Monophyly and extensive extinction of advanced
eusocial bees: Insights from an unexpected
Eocene diversity
Michael S. Engel
Division of Entomology, Natural History Museum and Biodiversity Research Center and Department of Ecology and Evolutionary Biology,
1460 Jayhawk Boulevard, University of Kansas, Lawrence, KS 66045-7523
Communicated by Charles D. Michener, University of Kansas, Lawrence, KS, December 20, 2000 (received for review November 1, 2000)
Advanced eusociality sometimes is given credit for the ecological
success of termites, ants, some wasps, and some bees. Comprehensive
study of bees fossilized in Baltic amber has revealed an
unsuspected middle Eocene (ca. 45 million years ago) diversity of
eusocial bee lineages. Advanced eusociality arose once in the bees
with significant post-Eocene losses in diversity, leaving today only
two advanced eusocial tribes comprising less than 2% of the total
bee diversity, a trend analogous to that of hominid evolution. This
pattern of changing diversity contradicts notions concerning the
role of eusociality for evolutionary success in insects.
Apidae ͉ Baltic amber ͉ Hominidae ͉ Paleogene
The origin of sterile-worker castes in insects has long been
an arena for debate in evolutionary biology and was cited
by Darwin (1) as one of the more troubling biological phenomena
for his ‘‘species theory.’’ Eusociality, the category of
societies characterized by, among other features, nonreproducing
individuals (i.e., workers) that cooperate to raise the
brood of another female (i.e., queen), is a trait that has
appeared in numerous lineages from mammals (2) to shrimp
(3) but is best known and developed within the insects. Some
eusocial insect societies possess a further level of complexity
through morphological specialization of the worker caste.
Advanced eusocial societies, those societies with a morphologically
differentiated worker caste, are among the most
dominant and conspicuous groups of insects in the world and
include the termites, ants, social wasps (vespids), and some
bees. It is significant that these insects, at least the first two,
are arguably the most ecologically dominant groups of their
class (4–8). The ecological success and diversity of the termites
(Ϸ2,700 spp.), ants (Ϸ16,000 spp.), and social wasps (Ϸ1,000
spp.) have been attributed to their advanced eusocial behavior
as it has been also for the advanced eusocial bees (6, 7). The
advanced eusocial insects are highly efficient foragers; the
division of labor allows task specialization, rapid recruitment
to both floral and nonfloral resources, mobilization of colony
defense, and support for large colony sizes (6).
The bees (with Ϸ20,000 spp.; ref. 8) as a whole are more
diverse than the other advanced eusocial lineages. They are
known for their dominant role as pollinators of flowering plants
(9–11), thereby sustaining major ecological systems worldwide.
The great majority of bees, however, are solitary. Advanced
eusocial behavior is found in only a single group of floral
generalist bees: the corbiculate Apinae, so named for the
modification of the hind tibia into a corbicula or ‘‘pollen basket.’’
The corbiculate bees comprise four living tribes and include the
most commonly recognized of all bees: the orchid bees
(Euglossini), the bumble bees (Bombini), the stingless bees
(Meliponini), and the honey bees (Apini). Of these four, only
the stingless bees and honey bees exhibit advanced eusocial
behavior. Although the bumble bees are eusocial also, the only
morphological difference between queen and worker is size.
The orchid bees are not eusocial, although some species are
communal.
Monophyly of a clade consisting of the two advanced eusocial
tribes has been debated in recent years on the basis of morphological
and molecular evidence (12–14). Associated with the
conflict over corbiculate bee relationships have been various
estimates for the number of times eusociality has arisen. The
corbiculate bees are a derived clade of the family Apidae, itself
the most derived group of the seven families of bees (15, 16).
Based on the derived position of the corbiculate apines among
other bees, some authors have suggested that complex social
behavior in bees is a rather recent trait, having its origins and
rapid diversification merely 30–40 million years ago (17, 18). A
paleontological perspective is critical not only for understanding
the past diversity of life on Earth, episodes of extinction and
radiation, the origin of evolutionary novelties (i.e., synapomorphies),
and paleoecologies but also for the reconstruction of
hierarchical relationships among taxa (19–22). Thus, a study of
the geological diversity of social bees potentially can provide
insights into the evolution of this group of bees as well as resolve
the contention over their phylogeny.
