ECOTROPICA 17: 15–20, 2011

© Society for Tropical Ecology

IS THE RESILIENCE OF EPIPHYTE ASSEMBLAGES TO HUMAN

DISTURBANCE A FUNCTION OF LOCAL CLIMATE?
Florian A. Werner1,2*, Nils Köster3,4, Michael Kessler1,5 & S. Robbert Gradstein1,6
1
Albrecht-von-Haller Institut für Pflanzenwissenschaften, Universität Göttingen,
Untere Karspüle 2, 37073 Göttingen, Germany
2
current address: Funktionelle Ökologie, Institut für Biologie und Umweltwissenschaften,
Universität Oldenburg, Carl-von-Ossietzkystr. 9-11, 26111 Oldenburg, Germany
3
Nees Institut für Biodiversität der Pflanzen, Universität Bonn,
Meckenheimer Allee 170, 53115 Bonn, Germany
4
current address: Botanischer Garten und Botanisches Museum Berlin-Dahlem, Freie Universität Berlin,
Königin-Luise-Straße 6-8, 14195 Berlin, Germany
5
current address: Institut für Systematische Botanik, Universität Zürich,
Zollikerstr. 107, 8008 Zürich, Switzerland
6
current address: Muséum National d’Histoire Naturelle, Départment Systématique et Evolution,
57 rue Cuvier, 75005 Paris, France

Abstract. Field studies have shown a wide array of responses of vascular epiphyte diversity to human disturbance-assem-
blages of disturbed habitats range from largely unchanged to severely impoverished when compared with intact forest. This
variability is not well understood. We explored the hypothesis that the relative impoverishment of disturbed-habitat epiphyte
assemblages is a function of local climate, by analyzing the available literature on epiphyte diversity on isolated trees as a
model system. We found that assemblages of moist and moderately seasonal areas experience considerably stronger impov-
erishment than those of aseasonally wet or distinctly dry areas. We argue that the integrity of the vertical microclimatic
gradient is more crucial for the maintenance of epiphyte diversity in moderately seasonal forests than in distinctly dry or
aseasonally wet forests.
Key words: diversity, fragmentation, isolated trees, land-use change, mesoclimate, species density, species richness, tropical forest,
vascular epiphytes

INTRODUCTION isms is therefore critical for conservation (Gascon

et al. 1999, Laurance 2006).
There is much discussion in conservation biology
Case studies in secondary and degraded forests
about the extent to which anthropogenic habitats can have yielded a wide range of responses of epiphyte
allow the persistence and migration of organisms diversity to habitat disturbance. While some studies
living naturally in primary forest (Laurance 2006, have found vascular epiphyte assemblages signifi-
Gardner et al. 2007). Tropical forests continue to be cantly impoverished (Barthlott et al. 2001, Krömer
converted, degraded, and fragmented at a high rate, & Gradstein 2003, Wolf 2005), others found them
and the challenge of global climate change will re- as diverse as those of undisturbed forests (Hietz-
quire alterations in species migrations of unprece- Seifert et al. 1996, Larrea & Werner 2010). The re-
dented magnitude and rapidity (Bush 2002, Malhi sponse to habitat alteration even varies greatly
& Phillips 2004). Understanding how alteration and within single taxa. Bromeliads, for instance, were less
fragmentation of their primary habitat affects organ- species-rich in some secondary forests (Krömer &
Gradstein 2003) but considerably more species-rich
in others (Barthlott et al. 2001). This divergence of
patterns, however, is unsurprising considering that
* e-mail: [email protected] non-primary forests differ greatly in age and struc-

