardware,

software, geographic data and personnel designed to efficiently capture, store, update,
manipUlate, analyse and display allforms of geographically referenced information" (ESRI
1990). Within a GIS distinct classes of data are stored in different layers (see Burrough
1985, p.20). If the data layers within a GIS relate to the same geographic area, they can be
overlaid in any combination and it becomes possible to perform spatial analysis on them.
Many commonly used computer programmes such as Minitab and Autocad can handle
simple geographic or spatial data but are not considered to be Geographical Information
Systems. Whilst modern Computer Aided Design (CAD) systems are capable of placing
different types of geographic features on individual layers and linking these layers with
attributes in a database, they are merely graphic systems (Cowen 1988). Automated mapping
systems are capable of more complex linkages between geographic and display data, but are
still "restricted to the functions of data retrieval, classification, and automatic
symbolisation ....... a computer mapping system still is not a GIS" (Cowen 1988, p.1552).
To distinguish itself from computer mapping systems, a GIS must be capable of performing
the following four main tasks (Scholten & Lepper 1991):
1: Storing, managing and integrating spatially referenced data:
- spatial data relating to points, lines or areas;
- descriptive data relating to the attributes of the points, lines and polygons.
2: Spatial query - for example, searching for the areas in which a particular species is
found.
495
D. S. Edwards et at. (eds.!, Tropical Rainforest Research - Current Issues, 495-502.
496 JOHN T. WILLS
3: Geographic analysis - for example, statistical analysis of the relationship between data
sets on tree girth and altitude.
4: Displaying data in the form of high quality maps.
AIMS
The primary aim of this project was to create a GIS which could be used as an educational
tool for students and members of the public visiting the field centre, as well as being useful
to visiting scientists to illustrate the research work previously conducted. It was hoped that
map print-outs would form a useful means of information display for publications and that
the data contained in the GIS could be used for detailed spatial analysis by other researchers.
METHODOLOGY
Hardware & Software configurations

This project used a vector based GIS: PC ARC/INFO version 3.4D on an IBM compatible
386 PC. Digitising was done on a Summagraphics Microgrid II digitiser, whilst maps were
plotted on a Hewlett Packard HP7475 plotter. A Toshiba 2200 8086 laptop was used for
database entry in the field. Some data was also transferred to a raster based image
processing and GIS package: IDRISI version 3.0, to produce digital elevation models, slope
angle and slope aspect maps.
Topographical and related data

The first stage in establishing a GIS is the creation of an accurate topographical
representation of the world's surface. To this end, 1:50000 scale Ordnance survey maps of
the project area detailing contours, rivers, streams and the Rectified Skew Orthographic
(RSO) grid system were digitised. After discussions with project members, data layers for
the paths, sub-camps and project boundary (based on contour interpretation and estimation)
were created by drawing in on the Ordnance Survey maps and re-digitising (Figure 1).
To up-date the existing maps and provide accurate positions for specific locations a
Trimble "Transpack" Global Positioning System (GPS) was used. GPS is a US Department
of Defence satellite-based navigation system which began development in the late 1970's.
The system consists of twenty-four satellites in high, geosynchronous orbits arranged in six
orbit planes. Each orbit plane which contains four satellites is inclined to the equator an
angle of 55 degrees. This means that all areas of the earth have at least four satellites visible
at anyone time. Ground-based GPS receivers intercept signals from the satellites to calculate
the user's position (Cross 1991). Before using the GPS it is necessary to select the
appropriate map projection for the area of the world in which the fixes are to be taken; in
Brunei's case this is the Timbali projection. This enables the GPS to convert the locational
information from World Geodetic System 1984 (WGS-84) coordinates into the correct
latitude and longitude for different map projections (Brown 1992).
Existing data
After completing the topographic coverages a search was made for existing data relating
to the project area. Much of Brunei has been covered by detailed geological and soil
surveys, but due to its remoteness the project area has not been included in these surveys,
the only data available being of a very coarse resolution. The "Temburong Renewable
Resources Study" (Anderson & Marsden 1983) was found to be the only comprehensive
dataset available. This survey covered the whole country and provides detailed information
Geographical Information System 497
[ZQJ Rivers
[2SZJ Project Area
~ Paths
o 2 3k
Figure 1. Example of the output from the GIS. Included are data layers of
rivers, paths, the project boundary and the sub-camps.
on vegetation type in Southern Temburong district. The relevant 1:50000 maps were
digitised and exist as a separate data layer.
Project data
With over seventy scientists conducting research in the project area, the volume of data
produced was substantial and spans a wide range of scientific disciplines. To enable a
variety of users to make effective use of the system it was therefore necessary to record the
data produced in a concise and understandable form.
Some data was received in the form of written reports from scientists who had completed
their field studies whilst the remaining scientists were interviewed. Once it was decided that
a particular research project would require GIS input, a specific request for data was made:
for example, the points or areas sampled and the species found there. This data was entered
into specific subject database files containing spatial data (e.g. species present) and attribute
data (e.g. an identification number linking the data to a particular sampling point or area).
Some files relate to the work of individual scientists while others are the combination of
results from two or more scientists working in the same field. The files were modified and
updated as required.
Each database file was stored as a different data layer to allow clear viewing of the results
and to avoid confusion. There are six background coverages relating to basic topology and
the project's infrastructure. Twenty four coverages relate to the scientific results. This series
of coverages accurately reflects the complexity of the real world in a clear, intelligible way.
498 JOHN T. WILLS
RESULTS
The following information was entered into the GIS:
Topographic data
1 -
Project boundary
2 -
Contours
3 -
Rivers and streams
4 -
Paths
5 - Pondoks (the field centre and sub-camps in the forest)
6 - Grid reference system (Rectified Skew Orthographic)
Existing data
1 - Vegetation Type (Temburong Renewable Resources Study 1983)
Project data
A: Plants 1- Plants of Belalong (extracted from log book records)
2 - Plant growth environment study (D. Mitchell)
3- Pteridophyte collection (D. Edwards)
4- Ethnobotanical survey (K. Abu Salim)
B: Invertebrates 1- Invertebrates of Belalong (extracted from log book records)
2 - Phasmid survey (P. Bragg & I. Abercrombie)
3 - Stalk eyed fly study (D. Burkhardt & I. De la Motte)
4 - Brachyceran Diptera study (M. Drake)
5 - Microlepidoptera study (G. Robinson)
6 - Woodlice study (1. Medway)
C: Fish 1 - Fish survey (S. Choy)
D: Herps 1 - Frog survey (c. Reading)
2 - Reptiles of Belalong (I .Das & log book records)
3 - Amphibians of Belalong (I. Das & log book records)
E: Birds 1-4 - Birds of Belalong (c. Mann, D. Wells & log book records)
F: Mammals 1 - Mammals of Belalong (J. Charles & log book records)
2 - Bat survey (c. Francis)
3 - Mammal transect survey (L. Bennett)
G: Physical data 1 - Waterlevel records of S. Belalong (manual records)
2 - Rainguage records at KBFSC (manual records)
Because these coverages relate to the same base area it is possible to overlay them in any
desired combination and perform spatial analysis. For example, it is possible to show which
species are present in any sampling point or area on any of these coverages, and vice versa
in which sampling points or areas a particular species is found.
DISCUSSION
Spatial accuracy
The only maps available for the project area were at a scale of 1:50000, based on aerial
surveys flown some thirty years ago. The absence of more up-to-date maps is due to the
remoteness of the area, although a new aerial survey was flown in 1992 and will be used to
produce 1: 10000 maps. Some of the scientific results produced relate to very specific
locations and it was therefore felt that the detail of the base map should be supplemented to
provide accurate locations at a larger scale. This is not to say that the original 1:50000 maps
are inaccurate, rather that they are not of a sufficiently large enough scale to identify points
only a few metres apart accurately.
As previously mentioned, a GPS was used to fix the positions of the permanent scientific
plots, project sub-camps, the Field Studies Centre and streams not recorded on the Ordnance
Survey maps. The latitude-longitude positions produced by the GPS were converted into
RSO grid references with the help of the Land Survey Department in Brunei, by using a
computer programme.
It was found that in open locations unobscured by the forest canopy three or more
satellites could be tracked and the GPS used in two dimensional or three dimensional mode.
If the height of a given location was known, the best results were obtained by using the GPS
in two dimensional mode so that locational accuracy could be concentrated on the two
horizontal dimensions. The Trimble handbook states that an inaccuracy of five metres in the
figure entered for height will lead to a position error of approximately 15-20 metres,
although in the author's experience this error was sometimes considerably larger.
Under closed canopy or in semi-obscured locations it was necessary to use the GPS in two
dimensional mode permanently, as the physical obstruction of the trees limited the number
of satellites 'visible' and caused considerable position drift. By taking several readings it was
possible to obtain a good position cluster, of which the centre theoretically provides the best
location. Whilst some results exhibited a high degree of clustering, the positions cannot be
assumed to be totally reliable. The altimeter used had to be calibrated at a point of known
height, with the height of other points being derived by the altimeter applying the relationship
between increased altitude and decreasing air pressure (and vice versa). Tropical climates
are noted for their unstable bodies of air and large diurnal changes in air pressure (Briggs
& Smithson 1985). Thus, an altimeter which relies solely on air pressure to determine height
will produce different readings at the same point over time. Consequently, position readings
from the GPS taken over time at the same point show considerable drift, in line with the
findings on height-position error in the previous section. Whilst readings taken at the same
time may show a high level of clustering, the central point cannot therefore be taken as the
optimum location, unless the height value entered is known to be accurate. To assess the
effect of altitudinal error, readings were taken using altimeter height and observer's estimated
height (estimated from a given point on the Ordnance Survey map). There was frequently
a discrepancy between the two height values of up to 25 metres, with the resultant position
readings varying by up to 250 metres (see Figure 2).
GPS is therefore highly accurate in two dimensional mode where the height is accurately
known, although in closed canopy areas some persistence may be required for the machine
to track sufficient satellites and position drift may result from tree obstruction. It was more
difficult to obtain positions in valley side locations than in valley bottoms and on ridge tops,
although this will depend on the position of satellites at the time the fix is taken. As an
indicator of accuracy, the machine gives its own measure of the precision of each fix,
displaying +/-30 metres to +/-1000 metres with each reading. More accurate positions can
therefore be obtained by recording only those with the highest accuracy rating. Inaccuracies
in the height value entered will lead to considerable errors in position location to such an
extent that the positions may not be valid. This problem appears to be the biggest stumbling
block to the further use of GPS in forested areas. Whilst it is suggested that the GPS would
be most useful in open areas such as desert or tundra, methods for lifting the GPS aerial
above the canopy, for example by balloon or telescopic pole, coupled with improved
altimeter accuracy could increase its usefulness in mountainous or forested areas.
The GPS data are currently being analysed to assess the precise relationship between the
number of readings taken and the degree of clustering. This analysis will show whether or
not the assumptions about clustering hold true.
500 JOHN T. WILLS
Figure 2. Distribution of GPS readings recorded from Bukit Belalong. Grid square = 1000 m x 1000 m.
The lack of spatial data

The majority of Geographical Information Systems relate to areas for which considerable
spatial data already exists; for example, data on land-use classes, geology or soil type. With
the exception of the vegetation type data and base maps, no detailed spatial data for this area
existed before the Brunei Rainforest Project took place. Whilst considerable spatial data now
exists as a result of the project, this relates to buffered areas around paths and streams (for
16 coverages with streams buffered at 15 m and paths at 60 m) and points or sampling
stations (7 coverages). As research in the project area continues and more data become
available, it may be possible to extrapolate them to produce species location maps or species
density predictions. However, this process is problematic, especially where existing species
data are poor and the species habitat parameters wide. Where the habitat parameters are
narrow, species presence estimates are more feasible. For example, in the case of a
particular fish (a species confined to small spatial areas) a strong correlation was found to
exist between stream order and altitude for existing data on the species of fish found at thirty-
nine sampling stations in the project area.
Data unification
Ideally the data pertaining to one subject area (e.g. plants) would be combined as a single
coverage to make the storage and presentation of data more logical. However, this would
necessitate the construction of complex database files and would only be possible where all
of the data had a comparable degree of locational accuracy and related to approximately the
same geographical area. For example, it might not be desirable to combine the data on
Microlepidoptera which relate to precise sampling locations with the other datasets on
Invertebrates which relate to much broader sampling zones.
The utility of Arc/Info

Arc/Info operates by using a large number of individual commands. Some of these are
very similar, have complex suffixes and can take a considerable time to learn. Arc/Info
proved to be an excellent platform for digitising, display and analysis, easily meeting the
requirements of this project. The map output was of a very high quality and it was possible
to produce clear, detailed maps with relative ease. Arc/Info uses Dbase files to store data;
for this GIS, Dbase III + was used. This version of Dbase has a limited number of fields.
In the case of the bird data, because there were over 200 species found in the project area
it was necessary to create four separate coverages. This is not essentially a fault with
Arc/Info and could be rectified by transferring the data into Dbase IV.
SUGGESTIONS FOR EFFECTIVE GIS CREATION
The following comments stem from the author's GIS experience and are relevant to the
creation of future GIS's:
Spatial accuracy
Ideally the area covered by this GIS would have been mapped in detail before the
scientific studies began, thereby increasing the accuracy of the data produced. For example,
data could then be expressed as species 'X' present at a particular grid reference. This
would greatly increase the value of the GIS for detailed analysis and modelling work:
maximising spatial accuracy increases the effectiveness of a GIS.
Data protocols
The establishment of data protocols prior to the start of the project would have simplified
data entry into the GIS. For example, a GIS data entry form could have been created for
project members to complete with their final reports, thereby reducing the time needed to talk
to all members individually. Creating data protocols formalises the system of data input and
ensures that all data are compatible with the requirements of the GIS.
Following closely defined aims

The primary aim of this project was to provide a GIS for educational use. The planned
uses of a system will influence the data included, the levels of accuracy required and the
data format. Unless the aims are outlined in advance, the final system may not meet the user
requirements. It is therefore essential that all GIS's are created with specific aims in mind.
Ease of use
It is essential that GIS's are user-friendly to operate, to enable the maximum number of
people to use the system with the minimum training and specialist knowledge. To this end,
the completed datasets for the Belalong GIS were entered into ARCVIEW, a display/analysis
package produced by the makers of ARC/INFO.
CONCLUSIONS
This project has created a large number of coverages and provides a good overview of the
area's topology and the research work conducted. In view of the educational aims of this
GIS, the problems of spatial accuracy were not of paramount importance. Whilst the
effectiveness of the GPS was limited due to physical obstruction and altimeter problems, its
use in forested areas would be greatly increased by following the suggestions outlined.
Arc/Info proved to be a very effective package and had no significant limitations.
Maximising spatial accuracy, establishing data protocols, following closely defined aims and
ensuring ease of use were found to be very important factors in creating an effective GIS.
502 JOHN T. WILLS
ACKNOWLEDGEMENTS
I thank the Land Survey Department of Brunei Darussalam for converting the GPS positions
into RSO grid co-ordinates (in particular Dy. Mas Suria Watia bte Hj Watib) and Brunei
Shell Sdn Bhd for offering to provide a Landsat TM image of the project area. I am grateful
to all Brunei Rainforest Project members for providing their data and co-operation, and to
all of my sponsors, Universiti Brunei Darussalam and the Royal Geographical Society for
making my work in Brunei Darussalam possible.
LITERATURE CITED
ANDERSON, J. & MARSDEN, D. 1983. The Temburong Renewable Resources Study. Forestry Department,
Bandar Seri Begawan, Brunei Darussalam.
BRIGGS, D. & SMITHSON, P. 1985. Fundamentals Of Physical Geography. Unwin Hyman, London.
BROWN, A. 1992. The GPS Coordinate system explained. GIS Europe 1:62-63.
BURROUGH, P. A. 1986. Principles of Geographical Information Systems for Land Resources Assessment.
Clarendon Press, Oxford.
COWEN, D. 1988. GIS versus CAD versus DBMS: What Are the Differences? Photogrammetric Engineering and
Remote Sensing 54:1551-1555.
CROSS, P. A. 1991. GPS for GIS's. Mapping Awareness & GIS Europe 5:30-34.
ENVIRONMENTAL SYSTEMS RESEARCH INSTITUTE (ESRI) 1990. Understanding GIS; The ARC/INFO
Method. ESRI, Redlands, California.
SCHOLTEN, H. J. & LEPPER, M. J. C. De 1991. The benefits of the application of Geographical Information
Systems in public and environmental health. World Health Statistics Quarterly 44.
Tropical rain forests in Sri Lanka: characteristics, history of human
impact, and the protected area system
W. ERDELEN
Department of Biogeography, University of the Saarland, Postfach 1150, D-6604J

Saarbriicken, Germany
ABSTRACT. Natural forests cover about 20% of the tropical island of Sri Lanka. However, distribution
in the major hygroclimatic zones is biased towards the drier parts of the island. Forest coverage is about
30% in the dry zone, but only 9% in the wet zone where Sri Lanka's rain forests occur. They contain
the most diverse communities of the island with the highest concentration of endemic species. Over the
past centuries rain forests have been reduced to small and more or less isolated fragments. To a large
extent this has been a result of human dispersion and concomitant impact on natural landscapes. Most
significant changes took place during the British colonial time (1796-1948) and after independence (1948).
Present population trends in the wet zone of Sri Lanka tend towards exacerbating impacts on the
remaining rain forests. Therefore, the existing protected area system should be extended with emphasis
on remaining rain forest fragments.
KEY WORDS: forest communities, vegetation history, forest reserves, conservation.
INTRODUCTION
Tropical rain forests have been a central focus of international conservation activities such
as FAO's Tropical Forestry Action Plan and the International Tropical Timber Agreement.
Another major and related concern has been the preservation of global biodiversity (e.g.
WCMC 1992). As the most complex of our extant ecosystems, both with regard to their
plant and animal communities, rain forests have played a key role in this context. Although
research on tropical forests has a long tradition, our knowledge about their structural and
functional characteristics is still relatively poor. Up to the 1960s the tropics had been
considered a region with a past of constant environmental conditions. Later evidence clearly
documented that many tropical rain forest ecosystems have experienced drastic changes in
geographic extension and species composition, particularly during the Quaternary (overviews
in Flenley 1979, Goudie 1985). In addition, the mosaic character of rain forests (Remmert
1991) and their intrinsic (patch) dynamics (Pickett & White 1985) have changed our views
of the 'constant' tropical environment. Moreover, substantial evidence has shown that
alternative interpretations if not changes in ecological paradigms may be required to explain
structural and functional characteristics of tropical ecosystems (overview in Waide & Lugo
1992).
A drastic reduction in extent of tropical forests, especially in the neotropics, has taken
place over the last decades. This also applies to Sri Lanka where, particularly in the wet
zone lowlands and in the montane areas, severe human impacts have occurred. The rain
forests were already fragmented in the middle of the last century (Ishwaran & Erdelen 1990).
Immediate measures are required to preserve these remnants and to extend the protected area
system with emphasis on forest fragments in the wet zone. This paper describes
characteristics of the rain forests of Sri Lanka as well as the history of human impacts and
gives an overview of the protected area system.
503
D. S. Edwards et al. (eds.), Tropical Rain/orest Research - Current Issues, 503-511.
504 w. ERDELEN
GEOGRAPHIC BACKGROUND
Sri Lanka is a tropical island, located southeast of the Indian peninsula, between longitudes
E 79°39' and 81°53' and latitudes N 5°54' and 9°52' and with an area of 65,610 km2 • The
island comprises three peneplains with average altitudes of 30 m, 500 m, and 1500-1800 m,
respectively (Adams 1929, Domros 1976). The central highlands in the southern part of the
island comprise the second and third peneplain (Figure 1). These highlands consist of three
different geographic units, the Knuckles mountains in the northeast (highest peak:
Gombaniya, 1906 m), the central massif (Pidurutalagala, 2524 m), and the Sabaragamuwa
ridges in the southwest (Beralagala, 1386 m).
The dominating climatic factors are the southwest monsoon (May-September) and the
northeast trade wind (November-February). The southwest monsoon causes most of the
precipitation in the southwestern parts of Sri Lanka. During the northeast trade wind, the
northern and eastern regions receive their maximum precipitation. Convectional rains
dominate during the intermonsoon periods. Based on seasonal and spatial differences in
rainfall, Sri Lanka is subdivided into two major country-specific hygroclimatic zones, the wet
zone with mean annual rainfall above 1904 mm and the dry wne with means below 1904 mm
(Domros 1974, Figure 1).
lo"N
i
I
7°
6" o 20 40 km
ad'!: '---'--'
Figure 1. Relief and major vegetation types of Sri Lanka (after Erdelen 1989). KN = Knuckles mountains, R =
Ritigala, SH = Sabaragamuwa hills. AI, A2 = Tropical thorn forest, B = semi-evergreen forest, C = intennediate
forest, 01 = rain forest below 900 m, 02 (stippled) = submontane rain forest (900-1500 m), 03 (black) = montane
rain forest (above 1500 m). Towns: Co = Colombo, J = Jaffna, M = Mannar, P = Polonnaruwa, T =
Trincomalee.
Tropical rain forests in Sri Lanka 505
Temperature, relative humidity and day length show little seasonal variation. However,
average annual temperatures, which are about 27°C at sea level, decrease about 0.50-0.65°C
per 100 m altitudinal difference. Monthly temperature means vary only little, but daily
temperature maxima and minima differ between 5 and lOoC. Relative humidity is high
throughout the year, annual means are between 77% (Trincomalee) and 85 % (N'Eliya) and
vary on a regional and, more markedly, daily scale (see Domros 1974 & 1976 for details).
THE FOREST COMMUNITIES OF SRI LANKA
Distribution of Natural Vegetation

The natural vegetation of the island, particularly the forest commumtles, has been
classified into various units as summarized, for instance, by Grel1er & Balasubramaniam (in
press). In the National Atlas of Sri Lanka (Somasekaram 1988), 11 major formations are
distinguished, including seral stages such as tropical savanna forests and grasslands which
in most cases represent secondary vegetation (Pemadasa 1984), as wel1 as edaphical1y
control1ed formations such as mangroves. For this overview a subdivision of the island into
7 zones with 6 types of natural vegetation seems sufficient (Figure 1; for more details see
Gaussen et at. 1968). These include tropical thorn forests (or monsoon scrub jungle,
Manilkara-Chloroxylon series,) in the driest parts of the island (NW and SE) with 5-7 dry
months and an annual precipitation (a.p.) of about 1000-1400 mm, semi-evergreen forests
(Chloroxylon-Berrya-Vitex-Schleichera series), mainly in the northeastern, eastern and
southern parts (3-6 dry months, a.p. 1400-2000 mm), an intermediate type between
semi-evergreen forests and rain forests (Filicium-Euphoria-Artocarpus-Myristica transitional
series), in the intermediate zone of the island (1-2 dry months, a.p. 1900-2000 mm), and the
rain forest communities in the southwestern part and the hills (no dry season, a.p. over 2000
mm; Figure 1). The latter consist of lowland forests (Doona-Dipterocarpus-Mesua series,
sea level to about 900 m), submontane forests (Doona-Calophyllum-Syzygium series,
900-1500 m), and montane forests (Syzygium-Gordonia-Miche/ia-Elaeocarpus series, above
1500 m). Floristic and physiognomic characteristics of these formations are discussed, for
instance, in Balasubramaniam et al. (in press), Gaussen et al. (1968), Gunatilleke & Ashton
(1987), Jayasuriya et al. (in press) and Werner (1982, 1984).
The Rain Forests of Sri Lanka

Typical rain forests or "tropical moist forests" (sensu WCMC 1992) occur only in
southwestern Sri Lanka and in the hill region, i.e. in the wet zone (Figure 1). This zone
represents the only area of aseasonal wet climate between western Malesia and eastern coastal
Madagascar (Ashton & Gunatilleke 1987). Even the most humid regions of South India
experience a relatively dry phase (Legris & Viart 1961, Pascal 1984) which may explain the
lack of endemism in certain taxa such as the dipterocarps in this region (Gunatilleke &
Gunatilleke, personal communication).
The rain forests of Sri Lanka are not only the most diverse and species-rich of the natural
forests but they also contain the highest proportions of endemic plant and animal taxa
(Baldwin 1991, Erdelen 1989, Gunatilleke & Ashton 1987). Amongst angiosperm plants the
dipterocarps are the most characteristic elements of these forests. Of the 56 species
occurring in Sri Lanka, 55 are endemic (Balasubramaniam 1985, but see also Ashton 1980).
A reexamination of endemism in the lowland wet zone revealed that endemism in the woody
flora, as calculated from 9 sites sampled, was 68% (Gunatilleke et al. 1987). Moreover, a
recent study of the understorey plants in Sinharaja forest has also documented not only their
high species richness but also their high percentage of endemics, almost 50% of all woody
and non-woody plants (DeZoysa et al. 1988).
506 W. ERDELEN
Studies on the forest communities have shown that in Sri Lanka 15 floristic regions may
be distinguished (Gunatilleke & Gunatilleke 199O). Excluding the coastal and marine belt,
11 (79%) of the remaining 14 regions are located in the wet zone and the hills (Ashton &
Gunatilleke 1987). Even within these floristic regions tree communities may vary
considerably in physiognomy and species composition. For instance, studies. of aerial
photographs of Sinharaja forest have shown that the tree communities do not only vary with
altitude but have also a very pronounced mosaic character (Merritt & Ranatunge 1959,
Erdelen, in press). Though the best studied of the Sri Lankan rain forests, its internal
dynamics are still poorly understood. Only recently studies have been launched on the
phenology of major tree taxa, their pollination system, and reproductive biology (e.g.
Dayanandan et ai. 1990, DeZoysa et ai. 1991).
Many plant and animal species have very restricted geographic ranges within the island.
This has been documented for instance for Solenostenwn elongatus (Labiatae) and other
endemics confined to Ritigala hill in the northern part of the island (Figure 1; for details see
Jayasuriya 1984) or the endemics restricted to Hindumkanda, an isolated hiIl in the
southwestern lowlands (GunatiIleke & GunatiIleke 1984). These very localized geographic
distribution patterns have not adequately been considered with regard to conservation of the
forest communities in Sri Lanka, particularly those of the wet rone. Restricted geographical
distribution is one feature which makes species susceptible to extinction, another is low
population density. The latter has been shown to apply for many of the better known tree
taxa and also for the only rain forest understory community studied more in detail, i.e. that
of Sinharaja forest (DeZoysa et ai. 1988). For instance, according to Gunatilleke et al.
(1987) in the 9 study sites mentioned above, 98% of the woody endemic subcanopy and
canopy species and 85 % of the understory species are either rare, vulnerable or endangered
in terms of the mCN Red Data Categories. This heterogeneity in plant and animal
distribution and the fact that many species may be characterized as naturally already prone
to extinction becomes an even more important factor to be considered for conservation
measures with respect to human population growth in the island.
The History of Forest Cover in Sri Lanka

The impact of man on extension and successional stages of natural forest communities in
Sri Lanka has been an important environmental factor since the Holocene. During
prehistoric times, the centers of human population were the southwest (wet zone) and the
peripheral areas of the central hills (Figure 2). In the transitional phase between prehistory
and protohistory, human settlements shifted towards the dry zone, mainly to the north-central
parts of the island. Population density increased between 250 BC and 1235 AD, the
culmination period of Singhalese culture. During the Late Historical Phase, beginning in the
13th century, the political, economic and cultural centres shifted from the north-central,
eastern and southeastern parts of the island towards the lowland areas of the wet zone, the
central hills, and the extreme northern part. This process was coupled with the decline of
the high cultures in the dry rone and the beginning of the colonial phases, i.e. the
Portuguese, the Dutch, and the British periods. In particular during the British period
(1796-1948) impacts on natural forests in the southwestern parts and the hills were severe
and extensive. Since Sri Lanka's independence in 1948, human impacts have reached new
scales. High population growth triggered shrinkage and fragmentation of the remaining
natural forests. Natural forest cover has been drastically reduced during this century. In
1900 forest cover was stiIl 70%, by 1948 it had already been reduced to 50% and now it
amounts to about 20% (Preu & Erdelen 1992). This process has not equally affected the
different vegetation types. As a result, extension of areas under natural forest is presently
highly skewed towards the dry zone where semi-evergreen forest is the most widespread
forest type. Wet zone lowland and hiIl forests have been reduced to small fragments.
Figure 2. Human settlement patterns in Sri Lanka during the prehistoric phase (top left), the protohistoric and early
historic phase (top right), the Anuradhapura and Polonnaruwa period (bottom left) and at present (bottom right).
Based on archeological maps, data on irrigation systems ("tanks"), and actual population densities (data from
Somasekaram 1988).
At present the natural forest cover amounts to about 30% in the dry zone, but only 9%
in the wet zone. About two thirds of Sri Lanka's population are already living in this latter
zone. Extrapolating from present figures, population density in the wet zone will be about
860 persons per km 2 by the end of this millennium (Erdelen 1993). It remains to be seen
what percentage of the formerly widespread rain forests will be left by that time.
SRI LANKA'S PROTECTED AREA SYSTEM
Sri Lanka has the highest biodiversity per unit area of all Asian countries (Baldwin 1991).
The existing extent of protected areas amounts to almost 8000 km 2 or about 12.2 % of the
total land area (mCN 1990). An additional 0.4% of the area is proposed for reserves.
The protected area system consists of National Parks (N= 10, Ruhuna and Yala East
combined), Strict Natural Reserves (3), Nature Reserves (2), one Jungle Corridor (Nilgala),
and Sanctuaries (49). In addition, international designations have been assigned to the
Hurulu Forest Reserve (Biosphere Reserve), Bundala Sanctuary (Ramsar Wetland), and
Sinharaja (Biosphere Reserve, World Heritage Site). Their assignment to the mCN
management categories and other details are given in mCN (1990). Despite this extensive
508 W. ERDELEN
10' II'
t·
I·
,.
,.
L--.2.,okm
Figure 3. Protected area system of Sri Lanka. Simplified after Collins et al. 1991. Only wet zone forests shown.
Sanctuaries: BU = Bundala, CH = Chundikulam, MA = Madhu Road, PT = Padaviya Tank, PW = Peak
Wilderness, SA = Seruvila-Allai, TNH = Trincomalee Naval Headworks, VRR = Victoria-Randenigala-Rantambe.
Jungle Corridor: NG = Nilgala. Nature Reserves: MG = Minneriya-Giritale, TI = Tirikonamadu. National Parks:
FP = Flood Plains, GO = Gal Oya, MO = Maduru Oya, SO = Somawathiya Chaitiya, UW = Uda Walawe, WG
= Wasgomuwa, WI = Wilpattu, YA = Yala. World Heritage Sites: SH = Sinharaja.
area coverage the different vegetational zones are not equally represented. Most of the
protected areas are located in the dry zone (Figure 3). In other words, the regions with
highest biodiversity and concentration of endemics are under-represented. This also applies
with regard to the size distribution of the different categories of protected areas (Figure 4).
Only two of the protected areas are over 1000 km 2 in extent, i.e. Wilpattu and Yala National
Park. The only larger tracts relevant for rain forest conservation are Sinharaja (c. 90 km2)
and the Peak Wilderness Sanctuary (c. 224 km2).
Except for the two international biosphere reserves mentioned, 47 forest reserves have
been declared National Biosphere Reserves (IUCN 1990) within the national man and
biosphere program (details in Bharathie 1979). Most of these reserves contain forests
degraded to various extent, but would nevertheless form an important component for the
establishment of better connectivity between the elements of the protected area system. Of
particular importance in this regard would be national MAB plots located in the wet zone.
Integration of these areas into the existing protected area system, however, would require
considerable strengthening of links between the two responsible bodies, the Department of
Wildlife Conservation and the Forest Department of Sri Lanka.
60
40
20
o
, 200 200-400 400-600 600-800 800-1000 , 1000
Size Class
Figure 4. Size distribution of protected areas in Sri Lanka (N = 64). Size classes in kIn2 •
Based on data in IUeN (1990).
CONCLUSIONS
Although our knowledge of the flora and fauna of Sri Lanka is relatively good compared to
that of other Asian countries, there is still a clear need for further research on species
composition and species distribution, in particular of understorey plants and invertebrate
animals. As already pointed out (Erdelen 1993), studies on population viability and
vulnerability of target species for conservation have not yet been carried out. Furthermore,
some wet zone species have not been recorded since the last century and may have become
extinct by now. One major ecosystem of this zone, the swamp forests, formerly distributed
in areas now under rice cultivation, has been reduced to a single locality.
Data on actual forest cover, even as used in the most recent publications, date back to
1986 or even earlier (e.g. Somasekaram 1988). Present trends in human population growth
underscore the need for more updated information on natural forest cover in the island. This
could easily be accomplished by means of interpretation of satellite imageries.
In Sri Lanka, conservation activities in the future should focus on effectively protecting
remaining forests in the more humid parts of the island which harbour not only highest
species diversities but also the highest numbers of endemics. To accomplish this, updated
information on actual forest cover and an ecological monitoring system are urgently required.
The need for new organizational frameworks for the sustainable use and conservation of
tropical ecosystems has repeatedly been pointed out (e.g. Erdelen & Ishwaran 1991). First
steps towards developing cooperative programs and launching new initiatives to improve the
environmental situation in Sri Lanka have already been undertaken (overview in Erdelen &
Preu 1990). This has been but a starting point, particularly with regard to the increasing
pressure on the remaining natural landscapes in the island.
ACKNOWLEDGEMENTS
I am grateful to the Deutsche Forschungsgemeinschaft (German Research Foundation) for

sponsoring my participation in the rain forest conference in Brunei Darussalam. I thank the
organizers for inviting me to contribute. I appreciate the comments of Ralf Peveling and an
anonymous reviewer on an earlier draft of the manuscript.
510 W. ERDELEN
LITERATURE CITED
ADAMS, D. 1929. The geology of Ceylon. Caruulian Journal of Research 1:425-511.

ASHTON, P.S. 1980. Dipterocarpaceae. pp. 364-423 in Dassanayake, D.S. & Fosberg,F.R. (eds). Handbook to
the Flora of Ceylon I. Oxford & ffiH Publishing Company, New Delhi.
ASHTON, P.S. & GUNATILLEKE, C.V.S. 1987. New light on the plant geography ofCeylon.L Historical plant
geography. Journal of Biogeography 14:249-285.
BALASUBRAMANIAM, S. 1985. Tree flora of Sri Lanka. pp. 58-67 in Jayatilleke, A. (ed.). Proceedings of the
International Conference on Timber Technology, Moratuwa, Sri Lanka.
BALASUBRAMANIAM, S., RATNAYEKE, S. & WHITE, R., in press. The Montane Forests of the Horton
Plains Nature Reserve. In Erdelen, W., Preu, C., Ishwaran, N. & Madduma Bandara, C.M. (eds). Ecology and
Landscape Management in Sri Lanka. Margraf Publishers, Weikersheim.
BALDWIN, M.F. (ed.) 1991. Natural Resources of Sri Lanka. Keells Business Systems Ltd., Colombo.
BHARATHIE, SRI K.P. 1979. Man and biosphere reserves in Sri Lanka. The Sri Lanka Forester 14:37-40.
COLLINS, N.M., SAYER, J. & WHITMORE, T. (eds) 1991. The conservation atlas of tropical forests. Asia and
the Pacific. Macmillan Press, London.
DAYANANDAN, S., ATTYGALLA, D.N.C., ABEYGUNASEKERA, A.W.W., GUNATILLEKE, LA.U.N. &
GUNATILLEKE, C. V .S. 1990. Phenology and Floral Morphology in Relation to Pollination of Some Sri Lankan
Dipterocarps. pp. 103-133 in Bawa. K.S. & Hadley, M. (eds). Reproductive Biology of Tropical Forest Plants.
UNESCO, Paris.
DEZOYSA, N.D., GUNATILLEKE, C.V.S. & GUNATILLEKE, LA. U.N. 1988. Diversity of understorey
vegetation in the Sinharaja rainforest. The Sri Lanka Forester 18: 121-130.
DEZOYSA, N.D., GUNATILLEKE, C.V.S. & GUNATILLEKE, LA. U.N. 1991. Comparative Phytosociology
of Natural and Modified Rain Forest Sites in Sinharaja MAB Reserve in Sri Lanka. pp. 223-233 in
Gomez-Pompa, A., Whitmore, T.C. & Hadley, M. (eds). Rain Forest Regeneration and Management.
UNESCO, Paris.
DOMROS, M. 1974. The Agroclimate of Ceylon. A Contribution Towards the Ecology of Tropical Crops.
Geoecological Research 2. Steiner Verlag, Wiesbaden.
DOMROS, M. 1976. Sri Lanka. Die Tropeninsel Ceylon. Wiss. Liinderkunden 12. Wiss. Buchgesellschaft,
Darmstadt.
ERDELEN, W. 1989. Aspects of the Biogeography of Sri Lanka. pp. 73-100 in Schweinfurth, U. (ed.).
Forschungen auf Ceylon III. Steiner Verlag, Wiesbaden.
ERDELEN, W. 1993. Conservation of Tropical Forest in Sri Lanka - Problems and Issues. pp. 123-144 in
Balakrishnan, M., Bie, S.W. & Borgstrom, R. (eds). Tropical Ecosystems. A Text Book of Tropical Ecology
and Conservation. Norwegian Centre for International Agricultural Development, Norway.
ERDELEN, W., in press. The Relevance of Scale and Hierarchy Concepts for Landscape Management - A Sri
Lankan Perspective. In Erdelen, W., Preu, C., Ishwaran, N. & Madduma Bandara, C.W. (eds). Ecology and
Landscape Management in Sri Lanka, Margraf Publishers, Weikersheim.
ERDELEN, W. & ISHWARAN, N. 1991. Information Inputs to Regulate Use and Conservation of Tropical
Ecosystems - The Need for New Organizational Frameworks. pp. 195-202 in Erdelen, W., Ishwaran, N. &
Miiller, P. (eds). Tropical Ecosystems - Systems Characteristics, Utilization Patterns, and Conservation Issues.
Margraf Publishers, Weikersheim.
ERDELEN, W. & PREU, C. 1990. Background and Outcome of an International and Interdisciplinary Symposium
on "Ecology and Landscape Management in Sri Lanka". GeoJournaI22:215-217.
FLENLEY, J.R. 1979. The equatorial rainforest: a geological history. Butterworths, London.
GAUSSEN, H., LEGRIS, P., VIART, M. & LABROUE, L. 1968. Explanatory notes on the vegetation map of
Ceylon. Government Press, Colombo.
GOUDIE, A. 1985. EnVironmental Change. Clarendon Press, Oxford.
GRELLER, A.M. & BALASUBRAMANIAM, S., in press. Physiognomic, Floristic, and Bioclimatological
Characterization of the Major Forest Types of Sri Lanka. In Erdelen, W., Preu, C., Ishwaran, N. & Madduma
Bandara, C.M. (eds). Ecology and Landscape Management in Sri Lanka. Margraf Publishers, Weikersheim.
GUNATILLEKE, C. V .S. & ASHTON, P.S. 1987. New light on the plant geography of Ceylon II. The ecological
biogeography of the lowland endemic tree flora. Journal of Biogeography 14:295-327.
GUNATILLEKE, C.V.S., GUNATILLEKE, LA. U.N. & SUMITHRAARACHCHI, B. 1987. Woody endemic
species in the wet zone lowlands of Sri Lanka and their conservation in botanic gardens. pp. 183-194 in
Bramwell, D., Hamman, 0., Heywood, V.B. & Synge, H. (eds). Botanic Gardens and the World Conservation
Strategy. Academic Press, London.
GUNATILLEKE, LA.U.N. & GUNATILLEKE, C.V.S. 1984. Distribution of endemics in the tree flora of a
lowland hill forest in Sri Lanka. Biological Conservation 28:275-285.
GUNATILLEKE, l.A. U.N. & GUNA TILLEKE. C. V .S. 1990. Distribution of floristic richness and its conservation
in Sri Lanka. Conservation Biology 4:21-31.
ISHWARAN, N. & ERDELEN, W. 1990. Conserving Sinharaja - An Experiment in Sustainable Development in

