Insectes soc.

47 (2000) 94–95
0020-1812/00/010094-02 $ 1.50+0.20/0 Insectes Sociaux
© Birkhäuser Verlag, Basel, 2000

Short communication

Colony relatedness in aggregations of Apis dorsata Fabricius

(Hymenoptera, Apidae)
B.P. Oldroyd 1, K.E. Osborne 1 and M. Mardan 2
1 School of Biological Sciences A12, University of Sydney, Sydney NSW 2066, Australia, e-mail: [email protected]
2 Plant Protection Department, Faculty of Agriculture, 43400 UMP Serdang, Selangor Dural Ehsan, Malaysia

Received 8 December 1998; revised 15 July and 18 October 1999; accepted 3 November 1999.

Summary. Apis dorsata colonies often form dense aggrega- A second important feature of A. dorsata biology is fre-
tions, with over 100 colonies sometimes seen in the same quent colony migration (Butani, 1950; Thaker and Tonapi,
tree. Reasons for these aggregations are unknown, but one 1961; Morse, 1970; Koeniger and Koeniger, 1980; Dyer and
reasonable hypothesis is that colonies form a related family Seeley, 1994). Migration follows a seasonal pattern, and may
group. Here we show that 7 adjacent colonies sampled from be related to available forage, or to predation (particularly by
a single branch of a tree (near Alor Setar in northern penin- humans). A further intriguing possibility is that migration
sular Malaysia) containing over 120 colonies were not relat- helps control levels of the parasitic mite Tropilaelaps clareae
ed as mother/daughter. Thus the notion that aggregations (Rinderer et al., 1994).
arise through splitting of the first-arriving colonies can be A third feature is the process of reproductive swarming.
rejected. This appears to be of two kinds (Lindauer, 1956). In the first
(which seems akin to that observed in A. mellifera) a queen
Key words: Relatedness, swarming, aggregations, migration, will leave the nest, and, as she flies slowly away, she is
Apis dorsata. followed by a swarm of workers. It is not known whether
these swarms form a temporary combless cluster, or whether
they move directly to a new nest site. The distance that these
swarms generally move is also unknown, but at least one such
Introduction swarm has settled more than 500 m from the natal nest (Lin-
dauer, 1956). In the second mechanism, known as “budding”,
Apis dorsata is the largest bee in the genus. Its range extends a group of workers gradually separates from the nest, and
from the Indian subcontinent to south-east Asia, including forms a new colony about 1 m from the old (Lindauer, 1956).
the Philippines (Ruttner, 1988). Like the dwarf bees A. florea Here we report on the relatedness of adjacent colonies
and A. andreniformis, its colonies are characterised by single in a very large aggregation. If the “budding” observed by
combs, which are usually suspended from tree branches Lindauer (1956) is common, then we would predict that
(Morse, 1970; Seeley et al., 1982). Less commonly, colonies daughter colonies would be found adjacent to mother colo-
nest on cliff overhangs or on human-built structures (Ruttner, nies. Another potential reason why we might expect closely
1988). related colonies to group together is the inclusive fitness
A. dorsata (and the related species A. laboriosa) is unique benefits of cooperative defence, which has been postulated
in that colonies are often found in dense aggregations (Lin- as the reason for aggregations of A. dorsata (Seeley et al.,
dauer, 1956; Seeley et al., 1982; Roubik et al., 1985; Under- 1982).
wood, 1990; Dyer and Seeley, 1994). While A. mellifera nests
tend to be aggregated (Oldroyd et al., 1995; McNally and
Schneider, 1996), aggregation is much more pronounced Materials and methods
in A. dorsata. Over 120 individual colonies may occur on
a single tree or rock face (pers. obs.). Aggregations are facul- We studied the relatedness of 7 adjacent A. dorsata colonies to deter-
mine whether neighbouring colonies are unrelated. The null hypothesis
tative, and individual colonies or small aggregations of that two queens are related as mother daughter can be unambiguously
2–3 colonies are more common than mass aggregations rejected if the two queens do not share an allele at any one locus. The
(pers. obs.). null hypothesis that two queens are related as half sisters cannot be
Insectes soc. Vol. 47, 2000 Short communication 95