15
WERNER ET AL.

ture, and in their degree of structural differences from pies are usually low and open and cannot provide
their various potential vegetations. Parameters of pronounced microclimatic buffering (Graham &
potential influence on epiphytes include type and Andrade 2004), it may be assumed that their epi-
degree of disturbance, host-tree age, vegetation struc- phytes must be sufficiently well-adapted to persistent
ture (e.g. stand height, canopy openness) and result- drought, even in situations of increased exposure
ing microclimate, substrate quality and quantity (e.g. such as experienced on ITs. To explore the hypoth-
availability of large horizontal limbs, humus and esis that the relative impoverishment of vascular epi-
bryophyte mats), dispersal constraints, and local phyte assemblages is a function of local climate, we
climate (mesoclimate). Many of these factors are compared studies on epiphyte diversity on ITs and
difficult to quantify, and information on very few of mature forest trees at Neotropical sites.
them has been regularly provided, much less ana-
lyzed, in papers addressing the epiphyte assemblages METHODS
of non-primary habitats. This circumstance has We screened the available literature on epiphyte di-
hampered comparisons, rendering land-use change versity on ITs and, where needed, gathered addi-
effects poorly understood and unpredictable. tional information on site characteristics by contact-
Isolated trees (hereafter termed “ITs”) scattered ing the respective authors. We restricted our survey
in anthropogenic land-use matrices constitute key- to studies on vascular epiphyte assemblages that give
stone structures that offer refuge, enhance connectiv- species numbers for entire trees. To ensure consis-
ity, and provide nuclei of regeneration for many tency, several studies had to be excluded from analy-
groups of forest organisms (Manning et  al. 2006, sis due to major differences in sampling area (Wil-
Zahawi & Augspurger 2006). Moreover, they offer liams-Linera et al. 1995, Nkongmeneck et al. 2002),
an excellent model system for studying the effects limited taxon sampling (Dunn 2000, Cascante-
human disturbance on epiphytes. ITs can be viewed Marín et al. 2006), or the lack of a local baseline (e.g.
as the “ultimate”, smallest possible forest fragment Gonçalvez & Waechter 2003, Poltz & Zotz 2011).
(Williams-Linera et al. 1995, Gove et al. 2009), be- Ultimately, seven studies allowed for meaningful
ing exposed to maximum edge effects and poten- comparisons and were included in our analysis
tially constrained colonization. Unlike degraded or (Hietz-Seifert et al. 1996, Flores-Palacios & García-
secondary forests, ITs do not differ between sites in Franco 2008, Nöske et al. 2008, Köster et al. 2009,
the relative degree of their physical exposure, neither Werner & Gradstein 2009, Larrea & Werner 2010,
do they differ from undisturbed forest in important Köster et al. in press) (Table 1). To characterize the
host parameters like age, surface area, or bark char- mesoclimate for each of the sites, we calculated an
acteristics, unless biased in size or taxonomic com- index of aridity, dividing monthly precipitation by
position. monthly potential evapotranspiration (PET) (UNEP
Epiphyte community changes following land-use 1992). PET was calculated following the Thornth-
change have often been attributed primarily to waite method (Thornthwaite & Mather 1957) as:
changes in microclimate (e.g. Barthlott et al. 2001, PETTh = 3.65 × 10 -4d(t)I -1[10T(t)]a
Hietz 2005, Nöske et al. 2008). Epiphytes are known where d(t) is the total monthly daylight in hours,
for their high and fine-tuned humidity requirements, and T(t) is the monthly mean temperature. The two
being linked more closely to the atmosphere than variables I and a are empirically derived functions of
terrestrial biota (e.g. Kreft et al. 2004, Poltz & Zotz mean annual temperature. Our analysis focused on
2011). The relative degree to which epiphyte assem- species density, here defined as the number of species
blages become species-poorer following disturbance per host tree of comparable size. The “relative
may be linked to local climate. In aseasonally wet impoverishment” of ITs is given as the percentage of
(perhumid) regions, epiphyte diversity on ITs may reduction in mean epiphyte species density on ITs
not be as strongly affected by desiccation stress as in relative to forest trees. For each of the studies we
areas of intermediate humidity, where a large share also calculated the qualitative Sørensen index as a
of the epiphytic flora depends on the maintenance measure of floristic similarity between forest and ITs
of a mesic microclimate by intact forest canopies from published species lists or unpublished own
during dry periods. In various arid regions of the data using Estimate 8.2 (Colwell 2006). Polynomial
Neotropics we observed that ITs often harbor a rela- regression was done with Xact 8 (Scilab, Hamburg,
tively intact epiphytic flora. Because dry forest cano- Germany).

16
 TABLE 1. Site characteristics and vascular epiphyte diversity on isolated trees (ITs).

Mean Precipit. Canopy Isolation No. of Effect of Total Relative Qual.