Sri Lanka. Ambio 19:237-244.
IUCN (INTERNATIONAL CONSERVATION UNION) 1990. IUCN Directory of South Asian Protected Areas.
Gland.
JAYASURIYA, A.H.M. 1984. Flora of Ritigala Natural Reserve. The Sri Lanka Forester 16:61-156.
JAYASURIYA, A.H.M., GRELLER, A.M., BALASUBRAMANIAM, S., GUNATlLLEKE, I.A.U.N.,
GUNATlLLEKE, C.V.S. & DASSANAYAKE, M.D., in press. Phytosociological Studies of Mid-Elevational
(Lower Montane) Evergreen Forests in Sri Lanka. In Erdelen, W., Preu, C., Ishwaran, N. & Madduma
Bandara, C.M. (eds). Ecology and Landscape Management in Sri Lanka. Margraf Publishers, Weikersheim.
LEGRIS, P. & VIART, M. 1961. Bioclirnates of South India and Ceylon. Institut Francais de PondicMry.
Traveaux de la Section Scientifique et Technique 3:165-179.
MERRITT, V.G. & RANATUNGA, M.S. 1959. Aerial photographic survey of Sinharaja forest. The Ceylon
Forester (New Series) IV: 103-156.
PASCAL, J-P. 1984. Les forets denses humides sempervirentes des Ghats Occidentaux de l'Inde. Ecologie,
structure, f/oristique, succession. Institut Francais de PondicMry. Traveaux de la Section Scientifique et
Technique 20. Pondichery.
PEMADASA. M.A. 1984. Grasslands. pp. 99-131 in Fernando, C.H. (ed.). Ecology and biogeography in Sri
Lanka. Martinus Nijhof, The Hague.
PICKETT, S.T.A. & WHITE, P.S. (eds) 1985. The Ecology of Natural Disturbance and Patch Dynamics.
Academic Press, New York.
PREU, C. & ERDELEN, W. 1992. Geoecological Consequences of Human Impacts on Forests in Sri Lanka. pp.
147-164 in Goldammer, J.G. (ed.). Tropical Forests in Transition: Ecology of Natural and Anthropogenic
Disturbance Processes. Birkhiiuser Verlag, Basel.
REMMERT, H. 1991. The Mosaic-Cycle Concept of Ecosystems - An Overview. pp. 1-21 in Remmmert, H.
(ed.). Tropical Forests in Transition: Ecology of Natural and Anthropogenic Disturbance Processes. Ecological
Studies, Volume 85. Springer-Verlag, Berlin.
SOMASEKARAM, T. (ed.). 1988. The National Atlas of Sri Lanka. Survey Dept., Colombo.
WAIDE, R.B. & LUGO, A.E. 1992. A Research Perspective on Disturbance and Recovery of a Tropical Montane
Forest. Pp. 173-190 in Goldammer, J.G. (ed.). Tropical Forests in Transition: Ecology of Natural and
Anthropogenic Disturbance Processes. Birkhiiuser Verlag, Basel.
WCMC (WORLD CONSERVATION MONITORING CENTRE) 1992. Global Biodiversity: Status of the Earth's
living resources. Chapman & Hall, London.
WERNER, W.L. 1982. The Upper Montane Rain Forests of Sri Lanka. The Sri Lanka Forester 15:119-135.
WERNER, W.L. 1984. Die Hiihen- und Nebelwiilder aUf der Insel Ceylon (Sri Lanka). Trop. subtrop.
Pflanzenwelt 46. Steiner Verlag, Wiesbaden.
A Conflict over Natural Resources between Highland and Lowland
Populations in Thailand
WARANOOT TUNGITTIPLAKORN 1 and BENGT-ERIK BENGTSSON2
Asian Institute of Technology, Natural Resources Programme, GPO Box 2754, Bangkok
10501, Thailand
ABSTRACT. The ongoing conflict between the highlanders and the lowlanders at Mae Soi, Jomthong, Chiang
Mai, was explored to identify its causes. The study shows that perceptions about the conflict by both groups
are at variance. Lowlanders perceive the conflict to be environmental, i.e. due to deforestation, water
reduction, and contamination. Highlanders attribute social, political and economic causes to the conflict. The
study also finds a high dependency on the forest for fuel wood, fodder and building materials, and a low
dependency on the forest for food products. It can be concluded that environmental effects of the agriculture
practices in the hill areas have reduced both the forested area and its quality. Water reduction has occurred
but it is not possible to single out one particular group as responsible for this change. Cash crops introduced
among the hill tribes to replace opium poppy cultivation need pesticides (both organochlorines and
organophosphates are used), but pesticide concentrations found in downstream river water in this study are
estimated to be within safe limits for drinking purposes. The conflict is both environmental and socio-
economic. Resolving the conflict thus requires a broader perspective than resolving environmental problems
alone.
KEY WORDS: environmental damage, deforestation, water reduction, pesticide contamination, socio-economic
conflict.
INTRODUCTION
Increasing conflict between different interest groups over natural resources has become
apparent in Thailand in recent years. This study highlights a prevailing natural resource
management problem, the environmental conflict between highlanders and lowlanders. By
using one of the most widely publicized cases, the study examines the nature of the conflict
from the perspective of the local people. The aim of this study is to have a better
understanding of the social and environmental aspects of conflicts over resources in northern
Thailand.
Of over seventeen million hectares of northern Thailand, more than two-thirds is upland
(> 300 m < 1000 m above sea level [asl]) or highland (> 1000 m as\), which contain most
of the country's remaining forests. The highland forests are the origins of tributaries of
major rivers. While the majority of the population is concentrated in valleys, the uplands are
home for over five hundred thousand ethnic minorities, widely known as Chao Khao or hill
tribes. These geographical and demographical characteristics create a situation susceptible
to conflict over natural resources, especially control over headwaters.
The situation in Mae Soi reflects such problems and has developed into a severe crisis.
Some data in this article are from "Highland-Lowland Conflict Over Natural Resources: A Case of Mae Soi. Chiang Mai: Thailand". in
Society and Natural Resources. Vol. 8, pp. 279-288. Taylor & Francis. 1995.
lWaranoot Tungittiplakorn is currently a PhD candidate at the University of Victoria, Geography Department,
Victoria, Canada.
2Bengt-Erik Bengtsson is an Associate Professor at the Institute of Applied Environmental Research, Stockholm
University.
513
514 WARANOOT TUNGITTIPLAKORN AND BENGT-ERIK BENGTSSON
The low landers demanded that Hmong hill tribes move out of the headwaters forest. The
Hmong's agriculture practices allegedly destroy the forest which in turn reduces stream flow
that feeds agriculture downstream. Supported by a conservationist monk and an affiliated
foundation, low landers marched up the mountain and erected a 12-kilometre fence, so
limiting access to the forest and arable land. Protests were staged and letters written.
Finally, the cabinet in 1989 announced that the Hmong of Pakluay village would be
relocated. Hmong villagers, however, argued that they had cooperated with the government
in crop replacement and reforestation programmes and were sceptical about the resettlement
scheme. They appealed to the Royal Northern Project, the Department of Public Welfare
and the Military. These are the agencies that have most contact with the hill tribes. The
conflict widened the gaps not only between highlanders and low landers but also between the
agencies involved. Unable to reach a consensus, government agencies have stalled the
resettlement programme.
METHODS
One hundred and fifty people in the lowlands and highlands were interviewed. We selected
the respondents by simple random sampling, using the voter list as a sampling frame. The
villagers were from four villages identified by the district officer as most actively involved
in the conflict. Interviews were done both in northern dialect and in Hmong language.
Some Hmong women refused to be interviewed. In the Hmong culture, contacts with
outsiders are more the realm of the men. In such cases we selected a replacement. We also
talked to staff of various agencies including the Tribal Research Institute, the Thai-Norwegian
Highland Development Project, the Forestry Department, the United Nations Program for
Drug Abuse Control, Agricultural Extension Office, Health Office, District Office, and the
Dharmmanart Foundation for Conservation and Rural Development. These interviews were
informal, using a few structured questions as guidelines. We compared the result of the
interviews between highlanders and lowlanders regarding dependency on forest and water and
their perceptions of the conflict .
The Health Officer from Jomthong hospital and the District Agricultural Extension Officer
have collected water from the Mae Soi and other streams routinely since 1990. The samples
were generally sent to the laboratory at the Ministry of Health. On 18 November 1992,
however, an additional set of water samples were taken from three sites around the Hmong
village for analysis at the Environmental Engineering laboratory of the Asian Institute of
Technology (AIT), Bangkok. Water from Mae Tia stream was collected one kilometre
downstream from the Hmong village and two and three kilometres upstream. Two other
samples were collected from the Inthanon area, where intensive cultivation of similar crops
occurs.
Table 1. Distribution of Respondents, Mae Soi, Chiang Mai, Thailand, 1992.
Location Village Name Total No. of Nwnber of Households

Households in Respondents represented (%)
village
Lowland Nong Khan 166 34 20.48
WangNamyad 128 46 35.94
Huay Ha 90 17 18.89
Highland Pakluay 82 53 64.63
Total 466 150 32.19
Natural resource conflicts in Thailand 515
The water samples that were sent to AIT were analyzed according to the following
procedure. One litre from each water sample was extracted three times with Hx: OCM, 1: 1.
The combined organic phase from each sample was evaporated, dried with sodium sulphate
and the extracts concentrated to 1 m!. The extracts were injected into a Hewlett Packard
5890 gas chromatograph (GC) with a mass selective detector and analyzed for pesticides and
PCBs. The column was 50 m x 0.12 mm Chrompac, CP-Sil 8CB with film thickness 0.25
p.m. The carrier gas was helium and the flow 0.5 ml/min. The oven was held at 60°C for
two minutes and raised by 200 min- 1 to 220°C. The temperature was then raised by 3° min- 1
to 280°C. The injector and detector temperatures were 200°C and 300°C respectively. Scan
mode was used.
The Study Area

Mae Soi is a subdistrict located south of Jomthong, approximately 80 km south-west of
Chiang Mai city. The four villages under study were Nong Khan, Wang Namyad, Huay Ha
and Pakluay. The first three are lowland villages found at the elevation 270-300 mas!.
Pakluay is a highland village located at 1410 m asl (Figure 1).
The Pakluay cultivation area is adjacent to the headwaters of three streams: Mae Soi, Mae
Tim and Mae Tia. The Mae Soi stream is central to the dispute between the low landers and
the highlanders. Orographic precipitation allows Pakluay to depend largely on rain for
agriculture; however farmers who own land adjacent to the streams transport water through
pipes to their fields. The quantity of water in these streams was "little" and in the dry
seasons many of them dried up completely (Anon 1987).
o 2 KM.
SYMBOL
HIGHWAY
DIRT ROAD
STREAM
LOWLAND 0-300 mASL
UPLAND 301-1000 mASL
HIGHLAND ABOVE 1000 mASL
Figure 1. The study area, Tam bon Mae Soi, Chiang Mai Province, Northern Thailand.
Traditionally, the Hmong are pioneer swiddeners who regularly migrate when productivity
of their lands decline. Opium poppy and swidden rice were their two main crops. The
village was settled approximately 25 years ago. The current land shortage does not allow
them to move to a new area. They cultivate secondary forest, rotate their fields and even
cultivate the grassland. Pakluay used to raise a few hundred livestock. However, because
of the theft of the cattle and horses during the early 1980's, they have abandoned this
activity. Pakluay shifted from opium and rice to vegetable growing in the mid-1980's.
The lowland rice farmers depend on the Mae Soi stream for their paddy. The villagers
manage irrigation communally. The majority of the people, however, engage in upland rain-
fed cultivation of such cash crops as garlic, red onion, soil bean, peanut and chili. In the
past, tobacco was an important crop, as evidenced by the presence of tobacco curing plants
still in the area. Wood milling and charcoal making were also important means of making
a living in Wang Namyad and Huay Ha, but are now greatly reduced. Approximately 300
cattle were raised in 1991, with the upland forest providing grazing grounds. Fruit trees,
especially longans, have become popular in recent years. Most lowlanders are landless or
near landless. There have been major land transactions in the area, and a large part of the
land has been transferred to outsiders.
Natural forest in the lowlands, like any other place in Thailand, has long been cleared.
Around the late 1930's, teak was being extracted from the valley by the Bombay-Burma
Company. Following the path of the British company, the remaining hardwood species were
logged by Thai timber companies. Around the early 1970's, defective and poor quality trees
were removed to be used as fuel wood to cure tobacco leaves. After the tobacco company
concession ended, villagers took over the forest and harvested young trees for charcoal
burning. By 1977 the forest in Mae Soi was cleared (Maxwell 1992). What remains is a
remnant of the original deciduous forest along the Mae Soi stream. This area was considered
sacred by the villagers in the past, though many people have now clearly abandoned their
belief in the spirits of the grove. This is reflected in the extent of utilization of the area.
The upland at the base of the mountain is dominated by bamboo which is a common pioneer
in the process of forest regeneration. Further up, at 700-1000 m asl, the deciduous forest
prevails with evergreen species along the water course (Maxwell 1992).
Pine and evergreen hardwood dominate the highland. This area has been less disturbed
than the lowland and original species still exist. However, a large area of the highland has
been converted to agricultural land by highland villagers. The old fields left fallow are
covered with fire resistant weeds such as Imperata cylindrica L. (Poaceae) and Eupatorium
adenophorum Spreg. (Compositae) (Maxwell 1992).
Changes in the Highland

In 1981 the Highland Agricultural and Marketing Programme (HAMP) operated jointly
by the Thai government and the United Nations began their work in the Pakluay Hmong
village. The aim was to reduce opium cultivation by substituting it with cash crops. This
project was not very successful because the new crops were not viable both from the
productive and economic aspects. After the termination of HAMP, the Thai-Norwegian
Highland Development Project (TN-HDP) took over the task.
Around 1985 the Thai government stepped up measures against opium growers (raiding
fields and arresting farmers). This had a substantial impact on the farmers and led them to
gradually change crops. The Hmong has an extensive network through clan-based relations.
According to some villagers, their agricultural innovations usually come from those relatives
who live in villages under the administration of the Royal Project in the Inthanon area. By
the mid 1980's, the Hmong in Pakluay found that cabbage grew well in their environment.
Most important of all, however, was the presence of marketing outlets for cabbage in
Jomthong. TN-HDP followed up the innovation by providing seeds, pesticides and fertilizers
to farmers. From 1985 onward, Pakluay has been transformed into a cabbage growing
village.
Cabbage has brought significant income into the village. The 1986 survey conducted by
the Department of Land Development estimated the profit to be 1,700-12,000 baht per rai.
It also reported the per capita income of Pakluay villagers in 1986/7 to be -27 to 3,263 baht
(Tansiri & Chondamrongkul 1987). The figure for northern rural villagers was 826 baht in
1988 (Anon 1993). Nevertheless, cabbage requires a tremendous amount of input. Most
Hmong farmers today claim that they are indebted to lowland merchants because they bought
inputs for cabbages on credit.
RESULTS
Dependency on Forest
Both highlanders and lowlanders depend heavily on the forest for firewood, building
materials and fodder (see Table 2). The highlanders use vegetation from the forest to
supplement maize for their pigs. The low landers also gather firewood from the forest and
take lumber from the local forest. However, lowland villages are closer to alternative
sources of building materials and fuel, thus depending on the forest for these products to a
lesser degree than the hill tribes.
Food obtained from the forest is only supplementary (Table 3). The Hmong hunt more
frequently. They also gather medicinal plants more than the lowlanders. Mae Soi villagers
in the lowlands collect bamboo shoots and mushrooms, which they sometimes sell. They
also gather vegetables and ant's eggs occasionally. It was not possible to obtain data about
the biomass of these products in this study. The significance of forest as a source of food
is small. Most people from both groups buy food from local vendors. Respondents from
both areas stated that forest resources have been greatly reduced in recent years.
Dependency on Water
Figure 2 shows the periods when water is needed in the highlands and the lowlands. The
most critical period of water shortage is from April to the beginning of May. The rain has
not yet arrived and water is needed both in the highlands and the lowlands. The highlands
need water in May mainly for the vegetable seedlings. The villagers draw water from the
stream and transport it through pipes to their seedlings beds.
The lowlands need water for the second cropping of rice in April and May. This is a
recent phenomenon facilitated by 'electrical irrigation', a new irrigation scheme in which
water is pumped from the Ping river. This should not affect the Mae Soi stream but may
decrease the available cattle grazing lands. Cattle are also affected by the water shortage in
the Mae Soi stream.
Table 2. Percentage of respondents obtaining resources from the Mae Soi forest, Thailand 1992. n: number of
respondents; * significant at 95% confidence level using Chi-square test of association.
Resources Highlands Lowlands
n % of respondents n % of respondents
Firewood* 53 100 73 63.0
Fodder 51 43.1 45 40.0
Building materials· 52 100 54 75.9

Table 3. Forest Food Collection Frequency in 1991192, Mae Soi. Thailand. 1992. (Multiple Responses); n: number
of respondents; * significantly different at 95 % confidence level using Chi-square test of association.
Food Resources Highland (n = 51) Lowland (n = 37)
mean min max S.D. mean min max S.D.
Mushroom* 0.8 0 7 1.6 2.1 0 15 3.3
Bamboo shoot* 1.2 0 5 1.4 6.9 0 100 17.9
Animal* 5.4 0 98 19.6 0.0 0 0 0.0
Honey 0.0 0 0 0.0 0.1 0 3 0.6
Vegetables* 0.0 0 0 0.0 1.4 0 20 3.5
Ant's eggs* 0.0 0 0 0.0 1.4 0 10 2.8
Medicinal plants* 6.1 0 97 17.7 0.6 0 9 1.7
Longan has become very popular in the lowland because of the demand from abroad. As
a result, many villagers are turning their upland gardens into fruit orchards. However,
longan must receive adequate amounts of water before fruiting (May-July). The quality of
the fruit largely depends on the amount of water the trees receive. Smaller fruits cannot be
exported and fetch a very low price, thus villagers who can afford pumps take water from
the stream in great quantities. Because of the increase in longan in the areas, water shortage
is likely to escalate.
There are also some complaints about the reforestation project which withdraws water
from the stream during the critical period. The reforestation project is carried out by the
conservation NGO in cooperation with the Royal Forestry Department. This has resulted in
some villagers having a negative attitude towards this project.
As summarized by Shinawatra (1985), various studies suggest that cultivation in the
highland reduces stream discharge. Her analysis of run-off, in the dry season, of major
rivers in Jomthong reveals that the volume of water was reduced from 109.26 million cubic
metres per year during 1921-1951 to 68.5 million cubic metres during 1970-1980. However,
it is difficult to single out a particular activity carried out by a particular group of people as
being responsible for this change.
:;,~1T
::::::::::::::::::::::::::::::::::::::::::;
......p.?~ ............
ijm~lg Highland
corn
=Il~
-
Jan Feb Mar Apr May Jun Jul Aug sep Oct Nov Dec
:mmmmmmmmmmmmrrmpaddy
onion chili-onion onion
Lowland
mmmm tree seedlings for refost~in
Rainy season crops
-
iiiiiiiiii Winter season crops
Dry season crops
Figure 2. Agricultural calendar for highlanders and lowlanders of Mae Soi, northern Thailand.
When asked how long the water in the Mae Soi streams had been reduced in volume, 71
per cent of the lowlanders indicated 3-6 years. This contrasts with the length of time the
conflict in Mae Soi has existed. The earliest evidence that we gathered indicated the year
1981 as marking the beginning of the conflict. The conflict has thus lasted for over a
decade. Water shortage may not be the original cause of the conflict.
Results of Water Sample Analysis

No peaks on the gas chromatograph that corresponded to standard solutions of pesticides
(HCB, Aldrin, Dieldrin, Lindane, Mirex, op-DDE, pp-DDD, pp-DDT) or PCBs (IU-28, IU-
52, IU-101 , IU-118, IU-138, IU-153, IU-180) could be detected. The detection limit for the
contaminants listed above in Scan mode was less than 0.5 ug/L river water.
In order to detect even lower concentrations of contaminants, water extracts were also
analyzed with a GC equipped with an electron capture detector. From the 1 ml extracts
dissolved, 0.2 ml was taken out, diluted to 1 ml and injected into a Hewlett Packard 5890
Gc. The instrument settings were the same as described under Methods with the exception
that nitrogen was used as carrier gas.
The extracts were compared with standard solutions of pesticides and pentachlorobiphenyls
(PCBs) but no pesticides or PCBs could be detected. The approximate detection limit was
0.006 ug/L river water but no contaminants could be detected. This means that no trace of
pesticides and PCBs could be found at any of the sampling sites near Mae Soi or Khun Klang
Hmong villages. To fully investigate pesticide pollution of river water, the sampling should
be carried out at all stages in the agricultural cycle.
Villagers in Pakluay claimed that in 1970 they were the first Hmong to apply chemical
fertilizers to poppy fields. They also claimed they were pioneers in the use of herbicides in
the poppy fields around 1976-7 (Prangkhiaw 1984). When they shifted to cabbage cultivation
they began using pesticides.
Several water analyses have been carried out to evaluate the water quality from
agricultural fields in the highlands. The first was done by TN-HDP and Office of National
Environment Board in February 1990. The Jomthong agricultural extension office and the
district hospital also tested water samples in January 1991 and 1anuary 1992. All the results
indicated that the levels of pesticide were within safe limits. Our water analysis was carried
out in response to a request from the hospital and the agricultural extension staff. The aim
was to compare the results of the AIT laboratory and those of the Ministry of Health
laboratory, but the Ministry water samples were lost and no comparison was possible.
The results of the water analyses only provide some indication, not proof, that there are
no pesticides being released from the highland fields. The water samples were all collected
at the same season, thus not taking into account any possible seasonal variations. Secondly,
there seem to be no control points or consistency in the selection of areas being sampled.
Potato, a secondary crop to cabbage in the highlands, needs more pesticides than cabbage.
Some farmers reported having to apply the chemicals every five days. Fields that are used
continuously also need more pesticides. Winter and dry season cabbages need more
pesticides than rainy season crops.
The lowlanders used pesticides most commonly in their onion and garlic fields. They
indicated that onions needed pesticides every 3-4 days. Moderate amounts of pesticides were
used on other crops. The surveys done by the agricultural extension and the hospital in
Jomthong indicated that pesticides are used most frequently in the lowlands in December and
January. This corresponds to the onion growing period.
Deforestation
The document obtained from the district office indicated that, since they settled in the area
around 1975, the Pakluay villagers have cleared at least 320 ha. of forest. After the
lowlanders fenced off part of the forest, the Hmong abandoned half of their fields. The
Department of Land Development reported that the area belonging to Pakluay in 1987 was
161.5 ha. It reported that the cultivated area covered 139.5 ha. (see Table 4). The size of
the Pakluay land has become quite stable since 1987, although, critics of the Hmong believe
that these figures are gross underestimates (Kaye 1990).
Nature of the conflict

Table 5 shows the divergence of the highlanders' and the lowlanders' perceptions of the
causes of the conflict. The lowlanders perceived the conflict as largely environmental,
caused by the Hmong' s destruction of forest and the contamination of the river. The
highlanders attributed the conflict to more diverse causes. Most of them believed that
the conflict stemmed from the incident when a cattle thief (incidently a low lander) was killed
in the Hmong village in 1984. About one-third of the interviewees mentioned the success
of their cash crops to be the cause of the lowlanders' resentment. An equal proportion of
the highland respondents thought the conflict originated from the lowlanders' concern over
the forest. Other reasons cited were ethnic difference and the intervention by the
conservationist monk.
Table 4. Pakluay Cultivation Area (in hectares). Source: Jomthong district 1990 *Tansiri, B. and P.
Chondamrongkul1987. **Jomthong district 1991. nr = not reported; HAMP: Highland Agricultural and Marketing
Progrannne; TRI: Tribal Research Institute; TN-HDP: Thai-Norway Highland Development Project; ONCB: Office
of Narcotic Control Board; PLO: Provincial Land Office.
Year Total Poppy Potato Cabbage Fallow Reported by
1983 nr 70.72 nr nr nr HAMP
1984 nr 78.08 nr nr nr HAMP
1985 ill 47.20 29.28 ill ill TRI
1986 345.76 1.44 20.96 37.92 ill TN-HDP/ONCB
1987 161.44 0 28 59.84 nr TN-HDP/DLD*
1988 148.32 0 36.64 90.88 20.80 TN-HDP/ONCB
1989 nr 0 37.60 93.92 ill TN-HDP/ONCB
1990 160.96 0 129.60 31.36 PLO**
1991 ill 0 139.44 ill PLO**
Table 5. Perceived Causes of Conflict, Mae Soi, Chiang Mai, Thailand.1992. (Multiple Responses)
Perceived causes Highland (n = 41) Lowland (n = 81)

n % ofResp. n % of Resp
Deforestation by the Hmong 12 28.6 60 81.4
Pesticide contamination by the Hmong 2.4 12 16.0
Murder of the lowland thief 22 54.8 0 0.0
Racial issue 10 23.8 1 1.2
Economic competition 13 30.9 0 0.0
Intervention by the conservationist monk 8 19.0 5 6.2
Others 4 9.5 5 6.2
DISCUSSION
The pressure to use the resources of the highlands and lowlands is high. The most serious
threat to the highlands comes from the expansion of agricultural land. The promotion of cash
crops in the hills have undesirable effects in terms of decreasing forested land. However,
Pakluay's lands seem to have been stable since 1989, although if the Hmong continue their
cultivation of the same area, the soil productivity is likely to decrease. The Hmong are
already beginning to realise the effect of soil degradation. They spend a great proportion of
their income buying fertilizers and pesticides to maintain productivity of their land.
Pakluay villagers still use open fires for cooking and heating. This method consumes a
large amount of wood, although in some houses more efficient clay stoves have been used
to save fuel. Such innovations should be followed up by government or non-government
agencies concerned about forest destruction.
One obstacle against initiating constructive measures in the highlands is the lingering
indecision on whether to move the Hmong village. Agencies involved express reluctance to
invest any energy and resources in conservation if the village is to be relocated. However,
there is little possibility of finding a suitable area for relocation in the lowlands. The
government may resort to forced eviction, which is disruptive to the Hmong and also to the
existing population of the destined site. The situation is likely to remain a stalemate.
The conflict in Mae Soi involves both environmental and socio-economics. The conflict
evolved around the forest and water issues. The absence of a management plan over
resource use has resulted in over-exploitation both of the highlands and the lowlands. The
low landers have become environmentally conservative as a result of the campaign by an
outside NGO and a conservationist monk. The Hmong, existing outside the Nation's and
Buddhism's realm, are largely excluded from this campaign. The work of international
projects in the highlands has caused some resentment by the low landers, who think they are
being bypassed. Currently, the Hmong are better off economically while the Mae Soi people
are mostly poor and landless. This situation is partly the result of environmental effects as
a result of forest destruction in the highlands. Equally, if not more important, the economic
substandard in the lowlands originates from larger political and economic pressures, such as
land speculation, that has occurred throughout the nation in the last few decades.
CONCLUSION
It can be concluded that environmental effects of agriculture in the highland have reduced
the forested area and its quality. The remnant forest is probably not extensive enough to
remain viable, and major conservation measures are needed to protect these remaining tracts
of forest.
The reduction in water flow has been confirmed from various sources. However, it is not
possible in the present study to single out deforestation as being responsible for this change.
Cash crops introduced to the highlanders need pesticides. The pesticide concentrations
according to other studies are too low to be considered hazardous for drinking purposes and
too low to be detected by the sensitive analytical methods used in the present study. A more
rigorous study is needed to address this question.
The study reveals that highlander and lowlander perceptions of the conflict are at variance.
Lowlanders perceive the causes of the conflict as mainly environmental. The Hmong have
allegedly caused forest destruction, water reduction and contamination of the water.
Highlanders suggest more diverse factors, such as economic factors and ethnic pressures,
compounded by local incidents.
Differences in perceptions amongst the lowlanders and highlanders concerning the causes,
seriousness and possible resolution of the conflict are serious obstacles to resolving the
present situation. The apparent environmental conflict is linked to social, economic and
political conflict. Simply addressing environmental issues may not solve the problem.
ACKNOWLEDGEMENTS
The project was financially supported by a scholarship from the Rockefeller Foundation and
research grants from the Canadian International Development Agency (CIDA) and the
Swedish International Development Authority (SIDA). We would like to thank the staff of
the various agencies we interviewed, and especially the villagers of Pakluay and Mae Soi,
whose kind cooperation made this study possible. Finally, Ms Charlotta Lindberg, AIT, is
thanked for her analysis of the water samples.
LITERATURE CITED
Anonymous 1993. Basic Information Regarding Women and Society in Northern Thailand. Faculty of Education,
Chiang Mai, Thailand.
Anonymous 1987. Report of Soil Survey and Land use Planning for the Thai-Norwegian Highland Development
Project, Ban Pak/uay (Mahin Luang 1,2,3) and Doi Lium Plot, Huay Manao Area, Amphoe Jomthong, Changwat
Chiang Mai. Department of Land Development, Bangkok (in Thai).
Kaye, L. 1990. Of Cabbage and Cultures. Far Eastern Economic Review, 13 December 1990.
Maxwell, J.F. 1992. Vegetation of the Mae Soi Conservation Area, Jomthong District, Chiang Mai Province.
Department of Biology, Faculty of Science, Chiang Mai University (mimeograph.)
Prangkhiaw, J. 1984. Summary of Socio-Economic Survey, Pakluay, Mae Soi, Jomthong, Chiang Mai (in Thai).
Shinawatra, B. 1985. Highland-Lowland Inter-relationships in Northern Thailand: A Study of Production,
Distribution and Consumption. PhD dissertation, Michigan State University, Ann Arbor.
Tansiri, B. & Chondamrongkul, P. 1987. Report of Soil Survey and Landuse Planning for the Thai-Norwegian
Highland Development Project, Ban Pa K/uay (Mahin Luang 1,2,3) and Doi Lium Plot, Huay Manao Area,
Amphoe Jomthong, Changwat Chiangmai. Department of Land Development, Bangkok (in Thai).
Rattan inventory: determining plot shape and size
M. C. STOCKDALE and H. L. WRIGHT
Oxford Forestry Institute, Department of Plant Sciences, University of Oxford, Oxford OXI
3RB, United Kingdom
ASTRACT. Increased recognition of the importance of non-timber forest products has led to a need to
develop inventory methods for these species. One of the most valuable non-timber forest products are
rattans, the stems of climbing palms belonging to the sub-family Calamoideae and the basis of a furniture
industry worth an estimated U.S.$ 6.5 billion per annum in international trade. This paper discusses the
methodology for determining plot shape and size for the inventory of rattan. This methodology involves
selecting a desired level of precision (e.g. a sampling error of 20%) and finding the plot shape and size
which minimises the cost, in terms of time, of achieving it. Rectangular plots, oriented parallel to the
direction of the slope, are more cost efficient than square plots; however, there are limits to
rectangularity, as plots which are very long and narrow are prone to errors in boundary identification and
area estimation. Thus, a minimum plot width of 5 m is recommended. The most cost efficient plot size
ranges from 0.0025 ha to 0.025 ha, and is positively correlated with the desired sampling error, the total
area under inventory, and the plot rectangularity. It also varies with the parameter(s) being assessed. The
sample size is then determined by the plot size.
KEY WORDS: Palmae, Calamoideae, plot sampling, cost efficiency, Brunei Darussalam.
INTRODUCTION
Inventory methods for tree species have been well developed and documented (e.g. F.A.O.
1973, Husch et al. 1972, Loetsch et al. 1973, Philip 1983). Increased recognition of the
importance of non-timber forest products has led to a need to develop inventory methods for
these species (Williams 1991). Rattans are one of the most valuable non-timber forest
products; their durable stems are the basis of a furniture industry worth an estimated U.S.$
6.5 billion per annum in international trade and are responsible for providing approximately
half a million jobs (Caldecott 1988, Manokaran 1990).
Rattans, belonging to the sub-family Calamoideae (family Palmae), comprise
approximately 600 species from 13 genera, and range from the forests of southern China
through to Australia and Fiji, with largest concentrations in Peninsular Malaysia, Borneo and
New Guinea, as well as perhumid equatorial Africa (Uhl & Dransfield 1987). These climbing
palms vary in growth form; some species occur as solitary stems whilst others produce
suckers, resulting in clumps of stems.
Rattan inventory, as in other resource inventories, requires decisions to be made on the
sampling pattern or design, the nature of the sampling units used, the population and
parameters to be included, and the most appropriate methods of measurement (Philip 1983).
This study focuses upon the nature of the sampling unit, which may take the form of a point
or a plot. Decisions about sampling units very much depend upon the nature of the
population under study.
Point sampling involves the measurement of distances from a point to individual plants,
from which the number of plants per unit area may be calculated. This method has the
advantage of being quicker than sampling with plots, but if the assumption of random
distribution is violated, the density of a population will be underestimated (Grieg-Smith
1983). Point sampling was not tested in this study as rattans often grow in dense thickets;
523
524 M.C.STOCKDALE AND H.L.WRIGHT
for example, Tandug (1978) found the frequency distribution of rattan in a plot in the
Philippines to be non-random, and best fitted by the negative binomial function. A study
(Stockdale, in prep.) of the frequency distribution of rattan in Brunei and Indonesia will
further investigate this issue.
Plot sampling may use plots of a variety of shapes and sizes. Inventories - where the aim
is to enumerate as precisely as possible the nature and extent of the resource - may require
different plot shapes and sizes from those required for ecological studies or growth modelling
(Vanclay, in press). In an inventory, it is desirable that plots maximize within-plot variation
in order to minimise between-plot variation. With plots for ecological studies or growth
modelling, however, within-plot homogeneity may be preferable. In addition to precision,
cost is an important consideration in the selection of plot size and shape as, at the level of
large-scale inventories, an inefficient sampling method can waste much time and money.
Larger plots incorporate greater within-plot variation than smaller plots, but they tend to
involve greater cost for the same precision (Philip 1983). As far as plot shape is concerned,
a rectangular plot is better than a square or circular plot at representing the variance of a
population, if the population has a linear trend and the long axis of the plot is oriented
parallel to this trend (Freese 1961). However, the disadvantage of rectangular plots,
compared to circles and squares, is that their lower perimeter to area ratios make them liable
to serious bias from errors in identification of borderline plants and area estimation (Philip
1983). Circular plots are not recommended for inventories in mixed tropical forests where
the undergrowth is dense, making it difficult to walk in an arc around a central point (Alder
& Synnott 1992). Thus, only rectangles and squares, of a range of sizes, were compared in
this study.
Plot sizes in rattan inventories have ranged from 5 m by 5 m (Tandug 1978) to 250 m by
250 m (Sharma & Bhatt 1982) or 10 m by 400 m (Siswanto & Soemarna 1988). Plot shapes
have been circular (Tomboc 1988), square (Sharma & Bhatt 1982, Siswanto & Soemarna
1988, Tandug & Lasmarias 1984) and rectangular (Siswanto & Soemarna 1988, Tandug &
Lasmarias 1984, Zieck 1967). Only Tandug's (1978) study, however, has tested the relative
efficiency of different plot shapes and sizes for rattan inventory; her results will be discussed
subsequently.
The optimal plot shape and size for an inventory can be determined by two related
approaches (Arvanitis & O'Regan 1967, Taafe 1979, Tandug 1978, Zeide 1980). One
approach is to set a desired level of cost (in this study represented by the time taken to
conduct the inventory) and choose the plot shape and size which maximises precision. The
second is to set a desired level of precision (in this study represented by the sampling error)
and choose the plot type which minimises cost. This study used the latter approach.
METHODS
The work was conducted at a 1.5 ha study plot near the Kuala Belalong Field Studies Centre
in the Temburong District of Brunei, as part of the University of Brunei Darussalam/Royal
Geographical Society's Brunei Rainforest Project. The long axis of the 300 m by 50 m plot
extended from just above the Belalong river floodplain (60 m a.s.l) to two-thirds of the
distance up the east ridge (210 m a.s.l.). The plot consisted of five parallel lines spaced 10
m apart, which were marked with tagged poles at 10 m intervals. Thus the study plot was
subdivided into one hundred and fifty subplots of dimensions 10 m by 10 m, as shown in
Figure 1. The terrain in the plot encompasses small ridges, valleys and plateaus, with slopes
ranging from 0° to 60°. The vegetation of this area is lowland mixed dipterocarp forest
(U.B.D.! R.G.S. 1991).
Rattan inventory 525
210 m a.s.1. 210 m a.s.l.
0) 0)
a. a.
.Q 0
-UJ
0
C
0
300m -
"iii
0
C
300m
:p
0
u
0)
U
0)
...... ......
'6 '6
60 m a.s.1. 60 m a.s.1.
SOm 50m
Figure 1. Grid map of 1.5 ha study plot. Figure 2. Grid map of 1.5 ha study plot, overlaid with 30
plots of rectangularity = 5 and size = 0.05 ha.
A 100% enumeration of all rattan species in the study plot was conducted by a team of
three people. One person recorded the data and walked up the marked line. The other two
people, on either side of the line, used 2.5 m poles to determine whether the central points
of rattan clumps were within 5 m of the line. Only rattan clumps mature enough to have
developed at least one stem were enumerated. Data recorded for each 10 m x 10 m subplot
were, for each species, the number of clumps and the number of stems, as well as time taken
to enumerate each subplot. On separate occasions, time taken to demarcate the tagged poles
and to walk between all poles in both directions was also recorded. Data on species are
vouchered by herbarium material housed in the Forestry Department of Brunei Darussalam
and the Royal Botanic Gardens, Kew, United Kingdom.
RESULTS
Enumeration
The results of a 100% enumeration of the 1.5 ha plot are listed in Table 1. There were
twenty-one species of rattan from four genera. All species together comprised 343 clumps
(228 clumps/ha) and 794 stems (529 stems/ha).
Determining plot shape and size: the coefficient of variation

The study plot grid was overlaid with all possible combinations of plot shapes and sizes
which could completely cover the plot. An example is shown in Figure 2, using a plot shape
or rectangularity (measured as the ratio of the long axis to the short axis) of 5 and a plot size
of 0.05 ha. The plots were always oriented so that their long axes were parallel to the
direction of the slope, thus encompassing the maximum amount of topographical variation.
Plot shapes ranged from rectangularities of 1 to 30, and plot sizes ranged from 0.01 ha to
0.75 ha.
For each plot shape/size combination, the mean number of clumps/ha and stems/ha were
calculated, together with their co-efficients of variation. The coefficient of variation is a
measure of variation and is calculated as the ratio of the standard deviation to the mean,
normally expressed as a percentage. As the total area covered was the same for each plot
Table 1. Results of 100% enumeration of rattans (family Palmae: sub-family Calamoideae) in a 1.5 ha plot in
Temburong District, Brunei Darussalam.
Species No. Clumps No. Stems

Calamus amplijugus J.Dransf. 6 16
Calamus comptus J .Dransf. 2 2
Calamus conirostris Becc. 18 30
Calamus f1abellatus Becc. 92 210
Calamus laevigatus Mart. var. laevigatus 2 2
Calamus muricatus Becc. 9 9
Calamus pogonacanthus Becc. 23 124
Calamus praetermissus J .Dransf. 6 24
Calamus sordidus J.Dransf. 18 88
Calamus zonatus Becc. 2 7
Ceratobolus concolor BI. 14 43
Daemonorops didymophylla Becc. 1
Daemonorops korthalsii BI. 7 7
Daemonorops longipes (Griff.) Mart. 4 4
Daemonorops oxycarpa Becc. 13 13
Daemonorops periacantha Miq. 105 105
Korthalsia debilis BI. 4 20
Korthalsia echinometra Becc. 3 13
Korthalsia hispida Becc. 8 48
Korthalsia furtadoana J. Dransf. 4 14
Korthalsia rostrata BI. 2 14
Total: 4 genera, 21 species 343 (228 per ha)
shape/size combination, the means were also the same; however, the coefficients of variation
varied with plot type.
Multiple regression analysis found the percentage coefficients of variation, or CV, of both
c\umps/ha and stems/ha estimates to be more significantly related to both plot shape and size
than to plot size alone:
Clumps/ha:
LOG IO CV = 3.18 - 0.13 LOG IO SHAPE - 0.28 LOG IO SIZE (la)
R2adj =0.95, F=144.9, P < 0.001
LOG IO CV= 2.98 - 0.30 LOG IO SIZE (I b)

R2adj =0.85, F=77, P < 0.001
Stems/ha:
LOG IO CV= 3.74 - 0.24 LOG IO SHAPE - 0.19 LOG IO SIZE (I c)
R2adj =0.64, F=13.6, P < 0.01
LOG IO CV = 3.37 - 0.23 LOG IO SIZE (I d)

R2adj =0.34, F=8.2, P < 0.02
Other workers who have conducted tree inventory studies and examined the relationship
between the coefficient of variation and plot size (but not plot shape) have also found
equations of the generalised form: LOGIOCV = h - g LOGIOSIZE to be significant. Their
250
(a)
200
l-rect=l -rect=4 - rect=20
150
'#.
:>
<3
100
50
0
0 0.02 0.04 0.06 0.08 0.1
plot size (ha)
250
(b)
200
150
l- rect=l -rect=4 - rect=20 I
'#.
:>
<3
100
50
0
0 0.02 0.04 0.06 0.08 0.1
plot size (hal
Figure 3. The coefficient of variation (C.V. %) vs. plot size and shape, for: (a) clwnpslha and (b) stemslha.
values for g ranged from 0.25 (Freese 1961, Prodan 1968) and 0.27 (Taafe 1979, Tardif
1965) to 0.44 (Zeide 1980), depending upon the degree of heterogeneity of the forest. A
value for g of 0.5 indicates a random distribution. Thus the value for rattan clumps, 0.30,
is comparable to that of some tree species, and the value for rattan stems, 0.23, is indicative
of its more clumped distribution.
Figure 3 depicts this relation between the coefficient of variation and plot shape and size,
for (a) c1umps/ha and (b) stems/ha. The variation between plots is greater for estimates of
stems/ha than for c1umps/ha, although this difference is less marked at higher plot sizes and
rectangularities. Plots of larger size or of higher rectangularity have lower variation between
plots. The coefficient of variation can thus be predicted for a given plot size and shape, and
following this, the sample size for a given precision calculated.
Determining plot shape and size: the precision