rejected so simply, and because of the limited scope of this study, we do Table 1. Inferred genotype (microsatellite length in base pairs) of
not consider this possibility further here. queens heading 7 adjacent A. dorsata colonies
Samples were collected from a “bee tree” situated in rainforest at
Pedu Lake near Alor Setar in northern peninsular Malaysia. This partic- Colony A14 A88 A107 B124
ular tree (Kompassia excelsa) was extremely large (estimated as 90 m),
and in 1995 was occupied by 127 colonies. Honey bees have occupied 1 164/167 97/124 212/284 216/216
this tree since at least 1964, and colonies are harvested annually by 2 167/170 97/146 173/216 216/218
Mr. Salleh Mohd. Nor, a local honey hunter, using traditional methods. 3 167/169 118/131 223/297 216/218
Mr. Nor reports that the tree is usually occupied by migrating A. dorsata 4 164/169 97/124 189/268 218/218
swarms during October and November. These swarms build combs and 5 169/Null 97/104 207/209 216/218
establish colonies, and some reproduction of colonies may occur. Colo- 6 170/173 106/124 187/212 216/?
nies begin to abscond in March, and the tree is usually deserted by April. 7 164/164 99/102 189/205 216/218
Mr. Nor provided us with brood from 7 sequential colonies along the
lowest branch (about 40 m above ground) of this tree. Samples were
frozen in liquid nitrogen for transport to the laboratory where they were
stored at –70°C. Acknowledgments
The microsatellite genotype of each of 10 bees was determined for
4 loci (A14, A88, A107 and B124 identified by Estoup et al., 1994). To Thanks to K. Parker for technical assistance, D. Low and A. Barron for
do this, we extracted DNA from the leg of an adult by boiling the ground critically reading the manuscript, and the Australian Research Council
tissue in 5% Chelex 100 resin for fifteen minutes (Walsh et al., 1991). for funding. Two anonymous reviewers and the Editor straightened out
We then amplified these DNA extractions using the polymerase chain some woolly thinking. Thanks for that.
reaction (PCR) with primers specific to the honey bee microsatellite
loci. In each case the reverse primer had been labelled by the manufac-
turer (Bresatec, Adelaide, Australia) with the flourescent dye HEX. References
PCRs were conducted in 10 ml volumes, of which 2 ml was the DNA
extraction. Each reaction contained 400 nM of each primer, 150 mM of Butani D.K., 1950. An Apis dorsata colony in New Delhi. Ind. Bee J. 12:
each dNTP, 1.5 mM MgCl2, 1 ¥ reaction buffer, and 0.45 units of 115.
Tth polymerase (Fisher Biotech, Perth, Australia). After an initial dena- Dyer F.C. and T.D. Seeley, 1994. Colony migration in the tropical
turing of the template DNA for 4 min at 94°C, samples underwent honey bee Apis dorsata F. (Hymenoptera: Apidae). Insectes soc. 41:
30 cycles of denaturation (30 s at 94°C), annealing (30 s at 55 °C), 129–140.
and extension (30 s at 72°C) followed by final elongation for 10 min Estoup A., M. Solignac and J.-M. Cornuet, 1994. Precise assessment of
at 72°C. PCR products were electrophoresed on an automated DNA the number of patrilines and of genetic relatedness in honey bee
fragment analyser (Corbett Research, Sydney), which, by reference to colonies. Proc. Roy. Soc. Lond. B 258: 1–7.
standard lanes, called the allele length in base pairs (bp). Each lane was Koeniger N. and G. Koeniger, 1980. Observations and experiments on
manually checked, and if necessary the automated call was adjusted to migration and dance communication of Apis dorsata in Sri Lanka.