Elevation Aridity
Study Site Region Fog temp. [mm height of host host trees distance species impoverishment Sørensen
[m] index
[°C] yr-1] [m] trees [yr] studieda to forestb richnessa [%]b index
Flores-Palacios San Andrés
& García- Tlalnel- E Mexico 1550 common 14.0 1650 2.07 30-45 85 88c n.s. 102d 17c – 0.56
Franco 2008 huayocan

Hietz-Seifert
Los Tuxtlas SE Mexico 100 rare 24.5 4700 3.41 30-35 28 165 ** 83 20d n.s. 0.74
et al. 1996

Köster et al.
Bilsa NW Ecuador 550 common 23.5 3100e 2.80 30-35 5-15 52 n.s. 318 36 ** 0.80
in press

Köster et al.
Otonga NW Ecuador 2000 common 15.6 2550 3.07 25-30 5-25 52 n.s. 322 36 ** 0.77
2009

Larrea & un-

SierrAzul NE Ecuador 2200 15.2 3500e 4.40 25-35 6 90 – 155 3 n.s. 0.68
Werner 2010 common

Nöske et al. un-

San Francisco SE Ecuador 2000 15.6 2200 2.53 20 5-30 21 n.s.f 253 82 ** 0.27
2008 common

Werner &
seas. not
Gradstein Jerusalén N Ecuador 2300 16.9 530 0.60 5-8 80 n.s. 8 4 n.s. 0.77
common applicable
2009
a
values for entire study (forest trees and ITs combined).
b
“Effect of distance to forest” is marked by asterisks where epiphyte species numbers on ITs drop with increasing distance from closed forest. “Relative impov-
erishment” is the reduction in mean epiphyte species density (species per tree) on ITs relative to forest trees; asterisks mark significant differences in species
density between habitat types. Significance levels: * p < 0.05, ** p < 0.01, n.s. = not significant. See original publications for details on statistical analyses.
c
excluding two forest plots where mean tree size was substantially smaller than in ITs. Species density values controlled for taxonomic identity of host trees at
the same site: 0.61 (Quercus spp., inner crowns only; Flores-Palacios & García-Franco 2004); 0.87 (Liquidambar macrophylla, entire trees taken at 0.3 m dbh;
Flores-Palacios & García-Franco 2006).
d deduced from the linear regression equations (species no. vs. dbh at dbh = 0.5 m) given by Hietz-Seifert et al. 1996.
e pers. comm. L. Carrasco (Bilsa) and H.F. Greeney (SierrAzul) respectively.
f
RESILIENCE OF EPIPHYTE ASSEMBLAGES

17
from Werner 2008.
WERNER ET AL.

RESULTS AND DISCUSSION rainfall of 530 mm, species density on ITs was similar
to adjacent forest (Werner & Gradstein 2009).
We found that responses of vascular epiphyte assem-
In spite of the limited number of studies available
blages on ITs differed widely between studies (Fig. 1).
In two aseasonally wet areas of SE Mexico and NE for this analysis, the data suggests that, as predicted,
Ecuador, species density on isolated trees was not the response of vascular epiphyte assemblages to dis-
significantly reduced relative to forest trees of similar turbance does indeed vary systematically with meso-
size (Hietz-Seifert et  al. 1996, Larrea & Werner climate. A strong, significant decrease in species
2010). In contrast, at two moist study sites in NW density on ITs was observed only in the studies with
Ecuador with a moderately pronounced dry season intermediate humidity conditions, whereas both the
but frequent fogs throughout the year, epiphyte di- aseasonally wet and distinctly arid sites were little
versity was distinctly lower on ITs compared with affected (Fig. 1, Table 1). Thus a continuous supply