One way of expressing the precision of an estimate is as the sampling error. This is
calculated as the ratio of the confidence interval to the mean; thus if there is a sampling error
of 20 %, the confidence interval is 20 % of the mean, at the given level of probability (Phil ip
1983).
528
M.C.STOCKDALE AND H.L.WRIGHT
500
(a)
400
g rect= l -rect = 4 - rect = 20 I
300
'"
. ",
'"
a. 200
~In
100
0
0 0.02 0.04 0.06 0.08 0.1
plot size (ha)
500
(b)
400
g rect = l -rect=4 - rect = 20 I
300
. III~
'"
a. 200
.,~
100
0
0 0 0. 2 0.04 0.06 0.08 0.1
plot size (ha)
Figure 4. The sample size (n) required for a sampling error (E%) of 20% (p=0.95) vs. plot size and shape, for (a)
clwnps/ha and (b) stems/ha.
The number of plots (n) of a certain shape and size required for a chosen percentage
sampling error (E) can be calculated from the equation:
(2)
where t is the student's 't' statistic and the coefficient of variation (CV) is calculated from
equations (la) and (lc).
Figure 4 displays, for the parameters of (a) clumps/ha and (b) stems/ha, the number of
plots needed to obtain a sampling error of 20%, at a 95% probability level, for a range of
plot shapes and sizes. More plots are needed for estimating stems/ha than for estimating
clumps/ha. Looking at shape, the higher the rectangularity of the plots, the fewer are
necessary for the same sampling error. Looking at size, the larger the plots, the fewer are
necessary to achieve the same precision. All of these plot number/shape/size combinations
depicted in the figure are equal in terms of precision; the optimal plot number/shape/size
combination is that which achieves this precision for the least cost.
The cost
The cost was approximated as the total time taken to conduct the inventory:
Ttotal = T trml + Tenumerate (3)
which is divided into time to travel between plots (TtraveJ and time to delineate and enumerate
the plots (Tenumerate) (Gambill et af. 1985, Taafe 1979).
The time to travel between plots:
(4)
is the product of the mean walking time per unit distance (T1), and the total mlllimum
distance between n randomly placed plots in a total area A, which is approximated by
imagining a square grid of area A, covered by a systematic sample of n points. The distance
between any two points will be (A/n)1I2; this value is then multiplied by the total number of
points (n) (Arvanitis & O'Regan 1967, Zeide 1980).
The time to delineate and enumerate the plots:
Tenumerate = T2 (nP) (5)
is the product of the mean time to delineate and enumerate a plot per unit area (T2), and the
total area delineated and enumerated, calculated by multiplying n plots by their plot size (P)
(Bonilla 1969, Tardif 1965).
70
(a)
60 rect - l -re ct a 4 - rect - 20 I
~
e
.,
,§
CD
50
40
.---
-
Ii
30
20
10
.. -
0
0 0.02 0.04 0.06 0.0s 0.1
plot size (ha)
70
(b)
60
l-rect = l
F 50
5
05. 40
CD
~ 30
'ii!
Ii 20
10
0
0 0.02 0.04 0.06 0.08 0.1
plot size (ha)
Figure 5. The total time taken to conduct an inventory of an area (A) of 200 ha at a sampling error (E%) of 20%
(p=0.95) vs. plot size and shape, for: (a) clumps/ha and (b) stems/ha.
Figure 5 displays, for the parameters of (a) clumps/ha and (b) stems/ha, the total time
taken to inventory an area of 200 ha with a sampling error of 20%, for a range of plot
shapes and sizes. For the same precision, enumerating stems is more costly than enumerating
clumps. Looking at shape, the higher the rectangularity of the plot, the more efficient it is
in terms of time. Looking at size, the most efficient plot sizes are those at which total time
taken approaches a minimum value. These optimal plot sizes vary with shape; the higher the
rectangularity of the plot the larger the optimal plot size.
The above figure indicates optimal plot size for a specific sampling error (20%) and total
area (200 ha). Tables 2 and 3 list the optimal plot sizes (and their corresponding sample
sizes) for a range of sampling errors and total areas, for clumps/ha and stems/ha
respectively, with rectangularity held constant at 4. The optimal plot size, which determines
the sample size, is shown to be positively correlated to both the sampling error and total
area. The optimal plot sizes for estimates of clumps/ha, ranging from 0.0025 ha to 0.01 ha,
are less than those for stems/ha, which range from 0.005 ha to 0.025 ha.
DISCUSSION
Plot shape
In this study, plots with the highest rectangularity were more cost efficient than plots of
lower rectangularities (Figure 5). This is because fewer rectangular plots than square plots
are needed for the same sampling error (Figure 4), and this in turn is because the higher the
rectangularity of the plot, the lower the C.V., or variation between plots (Figure 3). It is
likely that there is less variation between rectangular plots because their orientation (parallel
to the direction of the slope) permits them to encompass more of the variation within the
plots.
A study of rattan inventory in the Philippines by Tandug (1978) compared rectangular
plots with square plots, both of various sizes, and found a square plot to be most efficient.
No mention was made of the orientation of the rectangular plots, thus it is possible they were
randomly oriented and did not incorporate as much of the overall variation as they might
have done. Another explanation for the difference in results is that the Philippine site may
be flatter than the Brunei site; in flatter areas, the rectangularity and orientation of the plot
may not have as strong an effect as in areas where the topography is sharply dissected.
In theory, for the Brunei study area, the higher the rectangularity of the plot, the better;
in practice, there are limits to rectangularity. This is because non-sampling errors, such as
incorrect enumeration of plants on the plot boundaries, increase with an increasing ratio of
plot perimeter to area (Philip 1983). At its 'optimum' plot size of 0.025 ha (for E=20%,
A=200 ha), a rectangularity of 20 (thus 3.5 m by 70 m) will have very high non-sampling
errors. From field experience, a minimum width of 5 m would seem to allow the
non-sampling error to remain at an acceptable level. Thus for the above sampling error and
total area, a rectangularity of 4, with its optimal plot size of 0.01 ha, will have the more
workable dimensions of 5 m by 20 m.
Plot size
The total time taken to conduct an inventory approaches minima at plot sizes ranging from
0.0025 ha to 0.01 ha for clumps/ha estimates and from 0.005 ha to 0.025 ha for stems/ha
estimates (with a rectangularity of 4). The optimal plot size of Tandug's (1978) study, 0.01
ha (with a rectangularity of 1), is within this range. In Tandug's study, clumps/ha were
estimated; unlike the present study, this included the enumeration of seedlings.
The reason that the cost (in terms of time) approaches a minimum is that for plots smaller
than the optimal size, travel time becomes the dominant component of the total time taken,
Table 2. Optimal plot size (ha) and its sample size (n) for estimating clurnps/ha (rectangularity = 4).
Total Area (ha) Sampling Error (E%)

10 15 20 25 30
50 0.0025 0.0025 0.0025 0.005 0.005
(n = 442) (n = 197) (n = 113) (n = 50) (n = 35)
100 0.0025 0.0025 0.005 0.005 0.005
(n = 442) (n = 197) (n = 78) (n = 50) (n = 35)
200 0.0025 0.005 0.01 0.01 0.01
(n = 442) (n = 136) (n = 53) (n = 35) (n = 25)
400 0.005 0.01 0.01 0.01 0.01

(n = 300) (n = 92) (n = 53) (n = 35) (n = 25)
Table 3. Optimal plot size (ha) and its sample size (n) for estimating stems/ha (rectangularity = 4).
Total Area (ha) Sampling Error (E%)

10 15 20 25 30
0.005 0.01 0.01 0.01 0.01
50
(n = 472) (n = 121) (n = 68) (n = 44) (n = 30)
100 0.005 0.01 0.01 0.01 0.025
(n = 472) (n = 121) (n = 68) (n = 44) (n = 16)
200 0.01 0.01 0.01 0.025 0.025
(n = 272) (n = 121) (n = 68) (n = 24) (n = 16)
400 0.01 0.01 0.025 0.025 0.025
(n = 272) (n = 121) (n = 37) (n = 24) (n = 16)
since much larger numbers of these plots are necessary to achieve the same sampling error
(Figure 4). For plots larger than optimal size, the delineation and enumeration time becomes
increasingly significant, as an increasing fraction of the total area must be delineated and
enumerated to achieve the same sampling error (Figure 4).
The optimal plot size, which determines the sample size, is positively correlated with the
chosen sampling error, the total area to be inventoried, and the rectangularity (Figure 5,
Table 2). Thus, a decision about desired sampling error, plus external limits such as the total
area to be inventoried and a minimum plot width of 5 m (which will determine the
rectangularity), will enable an inventory planner to determine an optimal plot size.
The parameter also has an effect upon optimal plot size; the most efficient plot size for
estimates of the parameter clumps/ha is smaller than that for estimates of stems/ha, all else
being held constant (Figure 5, Tables 2 and 3). According to F.A.O. (1973), when several
parameters are measured in the same inventory, the most important parameter should be
selected for optimisation. If several parameters are equal in importance, the parameter with
the greatest variance should be used. On both counts, it is better to use the plot sizes optimal
for estimating stems/ha; it is the more important parameter in terms of assessing economic
potential of an area, and it also has the higher variance (Figure 3).
CONCLUSIONS
In the sharply dissected Brunei site, rectangular plots, oriented parallel to the direction of the
slope, are more cost efficient than square plots for the inventory of rattan. As errors in
boundary identification and area estimation increase with the degree of rectangularity,
however, a minimum width of 5 m is recommended. The most cost efficient plot size ranges
between 0.0025 ha and 0.025 ha, and can be selected from Tables 2 or 3 after determining
the desired sampling error, the total area under inventory, and the parameter(s) to be
estimated.
ACKNOWLEDGEMENTS
This research was funded by grants from the Canadian International Development Agency
and the Dennis Curry Charitable Foundation. We would like to thank Judith Stockdale, Impin
Lassah and Simpurai Anggat, for their hard work. This research was undertaken as part of
the Brunei Rainforest Project, organised by the University of Brunei Darussalam and the
Royal Geographical Society. Many thanks to the corporate patrons, Royal Brunei Airlines,
The Baring Foundation, DICAM, GreenCard Trust, Hongkong Bank, Morgan Grenfell, and
Nomura-NIMCO for their commitment to the Kuala Belalong Field Studies Centre. Thanks
also to Dr. Peter Kanowski and Dr. John Dransfield for commenting on drafts of this paper.
LITERATURE CITED
ARV ANTIS, L. G. & O'REGAN, W. G. 1967. Computer simulation and economic efficiency in forest sampling.
Hilgardia 38: 133·164.
BONILLA, J.A. 1970. Plot size. The size most useful for mensurational surveys. Silvicultura 19: 15-38.
CALDECOTT, J. 1988. Climbing towards extinction. New Scientist 118 (1616):62-66.
FREESE, F. 1961. Relation of plot size to variability. An approximation. Journal of Forestry 59:679.
GAMBILL, C.W., WIANT, H.V., Jr., & YANDLE, D.O. 1985. Optimum plot size and BAF. Forest Science
31:587-594.
GRlEG-SMITH, P. 1983. Quantitative plant ecology (3rd edition). Studies in Ecology Vol. 9. Blackwell Scientific
F.A.O. 1973. Manual offorestry inventory with special reference to mixed tropical forests. F.A.O., Rome.
HUSCH, B., MILER, C.l. & BEERS, T.W. 1972. Forest Mensuration (2nd edition). The Ronald Press Company,
New York.
LOETSCH, F., ZOHRER, F. & HALLER, K.E. 1973. Forest Inventory, vols I and II. BLV Verlagsgesellschaft,
Munchen.
MANOKARAN, N. 1990. The state of the rattan and bamboo trade. RIC Occasional Paper No.7.
PHILIP, M. S. 1983. Measuring trees andforests. Division of Forestry, University of Dar Es Salam.
PRODAN, M. 1968. Forest Biometrics. Pergamon Press, Oxford.
SHARMA, S. K. & BHATT, P.M. 1982. An assessment of cane potential of Baratang Island in South Andaman
Forest Division. Indian Forester 108: 270-282.
SISWANTO, B. E. & SOEMARNA, K. 1988. Metode inventarisasi rotan di K. P. H. Pontianak, Kalimantan Barat
(Rattan inventory method in Pontianak Forest District, West Kalimantan). Bulletin Penyelidikan Hutan (Forest
Research Bulletin) 503: 1-11.
TAAFFE, K.E., Jr. 1979. Computing optimum plot size for wildland inventories. Resource Inventory Notes
BLM-23:8-15.
TANDUG, L. M. 1978. Sampling method for inventory of Philippine rattan and its distribution. Sylvatrop Philippine
Forestry Research Journal 3:155-170.
TANDUG, L. M. & LASMARIAS, V.T. 1984. Determination of the most appropriate sampling design for the
inventory of Philippine rattan. P.C.A.R.R.D.-I.D.R.C. National Integrated Research Program on Rattan.
Terminal report (unpublished). 3 pp.
TARDIF, G. 1965. Some considerations concerning the establishment of optimum plot size in forest survey. Forestry
Chronicle 41:93-102.
U.B.D.lR.G.S. 1991. Information and joining instructions for European members (2nd edition). University of Brunei
Darussalam/Royal Geographical Society Brunei Rainforest Project 1991-1992.
UHL, N. W. & DRANSFIELD, J., 1987. Genera Palmarum: a classification ofpalms based on the works of Harold
E. Moore, Jr. The L. H. Bailey Hortorium and the International Palm Society, Lawrence, Kansas.
VANCLA Y, J., in press. Growth models for tropical moist forest. C.A.B. International.
WILLIAMS, T. 1991. Research needs for bamboo and rattan to the year 2000. Unpublished manuscript of the
Tropical Tree Crops Program of the International Fund for Agricultural Research, Washington D.C., U.S.A.
ZIECK, J. F. U. 1967. Minor forest products investigation - to investigate the potential for the economic exploitation
of rattans in the territory of Papua and New Guinea. Research Notes Ull-2. Forest Products Research Centre,
Port Moresby.
ZEIDE, B.B. 1980. Plot size optimisation. Forest Science 26:251-257.
Conservation of biodiversity in Kenya's indigenous forests: biological
currencies used in management planning
GLYN DAVIES
Kenya Indigenous Forest Conservation Project, P.O. Box 71493, Nairobi, Kenya
KEY WORDS: endemism, endangered species, habitat-dependence, indicator species, national parks.
INTRODUCTION
With the recent up-surge of interest in conserving biological diversity, especially through
effective conservation of the world's tropical forests, there has been a considerable increase
in the funding focussed on forest management and a change in the way development agencies
now carry out economic appraisals of natural biological systems. A wealth of new literature
has been produced over the past 5 years outlining the different levels at which economic and
financial analyses must be applied (e.g. Brown & Moran 1993, Pearce 1991), and it is an
appropriate time for biologists to make a greater contribution and so allow economic models
to be refined further.
In this context, one issue of applied tropical forest research which needs to be addressed
more vigorously is the need to present a clearer set of biological guidelines to use in
evaluating different areas of forest. Whenever biologists discuss forests, mention is invariably
made of a particular species of interest, often because it is rare and/or endemic. In other
words, it is not just the number of species which makes an area important for biodiversity
conservation, but there are biological qualities which are used when assessing biological
value. These are distinct from the economic values of 'direct use', 'indirect use' and 'option'
(sensu Pearce 1991). For example, Collar & Stuart (1985) compiled a book on the
threatened birds of Africa and related islands in which the importance of an area for bird
conservation was based on the number of rare species found within. Indeed, Collar (1986),
when reviewing the conclusions of the African Bird Red Databook, emphasised how few rare
species were currently, or potentially, of any economic value, but that the risk of extinction
was the main concern.
The primary aim of this paper is to describe biological qualities of species which are used
when developing conservation strategies. The magnitude of these qualities varies between
species, and so they can be used as 'currencies' when calculating the overall value of an
assemblage of species within a particular forest. Field data on tree/shrub, butterfly, bird and
mammal species from four of Kenya's indigenous forests, have been compiled, and using
these biological currencies the relative importance of each forest for biodiversity conservation
has been assessed.
KENYA'S INDIGENOUS FORESTS
Kenya covers an area of 582,644 km 2 which encompasses about 10,700 km 2 of lake area,
including two large lakes: Turkana (6,405 km2) and part of Victoria (3,785 km 2). Rainfall
is relatively low, with very few areas receiving 2,000 mm per annum - the mlllllTIUm
535
536
GLYN DAVIES
Fijure I. Ajroclimatic zones with 'potential for forest cover in Ke~ya.
requirement for rain forest (Richards 1952). However climatic conditions are sufficiently wet,
with rainfall spread over two seasons, to support forest vegetation over about 15 % of the
country (Figure 1). In these areas the ratio of average annual rainfall to average annual
potential evaporation exceeds 50% (Kenya Soil Survey, 1980).
Within these wetter parts of Kenya, forest cover a decade ago amounted to 1.255 million
ha (Doute et al. 1981). Almost all of this lies within the 2.2 million ha of gazetted Forest
Reserves (1.7 million ha) plus proposed Forest Reserves (0.5 million ha) . As is also
apparent, there are many areas of gazetted Forest Reserve which contain no forest vegetation.
The areas in which indigenous forest occurs also have the highest potential for agriculture ,
and support about 80% of Kenya's 24-30 million people, a population which is expanding
at about 3.7 % p.a. Not surprisingly, closed-canopy indigenous forests only occupy about 2 %
of the country because many areas have been cleared for agricultural development, some
have been burnt by pastoralists in the dry regions , and since the 1950's about 10% of the
forest estate has been converted to industrial plantations mostly comprising exotic trees. In
addition to total loss of forest cover, some areas of forest have been degraded through over-
exploitation of forest products: timber extraction; charcoal and fuelwood collection; carving
wood and polewood cutting; hunting and trapping of ungulates (Davies 1992).
As a consequence of these pressures the condition of Kenya's forests has been
deteriorating and the British Overseas Development Administration has given support to the
Conservation of biodiversity in Kenya 537
ETHIOPIA
TANZANIA
Figure 2. Gazetted forests of Kenya, showing those surveyed during the KIFCON project.
Kenya Government to establish a Kenya Indigenous Forest Conservation Project (KIFCON

Project). The main aim of this project is to assist the Ministry of Environment and Natural
Resources to develop a strategy for improved conservation management of Kenya's
indigenous forests. This has initially required multidisciplinary surveys to be conducted in
different forest areas, and this paper reports the results of Biodiversity Surveys carried out
in four forests (Figure 2, Table 2) by survey teams of the National Museums of Kenya,
working under contract to the Forestry Division/KIFCON Project.
Throughout this paper, therefore, a national perspective will be taken, since it is forest
management within Kenya which is the context of the development programme technical
surveys. One of the required outputs of the Biodiversity Surveys is to identify priority
forests, and priority areas within forests, which require protection to ensure against extinction
of plant or animal species. Furthermore, this biological information must be presented in
a form that allows evaluation of species protection in relation to forest-use by rural
communities, commercial enterprises etc. To develop such a multiple-use strategy for forest
management, the biological information must be understood by planners who may have no
biological training, and it must be applicable across a broad range of forests.
BIOLOGICAL CURRENCIES
The recent report on Global Biodiversity, compiled by the World Conservation Monitoring
538 GLYN DAVIES
Centre expands on many of the issues noted below, and discusses them more fully
(Groombridge 1992).
Biodiversity
Throughout the environmental literature the conservation of global biodiversity is cited
as a fundamental aim, yet there is seldom a clear understanding of what 'biodiversity' means.
This, in part, is because biological diversity (= biodiversity) refers to the total variety of
living organisms on earth (Stuart et al. 1990) and can be considered at three levels: a) the
gene; b) the species, of which the individuals are products of genetic and environmental
inter-play, and; c) the ecosystem, which is composed of species interacting with one and
other and with the non-biological components of their environment.
There is therefore ample scope for confusion, both in understanding and discussion,
depending on the level at which biodiversity is considered. A confusion exacerbated by the
direct linkages between the three levels - changes at anyone level will affect the other two.
In this paper, biodiversity will be considered at the level of the species, accepting current
taxonomic nomenclature. This is a natural way to consider whole-animal variation, although
it does have the short-coming of failing to take full account of the genetic differences
between species. Some species may be more closely related and hence share more genetic
characters than others, so biodiversity is greatest when distantly related species are combined
(Groom bridge 1992).
Furthermore, only four taxonomic groups will be considered in this paper: vascular plants
(especially woody shrubs and trees), butterflies, birds and mammals. These groups are
relatively well-studied in Kenya and there are reasonably accurate species lists for several
forest areas. Information on trees and shrubs is an immediate requirement for assessing the
vegetation of a forest, since these larger plants are cut/collected most frequently. Data on
different birds species was sufficiently detailed from short surveys to be analyzed for an
assessment of forest quality in terms of supporting rare or forest-dependent species etc. The
effects of hunting, however, do not show up in avifaunal surveys so larger mammals,
especially the ungulates, were also surveyed.
Information from other groups (e.g. amphibians/reptiles, other invertebrates, non-vascular
plants and microbes) was very sketchy, and could not be improved upon during the brief
KIFCON surveys, largely because of the great seasonal changes in insect populations and
detectability of the herpetofauna. Information on butterflies was supplied by Dr Steve
Collins, who has collected extensively in Kenyan forests, and gives distributional data which
were not seasonally biased.
Throughout Kenya, the number of tree and shrub species, calculated from a total of 1,532
named species in the East African Herbarium in 1980 and adding about 10% for new
specimens collected in the last decade, is estimated at 1,700 species. For all vascular plants,
a figure of about 6,000 species is the current estimate (C.Kabuye, personal communication).
For butterflies the current estimate is 875 species (S.Collins, personal communication) and
for birds 1,079 (L.Bennun, personal communication). A mammal checklist is currently being
drawn up, with best estimates at the present time of 305 species, but only 119 species if
rodents, shrews and bats are excluded.
Rarity
Loss of biological diversity, through extinction of species, is often a gradual process in
which a species becomes increasingly rare until, eventually, there are too few individuals to
sustain the population. Species threatened with extinction have always attracted conservation
attention, and mCN developed a set of categories for threatened species in decreasing order
of rarity: endangered, vulnerable and rare. Recently there has been much debate about these
categories, especially since there is an inevitable lack of information on most threatened

species.
A recent proposal has been made to improve the threatened species categories (Mace &
Lande 1991). For given populations or sUb-populations (of reproductive individuals) the
foJlowing definitions have been offered:
Critical: 50% probability of extinction within 5 years or 2 generations, whichever is
longer
Endangered: 20% probability of extinction within 20 years or 10 generations, whichever
is longer
Vulnerable: 10% probability of extinction within 100 years.
The threatened species referred to in this paper probably all fit the above three definitions,
although there is very little information on rates of reproduction to confirm which category
applies, a problem encountered all over Africa (Andre et ai. 1992). All that is known, or
in some cases inferred, is that the population is small and declining in Kenya.
Habitat-dependence
Not withstanding the many causes of species extinctions, whether through chance events
(e.g. storms) or direct causes (e.g. hunting), a central issue for forest conservation in any
country is the loss or degradation of forest habitats. Species which are dependent on forest
habitat are those which are being lost globally, and there is a need to focus attention on these
even though many can live at high population densities where there is suitable habitat. If the
forest in an area is lost or degraded, the total number of species may increase as a result of
invasion by savanna-preferring species. This would not constitute an improvement in
biodiversity conservation, however, since it merely indicates replacement of forest species
by savanna species many of which are already abundant outside forest environments.
A forest-dependent species is one which attains its highest population densities in forests
which have been little-modified by human activities. Ornithologists in Kenya have divided
forest species into two sub-categories: F species which are forest-specialists that only occur
where there is dense multi-storeyed forest, and f species which are forest-generalists that
occur on forest edges and in forest gaps (Lewis & Pomeroy 1989, Pomeroy 1989). Inone
Kenyan afro montane forest it was the populations of F species which declined in areas
greatly modified by past timber extraction (Bennun & Waiyaki 1992). A similar trend
applies to F birds in the largest coastal forest (J.Fanshawe, personal communication).
Forest-dependent mammals include antelopes (e.g. red duiker) which are rare outside
forest habitat and primates that do not occur in unforested areas (e.g. colobus monkeys).
Forest-dependent butterflies are largely those whose larvae feed on forest plants.
For the category of forest-dependence to have any meaning, it is necessary to define
'forest'. The most obvious definition, which is followed in this paper, is that of White
(1983) in his review of African vegetation:
"a continuous stand of trees; canopy varies in height from 10m to 50m and usually
consists of several layers; crowns of individual trees interdigitate or overlap, and are often
inter-laced with lianes. A shrub layer is normally present being densest where the tree
canopy is most open. The ground layer is sparse or absent, generally lacking grass cover. "
In the wettest parts of Kenya, where slopes and soil are suitable, forest vegetation is easily
identified using this definition. There are two main problems, however: first, there is a
gradual transition from forest formations to more open canopy woodland formations as
conditions become drier, making it difficult to define a cut-off point between forest and
woodland vegetation in a given area; second, interspersed in many forest areas are grassy
glades which obviously do not constitute forest vegetation but form an integral part of forest
area and cannot be ignored.
540 GLYN DAVIES
Figure 3. Floristic zones in Kenya (after White 1983). Key: 2 - Guineo-Congolian rainforest (drier type); Ila -
lowland forest/secondary grassland mosaic; 16a - coastal mosaic/Zanzibar-lnhambane; 16b - coastal mosaic/forest
patches; 19a - Afromontane forest and bamboo.
Endemism
The degree to which a species is restricted in its geographical range can also contribute
to its rarity. An endemic species is one which is restricted to a particular area, varying in
size from a mountain, to a country, to a continent. In a recent assessment of global bird
conservation, ICBP (Bibby et at. 1992) developed the concept of restricted-range (r-r) species
for land birds whose breeding range has never occupied more than 50,000 km 2 • On the basis
of two 2 r-r species occurring in an area it is designated an Endemic Bird Area (EBA) and
by focussing on EBA's there is no chance of double counting species which occur in different
countries, for example, but still within the same EBA.
By focussing on different biogeographical regions in Kenya in this paper (see below), this
problem of double counting is largely avoided.
BIOGEOGRAPHY OF KENYAN FORESTS
Following the phytogeographical zones to which White (1983) has assigned African
vegetation formations, Kenyan forests occur in three floristic zones: coastal, afro montane and
western (Figure 3).
A narrow coastal belt, running inland approximately 30 km from the shore-line, is the
northern part of the Zanzibar-Inhambane region which stretches from northern Mozambique
to southern Somalia (White 1983). This region contains many isolated forest blocks which
have been separated from one and other for protracted periods and therefore the species
within each show high levels of endemism, and because they are frequently small, high levels
of rarity also. Furthermore, the coastal area can be divided into two sub-regions north and
south of the Mombasa-Kilifi area.
The bulk of the country's forests are in the afromontane region, a region which stretches
from the highlands of Guinea and Liberia in the west to Ethiopia in the east and the
Drakensburg Mountains in the south. There are close relationships between plant species in
all these forests, although the region has been further sub-divided into seven afromontane
systems (White 1978); within Kenya all the forests are in the Imatong-Usambara system.
There is a further distinction between forests on the west side of the Rift Valley, which have
low levels of plant endemism, and east-of-Rift forests which have higher levels both amongst
plants (Beentje 1990) and birds (Lewis & Pomeroy 1989).
Finally, there is a small area in the Lake Victoria basin which is the eastern-most
extension of the Guineo-Congolian forests that are distributed from the Congo basin to
coastal west Africa. Although the area occupied by this type of forest in Kenya was
probably once much more extensive, it is now represented by a single small forest,
Kakamega.
By selecting forests from these different biogeographical regions, the species richness
patterns which are described below, especially for plants and birds, refer to many different
species. When more forests are considered, many from the same biogeographical region,
care will have to be taken to ensure against 'double counting' species which occur in more
than one forest.
BIODIVERSITY IN KENYAN FORESTS
Biogeographic Regional Overview

Lists have been drawn up of Kenyan species which are threatened at an international level,
for trees and shrubs (Beentje 1988), birds (Fanshawe & Bennun 1991) and mammals OUCN
1990). Excluding species found in savanna, dry bushland and high moorlands, the
distribution of threatened forest species can be extracted from these lists and related to the
different biogeographical zones in Kenya (Table 1).
From a botanical perspective, almost half of the threatened trees and shrubs are found in
coastal (38%) or riverine (9%) forests, with some overlap between the two areas. Similarly,
Table 1. Internationally threatened forest species in different regions of kenya. Sources: Beentje 1988, Fanshawe
& Bennun 1991, mCN 1990.
Trees/Shrubs Birds Mammals

Number of threatened species 159 71 9
% forest species in different regions:
Coast 38% 27% 55%
Taita Hills 11 % 4% 0%
Riverine 9% 3% 22%
Afromontane 12% 14% 22%
Western 0% 6% 0%
542 GLYN DAVIES
Table 2. Altitude, rainfall, gazetted area and vegetation cover of four indigenous Forest Reserves. Forest types after
Beentje 1990.
Arabuko/Sokoke Shimba Hills Mau Kakamega

Altitude (m a.s.l.) 50 150 2,150-2,500 1,600
Rainfall(mm/annum) 600-1000 1,150 1,000-2,000 2,000
Gazetted area(ha) 42,000 22,000 126,550 21,600
Vegetation:
grass + 3,000 11,346 4,250
shrub 13,300 7,800 23,800 1,500
forest 27,400 10,400 57,000 12,440
bamboo 32,100
Forest Types 4 6 4(+)
over a quarter of the threatened Kenyan birds are found therein, and 55 % of mammals (5
species) also occur in coastal forests or associated grasslands.
The afromontane forests are moderately important for mammals (2 species), plants and
birds, but the western forests hold few threatened birds, no threatened mammals and no
threatened trees/shrubs.
The Taita Hills, and adjacent Kasigau, occur in an outlier patch of the afromontane region
near Usambara mountains of Tanzania and cover only a small area « 1,000 ha), yet are of
surprisingly great importance for flora and birds. They account for 11 % of Kenya's
threatened plants and 4 % of threatened birds, including endemic species/subspecies.
This regional overview immediately shows the importance of coastal forests for the
conservation of biodiversity within Kenya, irrespective of the numbers of species found in
different forests. However, as indicated below, when nationally rare species and forest-
dependent species are considered, forests in other regions assume greater importance.
Focal forests
In the course of KIFCON Project work four forests have been surveyed for fauna and
flora (Table 2). These forests are located in all three main biogeographical regions. At the
coast, Arabuko-Sokoke (42,000 ha) and Shimba Hills National Reserve with the two
contiguous Mkongani Forest Reserves (22,000 ha) are located north and south of the
Mombasa-Kilifi biogeographical subdivision respectively. The West Mau, Southwest Mau
and Transmara Forest Reserves are contiguous and cover 126,550 ha in the west-of-Rift
afromontane region, and finally Kakamega (21,600 ha) occurs in the western region. Within
these forests, all species which are considered rare in Kenya, irrespective of their abundance
elsewhere in Africa, are noted in the comparisons.
As expected, the largest numbers of rare trees were in the coastal forests; Shimba having
19 species compared with Sokoke's 11, Mau's 9 (including 3 new KIFCON records for the
East African Herbarium) and 7 nationally rare species in Kakamega (Table 3). The overall
number of species was also very high in the coastal forests, especially considering their small
size - Shimba is probably the most botanically rich forest in Kenya. Mau and Kakamega are
of much lower importance in terms of conservation of higher plant biodiversity.
The butterflies show a similar trend, with the highest species numbers at the coast.
Sokoke has many more species than Shimba, but Shimba has more rare species. Kakamega
has a similar number of rare species and is exceptionally rich in forest-dependent species,
although overall species richness does not reach the levels of the coastal forests.
Kakamega forest comes into its own when birds are considered. For such a small forest,
the 351 species that have been recorded is extraordinarily high. Indeed, Kakamega is unique
Table 3. Threatened, forest-dependent and total species counts for four forests. (e) = estimate based on species
collected to date.
AlSokoke Shimba Hills Mau Kakamega

PLANTS
Threatened trees/shrubs II 19 9 7
All trees/shrubs 127 350 100 93
Vascular plant species (e) 208 400 273 135
BUTTERFLIES
Threatened 6 15 3 13
Forest-dependent 20 24 20 67
All 369 299 240(e) 263
BIRDS
Threatened 18 13 9 15
Forest-specialist 20 18 40 75
Forest-generalist 63 37 60 124
Non-forest 32 30+ 12 152
Aquatic 11 ? 5 ?
All 126 85 117 351
MAMMALS
Threatened 5 4 6 2
Forest-dependent 7 8 7 9
All 31 28 28 20+
in continental terms as a forest in which the lowland species of the Guineo-Congolian forests
intermix with those of the upland species of the afromontane zone (Collar & Stuart 1988).
Like the butterflies, a large proportion of the Kakamega birds are forest-dependent, and
there are 15 species threatened within Kenya. To this can be added the large number of
savanna birds found in long-standing forest glades and anthropogenic grasslands of more
recent origin, of which 54 are grassland dependent and 10 are rare in Kenya. The importance
of forest glades for these species stems from the scarcity of uncultivated or ungrazed
grasslands outside protected forest areas. This observation introduces a new category of
forest-dependent species, namely 'forest-grassland dependent' birds!
The coast forests support many rare birds: 18 in Sokoke, of which 5 are considered
international restricted-range species, and 13 in Shimba of which 2 are r-r species (Bibby et
at. 1992). There are also a moderate number of forest species and some open-country species
in the drier forest formations such as the Brachystegia woodland of Sokoke.
Mammals, excluding shrews, bats and rodents about which little is known, have been all
but eliminated from the small Kakamega forest by intensive hunting pressure from the
adjacent densely populated communities' hunters. The coastal forests fare better, and include
the savanna-dwelling sable antelope in the Shimba grasslands. However, it is the large Mau
forest area which has the greatest importance for mammal conservation because the large
area supports viable populations of some of the rarest larger species (e.g. Bongo and Yellow-
backed Duiker).
Overall, therefore, the coastal forests, with their high degree of endemism are important
for fauna and flora alike, but are especially important for their high plant species richness
and rare bird species. The Mau forest has even lower levels of plant species endemism than
reported from the east-of-Rift afromontane forests (Beentje 1990), and remarkably few plant
species considering the huge area involved, perhaps reflecting the negative effects of high
544 GLYN DAVIES
altitude on plant species richness. But it is very important for mammal conservation.
Kakamega is important for bird and butterfly species conservation.
Interestingly, and contrary to general continental trends (Bibby et ai. 1992, Groombridge
1992), forests which are important for one vertebrate group are not necessarily important for
another. Kakamega is very important for birds but of little importance for mammals; Mau
is important for supporting populations of rare mammals but only supports birds which are
widely distributed in other afro montane forests.
Comparing the total number of species in each forest weights the results in favour of
groups with the greatest numbers of species (i.e. plants and butterflies). To avoid this bias,
the following three species categories were considered: a) threatened, b) forest-dependent,
and c) all species, and each forest was assigned a rank (1 to 4), for the four taxonomic
groups for which data are presented (Table 3). The number of times each forest scored the
1st or 2nd ranks (i.e. highest numbers of species) was summed and the most 'important'
forest assessed (maximum score = 12). The results were: Shimba (8), Sokoke (7), Kakamega
(6) and Mau (4). Shimba's importance stems from its plant and butterfly species richness,
Sokoke's from plant, bird and mammal species and Kakamega from birds and butterflies.
Taking only threatened species, and using the same ranking procedure, the order changes
slightly (maximum score = 4): Sokoke (3), Shimba (2), Kakamega (2), Mau (1). If the
actual numbers of threatened species are compared, the order reverts to that of the previous
exercise: Shimba (51), Sokoke (40), Kakamega (37) and Mau (27), with Sokoke and
Kakamega changing places if the Kakamega rare grassland birds are included.
Since threatened species are of more immediate concern than nonthreatened species, the
latter ranking gives a useful priority list for forest conservation planning at a national level.
However, it must not be forgotten that there are a further 12 large forest reserves, and over
100 smaller forest reserves to be assessed.
PLANNING AND POLICY ISSUES
Assessing biodiversity priorities

Have we developed a set of biological currencies which can be used to answer the
question "What forests in Kenya are most important for biodiversity conservation?", a
question that the Government of Kenya, donor agencies and NGOs are all asking.
The species level comparison works well, at least for the taxonomic groups considered,
and is an easy concept for nonbiologists to grasp even if biologists debate the definition! The
taxonomic groups considered also provided enough information with which to compare
threatened, forest-dependent and total species counts despite the incompleteness of inventory
work in all four forests.
Comparison could be made between biogeographic regions, which is an essential
requirement. In this case, since all four forests occurred in different biogeographical regions
and many species were unique to each forest, Shimba, Sokoke and Kakamega are all
considered priority forests for conservation intervention.
There are other large forests which are also important: Mt Kenya, Aberdares, Mathews
Range etc., and prioritising conservation intervention in these, which occur in the same
biogeographical region, will require a more detailed assessment of shared species.
Two other factors will have to be introduced into the comparisons: viability of
populations, and other values of the forest. As indicated for the Mau forest, its large size
is particularly important for maintaining viable populations of larger mammals, which is
central to maintaining forest biodiversity in the long-term. Information on species'
abundance is essential to allow estimation of the minimum area of a particular habitat
required to support 200-2000 breeding individuals. The calculation must, however, be
sensitive to different species population dynamics - 2000 elephant hardly occur in any forest
areas!
Biodiversity value may not be the only, or even the main reason for conserving or
protecting an area of forest. Environmental importance, such as for controlling water
quality, needs to be factored into the comparative calculations along with the more
conventional use values of forest resources.
Using 'indicator'species
Given the very large numbers of animal and plant species in forests, and the lack of
information on distribution or abundance for many of them, comparison of different areas
has relied on using a sub-section of the biological community about which most population
information is available. The use of such 'indicator' species has become standard practice
with preliminary analyses showing that, at least at continental level comparisons, there is a
good relationship between the areas of high endemism for different vertebrate groups (e.g.
Bibby et al. 1992, Collar & Stuart 1988).
However, the broad scatter of data points which form a correlation at the continental scale
can cease to apply within a smaller national framework. In the four Kenyan forests, bird
species tend to be fewer in forests with most plant species. There was no obvious relationship
between bird and butterfly species numbers. Kakamega is very important for birds but
unimportant for mammals, and the converse applies in Mau.
The use of indicator species in analyses of biodiversity must be done with great caution,
therefore, especially if there are a few small areas scattered across more than one
biogeographical zone. In particular, the invertebrates account for the majority of the species-
level biological diversity in forests, yet there is a remarkable lack of information on the
populations of all but a few groups (e.g. termites, butterflies); only information on butterflies
could be used in these analyses.
In general, time and resources are very limited for biodiversity surveys for management
planning, as opposed to in-depth research, and methods need to be developed that allow
information to be gathered quickly (ca. 7 day) and analysed quickly (ca. I month), from
several forests, or several areas of the same forest. Standard survey methods could ideally
be developed for several invertebrate 'indicator' groups which would allow a fuller
assessment of the biodiversity importance of different forests. Development of these methods
should be done at a few well-studied sites where rapid appraisal survey results can be
calibrated with known population estimates from more thorough research.
Economic assessment of biodiversity