take account of gel warping. J. Apic. Res. 19: 21–34.
Queen genotype was inferred from the genotypes of the 10 workers Lindauer M., 1956. Communication among the honeybees and stingless
via the following rules: (1) Queens carry only two alleles and all work- bees of India. Bee Wld. 38: 3–14, 34–39.
ers must carry one of those two alleles. (2) If any worker is homo- McNally L.C. and S.S. Schneider, 1996. Spatial distribution and nesting
zygous, then her mother carries that allele. (3) If all 10 workers carry an biology of colonies of the African honey bee Apis mellifera scutellata
allele, then the queen is likely to be homozygous for that allele. (4) If all (Hymenoptera: Apidae) in Botswana, Africa. Env. Ent. 25: 643–652.
10 workers carry one of two alleles, then the queen is likely to be hetero- Morse R.A., 1970. Annotated bibliography on Apis dorsata. Internatio-
zygous for those two alleles. In combination, the rules will almost nal Bee Research Association, Gerrards Cross, UK. 25 pp.
always unambiguously define the queen genotype. Violations will occur Oldroyd B.P., A. Smolenski, S. Lawler, A. Estoup and R. Crozier, 1995.
when the queen carries a null allele, but by reference to other loci, this Colony aggregations in Apis mellifera. Apidologie 26: 119–130.
condition is readily distinguished from the alternative possibility of Oldroyd B.P., A.J. Smolenski, J.-M. Cornuet, S. Wongsiri, A. Estoup,
multiple maternity (Oldroyd et al., 1996). T.E. Rinderer and R.H. Crozier, 1996. Levels of polyandry and
intracolonial genetic relationships in Apis dorsata (Hymenoptera:
Apidae). Anls. Ent. Soc. Am. 89: 276–283.
Results and discussion Rinderer T.E., B.P. Oldroyd, C. Lekprayoon, S. Wongsiri, C. Boonthai
and R. Thapa, 1994. Extended survival of the parasitic mite Tropi-
laelaps clareae on adult workers of Apis mellifera and Apis dorsata.
No queens were related as mother/daughter, because all pairs J. Apic. Res. 33: 171–174.
of individuals carried different alleles at at least one locus Roubik D.W., S.F. Sakagami and I. Kudo, 1985. A note on distribution
(Table 1). Although we cannot exclude the possibility that and nesting of the Himalayan honey bee Apis laboriosa Smith
some queens studied were half sisters, none were mother/ (Hymenoptera: Apidae). J. Kans. Ent. Soc. 58: 746–749.
Ruttner F., 1988. Biogeography and Taxonomy of Honeybees. Springer-
daughter, and it is most likely that colonies were unrelated. Verlag, Berlin. 284 pp.
This indicates that there is no tendency for daughter colonies Seeley T.D., R.H. Seeley and P. Aratanakul, 1982. Colony defence stra-
to remain close to the mother colony. Thus reproductive tegies of the honeybees in Thailand. Ecol. Mono. 52: 43–63.
swarming via “budding”, as observed by Lindauer (1956) Thaker C.V., Tonapi K.V., 1961. Nesting behaviour of Indian honeybees.
may not be common. Further, mutual defence via aggrega- I. Differentiation of worker, queen and drone cells of the combs of
Apis dorsata Fabre. Bee Wld. 42: 61–62.
tion of highly related colonies also appears unlikely. Underwood B.A., 1990. Seasonal nesting cycle and migration patterns
We conclude that aggregations of A. dorsata arise by of the Himalayan honey bee Apis laboriosa. Nat. Geog. Res. 6:
unrelated colonies gathering at particular nest sites, not by 276–290.
budding of colonies. Walsh P.S., D.A. Metzger and R. Higuchi, 1991. Chelex (R)100 as a
medium for simple extraction of DNA for PCR-based typing from
forensic material. Biotechniques 10: 507.