forest trees. Both at Bilsa (Köster et al. in press) and of moisture at perhumid sites appears to facilitate
Otonga (Köster et al. 2009), species density on ITs suitable growth conditions for a large number of epi-
was 36% lower. At a slightly seasonal moist forest in phyte species on ITs, even though post-disturbance
SE Ecuador, species density on ITs was over 80% turnover in species composition may be considerable
lower (Nöske et al. 2008, Werner 2008). This species- (e.g. Larrea & Werner 2010; compare also Sørensen
rich site regularly experiences short but pronounced indices in Table 1). In contrast, much lower species
dry spells induced by foehn winds (Emck 2007). ITs numbers were found under moderately moist, slight-
at an even more seasonal and drier site in eastern ly seasonal conditions, where many taxa depend on
Mexico showed a moderate decline in species density the protection by an intact canopy during periods of
(Flores-Palacios & García-Franco 2008). Finally, in a drought. Where the forest canopy is opened up, vapor
perarid forest area in northern Ecuador with annual pressure deficits increase strongly during dry spells
(e.g. Laurance 2004), which may result in high mor-
tality among resident moist forest epiphytes (Werner
2011). In dry forests, where the vertical microclimate
gradient is poorly developed even in intact stands
(Graham & Andrade 2004), disturbance appears to
result in relatively low declines in species numbers.
We suggest that the effective adaptations to drought
which are essential for dry forest epiphytes promote
their persistence on ITs. This notion is supported by
the paucity of epiphytic understory specialists re-
corded for tropical dry forests. In coastal Ecuador, for
instance, Gentry & Dodson (1987) found 41 species
of understory specialists among vascular epiphytes in
a wet forest, but none in a seasonal, semi-deciduous
moist forest. A corresponding pattern has been found
by Hietz & Hietz-Seifert (1995a, b) in montane
eastern Mexico. However, the most fragile, shade- or
moisture-demanding taxa of any local community are
unlikely to flourish in the multiple edge environment
FIG. 1. The relationship between humidity (given as of ITs (e.g. Poltz & Zotz 2011), and the β-diversity
an index of aridity) and the relative impoverishment of IT assemblages tends to be reduced (Flores-Palacios
in mean epiphyte species density (species number per & García-Franco 2004, 2008; Hietz 2005, Larrea &
tree) on isolated trees (ITs) relative to trees in intact Werner 2010; but see also Köster et al. 2009).
forest. Case studies: (1) Werner & Gradstein 2009; Factors other than mesoclimate cannot readily
(2) Flores-Palacios & García-Franco 2008; (3) Nöske explain the observed hump-shaped pattern of species
et al. 2008; (4) Köster et al. in press; (5) Köster et al. impoverishment (Fig. 1). Differences in host-tree size
2009; (6) Hietz-Seifert et  al. 1996; (7) Larrea & between habitat types were controlled for by restrict-
Werner 2010. The stippled line reflects 2nd-degree ing sampling to a certain range of stem diameter at
polynomial fit (y = -0.11x² + 0.57x - 0.27; r² = 0.48, breast height (dbh), or accounted for by regressing
p = 0.13). species numbers against dbh (Hietz-Seifert et  al.