From the point of view of economic analyses, the existence value of those species which
will cease to exist soonest should be higher than for those which are abundant and less likely
to be lost. This obviously applies irrespective of the increased value of commercial species,
whose supply of individuals or products declines while demand is constant or increasing.
Furthermore, rare species should be more valuable whether they are attractive to policy
makers or not; 'sexy' species which attract funding are still more valuable if they are rare
than if they are common.
This separation of species according to their abundance or rarity has yet to be included
in economic analyses of existence value (e.g. Brown & Moran 1993). Part of the problem
is that biologists must supply economists with the appropriate information and secondly,
giving economic value to species stumbles on the unresolved question of "What is a species
worth?" . Perhaps it is better to retain the biological currency, and give high value to
threatened species without trying to reduce the value to economic currencies.
546 GLYN DAVIES
ACKNOWLEDGEMENTS
I would like to thank Universiti Brunei Darussalam for the invitation to participate in this
conference, and the Royal Geographical Society for assistance with my travel costs. The
work presented entirely reflects the results of KIFCON field surveys carried out by scientists
and technicians from the National Museums of Kenya, and I thank them for the field data
which have been included in this paper. In addition, Dr Steve Collins (Monsanto Ltd)
supplied a checklist of butterflies. The KIFCON Project and Overseas Development
Administration (UK) gave permission and funds to allow me to attend this conference.
LITERATURE CITED
ANDRE, H. M., LEBRUN, P. & NOTI, M. I. 1992. Biodiversity in Africa: a plea for more data. Journal of
African Zoology 106:3-15.
BEENTJE, H.I. 1988. Atlas of the Rare Trees of Kenya. Utafiti 13:71-124.
BEENTJE, H.J. 1990. The forests of Kenya. Proceedings of the twelfth plenary meeting of aetfat. Mitteilungen
Institut fur Allgemeine Botanik Hamburg 23a:265-286.
BENNUN, L.A. & WAIYAKI, E.M. 1992. Ornithological Surveys of the Mau Forest Complex. Unpublished
KIFCON/NMK Report, Nairobi.
BIBBY, C.I., COLLAR, N.J., CROSBY, N.I., HEATH, M.F., IMBODEN, C.H., JOHNSON, T. H. , LONG,
A. 1., STATTERSFIELD, A.I. & THIRGOOD, S.I. 1992. Putting biodiversity on the map: priority areas for
global conservation. International Council for Bird Preservation, Cambridge.
BROWN K. & MORAN D. 1993. Valuing Biodiversity: The Scope and Limitations of Economic Analysis.
Conference proceedings on the Convention on Biological Diversity: National Interests and Global Imperatives.
ACTS/UNEP, Nairobi.
COLLAR, N.I. 1986. Species are a measure of man's freedom: reflections after writing a Red Data Book on
African birds. Oryx 20:15-19.
COLLAR, N.J. & STUART, S.N. 1985. Threatened birds of Africa and related islands. ICBP/IUCN, Cambridge.
COLLAR, N.I. & STUART, S.N. 1988. Key forests for threatened birds in Africa. ICBP, Cambridge.
DAVIES, A.G. 1992. Deforestation: East and West Africa. pp. 23-34 in Global Environment concerns: Impact
on Kenya. Kenya Museum Society, Nairobi.
DOUTE, R., OCHANDA, N. & EPP, H. 1981. A forest inventory of Kenya using remote sensing techniques.
KREMU, Nairobi.
FANSHAWE, J. & BENNUN, L.A. 1991. Bird Conservation in Kenya: Creating a national strategy Bird
Conservation International 1:293-319.
GROOMBRIDGE, B. ed. 1992. Global Biodiversity: status of the world's living resources. Chapman and Hall,
London.
IUCN 1990. Red list of threatened species. International Union for Conservation of Nature and Natural Resources,
Gland.
KENYA SOIL SURVEY 1980. Exploratory soil map and agroclimatic zones of Kenya.
LEWIS, A. & POMEROY, D. 1989. A bird atlas Of Kenya. A.A.Balkema, Rotterdam.
MACE, G.M. & LANDE R. 1991. Assessing Extinction Threats: Toward a Reevaluation of IUCN Threatened
Species Categories. Conservation Biology 5:148-157.
PEARCE, D.W. 1991. An Economic Approach to Saving the Tropical Forest. In Helm, D. (ed.). Economic Policy
Towards the Environment. Blackwell, Oxford.
POMEROY, D. 1989. The making of inventories: an introductory survey of methods for trees, shrubs, mammals
and birds. Unpublished report, University of Makarere, Uganda.
RICHARDS, P.W. 1952. The tropical rainforest. Cambridge University Press, Cambridge.
STUART, S.N., ADAMS, R.I. & JENKINS M.D. 1990. Biodiversity in Sub-saharan Africa and its Islands:
Conservation, management and sustainable use. IUCN & The World Conservation Union.
WHITE, F. 1978. The afromontane region. Pp. 463-513 in Werger, M.J.A. (ed.). Biogeography and ecology of
southern Africa. Iunk, The Hague.
WHITE, F. 1983. The vegetation of Africa. UNESCO, Paris.
INDEX TO GENERA AND SPECIES
Genera and species listed in appendices are not indexed
Abarema filamentosa . . . . . . . . . . . . . . . . . . 483 minuta . . . . . . . . . . . . . . . . . . . . . . . . 444
Abrophyllum . . . . . . . . . . . . . . . . . . . . . .. 223 Arenga undulatifolia . . . . . . . . . . . . . . . . . . 444
Abryna regispetri . . . . . . . . . . . . . . . . . . . . 94 Argyrodendron peralatum . . . . . . . . . . . . . . . 373
Acacia. . . . . . . . . . . . . . . . . . . . . . . . . .. 11 Arhopala buddTia . . . . . . . . . . . . . . . . . . . . 292
Acalypha stipulacea . . . . . . . . . . . . . . . . . . 327 metamuta . . . . . . . . . . . . . . . . . . . . . . . 292
Acantho/epsls . . . . . . . . . . . . . . . . . . . . . . 120 Armigeres . . . . . . . . . . . . . . . . . . . . . . . 89, 97
Acanthomyrmex . . . . . . . . . . . . . . . . . . . . . 120 dolichocephalus . . . . . . . . . . . . . . . . . .. 93
Acer saccharum . . . . . . . . . . . . . . . . . . . . . 219 flavus .. . . . . . . . . . . . . . . . . . . . . . .. 93
Acmena acuminatissima . . . . . . . . . . . . . . . . 327 Artocarpus ovatus . . . . . . . . . . . . . . . . . . . . 328
Acronychia vestita . . . . . . . . . . . . . . . . . . . 219 Asyndetus . . . . . . . . . . . . . . . . . . . . . . . . . 185
Acropsilus . . . . . . . . . . . . . . . . . . . . . . . . . 185 Athertonia diversifolia .. . . . . . . . . . .. 224, 225
Acropyga . . . . . . . . . . . . . . . . . . . . . . . . . 120 Atta . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
Actyolepis puspa . . . . . . . . . . . . . . . . . . . . 292 Attalea acaulis . . . . . . . . . . . . . . . . . . . . . . 479
Acylophorus . . . . . . . . . . . . . . . . . . . 88-91, 93 funifera . . . . . . . . . . . . . . . . . . 477-479, 483
Adinandra . . . . . . . . . . . . . . . . . . . . . . .. 223 humilis . . . . . . . . . . . . . . . . . . . . . . . . 483
Aedes gombakensis . . . . . . . . . . . . . . . . . .. 93 Baccaurea kunst/eri . . . . . . . . . . . . . . . . . . . 155
Aenictus . . . . . . . . . . . . . . . . . . . . . . . . . . 120 Bactris . . . . . . . . . . . . . . . . . . . . . . . . . . . 483
Afromomum . . . . . . . . . . . . . . . . . . . . . .. 219 Baptista . . . . . . . . . . . . . . . . . 88-90, 93, 96, 97
Agelaea bomeensis . . . . . . . . . . . . . . . . . . . 441 Barbodes collingwoodi .... 309-312, 380, 381, 384
Aglaia ditfusa . . . . . . . . . . . . . . . . . . . . . . 327 Barringtonia . . . . . . . . . . . . . . . . . . . . . . . 429
iloilo . . . . . . . . . . . . . . . . . . . . . . . . . 328 Basilia hispida . . . . . . . . . . . . . . . . . . . . .. 91
llanosiana . . . . . . . . . . . . . . . . . . . . . . 327 Batraxis . . . . . . . . . . . . . . . . . . . . . . . . .. 93
longipes . . . . . . . . . . . . . . . . . . . . . . . . 328 Begonia . . . . . . . . . . . . . . . . . . . . .. 155, 167
luzoniensis . . . . . . . . . . . . . . . . . . . . . . 327 cyanescens . . . . . . . . . . . . . . . . . . . . .. 48
Ahaetulla fasciolatus . . . . . . . . . . . . . . . . . . 318 guttata . . . . . . . . . . . . . . . . . . . . . . . . 155
prasinus . . . . . . . . . . . . . . . . . . . 155,318 phoeniogramma . . . . . . . . . . . . . . . . . . . 155
Ahemia glandulosa . . . . . . . . . . . . . . . . . . . 328 Beilschmiedia bancroftii . . . . . . . . . . . . . . . . 224
Ajuga . . . . . . . . . . . . . . . . . . . . . . . . . . . 325 brunnea . . . . . . . . . . . . . . . . . . . . . . . . 224
Allnzia pedicellaris . . . . . . . . . . . . . . . . . . . 483 recurva . . . . . . . . . . . . . . . . . . . . . . . . 224
Aleurites cordata . . . . . . . . . . . . . . . . 221, 225 tooram . . . . . . . . . . . . . . . . . . . . . . . . 224
moluccana . . . . . . . . . . . . . . . . . . . . . . 225 Berenicomis comatus . . . . . . . . . . . . . . . . . . 155
Allactoneura .. . . . . . . . . . . . . . . . . . . . .. 93 Betula allegheniensis . . . . . . . . . . . . . . . . . . 220
Allantus integrifolia . . . . . . . . . . . . . . . . . . . 328 Bischofia javanica . . . . . . . . . . . . . . . . . . . 328
Allostethella guttata . . . . . . . . . . . . . . . . . . 92 Blastophaga wassae . . . . . . . . . . . . . . . . . . 285
Allotinus . . . . . . . . . . . . . . . . . . . . .. 293, 294 Boesenbergia burttiana . . . . . . . . . . . . . . 48, 50
Alocasia guttata . . . . . . . . . . . . . . . . . . . .. 48 Boiga drapiezii . . . . . . . . . . . . . . . . . . . . . 318
lowii . . . . . . . . . . . . . . . . . . . . . . . . . . 48 Borassodendron bomeense . . . . . . . . . . . . . . 444
Alphitonia . . . . . . . . . . . . . . . . . . . . . . . . . 223 Bowenia spectabilis . . . . . . . . . . . . . . . . . . . 221
petriei . . . . . . . . . . . . . . . . . . . . . . . .. 226 Brachyrhynchus . . . . . . . . . . . . . . . . . . . .. 92
Alpmia glabra . . . . . . . . . . . . . . . . . . . . .. 48 Brachystegia . . . . . . . . . . . . . . . . . . . . . . . 543
ATseodaphne longipes . . . . . . . . . . . . . . . . . 328 Bradysia . . . . . . . . . . . . . . . . . . . . . . . . .. 93
Alstonia . . . . . . . . . . . . . . . . . . . . . . . . . . 348 Bronchocela cristatella . . . . . . . . . . . . 318, 321
scholaris . . . . . . . . . . . . . . . . . . . . . . . 328 Bruguiera . . . . . . . . . . . . . . . . . . . . . . . . . 429
Amblyopone . . . . . . . . . . . . . . . . . . . . . . . 11 Bubo coromandus . . . . . . . . . . . . . . . . . . . . 156
Amblypsilopus . . . . . . . . . . . . . . . . . . . . . . 185 Bucephalandra motleyana . . . . . . . . . . . . . .. 48
Amegl1la . . . . . . . . . . . . . . . . . . . . . . . 59, 64 Buceros bicomis . . . . . . . . . . . . . . . . . . . . . 155
Amomum . . . . . . . . . . . . . . . . . . . . . . . . . 47 rhinoceros . . . . . . . . . . . . . . . . . . . . . . 155
Amphiesma flavifrons . . . . . . . . . . . . . . . . . . 318 Bufo asper . . . . . . . . . . . . . . . . . . . . . . . . 317
Amydrium medium . . . . . . . . . . . . . . . . . . . 48 divergens . . . . . . . . . . . . . . . . . . . . . . . 317
Andira . . . . . . . . . . . . . . . . . . . . . . . . . . . 483 juxtasper . . . . . . . . . . . . . . . . . . . 317,321
Andropadus curvirostris . . . . . . . . . . . . . . . . 106 Bulbitermes . . . . . . . . . . . . . . . . . . . . . . . . 300
Anguilla marmorata . . . . . . . . 309,380,381,384 Bvrsonima sericea . . . . . . . . . . . . . . . . . . . 483
Amsoptera thurifera . . . . . . . . . . . . . . . . . . 235 Calamus . . . . . . . . . . . . . . . . . . . . . . . 23, 167
Annona salzmanii .. . . . . . . . . . . . . .. 483,484 amplijugus . . . . . . . . . . . . . . . . . . . . . . 526
Anopheles asiaticus . . . . . . . . . . . . .. 89, 90, 93 comptus . . . . . . . . . . . . . . . . . . .. 442, 526
Anoplolepis . . . . . . . . . . . . . . . . . . .. 120, 294 comrostris . . . . . . . . . . . . . . . . . . . . . . 526
Anorrhinus galeritus . . . . . . . . . . . . . . . . . . 155 jlabellatus . . . . . . . . . . . . . . . . . . . . . . 526
Ansonia albomaculata . . . . . . . . . . . . . 317,321 laevigatus . . . . . . . . . . . . . . . . . . . . . . 526
leptopus . . . . . . . . . . . . . . . . . . . . . . . . 317 manan . . . . . . . . . . . . . . . . . . . . . . . . . 430
longtdigita . . . . . . . . . . . . . . . . . . . . . . 317 muricatus . . . . . . . . . . . . . . . . . . . . . . . 526
spinuliJer . . . . . . . . . . . . . . . . . . . 317,321 pogonacanthus . . . . . . . . . . . . . . . . . . . 526
Anthene emolus . . . . . . . . . . . . . . . . . . . . . 292 praeterrnissus . . . . . . . . . . . . . . . . . . . . 526
Anthocephalus chinensis . . . . . . . . . . . . . . . . 223 sordidus . . . . . . . . . . . . . . . . . . . . . . . 526
Aonyx cinerea . . . . . . . . . . . . . . . . . . 343, 427 zonatus .... . . . . . . . . . . . . . . . . . . . . 526
Aphaenogaster . . . . . . . . . . . . . . . . . . . 11, 120 Caleta roxus . . . . . . . . . . . . . . . . . . . . . . . 293
Aphanamixis perrottetiana . . . . . . . . . . . . . . 328 Callicarpa . . . . . . . . . . . . . . . . . . . . 226, 363
Aphantaulax . . . . . . . . . . . . . . . . . . . . . . . 92 Callosclurus . . . . . . . . . . . . . . . . . . . . . . . 74
Apis . . . . . . . . . . . . . . . . . . . . . . . . 59, 62-65 melanogaster . . . . . . . . . . . . . . . . . . . . 427
andreniformis .. . . . . . . . . . . . . . . . . 61, 62 notatus . . . . . . . . . . . . . . 177, 178, 343, 344
cerana . . . . . . . . . . . . . . . . . . . . 61-63, 65 prevostii . . . . . . . . . . . . . . . . 177, 178, 343
dorsata . . . . . . . . . . . . . . . . . . . . . . 59i 62 Ca{ophyllum . . . . . . . . . . . . . . . . . . . . . . . 412
koschevnikovi . . . . . . . . . . . . . . . . 59, 6 -65 inophylloides . . . . . . . . . . . . . . . . . . . . 429
mellifera . . . . . . . . . . . . . . . . . . 62, 63, 65 Calyptomena viridis . . . . . . . . . . . . . . . . . . 428
Aplopeltura boa . . . . . . . . . . . . . . . . . . . . . 318 Campnospermum . . . . . . . . . . . . . . . . . . . . 348
Aporosa lunata . . . . . . . . . . . 367, 368, 370, 371 Camponotus . . . . . . . . . . . . . 10, 120, 292, 339
Apostasia elliptica . . . . . . . . . . . . . . . . . . . 48 gigas . . . . . . . . . . 62,331-334,336-339,362
Apterygodon vittaturn . . . . . . . . . . . . . . . . . . 318 Camptandra . . . . . . . . . . . . . . . . . . . . . . . 155
Arachnothe ra crassirostris . . . . . . . . . . . . . . 107 Camptopteromyia fractipennis . . . . . . . . . . 90, 93
longirostris . . . . . . . . . . . . . . . . . . . . . . 107 Canarium . . . . . . . . . . . . . . . . . . . . . . . . . 462
Aralia elata . . . . . . . . . . . . . . . . . . . . . . . . 221 asperum . . . . . . . . . . . . . . . . . . . . . . . 328
Araucaria . . . . . . . . . . . . . . . . . . . . . . . . . 284 Caprrcorms sumatraensis . . . . . . . . . . . . . . . 156
Arborophila charltoni . . . . . . . . . . . . . . . . . 155 Carebara . . . . . . . . . . . . . . . . . . . . . . . . . 11
Ardisia . . . . . . . . . . . . . . . . . . . . . . . . . . . 328 Carpinus betulus . . . . . . . . . . . . . . . . . . . . 219
Areca catechu . . . . . . . . . . . . . . . . . . . . . . 444 Carya . . . . . . . . . . . . . . . . . . . . . . . 224, 226
547
548 GENUS AND SPECIES INDEX
Casearia . . . . . . . . . . . . . . . . . . . . . . . . . 483 minor . . . . . . . . . . . . . . . . . . . . . . . . . 89
arborea . . . . . . . . . . . . . . . . . . . . . . . . 217 Culicoides . . . . . . . . . . . . . . . . . . . . . . . . 93
commersoniana . . . . . . . . . . . . . . . . . . . 483 Cupania oblongifolia . . . . . . . . . . . . . . . . . . 483
fuligunosa . . . . . . . . . . . . . . . . . . . . . . 328 Cupha erymanthis . . . . . . . . . . . . . . . . . . . . 128
Castanea . . . . . . . . . . . . . . . . . . . . . . . . . 226 Curculigo racemosa . . . . . . . . . . . . . . . . . . 445
Castanopsis . . . . . . . . . . . . . . . . . . . . . . . . 284 vilosa . . . . . . . . . . . . . . . . . . . . . . . . . 445
i
Casuarina equisetifolia . . . . . . . . . . . . . . . . 429 Curcuma . . . . . . . . . . . . . . . . . . . . . . . . . 442
Casuarius . . . . . . . . . . . . . . . . . . . . . . . .. 224 Curetis bulis . . . . . . . . . . . . . . . . . . . . . . . 293
Cataulacus .. . . . . . . . . . . . . . . . . . . . . . . 120 santana . . . . . . . . . . . . . . . . . . . . . . . . 293
Candacaecilia nigroflava . . . . . . . . . . . . . . . 317 bianthUS gardneri . . . . . . . . . . . . . . . . . . 483
Celtis luzonica . . . . . . . . . . . . . . . . . . . . . . 328 clemys dentata . . . . . . . . . . . . . . . . . . . . 155
~ptis' : : : : : : : : : : : : : : : : : : : : : : : : 1~
Centromyrme.x . . . . . . . . . . . . . . . . . . . . .. 11
Cepora ludith . . . . . . . . . . . . . . . . . . . . . . 131
pactolicus . . . . . . . . . . . . . . . . . . . . . . 131 nopterus shim . . . . . . . . . . . . . . . . . . . . 427
Ceratobolus concolor . . . . . . . . . . . . . . . . . 526 rina . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
Ceratosolen bisulcatus . . . . . . . . . . . . . . . .. 285 rtandra . . . . . . . . . . . . . . . . . . .. 46-48, 221
gressitti . . . . . . . . . . . . . . . . . . . . . . . . 285 bai/eyana . . . . . . . . . . . . . . . . . . . . . . . 221
medlerianus . . . . . . . . . . . . . . . . . . . . . 285 pendulijlora . . . . . . . . . . . . . . . . . . . . . 48
nexilis . . . . . . . . . . . . . . . . . . . . . . . . . 285 Cyrtodiopsis . . . . . . . . . . . . . . . . . . . 170, 173
Cercropia . . . . . . . . . . . . . . . . . . . . . . . .. 11 dalmanni . . . . . . . . . . . . . . . . . . . . . . . 173
Cerpachys . . . . . . . . . . . . . . . . . . . . . . . . 11 quinqueguttata . . . . . . . . . . . . . . . 170, 173
Cervus unicolor . . . . . . . . . . . . . . . .. 343,427 whitei .. . . . . . . . . . . . . . . .. 170, 172, 173
Ceyx erithacus . . . . . . . . . . . . . . . . .. 101, 105 Cyrtostachys renda . . . . . . . . . . . . . .. 444, 446
Chaetocarpus . . . . . . . . . . . . . . . . . . . . . . 483 Dacalana . . . . . . . . . . . . . . . . . . . . . . . . . 132
Chaetogonopteron . . . . . . . . . . . . . . . 185, 188 Dacrycarpus cumingianus . . . . . . . . . . . . . . . 329
Chaetospama feae . . . . . . . . . . . . . . . . . . . 92 Daemonorops didymophylla . . . . . . . . . . . . . 526
Chapenna fusca . . . . . . . . . . . . . . . . . . . . . 317 korthalsii . . . . . . . . . . . . . . . . . . . . . . . 526
Charaxes dumfordi . . . . . . . . . . . . . . . . . . . 132 longipes . . . . . . . . . . . . . . . . . . . . . . . . 526
Chela . . . . . . . . . . . . . . . . . . . . . 380-382, 384 oxycarpa . . . . . . . . . . . . . . . . . . . . . . . 526
Cheritra freja . . . . . . . . . . . . . . . . . . . . . . 293 periacantha . . . . . . . . . . . . . . . . . . . . . 526
Chersodromia . . . . . . . . . . . . . . . . . . 185, 186 Daphne . . . . . . . . . . . . . . . . . . . . . . . . . . 325
Chiropodomys gliroides . . . . . . . . . . . . . . . . 156 Dasyhela assimilis . . . . . . . . . . . . . . . . . 89, 93
karlkoopmani . . . . . . . . . . . . . . . . . . . . 427 grata . . . . . . . . . . . . . . . . . . . . . . . 89, 93
Chisocheton pentandrus . . . . . . . . . . . . . . . . 328 Dasyprocta punctata ... . . . . . . . . . . . . . . . 226
Chrysopelea pelias . . . . . . . . . . . . . . . . . . . 318 Decamorium . . . . . . . . . . . . . . . . . . . . . .. 11
Chrysosoma . . . . . . . . . . . . . . . . . . . 185, 187 Delonix regia . . . . . . . . . . . . . . . . . . . . . . 462
Chrysotus . . . . . . . . . . . . . . . . . . . . . . . . . 185 Dendrelaphis caudolineatus . . . . . . . . . . . . . . 318
Cinnamomum laubatii . . . . . . . . . . . . . . . . . 224 Dendrocalamus pendulus . . . . . . . . . . . . . . . 97
mercadoi . . . . . . . . . . . . . . . . . . . . . . . 327 Dendrocnide moroides . . . . . . . . . . . .. 226, 228
Cirrochroa emalea . . . . . . . . . . . . . . . . . . . 131 Diacamma . . . . . . . . . . . . . . . . . . . . . . . . 120
tyche . . . . . . . . . . . . . . . . . . . . . . . . . 131 intricatum . . . . . . . . . . . . . . . . . . . 331-338
Clarias leiacanthus . . . . . . . . . . . . 380,381,384 rugosum . . . . . . . . . . . . . . . . . . . . . . . 362
teijsmanni . . . . . . . . . . . . 309, 380-382, 384 Dialium . . . . . . . . . . . . . . . . . . . . . . . . . . 429
Cleistocalyx operculata . . . . . . . . . . . . . . . . 327 Dianella . . . . . . . . . . . . . . . . . . . . . . . . . . 445
Clethra . . . . . . . . . . . . . . . . . . . . . . 223, 329 Diaphorus . . . . . . . . . . . . . . . . . . . . . . . . 185
Clidemia capitellata . . . . . . . . . . . . . . . . .. 483 Dicerorhinus sumatraensis . . . . . . . . . . . . . . 155
ClinodiploslS . . . . . . . . . . . . . . . . . . . . . .. 93 Dichapetalum . . . . . . . . . . . . . . . . . . . . . . 441
Clogmw . . . . . . . . . . . . . . . . . . . .. 89, 90, 93 Dicrurus hottentotus . . . . . . . . . . . . . . . . . . 428
Clubiona . . . . . . . . . . . . . . . . . . . . . . . . . 92 leaucophaeus . . . . . . . . . . . . . . . . . . . . 428
Clusia . . . . . . . . . . . . . . . . . . . . . . . . . . . 483 Dicuspiditermes nemorosus .... 299, 300, 302, 303
Coccoloba . . . . . . . . . . . . . . . . . . . . . . . . 483 Didymocarpus . . . . . . . . . . . . . . . . . . . . .. 46
Colobopsis . . . . . . . . . . . . . . . . . . . . 120,362 lanceolatus . . . . . . . . . . . . . . . . . . . . .. 48
Commersonia . . . . . . . . . . . . . . . . . . . . . . 223 Dinoderus . . . . . . . . . . . . . . . . . . . . . . . .. 90
bartramia . . . . . . . . . . . . . . . . . . . . . . . 227 Dioscorea . . . . . . . . . . . . . . . . . . . . . . . . . 463
Compterosmittia . . . . . . . . . . . . . . . . . . . .. 93 Diospyros . . . . . . . . . . . . . . . . . . . .. 413,441
Conarthrus ... . . . . . . . . . . . . . . . . . . . 90, 93 copelandii . . . . . . . . . . . . . . . . . . . . . . 328
Conomorpha oblongifolia . . . . . . . . . . . . . .. 483 philippense . . . . . . . . . . . . . . . . . . . . . . 328
Coproporus . . . . . . . . . . . . . . . . . . . . . . .. 93 Dip7azium . . . . . . . . . . . . . . . . . . . . . . . 47, 48
Copsychos saularis . . . . . . . . . . . . . . . . . . . 428 Dlplazium crenatoserratum . . . . . . . . . . . . .. 48
pyrrhopygus . . . . . . . . . . . . . . . . . . . . . 107 Dlplodiscus paniculatus . . . . . . . . . . . . . . . . 328
Coptotermes . . . . . . . . . . . . . . . . . . . . . . . 301 Dlpterocarpus . . . . . . . . . . . . . . . 413, 428, 429
Cordia nodosa . . . . . . . . . . . . . . . . .. 483, 484 Dlpterocarpus grandijlorus . . . . . . . . . . . . . . 235
superba . . . . . . . . . . . . . . . . . . . . . . .. 483 Dlscolampa . . . . . . . . . . . . . . . . . . . . . . . . 293
Coryfus . . . . . . . . . . . . . . . . . . . . . . . . . . 226 Discothyrea . . . . . . . . . . . . . . . . .. 11, 120, 362
Corymborkis . . . . . . . . . . . . . . . . . . . . . . . 445 Dissochaeta . . . . . . . . . . . . . . . . . . . . . . . 226
Coscinida . . . . . . . . . . . . . . . . . . . . . . . " 92 Dogania subplana . . . . . . . . . . . . . 155,318,319
Costus globosus . . . . . . . . . . . . . . . . . . . . . 447 Dolichoderus . . . . . . . . . . 10, 117, 119, 120, 122
speciosus . . . . . . . . . . . . . . . . . . . . . .. 48 thoracicus . . . . . . . . . . . . . . . . . . . . . . 122
Couepia impressa. . . . . . . . . . . . . . .. 483, 484 Donax . . . . . . . . . . . . . . . . . . . . . . . . . . . 446
Cratoxylon arborescens . . . . . . . . . . . . . . . . 413 Dorylus . . . . . . . . . . . . . . . . . . . . . . . . . . 11
arboreum . . . . . . . . . . . . . . . . . . . . . . . 328 Draco cristatellus . . . . . . . . . . . . . . . . . . . . 318
fomwsum . . . . . . . . . . . . . . . . . . . . . . . 328 melanopogon . . . . . . . . . . . . . . . . . 97,318
sumatranum . . . . . . . . . . . . . . . . . . . . . 328 quinquefasciatus . . . . . . . . . . . . . . . . . . 318
Crematogaster . . . . . . . .. 10, 120, 122, 292, 293 Dracontomelon edule . . . . . . . . . . . . . . . . . . 327
Criniger bres . . . . . . . . . . . . . . . . . .. 105, 107 Drapetis . . . . . . . . . . . . . . . . . . . . .. 185, 186
Crocldura fuliginosa . . . . . . . . . . . . . . . . . . 177 Drimys . . . . . . . . . . . . . . . . . . . . . . . . . . . 326
Crossopalpus . . . . . . . . . . . . . . . . . . . . . . . 185 piperita . . . . . . . . . . . . . . . . . . . . . . . . 329
Croton selowii . . . . . . . . . . . . . . . . . . . . . . 483 Drupiulia . . . . . . . . . . . . . . . . . . . . . . . . . 293
Cryptocarya . . . . . . . . . . . . . . . . . . . . . . . 328 ravindra . . . . . . . . . . . . . . . . . . . . . . . 292
ampla . . . . . . . . . . . . . . . . . . . . . . . . . 327 theda . . . . . . . . . . . . . . . . . . . . . . . . . 293
angulata . . . . . . . . . . . . . . . . . . . . . . . 224 Dryobalanops . . . . . . . . . . . . 193.. 228, 277.. 413
grandis .. . . . . . . . . . . . . . . . . . . . . . . 224 aromatlca . . . . . . . . . . . . 27/,280,34/-355
oblata . . . . . . . . . . . . . . . . . . . . . . . . . 224 beccarii . . . . . . . . . . . . . . . . . . 205-207.. 412
pleurosperma . . . . . . . . . . . . . . . . 224, 225 lanceolata . . . . . . . . . . . . . . . . . . . 36/-371
Crytophleps . . . . . . . . . . . . . . . . . . . . . . . . 185 Drypdes . . . . . . . . . . . . . . . . . . . 367-371,373
Culex . . . . . . . . . . . . . . . . . . . . . . . . . . . 93 DUiibanga moluccana . . . . . . . . . . . . . . . . . 328
GENUS AND SPECIES INDEX 549
Dundubia vaginata .. . . . . . . . . . . . . . . . ., 81 pubine rvis . . . . . . . . . . . . . . . . . . . . . . 328

Durio . . . . . . . . . . . . . . . . . . . . . . . . . . . 429 robusta . . . . . . . . . . . . . . . . . . .. 285, 286
[)yera costulata . . . . . . . . . . . . . . . . . . . . . 143 septica . . . . . . . . . . . . . . . . . 285, 286, 327
Dysoxylon arborescens . . . . . . . . . . . . . . . . . 328 trachypison . . . . . . . . . . . . . . . . . 285, 286
altissimum . . . . . . . . . . . . . . . . . . . . . . 327 variegata . . . . . . . . . . . . . . . . . . , 288, 328
Echinopla . . . . . . . . . . . . . . . . . . . . . . . . . 120 virgata . . . . . . . . . . . . . . . . . . . . 285, 286
Echinosorex gymnurus . . . . . . . . . . . . . . . . . 156 wassa . . . . . . . . . . . . . . . . . . . . , 285,286
Elaeocarpus . . . . . . . . . . . . . . . . 166, 167,225 Flagellaria indica . . . . . . . . . . . . . . . . . . . . 439
angustifolius . . . . . . . . . . . . . . . . . . . . . 225 Flindersia brayeleyana . . . . . . . . . . . . . . . . . 220
calomala " . . . . . . . . . . . . . . . . . . . . . 327 Flos morphina . . . . . . . . . . . . . . . . . . . . . . 132
Elaphe taeniura . . . . . . . . . . . . . . . . . . . . . 155 Fontainea picrosperma . . . . . . . . . . . .. 224, 225
Elaphropeza . . . . . . . . . . . . . 183,185-187,191 Forclp'omyta . . . . . . . . . . . . . . . . . . . . . 89-91
Elateriospermum tapos . . . . . . . . .. 139, 141, 143 blkanni . . . . . . . . . . . . . . . . . . . . . . . . 93
Elatostema . . . . . . . . . . . . . . . . . . . . . . . . 48 similans . . . . . . . . . . . . . . . . . . . . . . .. 93
Elephas maximus . . . . . . . . . . . . . . . . . . . . 155 swezeyanaadji.nis .... .. .. .. .. .. .... 93
Elettariopsis kerbyi . . . . . . . . . . . . . . . . . 48, 50 Fraxinus americana . . . . . . . . . . . . . . . . . . 220
Eleusis . . . . . . . . . . . . . . . . . . . . . . . . . .. 93 Fulvius . . . . . . . . . . . . . . . . . . . . . . . . . . 92
Embal/onura monticola . . . . . . . . . . . . . . . . 427 Galearia falva . . . . . . . . . . . . . . . . . . . . . . 441
Emesopsis . . . . . . . . . . . . . . . . . . . . . . . 91, 93 Galeopsis tetrahit . . . . . . . . . . . . . . . . . . . . 227
Emmotum affine . . . . . . . . . . . . . . . . . . . .. 483 Gamasomorpha . . . . . . . . . . . . . . . . . . . . . 92
Empis . . . . . . . . . . . . . . . . . . . . . . . . . . . 185 Garcinia binucao . . . . . . . . . . . . . . . . . . . . 328
Eniiiandra bessaphila . . . . . . . . . . . . . . . .. 224 Gastromyzon bomeensis ... 309,311,312,380-384
insignis . . . . . . . . . . . . . . . . . . . . . . . . 224 [asciatus . . . . . . . . . . . . . . . . . . . . 380-384
monothyra . . . . . . . . . . . . . . . . . . . . . . 224 lepidogaster .... . . . . . . . . . . . 309, 380-384
montana . . . . . . . . . . . . . . . . . . . . . .. 224 monticola . . . . . . 309,311,312,380,381,384
palmerstonii . . . . . . . . . . . . . . . . . . . . . 224 punctulatus . . . . . . . . . . . . 380,381,383,384
sankeyana . . . . . . . . . . . . . . . . . . . . .. 224 ridens . . . . . . . . . . . . . . . 309,380,381,384
xanthocarpa . . . . . . . . . . . . . . . . . . . . . 224 Gecinulus viridis . . . . . . . . . . . . . . . . . . . . 94
Enhydris punctata . . . . . . . . . . . . . . . . . . . . 318 Gehyra mutilatus . . . . . . . . . . . . . . . . . . . . 318
Enicurus. . . . . . . . . . . . . . . . . . . . .. 101, 105 Gekko smithii . . . . . . . . . . . . . . . . . . . . . . 3 18
&~ ......................... ~ Geniostoma . . . . . . . . . . . . . . . . . . . . . . . . 223
Eooxy lides tharis . . . . . . . . . . . . . . . . . . . . 292 Geonoma blanchetiana . . . . . . . . . . . . . . . . 483
Epicahris cumingiana . . . . . . . . . . . . . . . .. 328 Gigantochloa latifolia . . . . . . . . . . . . . . . . . 97
Epicypta . . . . . . . . . . . . . . . . . . . . . . . . 91,93 levis . . . . . . . . . . . . . . . . . . . . . . . . .. 97
Epipremnum falcifolium . . . . . . . . . . . . . . .. 48 ligulata . . . . . . . . . . . . . . . . . . . . . . .. 97
Eplrixanthes . . . . . . . . . . . . . . . . . . . . . 47, 48 scortechinii . . . . . . . . . . . 85, 87, 88, 91, 97
Erites argentata . . . . . . . . . . . . . . . . . . . . . 131 Glaniopsis gossei . . . . . . . . . . . . . . . . . 380-384
elegans .. . . . . . . . . . . . . . . . . . . . . . . 131 hamtsclii . . . . . . . . . . . . . . . . . . . . 380-384
Ervatamia . . . . . . . . . . . . . . . . . . . . . . . . . 328 Globba . . . . . . . . . . . . . . . . . . . . . . . . . . 155
Erythroxylum mattos-silvii . . . . . . . . . " 483, 484 brachyanthera . . . . . . . . . . . . . . . . . . . . 443
Etlingera . . . . . . . . . . . . . . . . . . . . . . . . . 47 Glochidion . . . . . . . . . . . . . . . . . . . . 223, 373
brevilabris . . . . . . . . . . . . . . . . . . . . .. 48 Gluta laxiflora . . . . . . . . . . . . . .. 139, 141, 143
ji.mbribracteata . . . . . . . . . . . . . . . . . . . 48
punicea . . . . . . . . . . . . . . . . . . . . . 48,442 Gly;~Ih' major: : : : : : : : : '309,' 3'80: 381', j§~
Euenkrates . . . . . . . . . . . . . . . . . . . . . . . . 92 Gmelina arborea . . . . . . . . . . . . . 277,278,280
Eugeissona triste . . . . . . . . . . . . . . . . . . . . 348 Gnamptogenys . . . . . . . . . . . . . . . . . . . . .. 11
minor . . . . . . . . . . . . . . . . . . . " 444, 446 Gnetum . . . . . . . . . . . . . . . . . . . . . . . . . . 441
Eugenia grandis . . . . . . . . . . . . . . . .. 228, 429 Goniothalamus amuyon . . . . . . . . . . . . . . . . 328
Eumorphus vitalisi . . . . . . . . . . . . . . . . . . . 93 Gonocephalus bomeensis . . . . . . . . . .. 318, 321
Eumyias thalassina . . . . . . . . . . . . . . . . . . . 108 Gonydactylus baluensis . . . . . . . . . . . . . . . . 318
Euodia xanthoxyloides . . . . . . . . . . . . . . . . . 219 consobnnus . . . . . . . . . . . . . . . . . . 318-320
Eupatorium . . . . . . . . . . . . . . . . . . . . . . . . 363 ingeri . . . . . . . . . . . . . . . . . . . . . . . . . 318
adenophorum . . . . . . . . . . . . . . . . . . . . 516 pubisulcus . . . . . . . . . . . . . . . . . . . . . . 3 18
Euphoria didyma . . . . . . . . . . . . . . . . . . . . 327 Gonystylus bancanus . . . . . . . . . . . . . . . 411-415
Euprenolepis . . . . . . . . . . . . . . .. 120, 361, 362 Graphlum eurypylus . . . . . . . . . . . . . . 130, 131
Eurema . . . . . . . . . . . . . . . . . . . 130, 131, 133 evemon . . . . . . . . . . . . . . . . . . .. 130, 131
andersoni . . . . . . . . . . . . . . . . . .. 130, 131 Graptomyza . . . . . . . . . . . . . . . . . . . . . 93, 94
simulatrix . . . . . . . . . . . . . . . . . .. 130, 131 Grewia multiflora . . . . . . . . . . . . . . . . . . . . 328
Eurya . . . . . . . . . . . . . . . . . . . . . . . . . . . 327 Gymnostoma .. . . . . . . . . . . . . . . . . . . . .. 27
Eurydiopsis argentifera . . . . . . . . . . . . 170, 173 Hamaxas . . . . . . . . . . . . . . . . . . . . . . . . . 92
EUryOPIS . . . . . . . . . . . . . . . . . . . . . . . . .. 92 Hampala bimaculata . . . .. 309-312,380-382,384
Eusideroxylon zwageri . . . . . . . . . . . . . . . . . 327 Harpegnathos . . . . . . . . . . . . . . . . . . . . . . 120
Exilisciurus exilis . . . . . . . . . . . .. 177, 178, 343 Helarctos . . . . . . . . . . . . . . . . . . . . . . . .. 62
Fagraea . . . . . . . . . . . . . . . . . . . . . . . . . . 223 Helarctos malayanus . . . . . . . . . . . . . . . . . . 342
Fagus sylvatica . . . . . . . . . . . . . . . . . . . . . 227 Helopeltis theobromae . . . . . . . . . . . . . . . . . 122
Faunis . . . . . . . . . . . . . . . . . . . . 130, 131, 133 Hemulactylus frenatus .. . . . . . . . . . .. 318, 321
gracilis . . . . . . . . . . . . . . . . . . .. 130, 131 Hemigalus derbyanus . . . . . . . . . . . . . . . . . 427
stomphax . . . . . . . . . . . . . . . . . .. 130, 131 Hemiphyllodactylus typus . . . . . . . . . . . . . . . 318
Felis marmorata . . . . . . . . . . . . . . . . . . . . . 343 Hennettea succosa . . . . . . . . . . . . . . . . . . . 483
Feltiella . . . . . . . . . . . . . . . . . . . . . . . . .. 93 Heosemys grandis . . . . . . . . . . . . . . . . . . . . 156
Ficedula narcissina . . . . . . . . . . . . . . . . . . . 156 spinosa . . . . . . . . . . . . . . . . . 156,318,319
Ficus . . . . . . . . . . . . . . . . . . 283, 285, 325, 483 Herpestes semitorquatus . . . . . . . . . . . . . . . . 343
ampelas . . . . . . . . . . . . . . . . . . . . . . . . 327 Hesperus perakensis . . . . . . . . . . . . . . . . . . 93
balete . . . . . . . . . . . . . . . . . . . . . . . . . 327 Hibiscus . . . . . . . . . . . . . . . . . . . . . . . . . . 429
botryocarpa . . . . . . . . . . . . . . . . . . . . . 328 Hipposideros galeritus . . . . . . . . . . . . . . . . . 427
carica . . . . . . . . . . . . . . . . . . . . . . . . . 288 Hlrtella bahiensis . . . . . . . . . . . . . . .. 483, 484
complexa . . . . . . . . . . . . . . . . . " 285, 286 Homalanthus . . . . . . . . . . . . . . . . . . . . . . . 328
congesta . . . . . . . . . . . . . . . . . . . . . . . 328 Homallotermes . . . . . . . . . . . . . . . . .. 300, 302
erythrosperma . . . . . . . . . . . . . . " 285, 286 foraminifer . . . . . . . . . . . . . . . . . , 299, 302
grossularoides . . . . . . . . . . . . . . . . . . . . 220 Homalomena cordata . . . . . . . . . . . . . . . . . 445
gul . . . . . . . . . . . . . . . . . . . . . . . . . . . 327 Homo sapiens . . . . . . . . . . . . . . . . . . . . . . 470
minahassae . . . . . . . . . . . . . . . . . . . . . 327 Hopea . . . . . . . . . . . . . . . . . . . . . . . . . . . 277
mollior . . . . . . . . . . . . . . . . . . . . 285, 286 Horaga syrinx . . . . . . . . . . . . . . . . . . . . . . 132
nodosa . . . . . . . . . . . . . . . . . . . . 285,286 Homstedtia reticulata . . . . . . . . . . . . . . . . . 442
nota . . . . . . . . . . . . . . . . . . . . . . . . . . 327 Hospitalitermes medioflavus . . . . . . . . . 299,300
pertusa . . . . . . . . . . . . . . . . . . . . . . . . 287 Hynos . . . . . . . . . . . . . . . . . . . . . . . 185, 186
Hydnocarpus . . . . . . . . . . . . . . . . . . . . . . . 428 Litsea . . . . . . . . . . . . . . . . . . . . . . . . . . . 413
sumatrana . . . . . . . . . . . . . . . . . . 141, 143 acuminata . . . . . . . . . . . . . . . . . . . . . . 217
Hydrablabes periops ... . . . . . . . . . . . . . . . 318 leefeana . . . . . . . . . . . . . . . . . . . . . . . 224
Hydrometra . . . . . . . . . . . . . . . . . . . . . . .. 95 perrotetii . . . . . . . . . . . . . . . . . . . . . . . 328
Hylobates . . . . . . . . . . . . . . . . . . . . . . . .. 74 Lobocheilus bo . . . . . . . . 309-312,380,381,384
klossii . . . . . . . . . . . . . . . . . . . .. 427, 429 Logania . . . . . . . . . . . . . . . . . . . . .. 291,293
lar . . . . . . . . . . . . . . . . . . . . . . . . . . . 155 maiayica . . . . . . . . . . . . . . . . . . . 293,294
muelleri . . . . . . . . . . . . . . . . . . .. 343, 344 maTTTWrata . . . . . . . . . . . . . . . . . . . . . . 294
Hylopetes sipora . . . . . . . . . . . . . . . . . . . . 427 Longipeditermes longipes . . . . . . . . 299,300,304
Hypoclinea . . . . . . . . . . . . . . . . . . . . . . . . 294 Lophatherium gracile . . . . . . . . . . . . . . . . . 445
thoracicus . . . . . . . . . . . . . . . . . . . . . . 294 Lophomyrmex . . . . . . . . . . . . . . . . . . . . . . 120
Hypolycaena erylus . . . . . . . . . . . . . . . . . . . 293 Lycodon subcinctus . . . . . . . . . . . . . . . . . . . 318
othona . . . . . . . . . . . . . . . . . . . . . . . . 292 Lygodium circinnatum . . . . . . . . . . . . . . . .. 47
Hypoponera . . . . . . . . . . . . . . . . . . . 120, 362 Lyonia . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
Hypothymis azurea . . . . . . . . . . . . . . . . . . . 428 Lyriothemis cleis . . . . . . . . . . . . . . . . . . . . 94
Hypsiprymnodon TTWSChatus . . . . . . . . . . . . . 224 Mabuya rudis . . . . . . . . . . . . . . . . . . . . . . 318
Hystnx brachyura . . . . . . . . . . . . . . . . . . . . 156 Macaca . . . . . . . . . . . . . . . . . . . . . . . . . . 74
llex . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 483 arctoides . . . . . . . . . . . . . . .. 153, 155, 156
Imperata cylindrica . . . . . . . . . . . . . . . . . . . 516 fascicularis . . . . . . . . . . . . . . . . . . . . . . 343
lnilalmus . . . . . . . . . . . . . . . . . . . . . . . . . 93 nemestrina . . . . . . . . . . . . . . . . .. 343, 344
Indotestudo elongata . . . . . . . . . . . . . . . . . . 156 pagensis . . . . . . . . . . . . . . . . . 74,427,429
Inga . . . . . . . . . . . . . . . . . . . . . . . . 219, 229 Macaranga . . . . . . . . . . . . . . . .. 122, 223, 373
lomys sipora . . . . . . . . . . . . . . . . . . . . . . . 427 gigantea . . . . . . . . . . . . . . . . . . . . . . . 278
IpoTTWea batatas . . . . . . . . . . . . . . . .. 271, 275 heynei . . . . . . . . . . . . . . . . . . . . . . . . . 223
Iridomyrmex . . . . . . . . . . . . . . . . . . . 120, 293 hypo/euca . . . . . . . . . . . . . . . . . . . . . . 363
cordatus . . . . . . . . . . . . . . . . . . . . . . . 11 trachyphylla . . . . . . . . . . . . . . . 367-371,373
!tara .. . . . . . . . . . . . . . . . . . . . . . . . . 82, 83 Macroglossus sobrinus . . . . . . . . . . . . . . . . . 427
Ixora . . . . . . . . . . . . . . . . . . . . . . . . . . . . 444 Macrognathusmaculatus . 309,311,380,381,384
1acq.ueTTWntia . . . . . . . . . . . . . . . . . . . . . . 363 Macropygia phasianella . . . . . . . . . . . . . . . . 428
lanudes . . . . . . . . . . . . . . . . . . . . . . . . . . 292 Macrotermes . . . . . . . . . . . . . . . . . .. 302, 303
lohora johorensis . . . . . . . . . . . . . . . . . . .. 95 carbonarius . . . . . . . . . . . . . . . . . 302, 304
luglans . . . . . . . . . . . . . . . . . . . . . . 224, 226 Malacopteron cinereum . . . . . . . . . . . . 105, 107
Kalophrynus intermedius . . . . . . . . . . . . . .. 317 magnum . . . . . . . . . . . . . . . . . . . . . . . 107
pleurostigma . . . . . . . . . . . . . . . . .. 94, 317 Mallotus . . . . . . . . . . . . . . . . . . . . .. 223, 373
Kenvoula luirwickei . . . . . . . . . . . . . . . . . . 427 multiglandulosus . . . . . . . . . . . . . . . . . . 327
Ketupa zey/onensis . . . . . . . . . . . . . . . . . . . 156 paniculatus . . . . . . . . . . . . . . . . . . . . . . 278
Koompasia . . . . . . . . . . . . . . . . . . . . . . . . 61 philippinensis . . . . . . . . . . . . . . . . . . . . 328
Kortlialsia debilis . . . . . . . . . . . . . . . . . . . . 526 Manihot utilissima . . . . . . . . . . . . . . . . . . . 441
echinometra . . . . . . . . . . . . . . . . . . . . . 526 Manis javanica . . . . . . . . . . . . . . . . . 156, 427
furtadoana . . . . . . . . . . . . . . . . .. 363, 526 Manouria emys . . . . . . . . . . . . . . . . . . . . . 319
hispida . . . . . . . . . . . . . . . . . . . . . . . . 526 Mapania . . . . . . . . . . . . . . . . . . . . . . . . . 167
rostrata . . . . . . . . . . . . . . . . . . . . . . . . 526 TTWnostachya . . . . . . . . . . . . . . . . . . . 48, 50
Krakatauia . . . . . . . . . . . . . . . . . . . . . . . . 185 palustris . . . . . . . . . . . . . . . . . . . . . 48, 50
Krytopterus . . . . . . . . . . . . . . . . . . 380-382, 384 sumatrana . . . . . . . . . . . . . . . . . . . . . . 443
Labntermes . . . . . . . . . . . . . . . . . . . . . . . . 300 Margattea . . . . . . . . . . . . . . . . . . . . . . . .. 92
Lacconectus krikkeni . . . . . . . . . . . . . . . . .. 94 Mastacembelus unicolor . . . . . . . . . 380,381,384
Landolphia .. . . . . . . . . . . . . . . . . . . . . .. 462 Maticora bivirgatus . . . . . . . . . . . . . . . . . . . 318
Lansium dubium . . . . . . . . . . . . . . . . . . . .. 328 intestinalis . . . . . . . . . . . . . . . . . . . . . . 318
Laportea /uzonensis . . . . . . . . . . . . . . . . . . . 328 Maxburetia rupicola . . . . . . . . . . . . . . . . . . 155
Laricus obscurus . . . . . . . . . . . . . . . . . . . . 427 Maxillaria . . . . . . . . . . . . . . . . . . . . . . . . 64
hosei . . . . . . . . . . . . . . . . . . . . . . . . . 343 Maxomys muelleri . . . . . . . . . . . . . . . . . . . . 175
Lasiochila goryi . . . . . . . . . . . . . . . . . . . .. 94 rajah . . . . . . . . . . . . . . . . . . . 175, 177-181
Lasiochilus . . . . . . . . . . . . . . . . . . . . . . .. 92 surifer . . . . . . . . . . . . . . . . . . . . 175, 177-181
Lasius . . . . . . . . . . . . . . . . . . . . . . . . . . . 120 whlteheadl . . . . . . . . . . . . . . . . . .. 177 -180
Laxita tenata . . . . . . . . . . . . . . . . . . . . . . . 13 2 Mayriella .. . . . . . . . . . . . . . . . . . . . . . . . 120
Leichrodes rufofulvus . . . . . . . . . . . . . . . . .. 93 Medetera . . . . . . . . . . . . . . . . . . . . . 185, 187
Leiocassis micropogon . . . . 310-312, 380, 381.. 384 Megaderma spasma . . . . . . . . . . . . . . . . . . . 427
Leopoldamys sabanus . . . . . . . . . . . . . . 17/-180 Megalabops sexguttata . . . . . . . . . . . .. 170, 173
Lepidotrigona . . . . . . . . . . . . . . . . . . . . . . 62 Megase/ia . . . . . . . . . . . . . . . . . . . . . . . .. 87
Lepironia articulata . . . . . . . . . . . . . . . . . . 443 animaloterga . . . . . . . . . . . . . . . . . . . . 93
Lepisanthes schizolepis . . . . . . . . . . . . . . . .. 328 deningi . . . . . . . . . . . . . . . . . . . . . . 93, 97
Leptanilla . . . . . . . . . . . . . . . . . . . . . . . .. 11 gombakensis . . . . . . . . . . . . . . . . . . . .. 93
Leptobrachium TTWntanum . . . . . . . . . .. 317, 321 humida . . . . . . . . . . . . . . . . . . . . . . 90, 93
Leptogenys . . . . . . . . . . . . . . . . . . . .. 11, 120 kovaci . . . . . . . . . . . . . . . . . . . . . . . .. 93
Leptolalax gracilis . . . . . . . . . . . . . . . 317,321 Megophrys baluensis . . . . . . . . . . . . . . 316,317
Leptosia nina. . . . . . . . . . . . . . . . . .. 128, 129 nasuta . . . . . . . . . . . . . . . . . . . . . . . . . 317
Leptospermum f/avescens . . . . . . . . . . . . . . . 329 Melaleuca . . . . . . . . . . . . . . . . . . . . . . . .. 68
Lestodiplosis . . . . . . . . . . . . . . . . . . . . . .. 93 Melastoma . . . . . . . . . . . . . . . . . . . . . . . . 223
synngopais . . . . . . . . . . . . . . . . . . . . .. 92 malabathricum . . . . . . . . . . . . . . . 227, 228
Leucaena leucocephala . . . . . . . . . . . . 271, 275 Melicope broadbentiana . . . . . . . . . . . . . . . . 219
Leucosyke capiteliata . . . . . . . . . . . . . . . . .. 328 triphylla . . . . . . . . . . . . . . . . . . . . . . . 328
Lexias dirtea . . . . . . . . . . . . . . . . . . . . . . . 131 Melims multiflora . . . . . . . . . . . . . . . . . . . . 255
Licg;j~da. :: : :: : : :: :: : :: : :: : : :: : :: m
Lichtwardtia . . . . . . . . . . . . . . . . . . . . . . . 185
Memecylon . . . . . . . . . . . . . . . . . . . . . . . . 327
Meranoplus . . . . . . . . . . . . . . . . . . . . . . . . 120
mucronatus . . . . . . . . . . . . . . . . . . . . . 292
Licuala . . . . . . . . . . . . . . . . . . . . . . 443, 444 Meristogenys poecilus . . . . . . . . . . . . . . . . . 317
orbicularus . . . . . . . . . . . . . . . . . . . . .. 444 Merizocera . . . . . . . . . . . . . . . . . . . . . . .. 92
ramseyi . . . . . . . . . . . . . . . . . . . . . . . . 23 Merremia . . . . . . . . . . . . . . . . . . . . . . . . . 363
spinosa . . . . . . . . . . . . . . . . . . . . . . .. 444 bomeensis . . . . . . . . . . . . . . . . . . . . . . 445
LindSayomyrtus raceTTWides . . . . . . . . . . . . .. 27 Mesophlebion falcatilobum . . . . . . . . . . . . . . 50
Linospadix . . . . . . . . . . . . . . . . . . . . . . . . 23 Mesorhaga . . . . . . . . . . . . . . . . . . . . . . . . 185
Liopeltis baliodeirus . . . . . . . . . . . . . . . . . . 318 Metaparaclius . . . . . . . . . . . . . . . . . . . . . . 185
tricolor . . . . . . . . . . . . . . . . . . . . . . . . 318 Metaphrynella sumiana . . . . . . . . . . . 80, 81, 317
Liphorrhopalum sessilis . . . . . . . . . . . . . . . . 285 Metroxylon sagu . . . . . . . . . . . . . . . . . . . . . 429
Llthocarpus . . . . . . . . . . . . . . . . . . .. 284,327 Miconia . . . . . . . . . . . . . . . . . . . . . . . . . . 229
densiflora . . . . . . . . . . . . . . . . . . . . . .. 219 albicans . . . . . . . . . . . . . . . . . . . . . . . . 483
550
Miconia amoena . . . . . . . . . . . . . . . . . . . . . 483 Oriolus chinensis . . . . . . . . . . . . . . . . . . . . 428