The Baltic-amber fauna, recently dated approximately 45
million years in age (middle Eocene), is the earliest diverse
fossil record of bees. Previously believed to consist of a few
enigmatic taxa (23), Baltic amber actually harbors a remarkable
array of taxa representing most higher clades of the
Apoidea. Indeed, bee fossils are rare in amber with one
specimen appearing for approximately every 5,000 inclusions
examined; however, I was able to find 152 specimens representing
38 extinct species and include them in cladistic analyses
with living families, tribes, and genera of corbiculate and
noncorbiculate bees (16). In addition, fossil bees from all other
deposits and ages were compared with the Baltic amber as well
as Recent faunas to gain an insight into changes of the fauna
and diversity through time. Compression fossils provide additional
data on ages and locations, but preservation in amber
has unique fidelity. These fossils are therefore more meaningful,
because bees are typically identifiable on the basis of
minute structures of the mouthparts, hairs, or leg spurs,
characters that are frequently not preserved in compression
fossils. Moreover, amber preserves internal skeletal features as
well as tissues (24), making it possible to compare some
internal characteristics of fossil and Recent bees. One result of
this paleontological work was the discovery of a hitherto
unexpected diversity of Eocene corbiculate bees. Although
corbiculate bee fossils are known also from a variety of other
deposits (e.g., refs. 25–27), their numbers in terms of species
are dramatically less than those from Baltic amber. These
The publication costs of this article were defrayed in part by page charge payment. This
article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C.
§1734 solely to indicate this fact.
Article published online before print: Proc. Natl. Acad. Sci. USA, 10.1073͞pnas.041600198.
Article and publication date are at www.pnas.org͞cgi͞doi͞10.1073͞pnas.041600198
PNAS ͉ February 13, 2001 ͉ vol. 98 ͉ no. 4 ͉ 1661–1664
EVOLUTION
amber fossils reveal interesting patterns of the overall temporal
changes in the composition of bee faunas as well as
compelling evidence for the monophyly of the advanced
eusocial bee tribes and implications for their diversification.
More recent corbiculate bee taxa (i.e., Oligocene or younger)
preserved as either amber inclusions or compression fossils
group cladistically not only within the four living tribes, but
some are within living genera (e.g., refs. 25 and 27).
Fig. 1. Phylogeny of the corbiculate bees (Apinae) as well as representative fossils from a variety of Tertiary deposits based on analysis of 51 adult morphological
characters of external and internal skeletal features. Topology presented is a strict consensus of two most parsimonious trees: tree length, 65; consistency index,
81; and retention index, 93 (16). From top to bottom (and left to right), the fossils are Euglossa moronei Engel, Genus A, Genus B, Genus D, Apis henshawi
Cockerell, Genus E, Genus G, Genus I, and Nogueirapis silacea (Wille). Eocene fossil taxa on the cladogram include genera A–I and the meliponine Kelneriapis.
Black lines indicate noneusocial (solitary or communal) species; yellow indicates primitively eusocial lineages (i.e., lacking a morphologically specialized worker
caste); red indicates advanced eusocial lineages. Fossils represented by individuals that are morphologically workers are indicated in green. Most fossil genera
are not monotypic (e.g., Genus B has five species, Genus C has four species) (16).
1662 ͉ www.pnas.org Engel
Cladistic analyses of Recent and fossil corbiculate bee taxa by
using 51 characters of both external and internal morphology
(tree length, 65; consistency index, 81; retention index, 93) (Fig.
1) indicate the presence in the Eocene of a previously unknown
diversity of extinct corbiculate clades. The data matrix of characters
and states on which the cladogram is based is available
(16), and specimens are deposited in public institutions. Nearly
all of these fossil lineages are represented by individuals exhibiting
the reduced metasoma and to a lesser degree the barbed
sting found only in the worker caste of advanced eusocial species,
especially well developed in the Meliponini for the former
character and Apini for the latter. Thus, although the fossils were
coded and analyzed as unknown for their social-character state,
all except one (Genus A in Fig. 1) apparently were advanced
eusocial forms. The facts that these extinct taxa group cladistically
with the honey bees and stingless bees on the basis of all
other traits, and the primitively eusocial bumble bees form their
sister group (Fig. 1), further support the interpretation that the
fossils also were members of advanced eusocial societies (28).
