Commentary

In search of ancestors

Ted R. Schultz*

Department of Entomology, MRC 188, National Museum of Natural History, Smithsonian Institution, Washington, DC 20560-0188

nts are arguably the greatest success to a symbiotic fungus, which they in turn Grimaldi and Agosti (ref. 1, table 1). In the Astory in the history of terrestrial meta- eat). Interactions with have shaped the Eocene as in the present, ants comprise a zoa. On average, ants monopolize 15–20% evolution of diverse organisms to an aston- significant proportion of the biota, of the terrestrial biomass, and in ishing degree. Ants participate in symbio- and ant species are easily assigned to extant tropical regions where ants are especially ses—some facultative, some obligate—with subfamilies and, in most cases, to extant abundant, they monopolize 25% or more. over 465 plant species in over 52 families (2), genera as well. Whatever the story of the But ants did not always run the world. They with thousands of species (3, 4), ants’ rise to dominance, it obviously un- do not appear in the fossil record until the and with as-yet-unknown numbers of fungi folded before the deposition of these Ter- mid-, and for more than the first and microorganisms (5, 6). Clearly, the tiary fossil remains. half of their history—a period spanning 60 study of most ecosystems must include the In the absence of Cretaceous ant fossils, to 80 million years—ants occupied a rela- study of the resident ant species. Because of the quest for ant origins relied until recently tively modest position in the terrestrial bio- their complex colony-level behaviors, ants on comparative studies of the morphology, sphere. To understand the factors, both serve as model organisms for the highly behavior, and ecology of the likeliest ‘‘most ecological and historical, that contributed to visible disciplines of behavioral ecology and primitive’’ extant species. Between 1951 and the rise of the ants, we require a clearer sociobiology, particularly in studies focused 1977, this quest drew a series of myrmecolo- picture of the stepwise evolution of the on the dynamics of kin selection, within- gists to Western Australia’s outback in major ant lineages. Now, Grimaldi and Ag- colony conflicts of interest, caste differen- search of the ant-origin ‘‘holy grail,’’ the osti (1) report in a recent issue of PNAS the tiation, and division of labor. species Nothomyrmecia macrops. Known remarkable discovery of a worker ant, pre- Given the strong interest in ants from a from two imperfectly labeled worker spec- served in amber for over 90 million years, range of biological disciplines, it is especially imens collected by members of a frontier that is clearly assignable to a modern ant important that we understand the unique excursion in 1931, N. macrops was subse- subfamily that contains many familiar extant evolutionary history of this group. Fortu- quently declared the world’s most primitive species, including carpenter ants. Combined nately, during the last decade, a solid foun- ant (19, 20). After years of failed efforts, with other paleontological and phylogenetic dation has been established for ant system- living colonies were finally located in 1977 information, this unexpected fossil strongly atics and, indeed, for ant biology in general (21), and N. macrops quickly became one of indicates that the diversification of many ant by Ho¨lldobler and Wilson (4) (exhaustive the most studied ants on the planet. subfamilies occurred earlier and more rap- biological overview), Bolton (7, 8) (taxo- Comparisons of Nothomyrmecia with ple- idly than previously suspected. nomic catalog and genus-level keys), Ward siomorphic (‘‘primitive’’) members of other Ants represent the family Formicidae in et al. (9) (exhaustive bibliographic data- ant lineages provided important hypotheses the insect order and, like the base), and Agosti et al. (10) (manual for about the morphology of the earliest ants yellow jackets, hornets, and paper to using ants as bioindicators). During this and, in 1966, ant systematists were able to which they are closely related, ants are sting- same time period, partial progress toward test those hypotheses directly. In that year, ing wasps. All ants are eusocial, that is, they reconstructing the subfamily-level phylog- live in colonies in which a wingless neuter eny of ants has been achieved in studies two rock hounds, Mr. and Mrs. Edmund daughter caste cooperates to raise subse- using extant species (11–13), but these stud- Frey, collected a relatively large piece of quent generations of their mother queen’s ies have been impeded by the lack of criti- amber from a clay embankment in Cliff- offspring. Like all of its descendants, the cally informative paleontological data. wood Beach, New Jersey. Although origi- ancestral ant was almost certainly eusocial, Ant paleontology has evolved as a sub- nating from a site that is, in modern times, with colonies made up of small bands of discipline of ant systematics ever since the less exotic than that of Dominican or even hunter–gatherers living in simple tempo- latter’s origins in the mid-19th century. Baltic amber, the New Jersey amber is far rary nests in the soil. From this modest Among the first ant fossils to receive serious older, most likely originating in the mid- beginning arose the current diversity of the attention were those found in Baltic amber Cretaceous Turonian stage (90–94 mya). family Formicidae, numbering over 9,500 (14, 15), which we now know originated in Peering backward in time through a window described species and an estimated 3,000 to the early Oligocene [Ϸ30 million years ago polished in the Frey’s amber nugget, Wilson 9,000 additional species as yet unknown to (mya)]. Except for indicating that ant dis- et al. (22, 23) became the first myrmecolo- science. Today ants occupy keystone posi- tributions have changed dramatically, the gists to examine an ant from the Cretaceous, tions in most terrestrial environments, serv- Baltic amber ant fossils are strikingly nearly three times as old as any previously ing as major conduits of energy and organic modern—so much so, in fact, that at least known ant fossil. Subsequently named Sphe- material. They are, for example, important one author was inspired to suggest that they turners of the soil, matching or exceeding might be fakes (16). Most other significant See companion paper on page 13678 in issue 25 of volume the activity of earthworms in this role. They sources of ants reveal a similarly 97. are among the leading predators of inver- modern ant fauna, including the North *E-mail: [email protected]. tebrates in most ecosystems, and in the American Florrisant (Colorado) and Green Ϸ Article published online before print: Proc. Natl. Acad. Sci. Neotropics they are the leading herbivores River (Wyoming) shales (Oligocene, 30 USA, 10.1073͞pnas.011513798. Article and publica- as well, with leaf-cutter ants taking more mya) (17), the Dominican amber (Miocene, tion date are at www.pnas.org͞cgi͞doi͞10.1073͞ than 15% of the fresh vegetation (feeding it Ϸ20 mya) (e.g., ref. 18), and others listed by pnas.011513798