pileata . . . . . . . . . . . . . . . . . . . . . . . . 483 xantonotus . . . . . . . . . . . . . . . . . . . ... 428
MicreTr/{Jis . . . . . . . . . . . . . . . . . . . . . . . . 185 Orthopodomyia anopheloides . . . . . . . . . . . 89, 93
Micrixalus baluensis . . . . . . . . . . . . . . . . . . 317 Oryza sativa . . . . . . . . . . . . . . . . 241, 246, 252
Microcerotermes . . . . . . . . . . . . . . . . . . . . 300 Ossaea retropila . . . . . . . . . . . . . . . . . . . . . 483
dubius . . . . . . . . . . . . . . . . . . . . . . . . . 304 Osteochilus spilurus . . . . . 309-312,380, 381, 384
Microdesmis puberula . . . . . . . . . . . . . . . . . 219 Ostrya virginiana . . . . . . . . . . . . . . . . . . . . 219
Microhyla bomeensis . . . . . . . . . . . . . . .... 317 Ottoma . . . . . . . . . . . . . . . . . . . . . . . . . . 483
petrigena . . . . . . . . . . . . . . . . . . . 317,318 Qxyceros . . . . . . . . . . . . . . . . . . . . . . . . . 441
Microvelia . . . . . . . . . . . . . . . . . . . . . . . . 95 Paedalgus . . . . . . . . . . . . . . . . . . . . . . . . . 120
Mikania micrantha . . . . . . . . . . . . . . . . . . . 278 Pagiantha plumerilifolia . . . . . . . . . . . . . . . . 328
Miletus . . . . . . . . . . . . . . . . . . . . . . . . . . 294 Palaquium . . . . . . . . . . . . . . . . . . . . 328,428
Mischopyga . . . . . . . . . . . . . . . . . . . . . . . . 185 [oxworthii . . . . . . . . . . . . . . . . . . . . . . 327
Misumenops . . . . . . . . . . . . . . . . . . . . . . . 95 luzoniensis . . . . . . . . . . . . . . . . . . . . . . 327
Monocarpia . . . . . . . . . . . . . . . . . . . . . . . 167 merrillii . . . . . . . . . . . . . . . . . . . . . . . . 327
Monomorium . . . . . . . . . . . . . . . . . . . . . .. 10 philippense . . . . . . . . . . . . . . . . . . . . . . 327
Moultonianthus leembrugianus . . . . . . .. 141, 143 Pandanus . . . . . . . . . . . . . . . . . . . . . . . . . 429
Muntiacus muntjac . . . . . . . . . . . . . . . 343, 344 Pangio mariarum . . . . . . . . . . 309,380,381,384
Musa . . . . . . . . . . . . . . . . 48,367-371,373,443 Panthera pardus . . . . . . . . . . . . . . . . . . . . . 156
acuminata . . . . . . . . . . . . . . . . . . . . . . 278 Paraclius ... . . . . . . . . . . . . . . . . . . . . . . 185
. gracilis . . . . . . . . . . . . . . . . . . . . . . . . 443 Paracrossochilus acerus . .. 309-312, 380-382, 384
sapientum . . . . . . . . . . . . . . . . . . . . . . 443 vittatus . . . . . . . . . . . . . . . . . 380,381,384
Musclcapa daurica . . . . . . . . . . . . . .. 101, 105 Paradoxurus hermaphroditus . . . . . . . . . . . . . 427
Mustela nudipes . . . . . . . . . . . . . . . . . . . . . 343 Paralaxita orphne . . . . . . . . . . . . . . .. 131, 132
. . . . . . . . . . . . . . . . . . . . . . .. 11 telesia . . . . . . . . . . . . . . . . . . . .. 131, 132
~
opne
otis muricela . . . . . . . . . . . . . . . . . . . . . 427 Paramapania . . . . . . . . . . . . . . . . . . . . . . . 443
rcia . . . . . . . . . . . . . . . . . . . . . . . . . . . 483 Paranticops delesserti . . . . . . . . . . . . . 128.. 129
rianthus liberica . . . . . . . . . . . . . . . . . . . 219 Parashorea macrophylla . . . . . . . . . . . . . 36/-371
rioneuron cyaneum . . . . . . . . . . . . . . . .. 46 malaanonan . . . . . . . . . . . . . . . . . . . . . 327
yristica elliptica . . . . . . . . . . . . . . . . . . . . 328 Paratopula . . . . . . . . . . . . . . . . . . . . . . . . 120
philippinensis . . . . . . . . . . . . . . . . . . . . 328 Paratrechina . . . . . . . . . . . . . . . . . . . . . . . 120
. . . . . . . . . . . . . . . . . . .. 120, 122 Parawaous . . . . . . . . . . . . . . . . . 380,381, 384
~ rmecodia . . . . . . . . . . . . . . . . . . . . . . . 11
rmeclna
Pariotrigona . . . . . . . . . . . . . . . . . . . . . .. 59
. . . . . . . . . . . . . . . . . . . .. 10,294 pendleburyi . . . . . . . . . . . . . . . . . . . 61, 62
~
rmica
rmoteras . . . . . . . . . . . . . . . . . . . . . . . . 120 Parrhinotermes . . . . . . . . . . . . . . . . . . . . . 300
rsine umbellata . . . . . . . . . . . . . . . . . . . 483 Paspalum conjugatum . . . . . . . . . . . . . . . . . 443
MYStrium . . . . . . . . . . . . . . . . . . . . . . . . . 120 Pavetta . . . . . . . . . . . . . . . . . . . . . . . . . . 444
¥ystrops . . . . . . . . . . . . . . . . . . . . . . . . . 478 Pedicellia fuscescens . . . . . . . . . . . . . . . . . . 327
Mystus . . . . . . . . . . . . . . . . . . . . . 380-382, 384 Pedostibes hosei . . . . . . . . . . . . . . . . . . . . . 317
baramensis . . . . . . . . . 309-312,380,381,384 rugosus . . . . . . . . . . . . . . . . . . . . . . . . 317
Nannosciurus melatonis . . . . . . . . . . . . .... 343 Pelastoneurus . . . . . . . . . . . . . . . . . . . . . . 95
Nanodromia . . . . . . . . . . . . . 183, 185-187, 191 Pelophryne brevipes . . . . . . . . . . . . . . . . . . 317
Nauclea ojJicinalis . . . . . . . . . . . . . . . . . . . 223 Pentagonica . . . . . . . . . . . . . . . . . . . . . . . 93
Nedda serrata . . . . . . . . . . . . . . . . . . . . .. 46 Pentan . . . . . . . . . . . . . . . . . . . . . . . . . . . 429
Nemacheilus . . . . . . . . . . . . . 380, 381, 383, 384 Pentaphragma . . . . . . . . . . . . . . . . . . . . . . 441
Nematabramis nieuwenhuisi . . . . . . . . . . . . . 310 Pentastruma . . . . . . . . . . . . . . . . . . . . . . . 120
steindachneri . . . . . . . 309-312, 380-382, 384 Pera . . . . . . . . . . . . . . . . . . . . . . . . . . . . 483
Neofelis nebulosa . . . . . . . . . . . . . . . . . . . . 342 Peri cnemis stictica . . . . . . . . . . . . . . . . . . . 94
Neogastromyzon . . . . . . . . . . . . . . . . . . 380-384 Pemis apivorus . . . . . . . . . . . . . . . . . . . . . 428
nieuwelihuisi . . . . . . . . . . . . 309-312, 380-384 Persea . . . . . . . . . . . . . . . . . . . . . . . . . . . 483
Neolamarkia cadamba . . . . . . . . . . . . . . . . . 223 Petinomys hageni . . . . . . . . . . . . . . . . . . . . 427
Neolitsea . . . . . . . . . . . . . . . . . . . . . .... 328 Phacaspis . . . . . . . . . . . . . . . . . . . . . . . . . 185
dealbata . . . . . . . . . . . . . . . . . . . .... 224 PhaceliJphrynium maximum . . . . . . . . . . . . .. 48
Neonauclea media . . . . . . . . . . . . . . . . . . . 327 Phaenicopliaeus curvirostris . . . . . . . . . . . . . 428
Neotrewia cumingii . . . . . . . . . . . . . . . . . . . 328 Phaeomeria minor . . . . . . . . . . . . . . . . . . . 429
Nepenthes . . . . . . . . . . . . . . . . . . . . . . . .. 85 Phaleria capitata . . . . . . . . . . . . . . . . . . . . 328
distallatoria . . . . . . . . . . . . . . . . . . . . . 97 Pheidole . . . . . . . 93, 117, 119, 120, 122,361,362
Nephelium . . . . . . . . . . . . . . . . . . . . . . . . 328 Pheidologeton . . . . . . . . . . . . . . . 119, 120,361
Nesobia mjobergi . . . . . . . . . . . . . . . . . . . . 317 Philautus tectus . . . . . . . . . . . . . . . . . . . . . 317
Neurigona . . . . . . . . . . . . . . . . . . . . . . . . 185 Philosepedon . . . . . . . . . . . . . . . . . . . . . 90, 93
Niemeyera prunifera '" . . . . . . . . . .. 224, 225 Phragmites karka .. . . . . . . . . . . . . . . . . . . 443
Niewiedia . . . . . . . . . . . . . . . . . . . . . . . . . 445 Phruiloneura . . . . . . . . . . . . . . .. 185, 188, 190
Ninox scutulata . . . . . . . . . . . . . . . . . 10 I.. 105 abbreviatus . . . . . . . . . . . . . . . . . 188, 190
Niviventer cremoriventer . . . . . . . . . . . . 17/-180 Phyllotrox tatianae . . . . . . . . . . . . . . . . . . . 478
Noahdendron nicholasii . . . . . . . . . . . . 21,23-27 Physop)'ga . . . . . . . . . . . . . . . . . . . . . . . . 185
Nomamyrmex . . . . . . . . . . . . . . . . . . . . . . 11 Picea aoles . . . . . . . . . . . . . . . . . . . . . . . . 227
Normanbya normanbyi . . . . . . . . . . . . . . . .. 23 Picus viridanus . . . . . . . . . . . . . . . . . . . . . 156
Notochefys platynota . . . . . . . . 155, 156,318, 319 Pinanga salicifolia . . . . . . . . . . . . . . . . . . . 444
Nyctixalus pictus . . . . . . . . . . . . . . . . . . . . 317 Pinophilus . . . . . . . . . . . . . . . . . . . . . . . . 93
Nypa fruticans . . . . . . . . . . . . . . . . . . . . . . 429 Pinus radiata . . . . . . . . . . . . . . . . . . . . . . . 468
Occitfozyga baluensis . . . . . . . . . . . . . 317, 321 Piper . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229
laevLS . . . . . . . . . . . . . . . . . . . . . . . . . 317 amalgo . . . . . . . . . . . . . . . . . . . . . . . . 219
Ochroma lagopus . . . . . . . . . . . . . . . . . . . . 373 Piptospatha burbidgei . . . . . . . . . . . . . . . . . 48
Octomeles sumatrana . . . . . . . . . . . . . . . . . 328 Pipturus arborescens . . . . . . . . . . . . . . . . . . 327
Odontomachus . . . . . . . . . . . . . . . . . . . . . . 120 Pisonia . . . . . . . . . . . . . . . . . . . . . . . . . . 483
rixosus . . . . . . . . . . . . . . . . . . . . . . . . 362 umbelliferae .. . . . . . . . . . . . . . . . . . . . 328
Odontoponera . . . . . . . . . . . . . . . . . . 120, 362 Pithecus aygula . . . . . . . . . . . . . . . . . . . . . 71
transversa . . . . . . . . . . . . 331-336A 338, 339 femoralls . . . . . . . . . . . . . . . . . . . . . 71, 72
Odontotermes . . . . . . . . . . . . 297,29,,301-304 Pitta baudi .. . . . . . . . . . . . . . . . . . . . . . . 105
Oecophylla smaragdina . . . . . . . . . . . . 293,295 sordida ... . . . . . . . . . . . . . . . . . . . . . 156
Oligodon annulifer . . . . . . . . . . . . . . . . . . . 318 Plagiolepis .. . . . . . . . . . . . . . . . . . . . . . . 120
Oligomyrmex . . . . . . . . . . . . . . . . . .. 120, 122 Plagiostacftys crocydocalyx . . . . . . . . . . . . . . 442
Omalanthus . . . . . . . . . . . . . . . . . . . . .... 223 strobiliJera . . . . . . . . . . . . . . . . . . . . .. 48
Ompak . . . . . . . . . . . . . . . . . . . . . 380-382, 384 Plagiozopelma . . . . . . . . . . . . . . . . . . . . . . 185
Ophiophagus hannah . . . . . . . . . . . . . . . . . . 318 Plfdymiscium blanchetii .. . . . . . . . . . . . . . . 483
(Jpisthotropis typica . . . . . . . . . . . . . . . . . . 318 Platypalpus . . . . . . . . . . . . . . 183, 185, 186, 191
Platythyrea . . . . . . . . . . . . . . . . . . . . . . . . 120 ehalconota . . . . . . . . . . . . . . . . .. 317, 321

Plectroctena . . . . . . . . . . . . . . . . . . . . . . . 11 erythraea . . . . . . . . . . . . . . . . . .. 317,321
Pleiostachya . . . . . . . . . . . . . . . . . . . . ... 229 hosei . . . . . . . . . . . . . . . . . . . . . . . . . 317
pruinosa . . . . . . . . . . . . . . . . . . . . . . . 228 ibanomm . . . . . . . . . . . . . . . . . . . . . . . 317
Plowrium . . . . . . . . . . . . . . . . . . . . . . . . . 223 kuhli . . . . . . . . . . . . . . . . . . . 317.. 320,1 321
Podocarpus . . . . . . . . . . . . . . . . . . .. 325, 326 palavanensis . . . . . . . . . . . . . . . 31/, 3 h-321
rotundus . . . . . . . . . . . . . . . . . . . . . . . 329 signata . . . . . . . . . . . . . . . . . . . . 317,321
Po/Yathia .. . . . . . . . . . . . . . . . . . . . . . . . 441 Rapala dieneces . . . . . . . . . . . . . . . .. 291, 293
Po[ypedates colletti . . . . . . . . . . . . . . . . . . . 317 iarbus . . . . . . . . . . . . . . . . . . . . . . . . . 292
leueomystax . . . . . . . . . . . . . . . . . 317,321 pheretima . . . . . . . . . . . . . . . . . . . . . . 292
maerotls . . . . . . . . . . . . . . . . . . . . . .. 317 Rapanea . . . . . . . . . . . . . . . . . . . . . . . . . . 483
otilophus . . . . . . . . . . . . . . . . . . . . . .. 317 Rasbora . . . . . . . . . . . . . 309-311,380,381,384
Polypedi/um convexum . . . . . . . . . . . . . . . 89, 93 argyrotaenia . . . . . . .. 309-312, 380-382, 384
Polyrhachis . . . . . . . . . . . . . . . . . 119, 120, 122 sumatrana . . . . . . . . . 309-311, 380-382, 384
bihamata . . . . . . . . . . . . . . . . . . . . . . . 362 Rattus exulans . . . . . . . . . . . . . . . . 175, 177-181
Polyseias . . . . . . . . . . . . . . . . . . . . . . . . . 223 leueopus . . . . . . . . . . . . . . . . . . . . . . . 68
murrayi . . . . . . . . . . . . . . . . . . .. 220, 226 rattus . . . . . . . . . . . . . . . . . . . . . . . . . 180
Ponera . . . . . . . . . . . . . . . . . . . . . . . . . . . 362 sabanus . . . . . . . . . . . . . . . . . . . . . . . . 427
Pongo . . . . . . . . . . . . . . . . . . . . . . . . . . . 74 surifer . . . . . . . . . . . . . . . . . . . . . . . . . 427
Pontia eryeinoides . . . . . . . . . . . . . . . . . . . 132 tiomanicus . . . . . . . . . . . . . . . . 177-181,427
Porterandia anisophylla . . . . . . . . . . . . . . . . 227 Ratufa affinis . . . . . . . . . . . . . 177, 178, 343, 344
Pothos . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 RhacopFiorus dulitensis . . . . . . . . . . . . . . . . . 317
Pouteria castanosperma . . . . . . . . . . .. 224, 225 everetti . . . . . . . . . . . . . . . . . . . . . . . . 317
Prenolepis . . . . . . . . . . . . . . . . . . . . . . . . 120 harrissoni . . . . . . . . . . . . . . . . . . . . . . 317
PresbytIS . . . . . . . . . . . . . . . . . . . . . . . 71-75 kajau . . . . . . . . . . . . . . . . . . . . . 317, 320
nigropalmatus . . . . . . . . . . . . . . . . . . . . 317
g~f:aUW : : : : : : : : : : : : : : : : : : : : : : :: ~T pardalis . . . . . . . . . . . . . . . . . . . . . . . . 317
catemana . . . . . . . . . . . . . . . . . . . . . .. 71 Rhaphidophora . . . . . . . . . . . . . . . . . . . . . 48
chrysomelas . . . . . . . . . . . . . . . . . . . . . 73 Rhelfhroseiums macrotis . . . . . . . . 177, 178, 343
comata . . . . . . . . . . . . . . . . . 71,72,74,75 Rhinomyias umbratilis .. . . . . . . . . . . . . . . . 104
femoralis . . . . . . . . . . . . . . . . . . . . . 71-75 Rhinoplax vigil . . . . . . . . . . . . . . . . . . . . . 155
fredericae . . . . . . . . . . . . . . . . . . . . . . 74 Rhizophora . . . . . . . . . . . . . . . . . . . .. 34, 429
frontata . . . . . . . . . . . . . . . . 73-75,343.. 344 Rhododendron . . . . . . . . . . . . 163, 325, 326, 329
me lalophos . . . . . . . . . . . . . . . . . . .. 11-74 tashiroi . . . . . . . . . . . . . . . . . . . . . . . . 217
obseura . . . . . . . . . . . . . . . . . . . . . . . . 156 Rhodomyrtus . . . . . . . . . . . . . . . . . . . . . . . 223
percura . . . . . . . . . . . . . . . . . . . . . . .. 71 Rhopalomastix . . . . . . . . . . . . . . . . . . . . . . 120
potenziani . . . . . . . . . . . 71,72,74,427,429 Rhoptromyrmex . . . . . . . . . . . . . . . . . . . . . 292
rubicunda . . . . . . . . . . . 71,73,74,343,344 wroughtonii . . . . . . . . . . . . . . . . . 293, 294
siamensis . . . . . . . . . . . . . . . . . . . .. 72-75 Rhytieeros undulatus . . . . . . . . . . . . . . . . . . 155
sumatrana . . . . . . . . . . . . . . . . . . . . . . 71 Rubus idaeus . . . . . . . . . . . . . . . . . . . . . . . 227
thomasi . . . . . . . . . . . . . . . . . . . . . . 71, 72 Ryparosa javanica . . . . . . . . . . . . . . . 21, 23-27
Prioneris cornelia . . . . . . . . . . . . . . . . . . . . 131 Sa7zmanma . . . . . . . . . . . . . . . . . . . . . . . . 483
philonome . . . . . . . . . . . . . . . . . . . . . . 13 1 Saurauia . . . . . . . . . . . . . . . . . . . . . . . . . 441
Prionopelta . . . . . . . . . . . . . . . . . . . . . . . . 120 andreana . . . . . . . . . . . . . . . . . . . . . . . 221
Proatta . . . . . . . . . . . . . . . . . . . . . . . . . . 120 latibractea . . . . . . . . . . . . . . . . . . . . . . 327
Procapritermes . . . . . . . . . . . . . . . . . 300, 303 Scheff/era . . . . . . . . . . . . . . . . . . . . . . . . . 442
Proceratium . . . . . . . . . . . . . . . . . . . . . . . 11 Schima . . . . . . . . . . . . . . . . . . . . . . . . . . 153
Prohamitermes . . . . . . . . . . . . . . . . . . . . . . 300 Schizaea dichotoma . . . . . . . . . . . . . . . . . . . 447
Proreus . . . . . . . . . . . . . . . . . . . . . . . . . . 92 Schizostachyum grande . . . . . . . . . . . . . . . . 97
Prosotas dubiosa . . . . . . . . . . . . . . . . . . . . 292 Schumannianthus . . . . . . . . . . . . . . . . . . . . 446
Protomyzon . . . . . . . . . . . . . . 380, 381, 383, 384 Sciaphila . . . . . . . . . . . . . . . . . . . . . . . 47,48
griswoldi . . . . . . . . . . . . . . . . . 380-382, 384 Sciapus . . . . . . . . . . . . . . . . . . . . . . . . . . 187
whiteheadi . . . . . . .. 309,380, 381, 383, 384 Seirtes . . . . . . . . . . . . . . . . . . . . .. 87, 89, 93
Prunus turneriana . . . . . . . . . . . . . . .. 224, 225 Sciums . . . . . . . . . . . . . . . . . . . . . . . . . . 225
Psalidomyrmex . . . . . . . . . . . . . . . . . . . . . 11 Seleria . . . . . . . . . . . . . . . . . . . . . . . . . . . 443
Psammottynastes pictus . . . . . . . . . . . . . . . . 318 Scrotochloa urceolata . . . . . . . . . . . . . . . . . 50
Pseudodiopsis . . . . . . . . . . . . . . . . . . . . . . 173 Sebastiania . . . . . . . . . . . . . . . . . . . . . . . . 483
detrahens . . . . . . . . . . . . . . . . . . . . . . . 173 Selaginella . . . . . . . . . . . . . . . . . . . . . . 47, 48
Pseudogobiopsis . . . . . .. 309,310,380,381,384 Semnopithecus auratus . . . . . . . . . . . . . . . .. 74
Pseudomyrmex . . . . . . . . . . . . . . . . . . . . .. 11 Senna alata . . . . . . . . . . . . . . . . . . . . . . . . 445
Pseudorabdion albonuchalis . . . . . . . . . . . . . 318 Sepedophilus . . . . . . . . . . . . . . . . . . . . . .. 93
longiceps . . . . . . . . . . . . . . . . . . . . . . . 318 Sequoia sempervirens . . . . . . . . . . . . . . . . . 219
Psittinus cyanums . . . . . . . . . . . . . . . . . . . . 428 Seseli libanotis . . . . . . . . . . . . . . . . . . . . . . 202
Psyehotria ereeta . . . . . . . . . . . . . . . . . . . . 483 Shorea . . . . . . . . . . . . . . 38, 277, 341, 428, 429
jambosioides . . . . . . . . . . . . . . . . . . . . . 483 albida . . . . . . . . . . . . . . . . . . . . . . . . . 411
mapourioides . . . . . . . . . . . . . . . . . . . . 483 curtisii . . . . . . . . . . . . . . . . . 277, 278, 280
Pteeticus longipennis . . . . . . . . . . . . . . . . .. 93 falciferoides . . . . . . . . . . . . . . . . . . 367-374
Pterisanthes grandis . . . . . . . . . . . . . . . . . . 167 laevis . . . . . . . . . . . . . . . 139, 141, 143, 144
Pternandra ... . . . . . . . . . . . . . . . .. 221, 223 leprosula . .. 277, 278, 280, 348, 367-371, 373
azurea . . . . . . . . . . . . . . . . . . . . . . . . . 226 macroptera . . . . . . . . . . . . . . . . . . . . . . 228
coeruleseens . . . . . . . . . . . . . . . .. 221, 227 pachyphylla . . . . . . . . . . . . . . . . . . . . . 412
eehinata . . . . . . . . . . . . . . . . . . . 221, 227 parvlJoZw . . . . . . 139, 141, 143,277,278,280
Pterocymbium macrocater . . . . . . . . . . . . . . 328 paueiflora . . . . . . . . . . . . . . . . . . . . . . 143
Pteropus hypomelanus . . . . . . . . . . . . . . . . . 427 Silurichthys hasselti . . . . . . . . . 309, 380, 381, 384
vampyms . . . . . . . . . . . . . . . . . . . . . . . 427 Silums . . . . . . . . . . . . . . . . . . . . . 380-382, 384
Pterospermum niveum . . . . . . . . . . . . . . . . . 328 Simias concolor . . . . . . . . . . . . . . . .. 427, 429
Ptiloeereus lowii . . . . . . . . . . . . . . . . . . . . 427 Sloanea javanica . . . . . . . . . . . . . . . . . . . . 167
Ptilona confinis . . . . . . . . . . . . . . . . . . . 90, 93 Smodicinodes kovaei . . . . . . . . . . . . . . . . .. 91
Ptyehozoon horsfieldii . . . . . . . . . . . . . 316,318 Soa . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
Ptyssiglottis psychotriifolia . . . . . . . . . . . . . . 48 Solanum daZlachyi . . . . . . . . . . . . . . . . . . . 226
Puntius binotatus . . . . . . . 309-312,380,381,384 polytrichum . . . . . . . . . . . . . . . . . . . . . 483
Pycnonotus nieuwenhuisi . . . . . . . . . . . 102, 106 Solenopsis . . . . . . . . . . . . . . . . . . . . 120, 122
reticulata . . . . . . . . . . . . . . . . . . . . 318
~
thon Solenostemon elongatus . . . . . . . . . . . . . . . . 506
uadmstmma . . . . . . . . . . . . . . . . . . 120, 122 Sonerila . . . . . . . . . . . . . . . . . . . . . . . . .. 48
uercus . . . . . . . . . . . . . . . . . . . . .. 224, 226 Spalgis . . . . . . . . . . . . . . . . . . . . . . . . . . . 293
anablythi ... . . . . . . . . . . . . . . . . . 317,320 Spermaphora miser . . . . . . . . . . . . . . . . . .. 92
Sphenomorphus cyanolaemus . . . . . . . . . . 318-320 parcearmatus . . . . . . . . . . . . . . . . . . . . 190