18
 RESILIENCE OF EPIPHYTE ASSEMBLAGES

1996) in all studies included in our comparison. Tree ACKNOWLEDGEMENTS

species may vary widely in their quality as epiphyte
We thank A. Flores-Palacios, J. García-Franco and
hosts (e.g. Reyes-García et al. 2008, Köster et al. in
particularly P. Hietz for valuable information on their
press). However, host-tree species composition was
studies. Additional site information was kindly pro-
controlled for in all but two studies where large
numbers of trees were randomly selected (Hietz- vided by L. Carrasco, H. Greeney, T. Krömer and G.
Seifert et al. 1996, Flores-Palacios & García-Franco Williams-Linera. Comments of two anonymous re-
2008). At the latter site at least, epiphyte diversity viewers substantially helped improve the manuscript.
patterns remained similar when controlling for host- We acknowledge the support of the Deutscher Aka-
tree species identity (Flores-Palacios & García- demischer Austauschdienst, Idea Wild, Minox, the
Franco 2004, 2006; Table 1), suggesting that host- Akademie der Wissenschaften und der Literatur in
tree identity was not a major source of variability. ITs Mainz, and the Deutsche Forschungsgemeinschaft.
which established themselves after forest clearance This is publication no. 249 of the Yanayacu Natural
may be expected to be poorer in epiphytes than History Research Group.
remnant ITs, because their flora depends more
strongly on colonization from distant sources. How- REFERENCES
ever, species density on ITs was unchanged at the
Barthlott, W., Schmitt-Neuerburg, V., Nieder, J. & S.
only site where ITs were regrowth (Jerusalén; Werner Engwald. 2001. Diversity and abundance of vascular
& Gradstein 2009). Time since isolation of ITs, epiphytes: a comparison of secondary vegetation and
matrix quality (pasture in all cases), or distance to primary montane rain forest in the Venezuelan Andes.
forest as a proxy of dispersal limitations are also pos- Plant Ecol. 152: 145–156.
sible predictors, but neither showed a relationship Bush, M.B. 2002. Distributional change and conservation
with differences in relative impoverishment between on the Andean flank: a paleoecological perspective.
the compared studies. Distance to forest had signifi- Glob. Ecol. Biogeogr. 11: 463–473.
cantly adverse effects on species density in one study Cascante-Marín, A., Wolf, J.H.D., Ostermeijer, H.G.B.,
(Hietz-Seifert et al. 1996) out of six for which this den Nijs, J.C.M., Sanahuja, O. & A. Durán-Apuy.
correlation was analyzed (Table 1). 2006. Epiphytic bromeliad communities in secondary
In conclusion, our results support the notion of a and mature forest in a tropical montane area. Bas.
general relationship between mesoclimate and distur- Appl. Ecol. 7: 520–532.
bance effects on species density in vascular epiphytes. Colwell, R.K. 2006. EstimateS: Statistical estimation of
Maximum impoverishment may be expected in moist species richness and shared species from samples. Ver-
areas prone to short dry spells, as at the study site of sion 8. User’s Guide and application available from
Nöske et al. (2008). At this site, more than 50% of http://purl.oclc.org/estimates (accessed February
well-established vascular epiphyte individuals tagged 2010).
in lower forest strata died within one year of their host Dunn, R.R. 2000. Bromeliad communities in isolated trees
trees being isolated in a fresh clearing (Werner 2011). and three successional stages of an Andean cloud forest
Our paper shows that the fate of epiphytes in in Ecuador. Selbyana 21: 137–143.
Emck, P. 2007. Klimadifferenzierung in Süd-Ecuador.
anthropogenic tropical habitats remains insuffi-
Ph.D. Thesis, Universität Erlangen, Germany.
ciently understood. The hypothesis of a hump-
Flores-Palacios, A. & J.G. García-Franco. 2004. Effects of
shaped relationship between mesoclimate and epi-
isolation on the structure and nutrient content of oak
phyte species losses still awaits rigid testing, for which
epiphyte communities. Plant Ecol. 173: 259–269.
the number of comparable studies presently is very
Flores-Palacios, A. & J.G. García-Franco. 2006. The relati-
low. Clearly, several factors other than climate (e.g. onship between tree size and epiphyte species richness:
dispersal limitations) can be important local deter- testing four different hypotheses. J. Biogeogr. 33:
minants of epiphyte assemblages in man-made 323–330.
landscapes (Cascante-Marín et al. 2006, Werner & Flores-Palacios, A. & J.G. García-Franco. 2008. Habitat
Gradstein 2008) and will need to be considered in isolation changes the beta diversity of the vascular
future studies. Advances in this field of research will epiphyte community in lower montane forest, Veracruz,
have important applications in conservation planning Mexico. Biodivers. Conserv. 17: 191–207.
and management, as well as in the prediction of Gardner, T.A., Barlow, J., Parry, L.W. & C.A. Peres. 2007.
global climate change effects on the diversity and Predicting the uncertain future of tropical forest species
functionality of tropical epiphyte communities. in a data vacuum. Biotropica 39: 25–30.