14028–14029 ͉ PNAS ͉ December 19, 2000 ͉ vol. 97 ͉ no. 26 Downloaded by guest on October 2, 2021 comyrma freyi and representing a new sub- frame bounded by the fossils a diversification of major ant body plans family, the , the New Jer- on the one hand and by the Eocene ant during the first 30 million years of ant evo- sey specimen was found to possess many fossils on the other. By demonstrating that lution; the rarity of ant fossils of this age (but not all) of the characters that had been the subfamily was contempora- relative to those of other implies, predicted for a truly primeval ant. Unfortu- neous with the Sphecomyrminae, however, however, that this diversification of form nately, some doubt remained about whether Grimaldi and Agosti (1) render this view was not accompanied by an increase in S. freyi was a true ant because of uncertainty untenable. In fact, because of Grimaldi et al. ecological dominance (ref. 1, table 1). In about the presence of the antibiotic- (13), we know that the extant subfamily contrast, by the mid-Eocene (roughly 50 producing metapleural gland, a defining was also contemporaneous; more mya), ants had achieved their current levels feature of ants. doubtfully, the (28) and the of abundance, with nearly all extant subfam- Despite this discovery, the New Jersey (25) may have been present. ilies and many extant genera in place. This increase in abundance suggests an explosive amber remained understudied for nearly 20 This unexpected information allows the as- radiation of ants during the second third of years until in 1986 D. Grimaldi took up the signment of minimum dates to several of the their history. task, aided by a number of fossil collectors most basal nodes of the cladogram for the What ecological and historical factors ac- and American Museum of Natural History family Formicidae (ref. 1, figure 2) and thus count for these two broad phases of ant volunteers. As a result, Grimaldi et al. (13) provides a calibrated standard for future ant diversification, the first resulting in the prin- described additional specimens of Spheco- systematic research. Combined with addi- cipal morphological differences between myrma, including a winged male, and ban- tional fossil data and phylogenetic informa- major lineages, and the second resulting in ished lingering doubts about the formicid tion for the Hymenoptera, this information ecological dominance? This question will be affiliations of S. freyi by confirming the also permits the plausible dating of the answered only when reliable paleoecologi- presence of a metapleural gland. Four gen- origin of ants. Grimaldi and Agosti (1), cal data are combined with a well-supported era and eight species of the Sphecomyrmi- Agosti et al. (29), and Grimaldi et al. (13) phylogeny. Ideally, this phylogeny must in- nae are now known from New Jersey, Sibe- argue convincingly that the ancestral ant corporate character data from a variety of rian, and Canadian Cretaceous ambers, diverged from aculeate during the sources, including DNA sequences of extant including Cretomyrma (two species), Cretaceous, no earlier than 140 mya, and ant species and morphological characters of Dlusskyidris (one species), Baikuris (three more likely between 110 and 130 mya, for both extant and fossil forms. Only fossil species based on males), and Sphecomyrma the following reasons: (i) Ant-like fossils information, however, permits the assign- (two species). From this fossil record, we (including the ) originating be- ment of minimum dates to the internal may conclude that ‘‘the most primitive fore 115 mya are entirely unknown; (ii) the nodes of the phylogeny—even molecular known group of ants, the Sphecomyrminae, oldest known hymenopteran fossils date dating techniques must be calibrated by lived over much of the Northern hemisphere from the and are represented solely paleontological data. Indeed, since the dis- during middle and late Cretaceous times’’ by the most plesiomorphic hymenopteran covery of S. freyi, fossils have played an (24). A possibly earlier ant, Cariridris, ten- family, the Xyelidae; (iii) the earliest ac- increasingly influential role in ant phyloge- tatively assigned to the subfamily Myrmeci- uleate fossils, representing the most primi- netics. In the last 6 years, the standard- reference cladogram proposed in 1994 by inae (represented today by the Australian tive stinging wasps, appear in the late Juras-