SphyracepFzala . . . . . . . . . . . . . . . . .. 170, 173 pectinatus . . . . . . . . . . . . . . . . . . . . . . 190
Spilomis cheela . . . . . . . . . . . . . . . . . . . . . 428
Spondias . . . . . . . . . . . . . . . . . . . . . 225,226 ~:Jratu; : : : : : : : : : : : : : : : : : : : : : : : l~g
Stachyphrynium . . . . . . . . . . . . . . . . . . . .. 50 spatulifer . . . . . . . . . . . . . . . . . . . . . . . 190
Staurols /atopalmatus . . . . . . . . . . . . . . . . . 317 unclnatus . . . . . . . . . . . . . . . . . . . . . . . 190
natator . . . . . . . . . . . . . . . . . . . . 317,320 vanaartseni . . . . . . . . . . . . . . . . . . . . . 190
tuberilinguis . . . . . . . . . . . . . . . . . . . . . 317 Thecurus crassispinis . . . . . . . . . .. 177, 178, 343
Stilicopsis . . . . . . . . . . . . . . . . . . . . . . . 91,93 Themara hirtipes . . . . . . . . . . 169, 170, 172, 173
Stiloderus . . . . . . . . . . . . . . . . . . . . . . . .. 93 Theridion .. . . . . . . . . . . . . . . . . . . . . . .. 92
Stromanthe porteana . . . . . . . . . . . . . . . . . . 483 Thinolestris . . . . . . . . . . . . . . . . . . . . . . . . 185
Strumigenys . . . . . . . . . . . . . . . . . . . . 93, 120 Thinophilus . . . . . . . . . . . . . . . . . . . . . . . . 185
Subulitermes ... . . . . . . . . . . . . . . . . . . .. 300 Thorakochirus ... . . . . . . . . . . . . . . . . . .. 93
Suncus etruscus . . . . . . . . . . . . . . . . . 177, 181 Thottea penitilobata . . . . . . . . . . . . . . . . . . 46
Sundamys muelleri . . . . . . . . . . . . . . . . 177-181 Tilia cortlata . . . . . . . . . . . . . . . . . . . . . . . 219
Sundareonectes sabanus . . . . . . . . . . . . . 380-384 Toona calantas . . . . . . . . . . . . . . . . . . . . . 328
Sundasciurus fraterculus . . . . . . . . . . . . . . . . 427 sureni . . . . . . . . . . . . . . . . . . . . . . . . . 462
hippurus . . . . . . . . . . . . . . . . 177, 178, 343 Topomyia spathulirostris . . . . . . . . . . . . . . .. 93
lowii . . . . . . . . . . . . . . . . 156, 177, 178,343 Tor douronensis . . . . . . .. 309-312, 380-382, 384
tenuis . . . . . . . . . . . . . . . . . . . . . . . . . 343 Toxorhynchites . . . . . . . . . . . . . . . 89, 94, 96, 97
Surendra /lorimel . . . . . . . . . . . . . . . . . . . . 292 leicesteri . . . . . . . . . . . . . . . . . . . . . 89, 93
Sus barbatus . . . . . . . . . . . . . . . . . . . . . . . 343 magnificus . . . . . . . . . . . . . . . . . . . . .. 94
Swartzia apetala . . . . . . . . . . . . . . . .. 483, 484 metallicus . . . . . . . . . . . . . . . . . 88, 89, 93
Symphonia globulifera . . . . . . . . . . . . . . . . . 483 Tragulus javanicus . . . . . . . . . . . . . . . 156, 343
Symplocos . . . . . . . . . . . . . . . . . . . . . . . . 327 napu . . . . . . . . . . . . . . . . . . . . . . . . . . 156
Sympycnus . . . . . . . 183, 185, 187, 188, 190, 191 Trema . . . . . . . . . . . . . . . . . . . . . . . 223, 278
aechmophorus . . . . . . . . . . . . . . . . . . . . 190 tomentosa . . . . . . . . . . . . . . . . . . 227, 228
albiserratus . . . . . . . . . . . . . . . . . . . . . 190 Triehadenia maerantha . . . . . . . . . . . . . . . . 328
alipes . . . . . . . . . . . . . . . . . . . . . . . . . 190 Triehastoma malaceense . . . . . . . . . . . . . . . . 107
anachrostichus . . . . . . . . . . . . . . . . . . . 190 Triehomanes singaporeanum . . . . . . . . . . . 48, 50
aranelpes . . . . . . . . . . . . . . . . . . . . . . . 190 Trigona . . . . . . . . . . . . . . . . . . . . . 59, 62, 64
campslcnemOldeJ . . . . . . . . . . . . . . . . . . 190 apiealis . . . . . . . . . . . . . . . . . . . . . . 61, 62
candldlmanus . . . . . . . . . . . . . . . . . . . . 190 canifrons . . . . . . . . . . . . . . . . . . . . . .. 61
compressipes . . . . . . . . . . . . . . . . 189, 190 eollina . . . . . . . . . . . . . . . . . . . . . . 61, 62
ctenophorus . . . . . . . . . . . . . . . . . . . . . 190 drescheri . . . . . . . . . . . . . . . . . . . . . 61, 62
gigas . . . . . . . . . . . . . . . . . . . . . . . . . 190 erythrogastra . . . . . . . . . . . . . . . . . . . . 61
lctericus . . . . . . . . . . . . . . . . . . . . . . . 190 jimbriata . . . . . . . . . . . . . . . . . . . . . .. 61
leucotarsus . . . . . . . . . . . . . . . . . . . . . . 190 fuscibasis . . . . . . . . . . . . . . . . . . . . . .. 61
miritarsis . . . . . . . . . . . . . . . . . . . . . . . 190 fuscobalteata . . . . . . . . . . . . . . . . . . . . 61
nigrisquamis . . . . . . . . . . . . . . . . . . . . . 190 geissleri . . . . . . . . . . . . . . . . . . . . . . .. 61
pallipes . . . . . . . . . . . . . . . . . . . . . . . . 190 haematoptera . . . . . . . . . . . . . . . . . . 61,64
hobbyi . . . . . . . . . . . . . . . . . . . . . . . 61,64
f.r;if::ta : : : : : : : : : : : : : : : : : : : : : : : : l~g
sticticus . . . . . . . . . . . . . . . . . . .. 188, 190
itama . . . . . . . . . . . . . . . . . . . . . . . 61,62
laeviceps . . . . . . . . . . . . . . . . . . . . . .. 61
teuchophoroides . . . . . . . . . . . . . . . . . . 190 melanocephala . . . . . . . . . . . . . . 61, 62, 64
torrenticola . . . . . . . . . . . . . . . . . . . . . 190 melina . . . . . . . . . . . . . . . . . . .. 59,61,62
tricomiger . . . . . . . . . . . . . . . . . . . . . . 190 nitidi ventris . . . . . . . . . . . . . . . . . . . . . 61
Syndyas . . . . . . . . . . . . . . . . . . . . .. 185, 186 reepeni . . . . . . . . . . . . . . . . . . . . . . . . 61
Syneches . . . . . . . . . . . . . . . . . . . . . 185, 186 ru}lbasalis . . . . . . . . . . . . . . . . . . . . . . 61
Syzygium decipiens . . . . . . . . . . . . . .. 327, 328 terminata . . . . . . . . . . . . . . . . . . . . . .. 61
mainitense . . . . . . . . . . . . . . . . . . . . . . 327 thoraciea . . . . . . . . . . . . . . . . . . . . . .. 61
nitidum . . . . . . . . . . . . . . . . . . . . . . . . 327 ventralis . . . . . . . . . . . . . . . . . . . . . . . 61
polycephaloides . . . . . . . . . . . . . . . . . . . 327 Trimeresurus bomeensis . . . . . . . . . . . . . 318-320
simile . . . . . . . . . . . . . . . . . . . . . . . . . 327 Tripteroides . . . . . , . . . . . . . . . . . . . . . . . 93
Tachydromia . . . . . . . . . . . . . . . . . .. 185, 187 Tnssopelopia . . . . . . . . . . . . . . . . . . . . . 89, 93
Tamasesia . . . . . . . . . . . . . . . . . . . . . . . . 92 Tristania . . . . . . . . . . . . . . . . . . . . . . . . . 412
Tapine lIa . . . . . . . . . . . . . . . . . . . . . . . . . 92 Triunia erythrocarpa . . . . . . . . . . . . .. 224, 225
Tapinoma . . . . . . . . . . . . . . . . . . . .. 120, 362 Trix scopula . . . . . . . . . . . . . . . . . . . . . . . 132
Tapirira guianensis . . . . . . . . . . . . . .. 483,484 Tropidolaemus wagleri . . . . . . . . . . . . . . 318-320
Taraka . . . . . . . . . . . . . . . . . . . . . . . . . . . 293 Tropidophorus beccarii . . . . . . . . . . . . . . . . 318
Taxus . . . . . . . . . . . . . . . . . . . . . . . . . . . 326 brookei . . . . . . . . . . . . . . . . . . . . . . . . 318
sumatrana . . . . . . . . . . . . . . . . . . . . . . 329 Tupaia . . . . . . . . . . . . . . . . . . . . . . . . . . . 344
Technomyrmex . . . . . . . . . . . . . . . . . . . . . . 292 dorsalis . . . . . . . . . . . . . . . .. 177, 178, 343
Tectaria . . . . . . . . . . . . . . . . . . . . . . . . 47, 48 glis . . . . . . . . . . . . . . . . . . .. 178, 343, 427
lobii . . . . . . . . . . . . . . . . . . . . . . . . . . 48 gracilis ... . . . . . . . . . . . . .. 177, 178, 343
Tectona grandis . . . . . . . . . . . . . . . .. 462, 467 minor . . . . . . . . . . . . . . . . .. 177, 178, 343
Teleopsis . . . . . . . . . . . . . . . . . . 170, 172, 173 montana . . . . . . . . . . . . . . . . . . . . . . . 177
belalongensis . . . . . . . . . . . . . 170, 171, 173 pieta . . . . . . . . . . . . . . . . . .. 177, 178, 181
belzebuth . . . . . . . . . . . . . . . . . . . . . . . 170 tana .. . . . . . . . . . . . . . . . . . . .. 177, 178
burkhardti . . . . . . . . . . . . . . . . . . . . . . 173 Tylonyeteris pachypus . . . . . . . . . . . . . . . 91, 94
discrepans " . . . . . . . . . . . . . . . . . . . . 173 robustula . . . . . . . . . . . . . . . . . . . . . 91, 94
pallifacies . . . . . . . . . . . . . . . 170, 171, 173 Tvmpanophyllum . . . . . . . . . . . . . . . . . . 80, 81
Termes . . . . . . . . . . . . . . . . . . . . . . . . . . . 300 Vnearia . . . . . . . . . . . . . . . . . . . . . . . . . . 226
Terminalia foetidissima . . . . . . . . . . . . . . . . 328 Uranotaenia . . . . . . . . . . . . . . . . . . 89, 90, 93
phellocarpa . . . . . . . . . . . . . . . . . . . . . 429 Uromys eaudimae-ulatus . . . . . . . . . . . . . . . 225
Tetramorium . . . . . . . . . . . .. 11, 119, 120, 122 Urophyllum . . . . . . . . . . . . . . . . . 223, 226, 227
Tetraponera . . . . . . . . . . . . . . . . . 94, 120, 122 gIabrum . . . . . . . . . . . . . . . . . . .. 226, 227
Teuchophorus . . . . . . . . . . . . 185, 188, 190, 191 hirsutum . . . . . . . . . . . . . . . . . . . . . . . 221
conspicuus . . . . . . . . . . . . . . . . . . . . . . 190 Uthina luzonica . . . . . . . . . . . . . . . . . . . . . 92
costalis .. . . . . . . . . . . . . . . . . . . . . . . 190 Vaccinium . . . . . . . . . . . . . . . . . . . . 223,325
denticulatus . . . . . . . . . . . . . . . . . . . . . 190 Varanus salvator . . . . . . . . . . . . . . . . 318, 321
digitatus . . . . . . . . . . . . . . . . . . . . . . . 190 Vatica endertii . . . . . . . . . . . . . . . . . . . . . . 143
femoratus . . . . . . . . . . . . . . . . . . . . . . . 190 odorata . . . . . . . . . . . . . . . . . . .. 141, 143
fuscicornis . . . . . . . . . . . . . . . . . . . . . . 190 Virola surinamensis . . . . . . . . . . . . . . . . . . . 229
grandior . . . . . . . . . . . . . . . . . . . . . . . 190 Vollenhovia . . . . . . . . . . . . . . . . . . . . . . . . 120
laingensis . . . . . . . . . . . . . . . . . . . . . . . 190 Wasmannia . . . . . . . . . . . . . . . . . . . . . . . . 120
Willowsiella . . . . . . . . . . . . . . . . . . . . . . . 120
Xanthophytum villarii . . . . . . . . . . . . . . . . . 429
Xylobium . . . . . . . . . . . . . . . . . . . . . . . . . 64
Xylocopa myops . . . . . . . . . . . . . . . . . . . 59, 64
Xylodiplosis . . . . . . . . . . . . . . . . . . 89,92, 93
Xylopta malayana . . . . . . . . . . . . . . . . . . . . 412
Yumbera . . . . . . . . . . . . . . . . . . . . . . . . . 188
Zingiber zerumbet . . . . . . . . . . . . . . . . . . . . 155
Zonocerus . . . . . . . . . . . . . . . . . . . . . . . .. 458
Zosterops . . . . . . . . . . . . . . . . . . . . . . . . . 152
GENERAL INDEX
Abbott's Booby . . . . . . . . . . . . . . . . . . . . . 472 - species richness & spatial distribution . 199-204
Abundance - of ants . . . . . . . . . . . . . . . . . . 361 Apocynaceae . . . . . . . . . . . . . . .. 142, 143, 328
- of arthropods . . . . . . . . . . . . . . . . . . . . 7 Apoh Kayan plateau, Kalimantan . . . . . . . . . . 421
- of birds . . . . . . . . . . . . . . . . . . . . . .. 106 Appalachian Mountains . . . . . . . . . . . . . 199,204
- of fish . . . . . . . . . . . . . . . . . . . . . . . . 377 Apparent quantum yields . . . . . . . . 369, 370, 373
- of termites . . . . . . . . . . . . . . . . .. 12, 297 Aquifoliaceae . . . . . . . . . . . . . . . . . . . . . . 483
Acanthaceae . . . . . . . . . . . . . . . . . . . . . 46, 50 Arabuko-Sokoke Forest Reserve, Kenya . . . . . 542
Acanthocephala ... . . . . . . . . . . . . . . . . . .. 5 Araceae/Aroids . . . . . . . . . . 43,47-50, 167,445
~cari . . . . . . . . . . . . . . . . . . . . . . . . . . . 5,89 Arachnida . . . . . . . . . . . . . . . . . . . . . . . . 5,92
Acarina .. . . . . . . . . . . . . . . . . . . .. 6, 14, 92 Aradidae . . . . . . . . . . . . . . . . . . . . . . . . . . 92
Aceh, Sumatra. . . . . . . . . . . . . . . . . . . . . . 425 Araneae . . . . . . . . . . . . . . . . . . . . . . . . . 5, 92
Acoustic communication . . . . . . . . . . . . . . . . 77 Araneida . . . . . . . . . . . . . . . . . . . . . . . . . . 23
Acoustic communities . . . . . . . . . . . . . . . 77, 80 Arc/Info GIS . . . . . . . . . . . . . . . . . . . . . .. 500
Acoustic guilds . . . . . . . . . . . . . . . . . 77, 80, 83 Arctiidae . . . . . . . . . . . . . . . . . . . . . . . 90, 93
Actinidiaceae . . . . . . . . . . . . . . . . . . . . 221, 327 Arecaceae (see also Palmae) .. . . .. 117, 142,483
Activity patterns, snakes & frogs . . . . . . . . . . 320 Argus pheasants . . . . . . . . . . . . . . . . . . . .. 154
Adaptation - to shade by plants . . . . . . . . 220, 374 Aristolochiaceae . . . . . . . . . . . . . . . . . . . . . . 46
- to light by plants .. . . . . . . .. 216,367,374 Arthropods . . . . . . . . . . . . .. 1-20, 21-28, 85-99
Africa . . . . . . . . . . . . . . . . . 219,411, 523, 535 - canopy studies . . . . . . . . . . . . . . 22, 205-208
African honeybee . . . . . . . . . . . . . . . . . . . . . 65 - species richness . . . . . . . . . . . . . . 9, 26, 194
Afromontane forest in Kenya . . . . . . . . . . 539, 542 Asian elephant . . . . . . . . . . . . . . . . . . . . .. 155
Agamids/Agamidae . . . . . . . . . . . . . . . . 318,321 Asian two-horned rhinoceros. . . . . . . . . . . .. 155
Agaonidae . . . . . . . . . . . . . . . . . . . . . . . . 283 Assassin bug . . . . . . . . . . . . . . . . . . . . . . . . 91
Agenda 21 . . . . . . . . . . . . . . . . . . . . . . . . 417 Asynchronous flowering in figs . . . . . . . . 283, 288
Agouti . . . . . . . . . . . . . . . . . . . . . . . . . . . 226 Atelestidae . . . . . . . . . . . . . . . . . 183, 184, 186
Air Hitam, Pontianak, Kalimantan . . . . . . 412, 413 Atlantic rain forests . . . . . . . . . . . . . . . . . . 478
Alangiaceae . . . . . . . . . . . . . . . . . . . . . . . 142 Australia . . . . . . . . . 21,67, 151, 186, 189,219,
All Taxa Biodiversity Inventories . . . . . . . . . . 209 221,225, 294, 332, 467, 473, 523
Allochthonous plant families ... . . . . . . . . . . 325 Australian Aborigines . . . . . . . . . . . . . . 467,473
Allopatry - in birds . . . . . . . . . . . . . . . . . .. lSI Austronesian people . . . . . . . . . . . . . . . . . . 435
- in Presbytis . . . . . . . . . . . . . . . . . . . . . 71 Autochthonous plant families . . . . . . . . . . . . . 325
Alluvial forest . . . . . . . . . . . . . . . . . . . 102, 253 Avifauna . . . . . . . . . . . . . . . . 101-116, 151-152
Altitudinal changes in forests . . . . . . . . .. 241-257 Bagridae . . . . . . . . . . . . . . . . . . . . . . . . . 380
Altitudinal distribution of plant families . .. 325-330 Bahia, Brazil . . . . . . . . . . . . . . . . . . . . 477,478
Altitudinal zonation of birds . . . . . . . . . . . . . lSI Bali . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
Amani Forest . . . . . . . . . . . . . . . . . . . . . . 107 Balikpapan, Kalimantan . . . . . . . . . . . . . . . . 454
Amazon/Amazonia . . . . . . . . 40,50,84,411,482 Balitoridae . . . . . . . 309,377,380,382,383,385
Amazon basin peoples . . . . . . . . . . . . . . . . . 460 Balsaminaceae . . . . . . . . . . . . . . . . . . . . . . 156
Amazonian forest . . . . . . . . . . . . . . . . . . . . 419 Bamboos . . . . . . . . . . . . . . . . . . 167,446,516
Amazonian Indians . . . . . . . . . . . . . . . . 459,478 - phytotelmata . . . . . . . . . . . . . . . . . . . 85-99
Amazonian terra firme rainforest ... . . . . . .. 357 - shoots (as food) . . . . . . . . . . . . . . . . .. 517
Amboseli National Park, Kenya . . . . . . . . . . . 418 Bananas . . . . . . . . . . . . . . . . . . . . . . . 442,443
Amphibians/Amphibia . . . . . . . . . . . . . . . 5,315 Bangar, Brunei Darussalam . . . . . . . . . . . . . . 389
Anacardiaceae . . . . . . 139, 141-143, 325, 327, 483 Bangi forest, Selangor . . . . . . . . . . . . . . . .. 121
Anguillidae . . . . . . . . . . . . . . . . . . . . . . . . 380 Barbets . . . . . . . . . . . . . . . . . . . . . . . 104, 154
Annelida . . . . . . . . . . . . . . . . . . . . . . . . . 4, 5 Barking deer . . . . . . . . . . . . . . . . . . . . . . . 343
Annonaceae . . . . . . . . . . 142,293,325,328, 483 Barkly Tablelands, Australia . . . . . . . . . . . . . . 68
Annual litter production . . . . . . . . . . . . . . . . 303 Barringtonia Forest . . . . . . . . . . . . . . . . . .. 429
Ant-plant symbioses . . . . . . . . . . . . . . . 11, 295 Barro Colorado Island, Panama . . . . . . . . . . . 217
Antelopes . . . . . . . . . . . . . . . . . . . . . . . .. 539 Basalt-derived soils . . . . . . . . . . . . . . . . 223, 226
Anthocoridae . . . . . . . . . . . . . . . . . . . . . . . . 92 Bataguridae/Batagurid turtles . . . . . . . . . . 155, 318
Anthropogenic grasslands . . . . . . . . . . . . 284, 543 Batflies . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
Ants . . . . . . . . . 1,9,92,94, 117, 193,331,459 Bats . . . . . . . . . . . . . . . . . . 67,69,91,94,427
- abundance . . . . . . . . . . . . . . . . . . . . . 361 Batu Apoi Forest Reserve . . . . . . . 29, 43, 50, 125,
- and lycaenid butterflies .... . . . . .. 291-296 133, 139,242,297,302,307,315,367,387
- as indicator species . . . . . . . . . . . . 117, 357 Bawangling evergreen tropical forest ecosystem . 489
- as predators . . . . . . . . . . . . . . . . . . . . . 91 Bearded pig . . . . . . . . . . . . . . . . . . . . . . . 343
- diversity . . . . . . . . . . . . . 117-123,357-366 Beech woodland . . . . . . . . . . . . . . . . . . . . . . 16
- ecology . . . . . . . . . . . . . . . . . . . . . . . 357 Bees . . . . . . . . . . . . . . . . . . . . . . . . . . . 59-66
- gardens . . . . . . . . . . . . . . . . . . . . . . . . 12 - halictid . . . . . . . . . . . . . . . . . . . . . . . . 64
- guilds .. . . . . . . . . . . . . . . . . . . . . . . 332 - honey . . . . . . . . . . . . . . . . . . . 61,63,64
- in bamboo internodes . . . . . . . . . . . . . 91-94 - megachilid . . . . . . . . . . . . . . . . . . . . . . 64
- leaf-cutter . . . . . . . . . . . . . . . . . . . . . . 12 - social . . . . . . . . . . . . . . . . . . . . . . . 59-66
- nests . . . . . . . . . . . . . . . . . . . . . . 331-340 - stingless . . . . . . . . . . . . . . . . . . 59, 62, 64
- role in forest dynamics . . . . . . . . . . . . 9-12 - sweat . . . . . . . . . . . . . . . . . . . . . . . . . 64
- species accumulation curves . . . . . . . 205, 206 Beetles - species richness in Borneo . . . . . 208, 209
- species richness . . . . . . . . . . . . . . . . .. 206 Begoniaceae . . . . . . . . . . . . . . . . . . 43, 50, 156
- species-abundance distribution . . . . . . . .. 205 Behaviour - of fish .... . . . . . . . . . . . . . . . 307
- weaver ants . . . . . . . . . . . . . . . . . . . . 293 - of stalk-eyed flies. . . . . . . . . . . . .. 169-174
Anura . . . . . . . . . . . . . . . . . . . . . . . . . . . 317 Belalong, Brunei Darussalam . . . . . 35, 59, 60, 64,
sounds . . . . . . . . . . . . . . . . . . . . . . . . . 83 101, 139, 170, 175, 178, 194, 205, 206,
Anuradhapura period, Sri Lanka . . . . . . . . . . 507 261,268,298,307, 315, 368, 377, 419
Aphids . . . . . . . . . . . . . . . . . . . . . . . . . . . 187 Belalong River. . . . . . . . .. 29,48,307,377, 524
Apiaceae . . . . . . . . . . . . . . . . . . . . . . . . . 193 Belalong-Temburong watershed . . . . . . . . . . . 378
555
556 GENERAL INDEX
Benthic fish species . . . . . . . . . . . . . . . . 309,310 Butterflies . . . . . . . . . . . . . . . . 125-138, 291-296
Bequest values (of trees) . . . . . . . . . . . . . . . 418 - diversity . . . . . . . . . . . . . . . . 125-138, 538
Berger-Parker Index . . . . . . . . . . . . . . . . .. 308 - phenology . . . . . . . . . . . . . . . . . . . . . 129
Beta diversity . . . . . . . . . . . . 133, 151, 193, 195 Cabbages . . . . . . . . . . . . . . . . . . . . . . . . . 516
Beta rarity . . . . . . . . . . . . . . . . . . . . . . . . 195 Caesalpiniaceae . . . . . . . . . . . . . . . . . . . . . 293
Bioacoustic diversity . . . . . . . . . . . . . . .. 77, 84 Calamoideae . . . . . . . . . . . . . . . . . . . . . .. 523
Bioassay, using rice . 241, 246, 250, 252, 255, 256 Cameroon . . . . . . . . . . . . . . . . . . . . . . . . . . 4
Biodiversity - conservation . . . . . . . . . . . . .. 539 Camphor . . . . . . . . . . . . . . . . . . . . . . . . . 347
- Convention . . . . . . . . . . . . . . . . . . . . 2, 15 Canada . . . . . . . . . . . . . . . . . . . . . . . . . . 272
- inventories . . . . . . . . . . . . . . . . . . . . . . 40 Cannibalism in Culicidae . . . . . . . . . . . . . . . . 97
- priorities . . . . . . . . . . . . . . . . . . . . . . 544 Canopy - effect of altitude . . . . . . . . . . . . . . 247
- surveys . . . . . . . . . . . . . . . . . . . . 537, 545 - fogging. . . . . . . . . . . . . . .. 193, 194, 196
- values . . . . . . . . . . . . . . . . . . . . . . .. 545 - gaps . . . . . . . . . . . . . . . . 215-217,225-229
Biogeographical zones in Kenya. . .. 540, 541, 544 - interception . . . . . . . . . . . . . . . . . . . . 394
Biogeography .... 67,69,86, 151,377,425,433 - openings . . . . . . . . . . . . . . . . . . . . . . 412
Biological currencies. . . . . . . . . .. 535, 544, 545 - tree species . . . . . . . . . . 268,273,367,368
Biological ethnobotany . . . . . . . . . . . . . . . . . 457 Cape Tribulation, Australia. . . . . . . . . . .. 21, 23
Biological inventory . . . . . . . . . . . . . . . . . . 209 Cape York Peninsula, Australia . . . . . . .. 67, 151
Biomass - of ants . . . . . . . . . . . . . . . . . . 9, 357 Carabidae . . . . . . . . . . . . . . . . . . . . . . . . . . 93
- of arthropods . . . . . . . . . . . . . . . . . . . . 7 Carbon cycle, role of termites . . . . . . . . . . . . . 14
- of fish . . . . . . . . . . . . . . . . . . . . . 308, 516 Carbon sinks . . . . . . . . . . . . . . . . . . . . 468,472
- of termites . . . . . . . . . . . . 12, 301, 302, 357 Carnivory, in Lycaenidae . . . . . . . . . . . . . .. 293
Bioregionalism . . . . . . . . . . . . . . . . . . . . . . 474 Cash crops . . . . . . . . . . . . . . . . . . . . . . . . 516
Biosphere Reserves . . . . . . . . . . . . . . . . . .. 507 Cassava . . . . . . . . . . . . . . . . . . . . . . . . . . 458
Biphydidae . . . . . . . . . . . . . . . . . . . . . . . . . 93 Cassowaries . . . . . . . . . . . . . . . . . . . . . . . 459
Birds .... 69,80,101-116,151-152,153,155,195 - and seeds . . . . . . . . . . . . . . . . . . . . . . 224
- abundance . . . . . . . . . . . . . . . . . . . . . 106 Catchment basins ..... '. . . . . . . . . . . . 393, 426
- and ants . . . . . . . . . . . . . . . . . . . . . . . . 12 Cats . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
- breeding . . . . . . . . . . . . . . . 101, 102, 109 Cecidomyiidae . . . . . . . . . . . . . . . . . . . . . 89-93
- diet preferences . . . . . . . . . . . . . . . . .. 104 Celastraceae . . . . . . . . . . . . . . . . 142,441,446
- dispersal of fruits & seeds . . . . . . . . . . . 224 Central America . . . . . . . . . . . . . . . . . . 219,225
- diversity . . . . . . . . . . 101-116, 151-152, 538 Ceratocombidae . . . . . . . . . . . . . . . . . . . . . . 92
- malaria .... . . . . . . . . . . . . . . . . . .. 152 Ceratopogonidae . . . . . . . . . . . . . . . . . . . 89-93
- palaearctic migrants .. . . . . . . . . . . . .. 101 Chaetognatha . . . . . . . . . . . . . . . . . . . . . . .. 5
- ringing . . . . . . . . . . . . . . . . . . . . . . . 109 Chance effects, and species richness . . . . . . . . 215
- vertical stratification. . . . . . . . . . . . . .. 101 Channel processes . . . . . . . . . . . . . . . . . . . 394
Birds of paradise . . . . . . . . . . . . . . . . . . . . 288 Chao Khao tribes, Thailand . . . . . . . . . . . . . 513
Bischofiaceae . . . . . . . . . . . . . . . . . . . . . . 328 Chao's method of estimating
Blattelidae . . . . . . . . . . . . . . . . . . . . . . . . . 92 species richness. . . . . . . . . . .. 198,206,207
Blattodea . . . . . . . . . . . . . . . . . . . . . . . 24, 92 Chelisochidae . . . . . . . . . . . . . . . . . . . . . . . 92
Blue Mountains, Jamaica . . . . . . . . . . . . . . . 223 Chemical communication . . . . . . . . . . . . . . . 186
Boidae . . . . . . . . . . . . . . . . . . . . . . . . . . . 318 Chestnut-necklaced partridge . . . . . . . . . . . .. 155
Bombacaceae . . . . . . . . . . . . . . . . . . . 142, 325 China . . . . . . . . . . . . . . . . . . . . . . . . 489, 523
Bongos . . . . . . . . . . . . . . . . . . . . . . . . . . 543 Chinese-German Coordinated Ecological Research
Bootstrap method of estimating species richness 197, Programme . . . . . . . . . . . . . . . . . . . . . 489
198,207 Chironomidae . . . . . . . . . . . . . . . . . . . . 89, 93
Boraginaceae . . . . . . . . . . . . . . . . . . . . . . . 483 Chloroxylon-Berrya-Vitex-Schleichera series ... 505
Borneo ..... 29,30,43,59,71,72,74,75, 101, Chordata . . . . . . . . . . . . . . . . . . . . . . . . . 4, 5
102,117,125,139,163.166,169,175,178,194, Christmas Island . . . . . . . . . . . . . . . . . . . . . 472
208,223,259,292,297,315,331,341,367,377, Chrysobalanaceae . . . . . . . . . . . . . . . . . 142, 483
387,388,411,417,420,435,463,495,523 Chrysomelids . . . . . . . . . . . . . . . . . . . . . . . 25
Boxing (tapping) . . . . . . . . . . . . . . . . . 235-237 Cicadidae/Cicadas . . . . . . . . . . . . . 77, 80-82, 84
Brachiopoda ... . . . . . . . . . . . . . . . . . . . .. 5 Civets . . . . . . . . . . . . . . . . . . . . . . . . . . . 427
Brachystegia forests . . . . . . . . . . . . . . . . 4, 543 Cladograms . . . . . . . . . . . . . . . .. 311, 312, 383
Brazil . . . . . . . . . . . . . . . 9, 133, 268, 477, 478 Clariidae . . . . . . . . . . . . . . . . . . . . . . . . . 380
Brenthidae . . . . . . . . . . . . . . . . . . . . . . . . . 93 Clay soils . . . . . . . . . . . . . . . . . . . . . . . . . 259
Brillouin Index . . . . . . . . . . . . . . . . . . . . . 308 Cleridae . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
Britain. . . . . . . . . . . . . . . . . . .. 193, 199, 200 Clethraceae . . . . . . . . . . . . . . . . . . . . . . . . 329
British Overseas Development Administration .. 536 Clouded leopards . . . . . . . . . . . . . . . . . 342, 343
Brunei Darussalam .... 4,9,22,29,43,59, 101, Clubionidae . . . . . . . . . . . . . . . . . . . . . . . . 92
125, 139, 163, 169, 175, 193,241,259, Clusiaceae . . . . . . . . . . . . . . 142, 325, 328, 483
297,307,315,331,367,377,387,417, Cnidaria . . . . . . . . . . . . . . . . . . . . . . . . . .. 5
435, 495, 523 Co-evolution, in fish . . . . . . . . . . . . . . . . . . 312
Bryophytes . . . . . . . . . . . . . . . . . . . . . . .. 253 Coastal forests in Kenya. . . . . . . . . . . . . . .. 542
Bufonidae . . . . . . . . . . . . . . . . . . . . . . . .. 317 Coastal hill dipterocarp forest . . . . . . . . . . . . 221
Bukit Belalong ... 48, 126, 241, 253, 391, 394, 500 Cobitidae . . . . . . . . . . . . . . . . . . . . . . . .. 380
- forest profile diagrams . . . . . . . . . . . . . 248 Coccinellidae . . . . . . . . . . . . . . .. 24, 25, 90, 93
Bukit Lagong Forest Reserve, Malaysia . . . . . . 347 Coenagrionidae . . . . . . . . . . . . . . . . . . . . . . 94
Bukit Lesong, Brunei Darussalam . . . . . . . . . . 394 Coexistence, in ants . . . . . . . . . . . . . . . 331, 332
Bukit Pagon, Brunei Darussalam . . . . . . . . . . 394 Coleoptera .. 5,8,9, 14,24,87,93, 133, 194,310
Bukit Rongkit, Malaysia. . . . . . . . . . . . . . .. 155 Collembola ..... 5, 6, 12, 14, 24, 88-92, 183, 186
Bukit Teraja, Brunei Darussalam . . . . . . . . . . 167 Colobinae . . . . . . . . . . . . . . . . . . . . . . . . . . 71
Bukit Timah Nature Reserve, Singapore . . . . . . 221 Colobus monkeys . . . . . . . . . . . . . . . . . . .. 539
Bukit Tudal, Brunei Darussalam. . .. 177, 178, 180 Colonisation - of logged forests . . . . . . . . . . . 344
Bukit Wang Mu Forest Reserve, Malaysia .... 153 - of bamboo internodes . . . . . . . . . . . . . . . 86
Bundala Sanctuary, Sri Lanka ... . . . . . . . .. 507 Colonizers and seed size . . . . . . . . . . . . . . . 223
Burseraceae . . . . . . . . . . . . . . . . 141, 142,328 Colubridae . . . . . . . . . . . . . . . . . . . . . . . . 318
GENERAL INDEX 557
Combretaceae . . . . . . . . . . . . . . . . . . . . . . 328 - of subterranean termites . . . . . . . . . . . . 299

Commensals . . . . . . . . . . . . . . . . . . . . . . . 321 - of termite nests . . . . . . . . . . . . . . . . . . 298
Commonwealth Forestry Conference . . . . . . . . 493 - of termites . . . . . . . . . . . . . . . . . . 297, 303
Communication, chemical, in empidoid flies ... 191 - of trees . . . . . . . . . . . . . . . . . 139-142, 489
Communities - age structure in birds . . . . . . .. 105 - of treeshrews . . . . . . . . . . . . . . . . . . . 344
- diversity, in stream fish . . . . . . . . . . . . . 385 Denver, Colorado . . . . . . . . . . . . . . . . . . . . 272
- of bamboo internodes ... . . . . . . . . . . . . 86 Dermaptera. . . . . . . . . . . . . . . . . . .. 5, 11,92
- organization of herpetofauna . . . . . . . . . . 315 Detritivores . . . . . . . . . . . . . . . . . . . . . . . . 297
Compensation points . . . . . . . . . . . . . . . 219,374 Diaphorinae . . . . . . . . . . . . . . . . . . . . . . . 185
Competition - for water . . . . . . . . . . . . . 216, 229 Diel cycles in ants . . . . . . . . . . . . 333, 337, 338
- in murids . . . . . . . . . . . . . . . . . . . 176, 178 Diel-temporal resource use . . . . . . . . . . . . . . 319
- in small mammals . . . . . . . . . . . . . . . . 175 Diel-time analysis . . . . . . . . . . . . . . . . . 315,316
- male, in stalk-eyed flies . . . . . . . . . . . .. 169 Diet - preferences of birds . . . . . . . . . . . . . . 104
Competitive hierarchy model . . . . . . . . . . . . . 332 - overlap between fish species . . . . . . . . . . 313
Compositae . . . . . . . . . . . . . . . . . . . . . . . . 516 Digital elevation model . . . . . . . . . . . . . . . . 406
Congo . . . . . . . . . . . . . . . . . . . . . . . . 411,541 Dilleniaceae . . . . . . . . . . . . . 139, 141, 142,428
Conservation . . . . . . . 67, 215, 230, 288, 322, 345, Dioecious fig species . . . . . . . . . . . . . . . . . . 284
417, 418, 425, 433, 477, 478, 503, 509, 535 DiopsidaelDiopsid flies . . . . . . . . . . . . . 169-174
- areas . . . . . . . . . . . . . . . . . . . . . . . . 422 Diplopods . . . . . . . . . . . . . . . . . . . . . . . . . . 92
- arguments for . . . . . . . . . . . . . . .. 467-475 Diptera . . . . . . . . . . . 5, 9, 87, 93, 169, 183, 184
- community-level . . . . . . . . . . . . . . . . . 283 Dipterocarpaceae . .. 117, 139, 141-143, 163, 193,
- of resources ., . . . . . . . . . . . . . . . . .. 461 205, 235, 247, 277, 325, 327, 341, 347, 428
- strategies . . . . . . . . . . . . . . . . . . . . . . 535 Dipterocarps . .. 167, 175,223, 229, 242, 277, 505
Continuum concept of community organisation . 195 - seedlings . . . . . . . . . . . . . . . . . . . . . . 229
Convention on Biological Diversity . . . . . . 209,417 - seeds, as food for ants. . . . . . . . . . . . . . 336
Cooperative Research Centre for Tropical Direct use values of forest . . . . . . . 417, 418, 535
Rainforest Ecology . . . . . . . . . . . . . . . . . . 21 Diseases of man, from forests . . . . . . . . . . . . 470
Copepoda . . . . . . . . . . . . . . . . . . . . . .. 87, 92 Disjunct distribution in Presby tis . . . . . . . . . . . 72
Corylophidae . . . . . . . . . . . . . . . . . . . .. 24, 93 Dispersal - barriers in fish . . . . . . . . . . . . . 383
Coscinida . . . . . . . . . . . . . . . . . . . . . . . . . . 92 - centre for Presby tis . . . . . . . . . . . . . . . . 71
Cost efficiency of inventory methods . . . .. 523-533 - of figs . . . . . . . . . . . . . . . . . . . . . . . . 288
Costa Rica . . . . . . . . . . . . 16,39,209,216, 217, - of fish . . . . . . . . . . . . . . . . . . . . . . . . 385
219,227, 229, 255, 357, 373, 418 - of plant propagules . . . . . . . . . . . . . 329, 330
Cote d'Ivoire . . . . . . . . . . . . . . . . . . . . . . . 219 - of seeds. . . . . . . . . . . . . . .. 215, 225, 226
Crab spiders . . . . . . . . . . . . . . . . . . . . . . . . 91 - of seeds by ants . . . . . . . . . . . . . . . . . . . 11
Crn~ ............................ ~ - of seeds by squirrels . . . . . . . . . . . . . . . 224
Crickets . . . . . . . . . . . . . . . . . . . . 77, 80, 82-84 Dispersion of ants nests .... . . . . . . . . . . . . 333
Critical populations of species . . . . . . . . . . . . 539 Distribution - of ants . . . . . . . . . . . . . . . . . . . 10
Crocker Range, Sabah . . . . . . . . . . . . . .. 34, 37 - of birds . . . . . . . . . . . . . . . . . . . . 101,151
Crop genetic resources . . . . . . . . . . . . . . . . 478 - of freshwater fish. . . . . . . . . . . . . . . . . 377
Crustacea/Crustaceans . . . . . . . . . . . 5, 87, 92, 95 - of Kapur Forest . . . . . . . . . . . . . . . . . . 348
Crypteroniaceae . . . . . . . . . . . . . . . . . . . .. 142 - of termites in forests . . . . . . . . . . . . . . . . 12
Cryptic pioneer . . . . . . . . . . . . . . . . . . . . . 219 - patterns of tree families . . . . . . . . .. 325-330
Ctenophora . . . . . . . . . . . . . . . . . . . . . . . .. 5 DIVERSITAS programme . . . . . . . . . . . . . . 209
Cucujidae . . . . . . . . . . . . . . . . . . . . . . . . . . 93 Diversity - alpha . . . . . . . . . . . 29,31,35,37,49
Culicidae . . . . . . . . . . . . . . . . . . . . . . . . 91-93 - beta . . . . . . . . . . . . . . . . . . 133,151,193
Cultural diversity . . . . . . . . . . . . . . . . . . . . 422 - bioacoustic . . . . . . . . . . . . . . . . . . . . . . 84
Cultural identity . . . . . . . . . . . . . . . . . . . . . 457 - changes after logging . . . . . . . . . . . . . . 341
Cultural values of trees . . . . . . . . . . . . . . . . 467 - ecosystem . . . . . . . . . . . . . . . . . . . . . . 2
Curculionidae/Curculionoidea . . . . . . . . 8, 24, 93 - gamma . . . . . . . . . . . . . . . . . . . . . . . . 49
Customary rights . . . . . . . . . . . . . 417, 452, 453 - genetic . . . . . . . . . . . . . . . . . . . . . . . . 2
Customary tenure . . . . . . . . . . . . . . . . . . . . 422 - gradients . . . . . . . . . . . . . . . . . . . . .. 152
Cycads . . . . . . . . . . . . . . . . . . . . . . . . . . 221 - indices . . . . . . . . . . . . . . 32, 106, 107, 482
Cyclones . . . . . . . . . . . . . . . . . . . . . . . . . 220 - landscape-level . . . . . . . . . . . . . . . . . . . 47
Cylapomorpha . . . . . . . . . . . . . . . . . . . . . . . 92 - local . . . . . . . . . . . . . . . . . . . 77, 169, 170
Cyperaceae . . . . . . . . . . . . . . .. 43,47, 50, 443 - McIntosh's Index . . . . . . . . . . . . . . 105, 107
Cyprinidae . . . . . . . . . . . . . . 309,377,380,381 - regional . . . . . . . . . . . . . . . . . . . . . . . . 77
Dammar . . . . . . . . . . . . . . . . . . . . . . . . . . 235 - Shannon-Wiener index . . . . . . . . . . . . . 308
Damselflies . . . . . . . . . . . . . . . . . . . . . . . . . 94 - Simpson's Index . . . . . . . . . . . . . . . . . 106
Danum Valley Conservation Area, Sabah . . .. 217, - Williams' alpha . . . . . . . . . . . . . . .. 34, 36
226, 268, 357 Diversity - of ants . . . . . . . . . 9-12, 117-123,357
Dark respiration rates .... 367,369,370,373,374 -of birds . . . . . . . . . . . . . 101,116,151,152
Databases, for plant taxonomy. . . . . . . . . . .. 164 - of butterflies . . . . . . . . . . . . . . . .. 125-138
Datiscaceae . . . . . . . . . . . . . . . . . . . . . . . . 328 - of empidoid flies ... . . . . . . . . . .. 183-192
Declaration on Environment and Development . 417 - of herbs . . . . . . . . . . . . . . . . . . . . . . 43-57
Decomposition processes . . . . . . . . . . . . . . . 297 - of herpetofauna . . . . . . . . . . . . . .. 315-323
Deep Ecology . . . . . . . . . . . . . . . . . . . 467, 472 - of Macrolepidoptera in Britain . . . . . . 30, 35
Deep-seated landslides/slope failures . . . . . 405, 406 - of Microlepidoptera . . . . . . . . . . . . . . 29-42
Deer . . . . . . . . . . . . . . . . . . . .. 343,427,459 - of small mammals . . . . . . . . . . . . . 175-182
Deforestation . . . . . . . . . . 357,452,455,513,519 - of termites . . . . . . . . . . . . . . 12-14,297-305
in Southeast Asia . . . . . . . . . . . . . . . . . . 417 - of trees . . . . . . . . . . . . . . . . . . . . 139-150
Density - of barking deer . . . . . . . . . . . . . . . 344 Doi Inthanon Park, Thailand . . . . . . . . . . . . . 421
- of kapur stands . . . . . . . . . . . . . . . . . . 354 Dolichoderinae . . . . . . . . 117, 119-122,293,294
- of mammals after logging . . . . . . . . 341-346 DolichopodidaelDolichopodinae .. 95, 183-188, 191
-ofmurids . . . . . . . . . . . . . . . 176,180,181 Domatia . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
- of primates . . . . . . . . . . . . . . . . . . 343, 344 Dominance, of tree families . . . . . . . . . . 139-142
- of squirrels . . . . . . . . . . . . . . . . . . 343, 344 Doona-Ca/ophyllum-Syzygium series .... . . .. 505
558 GENERAL INDEX
Doona-Dipterocarpus-Mesua series . . . . . . . .. 505 - processes . . . . . . . . . . . . . . . . . . . 260, 268
Dorylinae . . . . . . . . . . . . 117, 119-121,359,360 - rates. . . . . . . . . . . . . . . . .. 268,271,272
Doves, fruit . . . . . . . . . . . . . . . . . . . . . . . 288 Erythroxylaceae . . . . . . . . . . . . . . . . . . . . . 483
Dragonflies . . . . . . . . . . . . . . . . . . . . . . . . . 94 Ethiopia . . . . . . . . . . . . . . . . . . . . . . . . . . 541
Drainage, of peat swamp forest . . . . . . . . . . . 414 Ethnobotany . . . . . . . . . . 421,435,457,459,462
Drainage area in streams . . . . . . 377-379, 383-385 - categories/classificiation . . . . . . . . . 435-450
Drainage density . . . . . . . . . . . . . 393,397,404 Ethospecies . . . . . . . . . . . . . . . . . . . . 78, 81, 83
Drakensburg Mountains, Kenya . . . . . . . . . . . 541 Eucalyptus forests . . . . . . . . . . . . . . . . . . . .. 4
Drosophilidae . . . . . . . . . . . . . . . . . . . . . . 169 Eumetazoa . . . . . . . . . . . . . . . . . . . . . . . . . 5
Dry rainforests . . . . . . . . . . . . . . . . . . . . . 186 Euphorbiaceae . . . .. 117, 139, 141-143, 167,225,
Dryopteridaceae . . . . . . . . . . . . . . . . . . . . . . 47 325,327,328,428,482,483
Dumoga-Bone Park, Sulawesi . . . . . . . . 8,9, 421 Europe . . . . . . . . . . . . . . . . . . . 217,220,223
Durians . . . . . . . . . . . . . . . . . . . . . . . . . . 429 Evapotranspiration . . . . . . . . . . . . 391, 394, 403
Dusuns . . . . . . . . . . . . . . . . . . . . . . . 435-449 Evenness index, ant distribution . . . . . . . . . . . 360
- ethnobotany . . . . . . . . . . . . . . . . . 437, 447 Evolution . . . . . . . . . . . . . .. 172, 287, 291, 425
Dytiscidae . . . . . . . . . . . . . . . . . . . . . . . . . 94 - of ant-plant interactions . . . . . . . . . . . . . 294
Earthworms . . . . . . . . . . . . . . . . . . . . . . . . 90 - of biodiversity . . . . . . . . . . . . . . . . . . . 295
Earwigs . . . . . . . . . . . . . . . . . . . . . . . . . . . 90 - of dioecy in figs . . . . . . . . . . . . . . . . . 287
East Africa . . . . . . . . . . . . . . . . . . . . . 105, 109 - of diversity in the Lycaenidae . . . . . . . .. 291
East Kalimantan . . . . . . . . . . . . . . . . . . . .. 452 - of gynodioecious figs . . . . . . . . . . . . . . 287
Ebenaceae . . . . . . . . . . . . . . 141-143, 163,328 - of hilIslopes . . . . . . . . . . . . . . . . . . . . 404
EOOla Zaire . . . . . . . . . . . . . . . . . . . . . . . . 470 - of large seeds . . . . . . . . . . . . . . . . . . . 225
Echinodermata . . . . . . . . . . . . . . . . . . . . . .. 5 Evolutionary pressures . . . . . . . . . . . . . . . . . 284
Echiura . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 Existence values of forests . . . . . . . 417, 418, 545
Ecology - balance . . . . . . . . . . . . . . . . . . . . 461 Exploitation of Indonesian forest . . . . . . . 451-456
- diversity . . . . . . . . . . . . . . . . . . . . . . . 64 Extinction. . . . . . . . .. 67-69, 152, 341, 345, 469
- indices for fish . . . . . . . . . . . . . . . . . . 307 Extractive reserves/cycles . . . . . . . . . . . . 477, 478
- preferences of flowering plant families ... 329 Extrafloral nectar/nectaries . . . . . . . . . . . 294, 363
- response . . . . . . . . . . . . . . . . . . . . . . 325 Eyestalks of diopsid flies . . . . . . . . . . . . . . . 172
- role of peat lands . . . . . . . . . . . . . . . . . 414 F species (of birds) . . . . . . . . . . . . . . . . . .. 539
- segregation . . . . . . . . . . . . . . . . . . 311, 339 Fabaceae . . . . . . . . . . . . . . . . . . 142, 293, 483
- segregation model . . . . . . . . . . . . . . . . 313 Factor of safety (of slopes) . . . . . . . . 399-401, 403
- shade tolerance . . . . . . . . . . . . . . . . . . 216 Fagaceae . . . . . . . . . . . . 142, 326, 327, 329, 330
- tolerance . . . . . . . . . . . . . . . . . . . 219, 226 Family Importance Value (of trees) ... 139,141-143
- zones of freshwater fish. . . . . . . . . . . . . 383 FAO Tropical Forestry Action Plan . . . . . . . . 503
Economics - analyses . . . . . . . . . . . . . . . . . 545 Feeding behaviour - of bees . . . . . . . . . . . . . . 59
- botany . . . . . . . . . . . . . . . . . . . . . . . . 457 - of fish . . . . . . . . . . . . . . . . . . . . . . . . 307
- factors . . . . . . . . . . . . . . . . . . . . . . . 417 Feeding guilds . . . . . . . . . . . . . . . . . . . . .. 127
- values and valuation ..... 417,457,467, 535 Ferns . . . . . . . . . . . . . . . . . 163, 167,442,443
Ecosystem hierarchy . . . . . . . . . . . . . . . . . . 490 Fertilisation (nutrients) . . . . . . . . .. 241, 246, 253
Edaphic factors . . . . . . . . . . . . . . . . . . . . . . 49 Fibres (piassava) . . . . . . . . . . . . . . . . . 477-487
Elaeocarpaceae . . . . . . . . . . . . . . . . . . . . . 327 Figs, pollination/reproduction . . . . . . . . . 283-289
Elapidae . . . . . . . . . . . . . . . . . . . . . . . . . . 318 Fiji . . . . . . . . . . . . . . . . . . . . . . . . . . 152, 523
Elephants . . . . . . . . . . . . . . . . . . . . . . . . . 545 Filicium-Euphoria-Artocarpus-Myristica
Elevation - effects on fish distribution .... 383, 385 transitional series . . . . . . . . . . . . . . . . .. 505
- effects on plant distribution . . . . . . 325 - 330 Finnish taiga ant community . . . . . . . . . . . .. 339
Elongated tortoise . . . . . . . . . . . . . . . . . . .. 156 Fire resistant weeds . . . . . . . . . . . . . . . . . . 516
Emergent tree species .... 247, 347, 367, 368, 373 First-order jackknife method . . . . . . . . . . 206, 207
Empidoidea/Empididinae/Empidoid flies .. 183-192 Fish . . . . . . . . . . . . . . . . . . . 307-314, 377-386
Endangered species/populations . 415, 535, 538, 539 - assemblages . . . . . . . . . . . . . . . . . . . . 311
Endau Rompin, Peninsular Malaysia . . . . . . .. 132 - distribution. . . . . . . . . . . . . . . . . . . . . 377
Endemic Bird Areas . . . . . . . . . . . . . . . . . . 540 - morphology . . . . . . . . . . . . . . . . . 308,310
Endemics - birds . . . . . . . . . . . . . . . . . . . . 428 - species diversity . . . . . . . . . . . . . . . . . 312
- dipterocarps in Sri Lanka . . . . . . . . . . .. 505 F1acourtiaceae . . . . . . . . . . .. 141-143,328,483
- Lycaenidae . . . . . . . . . . . . . . . . . . . . . 292 Flora Malesiana . . . . . . . . . . . . . . . . . . . .. 163
- plant species . . . . . . . 155,411,425, 503, 508 Floristic zones in Kenya . . . . . . . . . . . . . . .. 540
- primates . . . . . . . . . . . . . . . . ... 429, 430 Flowering synchrony, in figs . . . . . . . . . . . . . 288
Endemism . . . . . . . . . . . . . 9,59,427, 535, 540 Focal forests in Kenya . . . . . . . . . . . . . . . .. 542
- in Mentawai mammals . . . . . . . . . . . . . 427 Fogging . . . . . . . . . . . . . . . . . . . . . .. 10, 205
- in Sri Lanka . . . . . . . . . . . . . . . . . . .. 505 Food bodies as food sources for ants . . . .. 11, 363
- in vertebrate groups . . . . . . . . . . . . . .. 545 Food webs in bamboo internodes . . . . . . 85, 91, 96
Endomychidae . . . . . . . . . . . . . . . . . . .. 90, 93 Foraging tactics in fish . . . . . . . . . . . . . . . . 307
England . . . . . . . . . . . . . . . . . . . . . . . 105, 107 Foreign labourers in forestry . . . . . . . . . . . . . 452
Entoprocta . . . . . . . . . . . . . . . . . . . . . . . . . 5 Forest - conservation . . . . . . . . . . . . . . . . . . 452
Environment - damage . . . . . . . . . . . . . . . . . 513 - conservation planning . . . . . . . . . . . . . . 544
- effects of agriculture. . . . . . . . . . . . . .. 521 - conversion . . . . . . . . . . . . . . . . . . 414,415
- evaluation . . . . . . . . . . . . . . . . . . . . . 467 - destruction in Indonesia . . . . . . . . . . . . . 456
- management . . . . . . . . . . . . . . . . . . . . 417 - destruction in Thailand . . . . . . . . . . . . . 521
- preservation . . . . . . . . . . . . . . . . . . . . 474 - economic rent . . . . . . . . . . . . . . . . . . . 454
- values . . . . . . . . . . . . . . . . . . . . . . . . 478 - exploitation . . . . . . . . . . . . . . . . . 451-455
Ephemeroptera . . . . . . . . . . . . . . . . . . . . . 310 - financial returns . . . . . . . . . . . . . . . . . 419
Epidermal organs of lycaenidae . . . . . . . . . . . 293 - food products . . . . . . . . . . . . . . . . . . . 518
Epiphytes . . . . . . . . . . . . . . . . . . . . . . 428, 429 - legislation . . . . . . . . . . . . . . . . . . . . . 455
- effects on arthropod diversity . . . . . . . . . . 27 - plantations ... . . . . . . . . . . . . . .. 277-281
Equilibrium status of hillslopes . . . . . . . . . . . 407 - population dynamics . . . . . . . . . . . . . . . 411
Ericaceae . . . . . . . . . . . . . . . . .. 325, 329, 330 - products . . . . . . . . 417-419,435,451-2,454
Erosion ... 260,271,272,303,397,398,404,406 - profile diagrams . . . . . . . . . . . . . . . . . 248
GENERAL INDEX 559
Forest (cont.) - regeneration . . . . . . . . . . . . . 516 - utilisation by fish . . . . . . . . . . . . . . . . . 308