19
WERNER ET AL.

Gascon, C., Lovejoy, T.E., Bierregaard, R.O., Malcolm J.R., Laurance, W.F. 2006. Have we overstated the tropical
Stouffer P.C., Vasconcelos, H.L., Laurance W.F., Zim- biodiversity crisis? Trends Ecol. Evol. 22: 65–70.
merman B., Tocher M. & S. Borges. 1999. Matrix Malhi Y. & O.L. Phillips. 2004. Tropical forests and global
habitat and species richness in tropical forest remnants. atmospheric change: a synthesis. Phil. Trans. Roy. Soc.
Biol. Conserv. 91: 223–229. B 359: 549–555.
Gentry, A.H. & C.H. Dodson. 1987. Diversity and bio- Manning, A.D., Fischer, J. & D.B. Lindenmayer. 2006.
geography of Neotropical vascular epiphytes. Ann. Scattered trees are keystone structures – implications
Miss. Bot. Gard. 74: 205–233. for conservation. Biol. Conserv. 132: 311–321.
Gonçalvez, C.N. & J.L. Waechter. 2003. Aspectos florísti- Nkongmeneck, B.-A., Lowman, M.D. & J.T. Atwood.
cos e ecológicos de epífitos vasculares sobre figueiras 2002. Epiphyte diversity in primary and fragmented
isoladas no norte da planície costeira do Rio Grande forests of Cameroon, Central Africa: a preliminary
Do Sul. Acta Bot. Bras. 17: 89–100. survey. Selbyana 23: 121–130.
Gove, A.D., Majer, J.D. & V. Rico-Gray. 2009. Ant assem- Nöske, N.M., Hilt, N., Werner, F.A., Brehm, G., Fiedler,
blages in isolated trees are more sensitive to species loss K., Sipman, H.J.M. & S.R. Gradstein 2008. Distur-
and replacement than their woodland counterparts. bance effects on epiphytes and moths in a montane
Bas. Appl. Ecol. 10: 187–195. forest in Ecuador. Bas. Appl. Ecol. 9: 4–12.
Graham, E.A. & J.L. Andrade. 2004. Drought tolerance Poltz, K. & G. Zotz. 2011. Vascular epiphytes on isolated
associated with vertical stratification of two co-occur- pasture trees along a rainfall gradient in the lowlands
ring epiphytic bromeliads in a tropical dry forest. Am. of Panama. Biotropica 43: 165–172.
J. Bot. 91: 699–706. Reyes-García, C., Griffiths, H., Rincón, E. & P. Huante.
Hietz, P. 2005. Conservation of vascular epiphyte diversity 2008. Niche differentiation in tank and atmospheric
in a Mexican coffee plantation. Conserv. Biol. 19:
epiphytic bromeliads of a seasonally dry forest. Biotro-
391–399.
pica 40: 168–175.
Hietz, P. & U. Hietz-Seifert. 1995a. Composition and
Thornthwaite, C.W. & J.R. Mather. 1957. Instructions and
ecology of vascular epiphyte communities along an
tables for computing potential evapotranspiration and
altitudinal gradient in central Veracruz, Mexico. J. Veg.
the water balance. Drexel Institute of Technology, La-
Sci. 6: 487–498.
boratory of Climatology, Publications in Climatology
Hietz, P. & U. Hietz-Seifert. 1995b. Structure and ecology
10: 185–311.
of epiphyte communities of a cloud forest in central
UNEP 1992. World Atlas of Desertification. Edward Ar-
Veracruz, Mexico. J. Veg. Sci. 6: 719–728.
nold, London.
Hietz-Seifert, U., Hietz, P. & S. Guevara. 1996. Epiphyte
Werner, F.A. 2008. Effects of human disturbance on diver-
vegetation and diversity on remnant trees after forest
sity and ecology of vascular epiphyte assemblages in the
clearance in southern Veracruz, Mexico. Biol. Conserv.
75: 103–111. Andes of Ecuador. Ph.D. Thesis, Universität Göttingen,
Köster, N., Friedrich, K., Nieder, N. & W. Barthlott. 2009. Germany.
Conservation of epiphyte diversity in an Andean lands- Werner, F.A. 2011. Reduced growth and survival of vascu-
cape transformed by human land use. Conserv. Biol. lar epiphytes on isolated remnant trees in a recent
23: 911–919. tropical montane forest clear-cut. Bas. Appl. Ecol. 12:
Köster, N., Nieder, J. & W. Barthlott. In press. Effect of 172–181
host tree traits on epiphyte diversity in natural and Werner, F.A. & S.R. Gradstein. 2008. Seedling establish-
anthropogenic habitats in Ecuador. Biotropica. ment of vascular epiphytes on isolated and enclosed
Kreft, H., Köster, N., Küper, W., Nieder, J. & W. Barthlott. forest trees in an Andean landscape, Ecuador. Biodivers.
2004. Diversity and biogeography of vascular epiphytes Conserv. 17: 3195–3207.
in Western Amazonia, Yasuní, Ecuador. J. Biogeogr. 31: Werner, F.A. & S.R. Gradstein 2009. Diversity of epiphytes
1463–1476. of tropical dry forest along a gradient of human distur-
Krömer T. & S.R. Gradstein. 2003. Species richness of bance in the Andes of Ecuador. J. Veg. Sci. 20: 59–68.
vascular epiphytes in two primary forests and fallows in Williams-Linera, G., Sosa, V. & T. Platas. 1995. The fate
the Bolivian Andes. Selbyana 24: 190–195. of epiphytic orchids after fragmentation of a Mexican
Larrea, M. & F.A. Werner. 2010. Response of vascular cloud forest. Selbyana 16: 36–40.
epiphyte diversity to different land-use intensities in a Wolf, J.H.D. 2005. The response of epiphytes to anthro-
neotropical montane wet forest. For. Ecol. Manage. pogenic disturbance of pine-oak forests in the highlands
260: 1950–1955. of Chiapas, Mexico. For. Ecol. Manage. 212: 376–393.
Laurance, W.F. 2004. Forest-climate interactions in frag- Zahawi, R.A. & C.K. Augspurger. 2006. Tropical forest
mented tropical landscapes. Phil. Trans. Roy. Soc. restoration: tree islands as recruitment foci in degraded
Lond. B 359: 345–352. lands of Honduras. Ecol. Appl.16: 464–478.

20