These fossils therefore reinforce the phylogenetic relationships
of both monophyletic general-eusocial and advanced-eusocial
clades (Fig. 1). Advanced eusociality thus originated only once
within the bees, presumably sometime in the Late Cretaceous.
The origin of eusociality is fairly ancient in the bees because an
advanced eusocial corbiculate bee (a meliponine) is known from
latest Cretaceous deposits, presumably Maastrichtian (11), from
North America (26).
Another significant aspect of the diversity of Eocene advanced
eusocial bees is their extinction, not the simple loss of a few
species but the apparent loss of several higher clades of advanced
eusocial bees (Fig. 1). The success and diversification of the
advanced eusocial insects frequently is attributed to their sociality
(6); a general overview of the modern diversity of the ants,
termites, and wasps would seem to support this notion. The
advanced eusocial bees, however, comprise Ϸ380 spp., less than
2% of the total diversity of bees worldwide (8). This fact, plus the
status of advanced eusocial bees as seeming to be derived
phylogenetically recently (15), suggests that on the basis of the
extant fauna alone, bee eusociality is in nascent stages of
evolution. Indeed, advanced eusociality in bees would seem to
have had the same general effect that it has had in the ants,
termites, and wasps; that is, social cooperation has led to
ecological dominance. Honey bees (Apis), in particular, are
aggressive foragers and sometimes outcompete other native-bee
species foraging for the same floral resources (29–33). In areas
that the western honey bee, Apis mellifera, has been introduced,
the diversity of native bees has become diminished somewhat,
with some species competitively excluded, including other advanced
eusocial lineages (29, 30, 32, 34–36). In fact, local
extinction of stingless-bee colonies was one of the predicted
likely outcomes of the competitive interactions between introduced
Apis and native Meliponini (36). Apis mellifera introduced
into Australia has been shown to have a negative impact not only
on other bee species but also on some nectar-feeding bird
populations (37). Interestingly, in the most ecologically dominant
and aggressive advanced-eusocial lineage (i.e., the honey
bees, genus Apis), there are only seven extant species (38). The
generally less aggressive stingless bees are more diverse, and
their maximal diversity occurs in areas where Apis species are not
native (i.e., South and Central America; refs. 8 and 39).
An epoch-by-epoch overview of the geological history of the
corbiculate bees including all known amber and compression
fossils throughout the world shows a marked decrease in diversity
of clades between the Eocene and more recent epochs (16).
Today there are only two tribal-level clades of advanced eusocial
bees. In the Eocene, there were at least three in central Europe
alone, and the cladogram suggests that the lineage that eventually
gave rise to the Apini in the Oligocene was present as well
(Fig. 1). There was at least a 50% loss of suprageneric clades
among the advanced eusocial bees. At infratribal levels, a
30–33% loss of diversity can be documented between the Eocene
and subsequent faunas. This loss is borne out not only by the
Eocene Baltic-amber fauna but also by Eocene compression
fossils of similar taxa in other deposits (e.g., ref. 40) and
unknown in post-Eocene sediments. Beginning in the Oligocene,
compression and amber fossil bees are distinctly modern in
appearance (16, 23, 25, 27). Amber potentially can give a biased
picture of a particular fauna by capturing primarily those taxa
that collect resin for nest construction. The picture of corbiculate
bee Eocene diversity in the Baltic area is likely quite accurate
because living corbiculate bees collect resins and are therefore
more likely to be preserved in amber than are bees of other
lineages (e.g., sweat bees, digger bees). Thus, considering the
corbiculate bees alone, if any bias is indeed present, then it is
oversampling the corbiculate bee fauna. The observed diversity
can be only an underestimate of the actual diversity, thereby
making these estimates of changes in the corbiculate bee fauna
probably conservative.