- reserves . . . . . . . . . . . . . . . . . . . . 503, 536 Hainan, China . . . . . . . . . . . . . . . . . . . . . . 489
- resources . . . . . . . . . . . . . . . . . . . 453, 517 Hansa Bay, Papua New Guinea . . . . . . . . 184, 190
- structure . . . . . . . . . . . . . . . . . . . . . . 347 Hanunoo people, Southeastern Mindoro . . . . . . 460
- sustainability . . . . . . . . . . . . . . . . . . . . 452 Hawaiian Islands . . . . . . . . . . . . . . . . . 152,217
Formicidae . . . . . . . . . . . . . . . . 6,93,117,206 Heath forest (kerangas) .. 13,38,44, 412, 481, 489
Formicinae . . . . . 117, 119-122,332,357,359,360 Hemichordata . . . . . . . . . . . . . . . . . . . . . . . 5
French Guiana . . . . . . . . . . . . . . . . . . . . . . . 64 Hemiptera . . . . . . . . . . . 5, 8, 9, 14, 92, 93, 310
Frequency distribution - of trees . . . . . . .. 139-141 Herbivory, role of ants . . . . . . . . . . . . . . . . . 10
- of rainfall events .. . . . . . . . . . . . . . . . 394 Herbs . . . . . . . . . . . . . . . . . . . . . . 43-57,222
Freshwater fish communities . . . . . . . . .. 307-314 Herpetofauna . . . . . . . . . . . . . . . . . . .. 315-323
Freshwater swamp forest ... . . . . . . 411-415, 429 Hesperiidae . . . . . . . . . . . . . . . .. 125, 127, 131
Frogs - litter . . . . . . . . . . . . . . . . . . . . . . . 322 Heteropoda . . . . . . . . . . . . . . . . . . . . . . . . . 92
- tree . . . . . . . . . . . . . . . . . . . . . . . . . 322 Heterostyly in figs . . . . . . . . . . . . . . . . . . . 284
- vocalisation . . . . . . . . . . . . . . . . . . 80, 82 Hexapoda . . . . . . . . . . . . . . . . . . . . . . . . . . 5
Frugivorous animals . . . . . . . . . . . . . . . 283, 288 Highland forests in Thailand . . . . . . . . . . . . . 513
Fruit doves . . . . . . . . . . . . . . . . . . . . . . .. 288 Hillslope characteristics . . . . . . . . . . . .. 387-409
Functional value of trees . . . . . . . . . . . . . . . 468 Hindumkanda, Sri Lanka . . . . . . . . . . . . . .. 506
Gallery rainforest . . . . . . . . . . . . . . . 67, 68, 133 Histeridae . . . . . . . . . . . . . . . . . . . . . . . . . . 93
Gamma-rarity . . . . . . . . . . . . . . . . . . . . . . 195 Hmong hill tribes, Thailand . . . . . . . . . . . . . 514
Gap seedlings . . . . . . . . . . . . . . . . . . . 370, 374 HomopteralHomopterans . 5, 10,24,293, 294, 310
Gap-phase species. . . . . . . . . .. 50, 227, 367-375 Honeybees . . . . . . . . . . . . . . . . . . . . 61,63, 64
Gas exchange . . . . . . . . . . . . . . . 369, 370, 374 Honeycreepers . . . . . . . . . . . . . . . . . . . . .. 152
Gastromyzontins . . . . . . . . . . . . . . . . . 309,310 Hooded pitta . . . . . . . . . . . . . . . . . . . . . .. 156
Gastrotricha . . . . . . . . . . . . . . . . . . . . . . . . 5 Hornbills . . . . . . . . . . . . . . . . . . 104, 153, 155
Gekkonidae . . . . . . . . . . . . . . . . . . . . . . . . 318 - bushy-crested . . . . . . . . . . . . . . . . . . . 155
Generalist plant species . . . . . . . . . . . . . . . . 327 - great . . . . . . . . . . . . . . . . . . . . . . . . . 155
Genetic resource . . . . . . . . . . . . . . . . . . . . 429 - helmeted . . . . . . . . . . . . . . . . . . . . . . 155
Genitalia of dolichopodids . . . . . . . . . . . 188, 189 - rhinoceros . . . . . . . . . . . . . . . . . . . . . 155
Geographical Information Systems . . . . . 316,406, - white-crowned . . . . . . . . . . . . . . . . . . 155
489-494, 495-502 - wreathed . . . . . . . . . . . . . . . . . . . . .. 155
Geographical variation in Presbytis ........ 71-76 Horton-Strahler stream classification . . . . . . . . 378
Geomorphology . . . . . . . . . . . 259,387, 392, 407 Host-specificity of fig pollinators . . . . . . . . . . 283
Gerlach-type runoff plots . . . . . . . . . . . . . . . 262 Host-tree specificity of arthropods . . . . . . . . . . 25
Germany . . . . . . . . . . . . . . . . . . . . . . . . . . 16 Huastec people, Mexico . . . . . . . . . . . . . . . . 463
Germination . . . . . . . . . . . . . . . . 217,223,227 Human commensals . . . . . . . . . . . . . . . 315,321
Gesneriaceae . . . . . . . .. 43, 46, 47, 50, 156, 221 Human immunodeficiency virus . . . . . . . . . . . 470
Ghana . . . . . . . . . . . . . . . . . . . . . . .. 10, 357 Hurulu Forest Reserve, Sri Lanka . . . . . . . . . 507
Gibbons . . . . . . . . . . . . . . . . . . . . . . . 154, 343 Hutchinsonian ratios . . . . . . . . . . . . . . . . .. 179
- Bornean . . . . . . . . . . . . . . . . . . . . . . 343 HybotidaelHybotinae . . . . . . . . . . . . . . . 183-192
- Kloss . . . . . . . . . . . . . . . . . . . . . . . . 429 Hydrograph . . . . . . . . . . . . . . . . . . . . 395, 397
- white-handed . . . . . . . . . . . . . . . . . . . 155 Hydrological processes . . . . . . . . . . . . . 401, 407
Gingers . . . . . . . . . . . . . . . . . . . . . . . 442, 443 Hydrology . . . . . . . . . . . . . . . . . 259, 387, 394
Glaciation . . . . . . . . . . . . . . . . . . . . . . . . . . 71 - influence of termites . . . . . . . . . . . . . . . . 14
Global Positioning Systems (GPS) . . . . . . 496, 499 Hydrophilidae . . . . . . . . . . . . . . . . . . . . 93, 94
Gnaphosidae . . . . . . . . . . . . . . . . . . . . . . . . 92 Hydrophorinae . . . . . . . . . . . . . . . . . . . 185, 187
Gnathostomulida . . . . . . . . . . . . . . . . . . . . . 5 Hymenoptera. .. 5,8,9, 14,24,93, 117,283,310
Gobiidae . . . . . . . . . . . . . . . . . . . . . . . . . 380 Hypericaceae . . . . . . . . . . . . . . . . . . . . . . 328
Gombak Field Station, Peninsular Malaysia .... 97 Hypoxidaceae . . . . . . . . . . . . . . . . . . . . . . 445
Gradient, of streams .. . . . . .. 377-379,383,385 Hystricidae . . . . . . . . . . . . . . . . . . . . . 177, 181
Gradients, topographical . . . . . . . . . . . . . . . 393 Ibans . . . . . . . . . . . . . . . . . . . . . . . . . . . . 421
Graminae/Grasses . . .. 50, 228, 442, 443, 445, 446 Icacinaceae . . . . . . . . . . . . . . . . . . . . . 142, 483
Green-backed Narcissus flycatcher . . . . . . . .. 156 Ichthyophiidae . . . . . . . . . . . . . . . . . . . . . . 317
Gross domestic product ... . . . . . . . . . . . . . 420 Importance Value Index for trees . .. 139, 140, 143
Growing stock patterns, trees . . . . . . . . . . . . 489 Improvement fellings . . . . . . . . . . . . . . . . . . 348
Growth performance, dipterocarps 277-281, 347-355 India . . . . . . . . . . . . . . . . . . . . . . . . . 303, 320
Growth rates of trees . . . . . 216,217,219,280,374 Indicator groups - of organisms . . . . . . . . . 15, 16
Gryllidae . . . . . . . . . . . . . . . . . . . . . . . 80, 84 - of species. . . . . . . . . . . . . .. 357, 535, 545
Guangdong, China . . . . . . . . . . . . . . . . . . . 489 Indigenous knowledge ... . . . . . . . . . . . 457, 459
Guilds, ants . . . . . . . . . . . . . . . . . . . . 331, 339 land use. . . . . . . . . . . . . . . . . . . . . . . . 477
Guineo-Congolian forests . . . . . . . . . . . . . .. 541 Indirect use values . . . . . . . . . . . . 417,418, 535
Gulf of Carpentaria, Australia . . . . . . . . . . . . . 68 Indonesia . . . . . . . . . . . . . . . . . . 6,8, 163,327,
Gunung Danum, Sabah . . . . . . . . . . . . . . . . 119 388, 398, 412, 421, 425, 451, 452, 455, 524
Gunung Leuser Park, Sumatra . . . . . . . . . . . . 421 Indonesian Borneo . . . . . . . . . . . . . . . . 411, 414
Gunung Mulu National Park, Sarawak .. 37,47,48, Indonesian Government legislation . . . . . . . . . 451
132, 139, 143,255,297,303-304, 298 392, 421 Indonesian National Development Plan ... 452, 453
Gunung Silam, Sabah . . . . . . . . . . . . . . 253, 255 Indonesian Selective Cutting System . . . . . . . . 412
Guyana . . . . . . . . . . . . . . . . . . . . . . . . . . 277 Infiltration rates . . . . . . . . . . . . . . . . . . 395, 407
Gymnophiona . . . . . . . . . . . . . . . . . . . . . . 317 Infinite slope analysis . . . . . . . . . . . . . . . . . 406
Gynodioecy in figs . . . . . . . . . . .. 283, 284, 287 Infinite slope model . . . . . . . . 387,398,399,407
Habitats - classification . . . . . . . . . . . . . . . . 384 Insectivoraiinsectivores . . . . . . 175, 177, 181,427
- disturbance and small mammals. . . . . . .. 175 Insect sounds . . . . . . . . . . . . . . . . . . . .. 77, 83
- diversity . . . . . . . . . . . . . . . . . . . . . . 322 Instrumental values of trees . . . . . . . . . . . 467, 468
- fragmentation . . . . . . . . . . . . . . . . . . . . 68 Integrated Multi-Sectoral Information System
- partitioning in streams . . . . . . . . . . . . . . 307 (IMSIS) . . . . . . . . . . . . . . . . . . . .. 489-494
- preferences of ants . . . . . . . . . . . . . . . . 339 Inter Tropical Convergence Zone . . . . . . . . . . 388
- preferences of empidoid flies . . . . . . . . . 187 Interception of rainfall . . . . . . . . .. 263, 264, 394
560 GENERAL INDEX
International Amazon Ecosystem MAB Project . 489 Las Tuxtlas Forest, Mexico . . . . . . . . . . . . . 217
International Tropical Timber Agreement . . . .. 503 Late-tolerant plants . . . . . . . . . . . . . . . . 219,220
International Tropical Timber Organisation Lathridiidae . . . . . . . . . . . . . . . . . . . . . . . . 93
(ITTO) . . . . . . . . . . . . . . . . . . . . . 456, 490 Latin America . . . . . . . . . . . . . . . . . . . . . . 477
Internode community of bamboos . . . . . . . . . 85-99 Latitudinal distribution of plants . . . . . . . . . . . 325
Interspecific interactions, of ants . . . . . . . 332-335 Latosols . . . . . . . . . . . . . . . . . . . . . . . . .. 260
Intrinsic value of a tree . . . . . . . . . . . . . . . . 472 Lauraceae/laurels . . .. 142,223-225, 325-330, 483
Inventories . . . . . . . . . . . . . . . 43,208,315,459 Leaf beetles . . . . . . . . . . . . . . . . . . . . . . . . 94
Inventory methods . . . . . . . . . . . . . . . . 321, 523 Leaf litter. . . . . . . . . . . . . . . . .. 255,267,304
Inventory of rattan . . . . . . . . . . . . . . . . . .. 523 - effect on sediment loss . . . . . . . . . . . . . 266
Irian Jaya . . . . . . . . . . . . . . . . . . . . . . 425, 451 - effect on sediment retention . . . . . . . . . . 265
Irrigation . . . . . . . . . . . . . . . . . . . . . . . . . 229 - role in preventing erosion . . . . . . . . . . . 259
Island biogeography . . . . . . . . . . . . . . . . . . 433 - rates . . . . . . . . . . . . . . . . . . . . . . . . . 249
Isopoda . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 Leaves - morphology. . . . . . . . . . . . . .. 367-370
Isoptera . . . . . . . . . . . . . . . . . . . . . . .. 5, 303 - specific leaf area . . . . . . . . . . . . . . 367, 370
ITTO Target 2000 . . . . . . . . . . . . . . . . . . . 493 - thickness . . . . . . . . . . . . . . . . . . . 368, 370
IUCN categories . . . . . . . . . . . . . 506, 507, 538 Lecythidaceae . . . . . . . . . . . . . . . . . . . . . . 142
Ivory Coast . . . . . . . . . . . . . . . . . . . . . . . . 272 Legislation, forests in Indonesia . . . . . . .. 451-456
Ixonanthaceae . . . . . . . . . . . . . . . . . . . 141, 142 Leguminosae . . . . . . . . . . . . . . . . . . . . 143,225
Jaccard's coefficient of similarity . . . . . . . . . . 383 Lentic zone . . . . . . . . . . . . . . 307,309,310,312
Jackknife method of estimating species richness . 198 Lepidopsocidae . . . . . . . . . . . . . . . . . . . . . . 92
Jamaica/Jamaican montane forests . . . . . . 223, 255 Lepidoptera .. 4, 5, 9, 10, 14,23, 27, 29, 93, 310
Japan. . . . . . . . . . . . . . . . . . . .. 217,220,225 - Lycaenidae and ants . . . . . . . . . . . . . . . . 12
Java . . . . . . . . . . 72,74,91,391,425,433,455 - Pyraloidea .. . . . . . . . . . . . . . . . . . . 29-42
Johor, Peninsular Malaysia . . . . . . . . . . . . . . 347 Leptanillinae . . . . . . . . . . . . . . . . . . . . . . . 359
Jungle Sanctuaries . . . . . . . . . . . . . . . . . . . 507 Levins index of niche breadth . . . . . . . . . 335, 338
Kadazan . . . . . . . . . . . . . . . . . . . . . . . . . . 421 Libellulidae . . . . . . . . . . . . . . . . . . . . . . . . . 94
Kakadu, Australia. . . . . . . . . . . . . . . . . . . . 473 Liberia . . . . . . . . . . . . . . . . . . . . . . . . . . 541
Kakamega Forest, Kenya ..... 107, 109,541,542 Life-Forms in folk classification . . . . . . .. 435-449
Kalimantan .. 75, 163,269,411,414,421,451,455 Light - compensation points ... 367,369, 370, 373
Kampong Batang Duri, Brunei Darussalam .... 242 - regimes . . . . . . . . . . . . . . . . . . . . 216, 374
Kampong Kapok, Muara, Brunei Darussalam ... 34 - tolerance . . . . . . . . . . . . . . . . . . . . . . 217
Kanching Forest Reserve, Malaysia ..... 347,348 Light-demanding plants . 216-218,220,222,26,227
Kapit Division, Sarawak . . . . . . . . . . . . . . . 341 Limestone hill flora . . . . . . . . . . . . . . .. 153,155
Kapuas river basin, Indonesia . . . . . . . . . . . . 415 Linaceae . . . . . . . . . . . . . . . . . . . . . . 141, 142
Kapur Forests, Malaysia . . . . . . . . . . . . 347-355 Lions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 418
Karimata island, Indonesia . . . . . . . . . . . . . . . 73 Lithocarpus-Castanopsis rain forest . . . . . . . . . 284
Katydids . . . . . . . . . . . . . . . . . . . 77,79, 80, 82 Lithosols . . . . . . . . . . . . . . . . . . . . . . . . . 260
Kayans . . . . . . . . . . . . . . . . . . . . . . . . . . 422 Litter - gaps . . . . . . . . . . . . . . . . . . . . 217,227
Kayap6 Indians . . . . . . . . . . . . . . . . . . . . . 463 - production . . . . . . . . . . . . . . . . . . 297, 347
Kelabits . . . . . . . . . . . . . . . . . . . . . . . . . . 422 - traps . . . . . . . . . . . . . . . . . . . . . . . . . 245
Kendall's rank correlation coefficient . . . . . . . 383 Litterfall . . . . . . . . . . . . 241, 245, 247, 249, 253
Kenya .. . . . . . . . . . . .. 105, 107-109,418, 535 - nitrogen concentrations .. . . . . . . . . . . . 256
Kenya Indigenous Forest Conservation Project . . 537 - rates . . . . . . . . . . . . . . . . . . . . . . . . . 268
Kenyahs . . . . . . . . . . . . . . . . . . . . . . . . . . 421 Lizards . . . . . . . . . . . . . . . . . . . . . . . 317,319
Kerangas (heath) forest .. 13,38,44, 412, 481, 493 - flying . . . . . . . . . . . . . . . . . . . . . . . . . 97
Key Ikeystone species . . . . . . . . . . . . . .. 15, 283 - monitor . . . . . . . . . . . . . . . . . . . . . . . 321
Khao Yai National Park, Thailand . . . . . . 419, 421 Llavac, Luzon, Philippines . . . . . . . . . . . . . . 237
Kimberley District, Australia . . . . . . . . . . . . . . 67 Local species richness .... . . . . . . . . . . . .. 193
Kinabalu National Park, Sabah .. 77, 117, 132,421 Loganiaceae . . . . . . . . . . . . . . . . . . . . 141, 142
Kinorhyncha . . . . . . . . . . . . . . . . . . . . . . .. 5 Logged-over forest . . . . . . . . . . . . . . . . . . . 413
Kiribati . . . . . . . . . . . . . . . . . . . . . . . . . . 152 Logging . . . . . . . . . . 341-346,411-414,419,420,
Koppen System . . . . . . . . . . . . . . . . . . . . . 428 425-433, 451-456
Kuala Belalong Field Studies Centre, - illegal . . . . . . . . . . . . . . . . . . . . . 421,454
Brunei Darussalam .... 44, 101, 125, 169, 175, - impact on Siberut . . . . . . . . . . . . . . 431, 432
262, 298, 307, 315, 332, 368, 387, 389, 524 Long-fallow forest management systems ... 481,484
Kuala Lompat, Pahang, Malaysia . . . . . . . . . . 119 Long-horned beetle . . . . . . . . . . . . . . . . . . . . 94
Kuper Range Wildlife Conservation Area, Papua Longans . . . . . . . . . . . . . . . . . . . . . . . . . . 518
New Guinea . . . . . . . . . . . . . . . . . . . . . 284 Longevities - of seeds . . . . . . . . . . . . . . . . . 226
La Selva Forest, Costa Rica . . . . . . 217,219,229 - of trees . . . . . . . . . . . . . . . . . . . . . . . 216
Labiatae . . . . . . . . . . . . . . . . . . . . . . . 325, 506 Loranthaceae . . . . . . . . . . . . . . . . . . . . . .. 167
Labiduridae . . . . . . . . . . . . . . . . . . . . . . . . 92 Lotic zones. . . . . . . . . . . . . . . .. 307,309,310
Lagriidae .. . . . . . . . . . . . . . . . . . . . . . . . . 93 Low thinning. . . . . . . . . . . . . . . . . . . . . . . 347
Laing Island, Papua New Guinea . . . . . . . . . . 184 Lower montane forest . . . . . . . . . . . . . . 241, 284
Lake Turkana, Kenya . . . . . . . . . . . . . . . . . 535 Lowland dipterocarp forest.. 30,34,77, 108, 117,
Lake Victoria, Kenya . . . . . . . . . . . . . . 535, 541 126,297,315,331,332,358,368,393,495,524
Land - ethics . . . . . . . . . . . . . . . . . . . . 467, 471 Lowland rain forest . . . . . . . . . . . . . . . . 215,241
- tenure . . . . . . . . . . . . . . . . . . . . . 417, 421 Luxembourg . . . . . . . . . . . . . . . . . . . . . . . 268
- use . . . . . . . . . . . . . . . . . . . . . . . . . . 492 Luzon, Philippines . . . . . . . . . . . . . . . . . . . 235
Landscape stability evaluation .. . . . . . . . . . . 406 Lycaenidae/Lycaeninae . 12, 125, 127, 131,291-296
Landslides . . . . . . . . . . . . . . . . . . 401,405-407 MacArthur & Levins' equation . . . . . . . . . . . 316
Langurs - dusky . . . . . . . . . . . . . . . . . . 154, 156 Macaques - Mentawai . . . . . . . . . . . . . . . . . 429
- maroon . . . . . . . . . . . . . . . . . . . . . . . 344 - pig-tailed . . . . . . . . . . . . . . . . . . . . . . 344
- Mentawai . . . . . . . . . . . . . . . . . . . . . . 429 - stump-tailed . . . . . . . . . . . . . . . . . 153, 155
- pig-tailed . . . . . . . . . . . . . . . . . . . . . . 429 Macrolepidoptera . . . . . . . . . . . . . . 9, 29, 30, 35
- white-fronted . . . . . . . . . . . . . . . . . . . 344 - in Borneo . . . . . . . . . . . . . . . . . . . . . . 39
Lapse rate on Bukit Belalong . . . . . . . . . . . . . 243 Macrotermitinae . . . . . . . . . . . . 297-299, 301-303
GENERAL INDEX 561
Madagascar . . . . . . . . . . . . . . . . . . . . . . . 505 Mixed dipterocarp forest . . . . . . . 13,38,43, 139,

Madang, Papua New Guinea . . . . . . . . . . . .. 183 140, 143, 175, 259, 260, 268, 341, 367, 377,
Mae Soi, Chiang Mal, Thailand . . . . . . . . . . . 513 413, 436, 489, 493
Magnesium concentrations in soil . . . . . . . . . . 261 Mkongani Forest Reserve, Kenya . . . . . . . . . . 542
Malaise traps . . . . . . . . . . . . . . . . . . . . . . .. 8 Moisture retention curves/fluxes . . . . . . . . 396, 397
Malay archipelago . . . . . . . . . . . . . . . . . . . . n Mole crickets . . . . . . . . . . . . . . . . . . . . . . . 82
Malay Peninsula. 46, n, 74, 91, 169, In, 220-223 Mollusca . . . . . . . . . . . . . . . . . . . . . . . . . 4, 5
Malaysia/Malaya .... 13,117,178,221,268,277, Mombasa-Kilifi area, Kenya . . . . . . . . . . . . . 541
291,345,347,411,452 Mongoose, collared . . . . . . . . . . . . . . . . . . 343
Male secondary sexual characters . . . . . . . . .. 186 Monitor lizards . . . . . . . . . . . . . . . . . . . . . 321
Malesia . . . . . . . . . . . . . . . . . 64,225,325, 505 Monsoons. . . . . . . . . . . . . . . . .. 358, 388, 389
Malpighiaceae . . . . . . . . . . . . . . . . . . . . .. 483 Monsoon rainforest . . . . . . . . . . . . . . . . . . 67-69
Maludam Peninsula peat swamp forest, Montane forest ... 13,38,217,241,283,325-330
Indonesia . . . . . . . . . . . . . . . . . . . . . . . 411 - profile diagram . . . . . . . . . . . . . . . . .. 248
Maluku, Indonesia . . . . . . . . . . . . . . . . . . . 451 - in Sri Lanka . . . . . . . . . . . . . . . . . . . . 505
Mammals/Mammalia . . . . . . 5, 67-69, 80, 341, 427 Monterey pine . . . . . . . . . . . . . . . . . . . . . . 467
- dispersal of fruits & seeds . . . . . . . . . . . 224 Moonrat . . . . . . . . . . . . . . . . . . . . . . . . . . 156
- diversity . . . . . . . . . . . . . . . 342, 343, 538 Moraceae . . . . . . . . . . . . 142, 283, 327, 328, 483
- marsupials . . . . . . . . . . . . . . . . . . . . . . 68 Mortality (of Kapur) . . . . . . . . . . . . . . . . . . 353
- monkeys . . . . . . . . . . . . . . . . . . . . . . . 71 Mosquito larvae . . . . . . . . . . . . . . 88, 89, 91, 95
- small mammal diversity . . . . . . . . . . 175-182 Moult cycles of birds . . . . . 101, 102, 105, 107, 109
- surveys .... . . . . . . . . . . . . . . . . . .. 153 Mount Banahao, Luzon, Philippines . . . . . . . . 325
Man and Biosphere Program . . . . . . . . . . . . . 425 Mount Kaindi, Papua New Guinea . . . . . . . . . 284
Management planning . . . . . . . 463,482, 535, 545 Mount Makiling, Luzon, Philippines . . . . . . .. 326
Mandor Nature Reserve, West Kalimantan .... 412 Mouse, pencil-tailed tree . . . . . . . . . . . . . . . 156
Mangrove forests .. 34,38,183,186,190,411,429 Mousedeer - greater . . . . . . . . . . . . . . . . . . 156
Manilkara-Chloroxylon series . . . . . . . . . . . . 505 - lesser . . . . . . . . . . . . . . . . . . . . . 156, 343
Manu, Peru . . . . . . . . . . . . . . . . . . . . . . . 221 Mozambique . . . . . . . . . . . . . . . . . . . . . . . 541
Marantaceae . . . . . . . . .. 50, 228, 325, 446, 483 Mudpuddling. . . . . . . . . . . . . . .. 127, 129, 133
Marbled cats . . . . . . . . . . . . . . . . . . . . . . . 343 MuridaelMurids . . . . . . . . . . . 175, 177, 178, 181
Marshall Islands . . . . . . . . . . . . . . . . . . . .. 152 Mushrooms . . . . . . . . . . . . . . . . . . . . . . . . 517
Marsupials . . . . . . . . . . . . . . . . . . 68, 177, 288 Mustelids . . . . . . . . . . . . . . . . . . . . . . . .. 177
Mass movement of hillslopes . . . . .. 397, 398, 406 Mutualism . . . . . . . . . . . 283,287,291,295,303
Mastacembelidae . . . . . . . . . . . . . . . . . . . . 380 Mycetophilidae . . . . . . . . . . . . . . . . . . . 90, 93
Mata Ayer Forest Reserve, Peninsular Malaysia3, 155 Mycorrhizae . . . . . . . . . . . . . . . . . . . . . . . 224
Material value of trees . . . . . . . . . . . . . . . . . 468 Myristicaceae . . . . . . 141-143,325,326, 328, 428
Mating behaviour of stalk-eyed flies . . . . . . . . 169 Myrmecophily . . . . . . . . . . . . . .. 291, 292, 294
Mau Forest Reserve, Kenya . . . . . . . . . . 107, 542 Myrmecoxeny . . . . . . . . . . . . . . . 291,293,294
Mbuti people, Zaire . . . . . . . . . . . . . . . . . . 462 Myrmicinae .... 117,119,120-122,357,359,360
McIntosh's Diversity Index . . . . . . . . 105-107, 109 Myrsinaceae . . . . . . . . . . 141, 142, 328, 482, 483
Medeterinae . . . . . . . . . . . . . . . . . . . . . . . 185 Myrtaceae . . . . . . . . 141-143,247,293,325-327,
Medicinal plants . . . . . . . . . . . . . . . . . . . . . 517 329, 330, 482, 483
Mediterranean Basin . . . . . . . . . . . . . . . . . . 223 Mysmeridae . . . . . . . . . . . . . . . . . . . . . . . . 92
Megabats ... . . . . . . . . . . . . . . . . . . . . . . 288 Namibia . . . . . . . . . . . . . . . . . . . . . . . . . . 332
Megatherms (of plants) . . . . . . . . . . . . . 325, 327 Nandi Forest, Kenya. . . . . . . . . . . . . . . . .. 107
Melastomataceae .... 43,50, 142,219,221,226, Nanga Medamit, Sarawak . . . . . . . . . . . . . . . 390
326, 327, 482, 483 Nanga Tayap, Kalimantan . . . . . . . . . . . 412-414
Meliaceae . . . . . . . . . . . . . .. 117, 142, 325-330 Nanga Tekalit, Sarawak . . . . . . . . . . . . . . . . 317
Mende people, Sierra Leone . . . . . . . . . . . . . 461 Nasutitermitinae . . . . . . . . . .. 297, 298, 300-302
Mendolong, Sabah . . . . . . . . . . . . . . . . . . . 391 National Biosphere Reserves . . . . . . . . . . 425, 508
Mentawai Islands, Indonesia . . . . . . . . 71,74,425 National Parks . . . . . . . . . 417, 421, 432, 507, 535
Merimbun village, Brunei Darussalam .... 436, 437 Natural resource management . . . . . . . . . . . . 513
Mersing Forest Reserve, Malaysia . . . . . . . . . 348 Nature Reserves . . . . . . . . . . . . . . . . . . . . . 507
Mesophyll vine forest . . . . . . . . . . . . . . . . . . 23 Nauru . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
Mesozoa . . . . . . . . . . . . . . . . . . . . . . . . . . 5 Nectar . . . . . . . . . . . . . . .. 59, 62, 65, 104, 184
Mexico . . . . . . . . . . . . . . . . . . . . . . . . . . . 64 Nectaries, extra-floral . . . . . . . . . . . . . . . . . . 11
Microhabitats . . . . . . . . . . . . . . . 315,331,363 Negative binomial method of estimating species
- analysis . . . . . . . . . . . . . . . . . . . . . . . 315 richness . . . . . . . . . . . . . . . . . . . . . 206, 207
- use . . . . . . . . . . . . . . . . . . . 315,316,320 Nemata . . . . . . . . . . . . . . . . . . . . . . . . . . 4,5
- utilisation by fish . . . . . . . . . . . . . . . . . 309 Nematodes . . . . . . . . . . . . . . . . . . . . . . . . . 4
Microhybotids . . . . . . . . . . . . . . . . . . . . . . 186 Nematomorpha . . . . . . . . . . . . . . . . . . . . . . 5
Microhylidae . . . . . . . . . . . . . . . . . . . . . . . 317 Nemertea . . . . . . . . . . . . . . . . . . . . . . . . . . 5
Microlepidoptera . . . . . . . . . . 9, 29-42, 133, 500 Nest distances in ants . . . . . . . . . . . . . . . . . 338
- in SE Asia . . . . . . . . . . . . . . . . . . . . . . 39 Nesting, in bees . . . . . . . . . . . . . . . . . . . . . . 61
- species richness . . . . . . . . . . . . . . . . . . . 31 Neurigoninae . . . . . . . . . . . . . . . . . . . . . .. 185
Micrometeorology of forest gaps . . . . . . . . . . 217 Neuroptera . . . . . . . . . . . . . . . . . . . . . . . .. 5
Micronesia .. . . . . . . . . . . . . . . . . . . . . .. 152 New Caledonia . . . . . . . . . . . . . . . . . . . . . 151
Microphoridae . . . . . . . . . . . . . .. 183, 184, 186 New Guinea . . . . . . . . . . . . . 151, 422, 425, 523
Microtherm (of plants) . . . . . . . . . . . . . . 325, 330 New Guinea forests . . . . . . . . . . . . . . . . . . 255
Migrant birds . . . . . . . . . . . . . . . . . . . 101, 105 New Hebrides . . . . . . . . . . . . . . . . . . . . " 109
Migration of fish . . . . . . . . . . . . . . . . . . . . 385 New Zealand . . . . . . . . . . . . . . . . . . . . 152, 186
Miletinae . . . . . . . . . . . . . . . . . . 291, 293, 295 Niche . . . . . . . . . . 307,315,321,326,331,339
Mimosaceae . . . . . . . . . . . . . . . . . . . . 293, 483 - breadth . . . . .. 307, 308, 310, 335, 336, 338
Mineral losses/sediment losses . . . . . . . . . 265, 267 - differentiation . . . . . . . . . . . . . . . . 151, 215
Miridae . . . . . . . . . . . . . . . . . . . . . . . . . . . 92 - distributions . . . . . . . . . . . . . . . . . . . . 331
Mist netting . . . . . . . . . . . . . . . . 101, 102, 108 - overlap . . . . . . . . . 307,308,311,316,321,
Mites . . . . . . . . . . . . . . . . . 89, 90, 92, 183, 186 322, 331, 335-339
562 GENERAL INDEX
Niche (conto) - overlap cladograms 311 0 0 0 0 0 0 0 0 0 224,271,272,274, 275, 283, 357, 417, 422
- parameters 335, 339
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Parasites - of ants 12
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
- segregation in Lycaenidae 291, 295 0 0 0 0 0 0 0 0 - of fig wasps 287

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Nieuwenhuis' bulbul 102 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Parazoa . . . . . . . . . . . . . . . . . . . . . . . . . . 5 0
Nigeria 0 0 0 0 13, 272, 458

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Parrots . . . . . . . . . . . . . . . . . . . . . . . . . . 288 0
Nitidulidae 0 0 0 0 93 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Pasoh Forest Reserve, Malaysia 13,297-304 0 0 0 0 0 0
Nitrification rates 246, 250, 255 0 0 0 0 0 0 0 0 0 0 0 0 0 Patch dynamics 0 503

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Nitrogen concentrations - in litterfall 241, 247 0 0 0 0 0 Pathogens 0 0 0 0 0 224

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
- in seeds 0 0 0 224 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Paussidae 0 0 0 0 0 12 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
- in soil 0 0 0 0 261 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Peak Wilderness Sanctuary, Sri Lanka 508 0 0 0 0 0 0 0
- on Bukit Belalong 249 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Pearson's correlation coefficient 129, 130 0 0 0 0 0 0 0 0
- losses 0 0 0 266, 267

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Peat swamp forest 38,411-414 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Noah Creek, Queensland, Australia 23 0 0 0 0 0 0 0 0 0 - as energy sources 414 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Nomadic lifestyle 462 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Pelage colour in Presby tis 71, 73 0 0 0 0 0 0 0 0 0 0 0 0 0
Non-pioneer species (of trees) 227,372 0 0 0 0 0 0 0 0 0 Pelagic species of fish 309, 310 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Non-timber forest products 419-421,477, 523 0 0 0 0 Pelobatidae 0 0 0 0 317

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
North America 217,220, 332

0 0 0 0 0 0 0 0 0 0 0 0 0 0 Penans 0 0 0 0 0 0 0 421 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
- deserts 0 0 0 332 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Peninsular Malaysia 88, 132, 153, 156, 180, 0 0 0 0 0
- temperate forests 332 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 277, 292, 297, 298, 304, 341, 345, 347, 373, 523
Northern Territory, Australia 67, 151 0 0 0 0 0 0 0 0 0 Penman evaporation/transpiration rates 251 0 0 0 0 0 0
Nuaulu ethnozoology 447 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Pentaphragmataceae 221 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Nuaulu people, Seram 462 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Pentatomidae 0 0 0 92

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Nuaulu, Seram 0459-4610 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Peradayan Forest Reserve, Brueni Darussalam 180 0
Nutmeg 00000000000000000000000000 229 Perlis, Malaysia 0 153

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Nutrients - in forest soils 241-257 0 0 0 0 0 0 0 0 0 0 0 0 Peru 0 0 0 0 0 0 9,221, 272, 419

0 0 0 0 0 0 0 0 0 0 0 0 0 0
- concentrations in streams 382 0 0 0 0 0 0 0 0 0 0 0 0 Pesticides - concentrations in Thailand 519 0 0 0 0 0 0 0
- losses 0 0 0 0 260
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 - contamination of streams 513 0 0 0 0 0 0 0 0 0 0 0 0
- role of termites in cycling 13 0 0 0 0 0 0 0 0 0 0 0 0 Pharmaceuticals 0 478

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
- supply 0 0 0 0 216
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Pharmacology 0 0458, 470
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Nyctaginaceae 325, 326, 328, 483

0 0 0 0 0 0 0 0 0 0 0 0 Phenology,ofbutterflies 125-133 000000000000
Nycterbiidae 0 0 0 91
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Pheromones 0 0 0186, 191
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Nymphalidae 0 0125, 127, 131

0 0 0 0 0 0 0 0 0 0 0 0 0 0 Philippines 235, 325, 429, 452, 524, 530
0 0 0 0 0 0 0
Oceania - forest bird diversity 151-152 0 0 0 0 0 0 0 0 0 Pholcidae 0 0 0 0 0 92, 95

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
- empidoid flies 185 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Phoridae/Phorids 87, 90, 93

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Ochnaceae . . . . . . . . . . . . . . . . . . . . . . . . 46 0 Phoronida 0 0 0 0 0 05 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Ochyroceratidae 0 92
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Phosphate mining 473
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Odonata 0 0 0 0 0 0 5, 310
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Phosphorus - concentrations in soil 261 0 0 0 0 0 0 0 0 0
Oecophoridae 0 0 4
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 - losses 0 0 0 0266, 267
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Oil of palosapis 0 235

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Photosynthesis 367,369,370-374
0 0 0 0 0 0 0 0 0 0 0 0
Olacaceae 0 0 0 0117,141,142
0 0 0 0 0 0 0 0 0 0 0 0 0 0 - efficiency 0 0 373
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Oleaceae 0 0 0 0 0 325
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Photosynthetically active photon flux density
Oligophagy 0 0 0 0 26
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 (PPFD) 0 0 0 0 367
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Onychophora 0 0 0 5
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Physico-chemical characteristics of forest
Oonopidae 0 0 0 0 92
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 streams 0 0 0 0 382
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Opiliones 0 0 0 0 0 5
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Physiological shade tolerance 216, 219, 226 0 0 0 0 0
Opium poppy 0 0 516

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Physiology of tree seedlings 367-375 0 0 0 0 0 0 0 0 0 0
Option values of forest 417, 418, 535 0 0 0 0 0 0 0 0 0 0 Phytotelmata 0 085, 86, 95, 97
0 0 0 0 0 0 0 0 0 0 0 0 0 0
Orchidaceae/Orchids 47, 50, 155 0 0 0 0 0 0 0 0 0 0 0 0 0 Pianka's Index 0 0 308

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Organic litter erosion 267 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Piassava palms & products 477-487 0 0 0 0 0 0 0 0 0 0 0
Organic matter - content of soils 261,263,266 0 0 0 Pieridae 0 0 0 0 125, 127, 129, 131
0 0 0 0 0 0 0 0 0 0 0 0
- losses 0 0 0 260, 266

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Pigs 0 0 0 0 0 0 0 250, 343, 459
0 0 0 0 0 0 0 0 0 0 0 0 0 0
Organic sediment losses 265 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 - bearded 0 0 0 343

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Orthic acrisols 0244, 259, 261

0 0 0 0 0 0 0 0 0 0 0 0 0 0 Pioneer trees 220, 221, 225-228, 367, 368, 372-374
0
Orthoptera 5, 9, 12, 24, 77, 80, 83,310, 336

0 0 0 0 Piperaceae 0 0 0 0 483
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Osteichthyes 0 0 0 5
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Pit vipers 0 0 0 0 0 320
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Otters 0 0 0 0 0 0 0 427
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Pitcher plants 0 0 95
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
- oriental small-clawed 343 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Pitfall trapping - of ants 117, 118, 122,358,361 0 0 0
Over-exploitation of forest products 536 0 0 0 0 0 0 0 0 - of herpetofauna 316 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Overlap of ant species 359 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 - of small mammals 176 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Owls - brown fish 156

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Plants - adaptations and ants 11 0 0 0 0 0 0 0 0 0 0 0 0 0 0
- dusky eagle 156

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 - domestication 462 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Oxygen pumps, trees as 472 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 - endemism 0 541

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Pachytroctidae 0 0 92
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Plantations 0 0 0 277, 280
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Pahang, Peninsular Malaysia 347 0 0 0 0 0 0 0 0 0 0 0 0 0 Plasticity of seedlings 367 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Palduay village, Thailand 514 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Platyhelminthes 0 05 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Palaearctic migrants (birds) 102, 109 0 0 0 0 0 0 0 0 0 0 0 Platystomatidae/Platystomatid flies 169, 191 0 0 0 0 0 0
Palaeomagnetic dating 392 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Plectoptera 0 0 0 0 310

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Palm oil plantations 425,433 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Poaceae 0 0 0 0 0 0 516

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Palmae/Palms 155, 166-7,429,442, 444, 446, 523

0 Podocarpaceae 0 325, 329
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Palosapis, resin yields 235-240 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Podsolic soils 0 0 428

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Panama 0 0 0 215,217,228,229,357
0 0 0 0 0 0 0 0 0 0 Pogonophora 0 0 0 05 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Pangolins 0 0 0 0 156, 427

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Pollination 0 0 0 283, 430
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Panthers 0 0 0 0 0 0 156
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 - in figs 0 0 0 0 283-289
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Panti Forest Reserve, Malaysia 348 0 0 0 0 0 0 0 0 0 0 0 Polonnaruwa period, Sri Lanka 507 0 0 0 0 0 0 0 0 0 0 0
Pantu, Sarawak 0 421

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Polygalaceae 0 0 0 142
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Papilionidae 0 125, 127, 129, 131

0 0 0 0 0 0 0 0 0 0 0 0 Polygonaceae 0 0 483
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Papua New Guinea 91, 183-186, 188, 189, 0 0 0 0 0 0 Polynesia 0 0 0 0 0151, 152
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
GENERAL INDEX 563
Ponerinae . . . . . . 117, 119-122,332,357,359,360 - Fund (Indonesia) . . . . . . . . . . . . . . . . . 455

Population density, of ants . . . . . . . . . . . . . . 363 - programmes . . . . . . . . . . . . . . . . . . . . 514
Population growth in Indonesia . . . . . . . . . . . 414 - projects in Thailand . . . . . . . . . . . . . . . 518
Population growth/densities in Sri Lanka ..... 506 Regeneration . . . . . . . . . . . . . 215, 216, 227, 349
Population of Kenya . . . . . . . . . . . . . . . . .. 536 - of kapur . . . . . . . . . . . . . . . . . . . 344-355
Population structure (of birds) . . . . . . . . . . .. 101 - of peat swamp forests . . . . . . . . 347,411-415
Porcupine - Malayan . . . . . . . . . . . . . . . . . . 156 Regeneration niche . . . . . . . . . . . . . . . . . . . 215
- thick-spined . . . . . . . . . . . . . . . . . 177, 343 Regional diversity of beetles in Borneo . . . . . . 208
Pore water pressure . . . . . . . . . . . . 387, 400-403 Regional species richness . . . . . . . . . . . . 193, 194
Poring Hot Springs, Sabah . . . . . . . . . 77,82, 117 Relict fauna . . . . . . . . . . . . . . . . . . . . . . . 433
Port Vila, Vanuatu . . . . . . . . . . . . . . . . . . . . 30 Relief - of the Batu Apoi Forest Reserve .. 394, 397
Posterity argument for conservation . . . . . . . . 468 - of Sri Lanka . . . . . . . . . . . . . . . . . . .. 504
Potassium, losses from erosion . . . . . . . . 266, 267 Reproductive behaviour of stalk-eyed flies .... 169
Potato, as crop in Thailand . . . . . . . . . . . . .. 519 Reproductive success in figs and fig-wasps . 283-289
Predation . . . . . . . . . . . . . . . . . . . .. 315, 385 Reptilia/Reptiles . . . . . . . . . . . . . . . . 5, 155,315
- by ants . . . . . . . . . . . . . . . . . . . . . . . . 11 - surveys of ... . . . . . . . . . . . . . . . . .. 153
- by wasps . . . . . . . . . . . . . . . . . . . . . . 283 Resettlement areas on Siberut . . . . . . . . . . . . 432
- of seeds . . . . . . . . . . . . . . . . . . . . . .. 224 Resettlement programmes in Thailand . . . . . . , 514
Preservation of global biodiversity . . . . . . . . . 503 Residual stand (after logging) . . . . . . . . . . . . 413
Prey diversity in fish . . . . . . . . . . . . . . . 308, 310 Resin yields in palosopis . . . . . . . . . . . . 235-240
Priapulida . . . . . . . . . . . . . . . . . . . . . . . . .. 5 Resources - depletion . . . . . . . . . . . . . . . . . 472
Primates . . . . . . . . . . 153, 155,341-343, 427, 539 - inventories . . . . . . . . . . . . . . . . . . . .. 523
Production forest . . . . . . . . . . . . . . ... 413, 451 - management . . . . . . . . . . . . . . . . . . . . 421
Profile diagrams of forest . . . . . . . . . . . . . . . 248 - overlap . . . . . . . . . . . . . . .. 332, 338, 339
Proteaceae . . . . . . . . . . . . . . . . . . . . . 142, 225 - partitioning . 84,291,295,307,311,331,332
Protected area system of Sri Lanka . . . .. 507 = 509 - utilisation . . . . . . . . . . . . . . . . . . . 315, 339
Protected forest in Indonesia . . . . . . . . . . . .. 414 Respiration rates (in plants) . . . . . .. 367, 373, 374
Proto-Malayan people . . . . . . . . . . . . . . . . . 433 Restinga (heath-like forests) . . . . . . . . . . . . . 478
Proto-terrestrial (plants) . . . . . . . . . . . . 47, 48, 50 Rhacophoridae . . . . . . . . . . . . . . . . . . . . . . 317
Protozoa . . . . . . . . . . . . . . . . . . . . . . 4, 5, 87 Rhamnaceae . . . . . . . . . . . . . . . . . . . . . . . 293
Prudential argument for conservation . . . . . . . 470 Rhaphiinae . . . . . . . . . . . . . . . . . . . . . . . . 185
Pselaphidae . . . . . . . . . . . . . . . . . . . . .. 12, 93 Rheophytic herbs . . . . . . . . . . . . . . . . . . . . . 48
Pseudococcids . . . . . . . . . . . . . . . . . . . . . . . 94 Rhinotermitidae " . . . . . . . . . . . . . . . . 297-303
Pseudomyrmicinae . . . . . . 117, \19-121,359,360 Rhizophoraceae .. . . . . . . . . . . . . . . . . . . . 325
Pseudophyllinae . . . . . . . . . . . . . . . . . . . . . . 81 Riau archipelago, Indonesia . . . . . . . . . . . . . . 72
Pseudoscorpiones .. . . . . . . . . . . . . . . . . . .. 5 Riau-Jambi forest, Indonesia . . . . . . . . . . . . . 451
Pseudoturnover . . . . . . . . . . . . . . . . . . 197, 205 Rice - bioassay . . . . . . . . . . . . . . 241, 246, 252
Psocoptera . . . . . . . . . . . . . . . 9, 23, 24, 90, 92 - cultivation . . . . . . . . . . . . . . . . . . . . . 517
Psychodidae . . . . . . . . . . . . . . . . . . . . . 90, 93 Rio de Janeiro, Argentina . . . . . . . . . . . . . . , 268
Pteridophytes . . . . . . . . . . . . . . . . . . . . . 47-50 Riparian forest, Brunei Darussalam . . . . . . . . , 175
Puerto Rico . . . . . . . . . . . . . . . . . . . . . . . 364 Ripple bugs . . . . . . . . . . . . . . . . . . . . . . . . . 95
Pyralidae . . . . . . . . . . . . . . . . . . . . . . . . . . 94 Rituals (relation to ethnobiology) . . . . . . . . . . 461
Quantum yield values . . . . . . . . . . . . . . . . . 374 River discharges . . . . . . . . . . . . . . . . . . . . 394
Queensland, Australia. 21,151,220,223,231,288 Riverine Forest, Siberut, Indonesia . . . . . . . . . 425
Quezon, Philippines . . . . . . . . . . . . . . . 235, 326 Rodentia/Rodents . . . . . " . . . . . . . . . . . 175, 177
Radiation, of empidoid flies . . . . . . . . . . . . . 183 Root competition . . . . . . . 215, 216, 223, 227-229
Rainfall - at Belalong .... 243,263,389,391,392 Root mats . . . . . . . . . . . . . . . . . . . . .. 267, 269
- at Danum Valley Field Centre . . . . . . . . 358 - effect on sediment retention . . . . . . . . . . 265
- data for Brunei Darussalam . . . . . . . 388-402 - role in preventing erosion . . . . . . . . . . . 259
- effect on erosion . . . . . . . . . . . . . . . .. 259 Rosaceae . . . . . . . . . . . . . . . . . . . . . . . . . 142
Rambai village, Brunei Darussalam . . . . . . . . 435 Rotifera . . . . . . . . . . . . . . . . . . . . . . . . . . 4, 5
RaminiRamin forest . . . . . . . . . . . . . . . 413-415 Rove beetles . . . . . . . . . . . . . . . . . . . .. 88, 89
Ramsar Wetlands . . . . . . . . . . . . . . . . . . . . 507 Royalties from timber . . . . . . . . . . . . . . . . . 420
Ranidae . . . . . . . . . . . . . . . . . . . . . . . . . . 317 Rubiaceae . . . . . . . .. 46,50, 142, 163,221,226,
Rapid Biodiversity Assessment . . . . . . . " 27, 321 327,429,444, 482, 483
Rare species and conservation . . . . . . . . . 538, 545 Runoff . . . . . . . . . . . . . . . . . . . . . 259,264-266
Rats . . . . . . . . . . . . . . . . . . . . . . . . . 154, 427 - and soil erosion ., 259-270,271-275,387-409
- black . . . . . . . . . . . . . . . . . . . . . . . . 152 Rural development in Indonesia . . . . . 451,452, 455
- brown spiny . . . . . . . . . . . . . . . . . . .. 177 Rutaceae . . . . . . . . . . . . . . . 326, 328, 441, 446
- dark-tailed tree . . . . . . . . . . . . . . . . . . 177 Sabah .. 34,77,97, 117, 132, 163,217,226,229,
- long-tailed giant . . . . . . . . . . . . . . . . . 177 253, 255, 264, 268, 341, 345, 357, 358, 420, 421
- Malaysian field ... . . . . . . . . . . . . . .. 177 Sabal Forest Reserve, Sarawak . . . . . . . . 489, 493
- Muller's . . . . . . . . . . . . . . . . . . . . . . 177 Sabiaceae . . . . . . . . . . . . . . . . . . . . . . 141,142
- Polynesian . . . . . . . . . . . . . . . . . . . .. 177 Sago Palm Swamp Forest . . . . . . . . . . . . . . . 425
- red spiny . . . . . . . . . . . . . . . . . . . 177, 178 Sagulubek Formation, Indonesia. . . . . . . . . .. 427
- white-tailed . . . . . . . . . . . . . . . . . 224-226 Salticidae/salticid spiders . . . . . . . . . . . . . 91, 92
- Whitehead's . . . . . . . . . . . . . . . . . . . . 177 Samarinda, Indonesia . . . . . . . . . . . . . . . . . 454
Rat kangaroo, musky . . . . . . . . . . . . . . 224, 225 Sambar deer . . . . . . . . . . . . . . . . . . . . . .. 343
Rattans . . . . . . . . . . . . . 428, 430, 439, 440, 523 San Carlos de Rio Negro, Brazil . . . . . . . . . . 489
- inventories . . . . . . . . . . . . . . . . .. 523-533 Santa Rosa National Park, Costa Rica . . . . . . . . 39
Recreation argument for conservation . . . . . . . 471 Sap beetles, as pollinators . . . . . . . . . . . . . . . 478
Recruitment (of seedlings) . . . . . . . . . . . . . . 354 Sapindaceae . . . . . . . . . . 142,293,327,328,483
Red duiker .. . . . . . . . . . . . . . . . . . . . . .. 539 Sapotaceae .... 142, 143,225,229,327,328,483
Red-yellow podzols . . . . . . . . . . . . . . . . . . . 341 Saprophytes, ground herbs . . . . . . . . . . . . . . . 47
Red/far-red ratio . . . . . . . . . . . . . . . . . 216, 226 Sarawak . . . . . . . . 10, 13,37, 109, 132, 143, 163,
Reduviidae ., . . . . . . . . . . . . . . . . . . . . . . . 93 268,297,298,317,341,387,392,411,412,
Reforestation . . . . . . . . . . . . . . . . . . . . 452, 455 420,421,422,489,492
564 GENERAL INDEX
Sauria . . . . . . . . . . . . . . . . . . . . . . . . . . . 318 - conflicts . . . . . . . . . . . . . . . . . . . . . . 513
Sayap-Kinabalu Park, Sarawak . . . . . . . . . . . 121 - development . . . . . . . . . . . . . . . . . . .. 492
Scandentia . . . . . . . . . . . . . . . . . . . . . . . . 177 Soils - bulk density . . . . . . . . . . . . . . . . 250, 259
Sciapodinae . . . . . : . . . . . . . . . .. 183, 185, 187 - chemical characteristics . . . . . . . . . . . . . 259
Sciaridae . . . . . . . . . . . . . . . . . . . . . . . 90, 93 - cohesion values . . . . . . . . . . . . . . . . . . 399
Scincidae . . . . . . . . . . . . . . . . . . . . . . . . . 318 - degradation in Thailand . . . . . . . . . . . .. 521
Scirtidae . . . . . . . . . . . . . . . . . . . . . . . . . . . 93 - hydrology . . . . . . . . . . . . . . 259, 262, 401
Sciuridae/Sciurids . . . . . . . . . . . .. 177, 178, 181 - infiltration capacities . . . . . . . . . . . . 269, 403
Scolytid beetles . . . . . . . . . . . . . . . . . . . . . 187 - moisture conditions . . . . . . . . . . . . 259, 403
Scorpiones . . . . . . . . . . . . . . . . . . . . . . . . . 5 - nitrate concentrations . . . . . . . . . . . . . . 256
Scrophulariaceae . . . . . . . . . . . . . . . . . . . . 325 - nitrification rates .. . . . . . . . . . . . . . . . 241
Seasonal patterns - of climate . . . . . . . . . 387, 407 - nutrients . . . . . . . . . . . . . . . 259,261, 268
- of sediment loss . . . . . . . . . . . . . . . . . . 265 - organic matter . . . . . . . . . . . . . . . . 259, 269
- of species occurrence . . . . . . . . . . . . . . 197 - phosphorus . . . . . . . . . . . . . . . . . . . . . 259
Seasonal soil suction . . . . . . . . . . . . . . . . . . 402 - potassium . . . . . . . . . . . . . . . . . . . . . . 259
Second-order jackknife method . . . . . . . . 206, 207 - profiles at Belalong . . . . . 260,401, 403-407
Sediment - loading to streams . . . . . . . . . . . . 269 - seed bank . . . . . . . . . . . . . . 215, 216, 226
- loss . . . . . . . . . . . . . . . . . . . . . . 259-268 - shear strength . . . . . . . . . . . . . . . . 404, 406
- monitoring . . . . . . . . . . . . . . . . . . . . . 271 - surface hydrology . . . . . . . . . . . . . . . . 260
- yields . . . . . . . . . . . . . . . . . 259, 265, 268 - tension . . . . . . . . . . 395, 397, 399, 401, 407
Seeds - dry mass values . . . . . . . . . . . . . . . . 222 Solanaceae . . . . . . . . . . . . . . . . . . . . . . . . 483
- mass . . . . . . . . . . . . . . . . . . . . . . . . . 215 Solar radiation at Belalong . 241, 243, 251, 253, 256
- predation . . . . . . . . . . . . . . . . ... 283, 284 Solomon Islands . . . . . . . . . . . . . . . . . . . .. 151
- production in figs . . . . . . . . . . . . . . 284, 287 Somalia . . . . . . . . . . . . . . . . . . . . . . . . .. 541
- set in figs . . . . . . . . . . . . . . . . . . . . . . 283 Sonagrams . . . . . . . . . . . . . . . . . . . . . . . 77-84
- size . . . . . . . . . . . . . . . . . . . . . . 215, 220 Sonneratiaceae . . . . . . . . . . . . . . . . . . . . . . 328
- soil bank . . . . . . . . . . . . . . . . . . . 215, 222 Soricidae . . . . . . . . . . . . . . . . . . . . . . . . . 177
Seedlings - establishment . . . . . . . . 217, 223, 355 South Africa . . . . . . . . . . . . . . . . . . . . . . . 294
- physiology . . . . . . . . . . . . . . . . .. 367-375 Space guilds in amphibians & reptiles . . . . . . . 320
- survival . . . . . . . . . . . . . . . . . . . . . . . 280 Space utilisation by fish . . . . . . . . . . . . . . . . 307
Segama River, Sabah . . . . . . . . . . . . . . . . . 358 Spartaeinae . . . . . . . . . . . . . . . . . . .. 88, 90-92
Selaginellaceae . . . . . . . . . . . . . . . . . . . . . . 47 Spatial information systems (GIS) . . . . . . . 490, 496
Selangor, Malaysia . . . . . . . . . . . . . . . . . . . 348 Speciation. . . . . . . . . . . . . .. 151, 189, 291, 295
Selective logging . . . . . . . . . . . . . . . . . 452, 455 - in stalk-eyed flies . . . . . . . . . . . . . . . .. 170
Semabat Agricultural Station, Brunei . . . . . 242, 389 Species - abundance curves . . . . . . . . . . . . . . . 49
Semongok, Sarawak . . . . . . . . . . . . . . . . . . 269 - area curves/relationships . . . . . . 195, 197-199
Seram, Indonesia . . . . . . . . . . . 1, 2, 6, 9, 14, 22 - associations in ants . . . . . . . . . . . . . 334, 335
Serow . . . . . . . . . . . . . . . . . . . . . . . . . . . 156 - conservation . . . . . . . . . . . . . . . . . . . . 419
Serpentes . . . . . . . . . . . . . . . . . . . . . . . . . 318 - diversity . . . . . . . . . 1-20, 67, 412, 482, 484
Setap shales . . . . . . . . . . . . . . . 44, 50, 139, 243 - of ants . . . . . . . . . . 9-12,117-123,357
Settlement patterns in Sri Lanka . . . . . . . . . .. 507 - of birds 101, 102, 105-107, 116, 151-152
Sexual dimorphism in stalk-eyed flies . . . . . . . 170 - of butterflies . . . . . . 125-138, 291, 295
Shade - influence on growth . . . . . . . . . . . . . 277 - of empidoid flies. . . . . . . . . .. 183-192
- tolerance . . . . . . . . . . . . 215,216,221,227 - of fish . . . . . . . . . . 307, 312, 377-386
Shade-loving species . . . . . . . . . . . . . . . . . . . 50 - of herbs . . . . . . . . . . . . . . . . . . 43-57
Shade-tolerant plants .. 216-218, 220-222, 226, 229 - of herpetofauna . . . . . . . . . . . 315-323
Shallow failure of slopes .. 398, 403, 404, 406, 407 - of macrolepidoptera in Britain ... 30, 35
Shannon's diversity index . . . . . . . . . . . . 395, 360 - of microlepidoptera . . . . . . . . . . 29-42
Shannon-Wiener diversity index. . .. 308, 482, 484 - of small mammals . . . . . . . 175-182,345
Shapiro-Wilk test for normality . . . . . . . . . . . 247 - of termites . . . . . . . . . . . . . . . . 12-14
Shear strength of soil . . . . . . . . . . . . . . . . . . 404 - of trees . . . . . . . . . . . . . . . . . 139-150
Shifting cultivation . . . . . . . . . . .. 341,453,454 - extinctions . . . . . . . . . . . . . . . . . . . .. 539
Shifting-mosaic dynamics, forest structure .... 489 - radiation . . . . . . . . . . . . . . . . . . . 183, 294
Shimba Hills National Reserve, Kenya . . . . . . 542 - in Sympycnus . . . . . . . . . . . . . . .. 188
Shrew - Savi's pygmy . . . . . . . . . . . . . . . . . 177 - richness. . . . .. 9,27,29,61, 117, 125, 127,
- South-east Asian white-toothed . . . . . . . . 177 132, 143, 151, 176, 193, 196, 198, 206, 207,
Siberut Island, Indonesia . . . . . . . . . . . . 425-433 215,230,295,357, 359, 377, 381, 382, 384,
Sienna Madre, Mexico. . . . . . . . . . . . . . . . . 463 385,477,482,485,489
Siluridae . . . . . . . . . . . . . . . . . . . . . . . . . 380 - similarity indices . . . . . . . . . . . . . . 412,413
Silvicultural practices . . . . . . . . . . . . . . 411, 414 - turnover . . . . . . . . . . . . 125, 127, 193, 195
Silvicultural requirements of dipterocarps .. 277-281 Spiders . . . . . . . . . . . . . . . . . . 90, 91, 95, 208
Simaroubaceae . . . . . . . . . . . . . . . . . . . . . . 328 Splash erosion . . . . . . . . . . . . . . . . . . . . . . 397
Similarity indices . . . . . . . . . . . . . . 32,412,413 Spodosolic soils . . . . . . . . . . . . . . . . . . . .. 484
Simpson's index of diversity . . . . .. 106, 482, 484 Spongiphoridae . . . . . . . . . . . . . . . . . . . . . . 92
Singapore . . . . . . . . . . . . . . . . 97,221,227,229 Springtails . . . . . . . . . . . . . . . . . . . . 88, 89, 90
Singhalese culture . . . . . . . . . . . . . . . . . . . . 506 Squirrels . . . . . . . . . . . . 154,225,341,343,427
Sinharaja Forest Reserve, Sri Lanka . . . . . 506, 507 - flying .. . . . . . . . . . . . . . . . . . . . . .. 155
Sipuncula . . . . . . . . . . . . . . . . . . . . . . . . .. 5 - giant . . . . . . . . . . . . . . . . . . . . . . 177,344
Sisorodae . . . . . . . . . . . . . . . . . . . . . . . . . 380 - horse-tailed . . . . . . . . . . . . . . . . . . . . 177
Slope characteristics . . . . . . . . . . . . . . . 387-409 - Low's . . . . . . . . . . . . . . . . . . . . . 156, 177
Snakes . . . . . . . . . . . . . . . . . . . . . . . . 155,317 - plain pygmy . . . . . . . . . . . . . . . . . . .. 177
- Oriental whip ... . . . . . . . . . . . . . . .. 155 - plantain. . . . . . . . . . . . . . .. 177,335,344
- pit viper . . . . . . . . . . . . . . . . . . . . . . 320 - Prevost's . . . . . . . . . . . . . . . . . . . . .. 177
- striped racer . . . . . . . . . . . . . . . . . . .. 155 - tufted ground . . . . . . . . . . . . . . . . . . . 177
Social bees . . . . . . . . . . . . . . . . . . . . . . . 59-66 Sri Lanka . . . . . . . . . . . . . . . . . . . . .. 503-512
Social forestry programmes . . . . . . . . . . . . . . 452 Stability analyses . . . . . . . . . . . . . 399, 403, 406
Socio-economics - benefits of Indonesian forest . 456 Stage hydrograph . . . . . . . . . . . . . . . . . 394, 395
GENERAL INDEX 565
Stalk-eyed flies . . . . . . . . . . . . . . . . . . 169-174 Thysanura . . . . . . . . . . . . . . . . . . . . . . . . . 5

Staphyleaceae . . . . . . . . . . . . . . . . . . . . . . 326 Tiliaceae . . . . . . . . . . . . . . . . . . . . . . 142, 328
Staphylinidae/Staphylinid beetles .8, 12,91,93, 195 Timah-Tasoh Dam, Malaysia . . . . . . . . . . . . . 156
Stemflow . . . . . . . . . . . . . . . . . . . . . . . . . 394 Timber - concessions on Siberut . . . . . . . . . . . 431
Stenophagy . . . . . . . . . . . . . . . . . . . . . . . . 322 - extraction . . . . . . . . . . . . . . 341, 432, 469
Sterculiaceae . . . . . . . . . . . . . . . . . . . . 142, 328 - harvesting in peat swamp forest . . . .. 411-415
Stoats . . . . . . . . . . . . . . . . . . . . . . . . . . . 152 - production in Indonesia . . . . . . . . . . . . . 454
Stomach contents of fish . . . . . . . . . . . . . 310, 312 - values . . . . . . . . . . . . . . . . . . . . . . . . 467
Stomatal conductance . . . . . . . 367, 369, 371-374 Time guilds in amphibians and reptiles . . . . . . 320
Stomatal density . . . . . . . . . . . . . . . 367-370, 373 Time utilisation in fish . . . . . . . . . . . . . . . . . 310
Stratification - of birds . . . . . . . . . . . . . . . .. 102 Tineidae . . . . . . . . . . . . . . . . . . . . . . . . . .. 9
- of butterlies . . . . . . . . . . . . . . . . . . . . 133 Tipulidae/Tipulid larvae . . . . . . . . . . . . ., 93, 94
Stratiomyidae/Stratiomyid beetles. . . . . . .. 90, 93 Tokelau Islands . . . . . . . . . . . . . . . . . . . . . 152
Streaked-breasted woodpecker . . . . . . . . . . . 156 Torrent loaches . . . . . . . . . . . . . . . . . . . . . 383
Streams - gradients . . . . . . . . . . . . . . . . 377, 382 Tortoises . . . . . . . . . . . . . . . . . . 155, 156,319
- orders . . . . . . . . . . . . 377-379,381,383-385 Tortri cidae .. . . . . . . . . . . . . . . . . . . . . . .. 9
- water quality . . . . . . . . . . . . . . . . . . . 382 Tourism . . . . . . . . . . . . . . . . . . . 419,425,433
Stumpage value . . . . . . . . . . . . . . . . . . . . . 454 Traditional medicines . . . . . . . . . . . . . . . . . 421
Subsistence practices . . . . . . . . . . . . . . . 460, 461 Traditional rights . . . . . . . . . . . . . . . . . . . . 422
Succession, in peat swamp forest . . . . . . . . . . 412 Transfrontier reserve . . . . . . . . . . . . . . . 153, 156
Sukadana, West Kalimantan . . . . . . . . . . 412, 413 Transmara Forest Reserve, Kenya . . . . . . . . . 542
Sulawesi . . . . . . 1,2,8,9,14,22,38,40,91,421 Transmigration . . . . . . . . . . . . . . 414,425,433
Sulphuric acid and resin production ..... 235-240 Tree ferns . . . . . . . . . . . . . . . . . . . . . . . . 217
Sumatra . . . . . . 64,71-75,102,411,425,433,455 Treeshrews . . . . . . . . . . . . . . . .. 154, 177,344
Sun bear . . . . . . . . . . . . . . . . . . . . . . 342, 343 - common . . . . . . . . . . . . . . . . . . . . . . 343
Sundaland/Sunda Shelf. . . . . . . . . . . . .. 75,425 - large . . . . . . . . . . . . . . . . . . . . . . . . . 177
Sungai Buloh Forest Reserve, Malaysia . . . . . . 277 - lesser . . . . . . . . . . . . . . . . . . . . . 177, 343
Sungai Gombak, Malaysia . . . . . . . . . . . . . . 269 - mountain . . . . . . . . . . . . . . . . . . . . .. 177
Sungai Liang, Brunei Darussalam . . . . . . . . .. 163 - painted . . . . . . . . . . . . . . . . . . . . . . . 177
Sungai Tenom, Malaysia . . . . . . . . . . . . . . . 269 - slender . . . . . . . . . . . . . . . . . . . . 177, 343
Sustainable forest management ..... 411,412, 415, - striped . . . . . . . . . . . . . . . . . . . . 177, 343
451,454, 456, 463, 477, 492 Trengganu, Malaysia . . . . . . . . . . . . . . . . . . 347
Swamp forests . . . . . . . . . 183, 186, 190,221,509 Trichoptera . . . . . . . . . . . . . . . . . . . . . . . . 310
Sweet potato . . . . . . . . . . . . . . . . . . . . 271, 275 Trionychidae/Trionychids . . . . . . . . . . . . 155, 318
Swidden rice . . . . . . . . . . . . . . . . . . . . . .. 516 Trophic interactions in bamboo internodes ... 85-99
Syconium . . . . . . . . . . . . . . . . . . . . . . . . . 283 Trophic resource use . . . . . . . . . . . . . . . . . . 321
Symbiosis, between lycaenids and ants .... 292, 295 Tropical Areas Bioindicators System . . . . . . . . . 15
Sympatry . . . . . . . . . . . . . . . . . . . . 71,75,307 Tropical Rainforest Cooperative Research Centre . 27
- in murids . . . . . . . . . . . . . . . . . . . . .. 178 Tropical thorn forests . . . . . . . . . . . . . . . . . 505
Symplocaceae . . . . . . . . . . . . . . . . . . . 326, 327 Tropudults . . . . . . . . . . . . . . . . . . . . . . . . 428
Sympycninae . . . . . . . . . . . . . . . . . . . . 184, 185 Tukano people, northwest Amazon . . . . . . . . . 460
Syrphidae/Syrphids . . . . . . . . . . . . . . . .. 93, 94 Tupaiidae/Tupaiids . . . . . . . . . 175, 177, 178, 181
Syzygium-Gordonia-Miche/ia-ElLIeocarpus series. 505 Turtles . . . . . . . . . . . . . . . . . . . . . . . . . . . 319
Taboos . . . . . . . . . . . . . . . . . . . . . . . . . . 422 - Asian leaf . . . . . . . . . . . . . . . . . . . . . 155
Tachinid flies . . . . . . . . . . . . . . . . . . . . . . 293 - giant pond . . . . . . . . . . . . . . . . . . . . . 156
Tachydromiinae . . . . . . . . . . . . . . . . . . 185, 186 - Malayan flat-shelled . . . . . . . . . . . . 155, 156
Taita Hills, Kenya . . . . . . . . . . . . . . . . . . . 542 - Malayan soft-shelled . . . . . . . . . . . . . . . 155
Tanypodinae . . . . . . . . . . . . . . . . . . . . . . . . 89 - spiny . . . . . . . . . . . . . . . . . . . . . . . . 156
Tanzania . . . . . . . . . . . . . . . . . . . . . . . . . 107 Tutong, Brunei Darussalam . . . . . . . . . . . . . . 435
Tapping, for resin . . . . . . . . . . . . . . . . 235-240 Ukong, Brunei Darussalam . . . . . . . . . . . 390, 391
Tardigrada . . . . . . . . . . . . . . . . . . . . . . . . . 5 Ulmaceae . . . . . . . . . . . . . . . . . . . . . . 142,328
Taxaceae . . . . . . . . . . . . . . . . . . 325, 329, 330 Ulu Gombak Field Centre, Malaysia ... 86, 88, 292
Teak . . . . . . . . . . . . . . . . . . . . . . . . . 467,516 U1u Segama, Sabah . . . . . . . . . . . . . . . . 269,345
Tectonic uplift . . . . . . . . . . . . . .. 387, 392, 407 Ulu Temburong National Park, Brunei . . . . . . 420
Tekam, Peninsular Malaysia . . . . . . . . . . . . . 345 UmbelliferaelUmbellifers .. 193, 199,200,201,204
Temburong Formation . . . . . . . . . . . . . . . . . 391 - local richness estimates . . . . . . . . . . . . . 201
Temperate deciduous forest .... 217,220,221,227 Understorey seedlings . . . . . . . . . . . . . . . . . 374
Temperature, difference with altitude . . . . . . . 241 Understorey species . . . . . . . . . . . . . . . . . . 219
Tenebrionidae/Tenebrionoidea . . . . . . .. 8, 90, 93 Ungulates. . . . . . . . . . . . . .. 153-155,342,344
Tensiometers . . . . . . . . . . . . . . .. 395, 396, 402 - trapping . . . . . . . . . . . . . . . . . . . . . . . 536
Tephritidae/Tephritid flies . . . . . . . . . 90, 92, 93, United Nations Conference on Environment and
169, 170, 171, 173 Development . . . . . . . . . . . . . . . . . 15,417
Termites .... 1, 6, 11, 12, 94, 250, 297-305, 357 United States. . . . . . . . . . . . . . . .. 64, 199, 272
Termitidae/Termitinae . . . . . .. 297, 298, 300-302 Universal Soil Loss Equation . . . . . . . . . . 271,272
Territoriality, of stalk-eyed flies . . . . . . . . . .. 169 Urticaceae . . . . . . . . . . . . . . . . . . . . . 327, 328
Testudines/Testudinidae . . . . . . . . . . . . . 318,319 Usambara mountains, Tanzania . . . . . . . . . . . 542
Tettigoniidae . . . . . . . . . . . . . . . . . . . 79, 80, 83 Valenya, Brazil . . . . . . . . . . . . . . . . . . . . . 479
Thai-Norwegian Highland Development Project. 516 Valley profiles . . . . . . . . . . . . . . . . . . . . . . 405
Thailand . . . . . . . . . 153,156,321,419,421,513 Values of forests . . . . . . . . . . . . . . . . . . . . 467
Thaleban National Park, Thailand . . . . . . 153, 156 Vanuatu . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
Theaceae . . . . . . . . . . . . . . . . . . . . . . 325-327 Varanidae . . . . . . . . . . . . . . . . . . . . . . . . . 318
Theridiidae . . . . . . . . . . . . . . . . . . . . . . . 90-92 Vascular plant diversity . . . . . . . . . . . . . . . . 538
Thinning treatments in silviculture . . . . . . 347-355 Vegetation heterogeneity . . . . . . . . . . . . . . . 489
Throughfall . . . . . . . . 259,262-264,267,363,394 Veliidae . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
Throughflow . . . . . . . . . . . . . . . . . . . . . . . 397 Venezuela . . . . . . . . . . . . . . . . . . . . . . . . 277
Thymelaeaceae . . . . . . . . . . . . . . 141, 142, 328 Verbenaceae . . . . . . . . . . . . . . . . . . . . 142,226
Thysanoptera . . . . . . . . . . . . . . . . . . . . . . 9, 24 Vernacular synonyms . . . . . . . . . . . . . . . . . 435
566 GENERAL INDEX
VertebrataNertebrates . . . . . . . . . . . . . 5,69,315
- sounds/vocalisation. . . . . . . 71-73, 77, 80, 83
Vertical stratification of birds . . . . . . . . . . . . 103
Viperidae . . . . . . . . . . . . . . . . . . . . . . . . . 318
Viruses . . . . . . . . . . . . . . . . . . . . . . . . 1,470
Vision of stalk-eyed flies . . . . . . . . . . . . . . . 169
Visual recognition in empidoids . . . . . . . . . .. 191
Viverrids . . . . . . . . . . . . . . . . . . . . . . . . . 177
Vocalisation - in birds . . . . . . . . . . . . . . . .. 108
- in PreslJytis . . . . . . . . . . . . . . . . . . . 71-73
Vo1can Barva, Costa Rica . . . . . . . . . . . . . . . 255
Wallace's Line. . . . . . . . . . . . . . . . . . . . .. 151
Wallacea . . . . . . . . . . . . . . . . . . . . . . . . . 425
Warm temperate rainforest . . . . . . . . . . . 217, 220
Wasps - life cycles . . . . . . . . . . . . . . . . . . . 283
- as parasites . . . . . . . . . . . . . . . . . . . . . 459
- as pollinators . . . . . . . . . . . . . . . . . . . 283
- predation in figs . . . . . . . . . . . . . . 283, 287
- reproductive success . . . . . . . . . . . . . . . 283
Water beetles . . . . . . . . . . . . . . . . . . . . . . . 95
Water measurers . . . . . . . . . . . . . . . . . . . . . 95
Water mites . . . . . . . . . . . . . . . . . . . . . . . . 87
Water quality . . . . . . . . . . . . . . . . . . . . . . 384
Water stress in trees . . . . . . . . . . . . . . . 372, 374
Water striders . . . . . . . . . . . . . . . . . . . . 88, 97
Watershed protection . . . . . . . . . . . . . . . . . . 419
Wau, Papua New Guinea . . . . . . . . . . . . . . . 284
Weasel, Malay. . . . . . . . . . . . . . . . . . . . . . 343
Weather stations at Bela10ng . . . . . . . . . . 241, 242
Weathering of soils . . . . . . . . . . . . . . . . . . . 393
Weathering rates . . . . . . . . . . . . . . . . . 404, 407
Weaver ants . . . . . . . . . . . . . . . . . . . . . . . 293
Web-building spiders . . . . . . . . . . . . . . . . . . . 95
Weevils, as pollinators. . . . . . . . . . . . . . . . . 478
West African savannas. . . . . . . . . . . . . . . . . 304
West Indies. . . . . . . . . . . . . . . . . . . . . . . . . 13
West Kalimantan . . . . . . . . . . . . 411,412,414
West Malaysia . . . . . . . 85, 86, 221, 229, 292, 393
Western Ghats, India . . . . . . . . . . . . . . . 303, 320
White-eyes ..... . . . . . . . . . . . . . . . . . .. 152
Wildlife - management. . . . . . . . . . . . . . . . . 422
- survival after logging . . . . . . . . . . . . . . 345
Winkler's sampling method. . . . . .. 359, 362, 363
Winteraceae . . . . . . . . . . . . . . . . 325,329,330
Wood-boring beetles . . . . . . . . . . . . . . . . . . . 90
Wood-lice . . . . . . . . . . . . . . . . . . . . . . . . . . 90
Woodchips, value of . . . . . . . . . . . . . . . . . . 467
Woodpecker . . . . . . . . . . . . . . . . . . . . . . . . 94
World Commission on Environment &
Development . . . . . . . . . . . . . . . . . 421, 469
World Conservation Monitoring Centre ..... . 537
World Geodetic System . . . . . . . . . . . . . . . . 496
World Heritage Sites. . . . . . . . . . . . . .. 21, 507
Xiaoliang barren coastal ecosystem . . . . . . . . . 489
Yellow-backed Duiker . . . . . . . . . . . . . . . . . 543
Zaire. . . . . . . . . . . . . . . . . . . . . . . .. 13,411
Zingiberaceae ..... 43, 47, 48, 50, 155,219,429
Zones of establishment of plant families . . . . . . 330
Zoning, in National Parks . . . . . . . . . . . . . . 419

Tropical Rainforest Research Current Iss

Uploaded by

Copyright:

Available Formats

Tropical Rainforest Research Current Iss

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Tropical Rainforest Research Current Iss

Uploaded by

Copyright:

Available Formats

TROPICAL RAINFOREST RESEARCH - CURRENT ISSUES

KLUWER ACADEMIC PUBLISHERS

Tropical raInforest research current issues / edited by D.S. Edwards

1. RaIn forests. 2. Rain forest ecology. 3. Rain forest

ISBN-13: 978-94-010-7255-7 e-ISBN-13: 978-94-009-1685-2

Published by Kluwer Academic Publishers,

Kluwer Academic Publishers incorporates

Sold and distributed in the U.S,A. and Canada

In all other countries, sold and distributed

Printed on acid-free paper

All Rights Reserved

No part of the material protected by this copyright notice may be reproduced or

Bandar Seri Begawan,

Biodiversity Division, Department of Entomology, The Natural History Museum, Cromwell

D. S. Edwards et at. (eds.), Tropical Rainforest Research - Current Issues, 1-20.

CONCEPTS OF ECOSYSTEM DYNAMICS

DISTURBANCE REGIME BIOTIC RESPONSE ECOSYSTEM STEPS

J02 - Hurric:mes Succession of communities

10' :-- Drought Cycles Popublioo

LOCAL FORCES REGIONAL FORCES

'ecological forces' = 'biogeographical forces'

THE FAUNAL COMPONENTS OF TROPICAL FORESTS

Abundance Biomass Sp. rich. Habitat Key

B = N(0.0305U62) (Rogers et al. 1976)

sampling programme, such as that described below, is required to determine arthropod

Par~itods Fresh Water

Plots A-C N. Sulawesi Sulawesi Author

TROPICAL FOREST DYNAMICS: TWO EXAMPLES OF KEY GROUPS

The contribution of ants to tropical forest dynamics

The contribution of termites to tropical forest dynamics

Localit~ Numbers m2 Biomass E m·2 Author

THE WAY FORWARD

R. L. KITCHING and 1. ZALUCKI

Canopy Arthropod Studies

Canopy Studies in Australia

-< -< -< -<

The Importance of Oligophagy

Current data shortcomings and Solow's analysis

The epiphyte problem

GADEN S. ROBINSON and KEVIN R. TUCK

Biodiversity Division, Department of Entomology, The Natural History Museum, Cromwell

KEY WORDS; Microlepidoptera, biodiversity, tropical rainforest, Brunei Darussalam.

Kuala Belalong site 2,

50 Rank 100 150

provide meaningful biological observations.

SPECIES WITH: Observed Expected (log series)

a(N/(N +a», a«N/(N +a»2J2), a«N/(N +a»3/3), etc.

1000 2000 3000

The localities studied

Numbers and diversity

SITE MICRO PYRAL ALL

SITE MICRO PYRAL ALL

SITE MICRO PYRAL ALL

Comparison of the sites

Commonness, rarity and taxonomic composition of the samples

The scale of diversity ?

How many species in the lowland forest?

stress that they are likely to be underestimates.

Similar species composition?