It seems that advanced eusociality has not fueled post-Eocene
diversification of the corbiculate bees. In fact, it is possible that
over geological expanses of time, aggressively foraging advanced-eusocial
species had a negative impact not only on
nonsocial species but also and especially on other advanced
eusocial lineages. A similar situation is reported from another
highly competitive, social species (albeit not eusocial) for which
the fossil record indicates aggressive exclusion of related social
genera and species that may have caused their extinction—
namely Homo sapiens (41). During the Pleistocene, a diversity of
related genera and now extinct species of Homo existed but
perhaps were competitively excluded by a single aggressive
species that is the only survivor of a prior hominid radiation (41,
42). As for the human example as well, competition perhaps is
not the sole explanation for the declining diversity of advanced
eusocial bees. Clearly the global climate during the Eocene was
warmer than that of today (e.g., tropical-like bee taxa in the
Baltic region). Such conditions may have been favorable for the
proliferation of these kinds of bees, although some advanced
eusocial bees do occur in cold-temperate habitats today. The
interaction of such factors across geological time may have led
to their decline.
Eusociality also seems not to have been correlated directly
with species diversity and abundance in the ants and termites.
Both of these groups originated in the Cretaceous, during which
time they remained rare and primitive while also being advanced
eusocial (43, 44). The fact that ants and termites became
abundant and diverse 50 million years or more after their origins
indicates that, as for the bees, eusociality alone cannot account
for their current success.
The pervasive factor in the success of the bees as a whole
probably has been their mutualistic association with flowering
plants, a relationship cultivated since the early mid-Cretaceous
(8–11, 16, 26). Advanced eusocial behavior is not in its nascent
stages in the bees but presently is a rudiment of what it once was
and seems to have been unimportant in the greater diversification
of bees. Eusociality in this context, despite its ecological
importance and the evolutionary interest associated with it, has
done little to inspire the diversity of one of its most highly
celebrated examples—the bees.
I thank David A. Grimaldi, Kumar Krishna, Valerie Krishna, L. Krishtalka,
Charles D. Michener, Molly G. Rightmyer, Jerome G. Rozen, Jr.,
and Thomas N. Taylor for encouragement and David A. Grimaldi,
Charles D. Michener, Molly G. Rightmyer, Jerome G. Rozen, Jr., and
four anonymous reviewers for reading earlier versions of the text. Steve
Thurston kindly prepared the color cladogram. Support for this research
was provided generously by Robert G. Goelet.
Engel PNAS ͉ February 13, 2001 ͉ vol. 98 ͉ no. 4 ͉ 1663
EVOLUTION
1. Darwin, C. R. (1859) On the Origin of Species by Means of Natural Selection or
the Preservation of Favored Races in the Struggle for Life (John Murray, London).
2. Sherman, P. W., Jarvis, J. U. M. & Alexander, R. D. (1991) The Biology of the
Naked Mole Rat (Princeton Univ. Press, Princeton).
3. Duffy, J. E. (1996) Nature (London) 381, 512–514.
4. Krishna, K. & Weesner, F. M. (1969) Biology of Termites (Academic, New
York), Vol. 1.
5. Krishna, K. & Weesner, F. M. (1970) Biology of Termites (Academic, New
York), Vol. 2.
6. Wilson, E. O. (1971) The Insect Societies (Belknap, Cambridge, U.K.).
7. Ho¨lldobler, B. & Wilson, E. O. (1990) The Ants (Belknap, Cambridge, U.K.).
8. Michener, C. D. (2000) The Bees of the World (Johns Hopkins Univ. Press,
Baltimore).
9. Barth, F. G. (1985) Insects and Flowers: The Biology of a Partnership (Princeton
Univ. Press, Princeton).
10. O’Toole, C. & Raw, A. (1991) Bees of the World (Blandford, London).
11. Grimaldi, D. (1999) Ann. Mo. Bot. Gard. 86, 373–406.
12. Schultz, T. R., Engel, M. S. & Prentice, M. (1999) Univ. Kans. Nat. Hist. Mus.
Spec. 24, 125–138.
13. Mardulyn, P. & Cameron, S. A. (1999) Mol. Phylogenet. Evol. 12, 168–176.
14. Koulianos, S., Schmid-Hempel, R., Roubik, D. W. & Schmid-Hempel, P. (1999)
J. Evol. Biol. 12, 380–384.
15. Roig-Alsina, A. & Michener, C. D. (1993) Univ. Kans. Sci. Bull. 55, 123–162.
16. Engel, M. S. (2001) Bull. Am. Mus. Nat. Hist., in press.
17. Seeley, T. D. (1985) Honeybee Ecology (Princeton Univ. Press, Princeton).
18. Winston, M. (1987) The Biology of the Honey Bee (Harvard Univ. Press,
Cambridge, MA).
19. Donoghue, M. J., Doyle, J. A., Gauthier, J., Kluge, A. G. & Rowe, T. (1989)
Annu. Rev. Ecol. Syst. 20, 431–460.
20. Gauthier, J., Kluge, A. G. & Rowe, T. (1989) Cladistics 4, 105–209.
21. Novacek, M. J. (1992) in Extinction and Phylogeny, eds. Novacek, M. J. &
Wheeler, Q. D. (Columbia Univ. Press, New York), pp. 46–88.
22. Grimaldi, D. & Cumming, J. (1999) Bull. Am. Mus. Nat. Hist. 239, 1–124.
23. Zeuner, F. E. & Manning, F. J. (1976) Bull. Br. Mus. Nat. Hist. Geol. 27,
149–268.
24. Grimaldi, D., Bonwich, E., Delannoy, M. & Doberstein, S. (1994) Am. Mus.
Novit. 3097, 1–13.
25. Engel, M. S. (1999) Am. Mus. Novit. 3272, 1–14.
26. Engel, M. S. (2000) Am. Mus. Novit. 3296, 1–11.
27. Camargo, J. M. F., Grimaldi, D. A. & Pedro, S. R. M. (2000) Am. Mus. Novit.
3293, 1–24.
28. Schultz, T. R., Cocroft, R. B. & Churchill, G. A. (1996) Evolution (Lawrence,
Kans.) 50, 504–511.
29. Roubik, D. W. (1978) Science 201, 1030–1032.
30. Roubik, D. W. (1983) Ecology 64, 971–978.
31. Eickwort, G. C. & Ginsberg, H. S. (1980) Annu. Rev. Entomol. 25, 421–446.
32. Buchmann, S. L. (1996) in The Conservation of Bees, eds. Matheson, A.,
Buchmann, S. L., O’Toole, C., Westrich, P., Williams, I. H. (Academic, New
York), pp. 125–142.
33. Sugden, E. A., Thorp, R. W. & Buchmann, S. L. (1996) Bee World 77, 26–44.
34. Gross, C. L. & MacKay, D. (1998) Biol. Conserv. 86, 169–178.
35. Sakagami, S. F. (1959) J. Anim. Ecol. 28, 51–68.
36. Roubik, D. W., Moreno, J. E., Vergara, C. & Wittmann, D. (1986) J. Trop. Ecol.
2, 97–111.
37. Paton, D. C. (1993) Bioscience 43, 95–103.
38. Engel, M. S. (1999) J. Hymenoptera Res. 8, 165–196.
39. Michener, C. D. (1990) Univ. Kans. Sci. Bull. 54, 75–164.
40. Lutz, H. (1993) Mainzer Naturwiss. Arch. 31, 177–199.
41. Tattersall, I. (2000) Sci. Am. 282, 56–62.
42. Tattersall, I. & Schwartz, J. H. (2000) Extinct Humans (Westview, New York).
43. Thorne, B. L., Grimaldi, D. &, Krishna, K. (2000) in Termites. Evolution,
Sociality, Symbiosis, and Ecology, eds. Abe, T., Bignell, D. & Higashi, M.
(Kluwer, Dordrecht, The Netherlands), pp. 45–60.
44. Grimaldi, D. & Agosti, D. (2000) Proc. Natl. Acad. Sci. USA 97, 13678–13683.
(First Published November 14, 2000; 10.1073͞pnas.240452097)
1664 ͉ www.pnas.org Engel