Baroni Urbani et al. (11) has been revised COMMENTARY Nothomyrmecia sic; and (iv) there is strong fossil evidence for bulldog ants and ), has been twice because of new fossil data (1, 13) and described from the Brazilian Aptian (110 a Cretaceous origin for the , which, once because of the discovery of an enig- mya) (13, 25). Earlier fossils in the family together with the and the Bethyli- matic extant species (12). We can only hope Armaniidae may or may not be ants (13, 26, dae, is the most likely sister group to the that this cycle of discovery and revision will 27); they are treated as a linking form ants. The only dissenting opinion, that of continue or even accelerate, so that our between ants and other stinging wasps by Crozier et al. (30) for a ant origin current clouded picture of ant phylogeny Grimaldi and Agosti (ref. 1, figure 2). based on a molecular clock, has been ques- will come increasingly into focus—in much After the discovery of Sphecomyrma,it tioned by Grimaldi et al. (13) and Rust and the same way, perhaps, that the initially was generally believed that the major lin- Andersen (31). vague but evocative form of a Cretaceous eages of ants (i.e., the subfamilies) must As pointed out earlier, Grimaldi and Ag- ant becomes progressively clearer when a have arisen gradually during the interval osti’s (1) conclusion that at least some ant nugget of amber is painstakingly prepared between 90 and 50 mya, i.e., during the time subfamilies diverged before 90 mya suggests and polished by D. Grimaldi and colleagues.

1. Grimaldi, D. & Agosti, D. (2001) Proc. Natl. Acad. 9. Ward, P. S., Bolton, B., Shattuck, S. O. & Brown, 18. Wilson, E. O. (1985) Science 229, 265–267. Sci. USA 97, 13678–13683. W. L., Jr. (1996) A Bibliography of Ant Systematics, 19. Brown, W. L., Jr. (1954) Insectes Sociaux 1, 2. Jolivet, P. (1996) Ants and Plants: An Example of University of California Publications in Entomol- 21–31. Coevolution (Backhuys, Leiden, The Netherlands). ogy (Univ. of California, Berkeley, CA), Vol. 116, 20. Brown, W. L., Jr. & Wilson, E. O. (1959) West. 3. Kistner, D. H. (1982) in Social Insects, ed. Hermann, pp. 1–417. Aust. Nat. 7, 25–30. H. R. (Academic, New York), Vol. 3, pp. 1–244. 10. Agosti, D., Majer, J. D., Alonso, L. E. & Schultz, 21. Taylor, R. W. (1978) Science 201, 979–985. 4. Ho¨lldobler, B. & Wilson, E. O. (1990) The Ants T. R., eds. (2000) Ants: Measuring and Monitoring 22. Wilson, E. O., Carpenter, F. M. & Brown, W. L., (Belknap, Cambridge, MA). Biological Diversity (Smithsonian Institution, Jr. (1967) Science 157, 1038–1040. 5. Schultz, T. R. & McGlynn, T. P. (2000) in Ants: Washington, DC). 23. Wilson, E. O., Carpenter, F. M. & Brown, W. L., Measuring and Monitoring Biological Diversity, eds. 11. Baroni Urbani, C., Bolton, B. & Ward, P. S. (1992) Jr. (1967) Psyche 74, 1–19. Agosti, D., Majer, J. D., Alonso, L. E. & Schultz, Syst. Entomol. 17, 301–329. 24. Wilson, E. O. (1985) Psyche 92, 205–216. T. R. (Smithsonian Institution, Washington, DC), 12. Ward, P. S. (1994) Syst. Entomol. 19, 159–175. 25. Branda˜o,C. R. F., Martins-Neto, R. G. & Vul- pp. 35–44. 13. Grimaldi, D., Agosti, D. & Carpenter, J. M. (1997) cano, M. A. (1990) Psyche 96, 195–208. 6. Mueller, U. G., Schultz, T. R., Currie, C. R., Am. Mus. Novit. 3208, 1–43. 26. Dlussky, G. M. (1983) Paleont. Zh. 3, 65–78. Adams, R. M. M. & Malloch, D. Quart. Rev. Biol., 14. Mayr, G. (1868) Beitr. Naturkunde Preussens 27. Wilson, E. O. (1987) Paleobiology 13, 44–53. in press. Physikal. Oekonom. Gesell. Koenigsberg 1, 1–102. 28. Dlussky, G. M. (1999) Paleont. Zh. 33, 409–412. 7. Bolton, B. (1994) Identification Guide to the Ant 15. Wheeler, W. M. (1914) Schr. Physikal. Oekonom. 29. Agosti, D., Grimaldi, D. & Carpenter, J. M. (1997) Genera of the World (Harvard Univ. Press, Cam- Gesell. Koenigsberg Preussen 55, 1–142. Nature (London) 391, 447. bridge, MA). 16. Crowson, R. A. (1965) Proc. 12th Int. Congr. 30. Crozier, R. L., Jermiin, L. S. & Chiotis, M. (1997) 8. Bolton, B. (1995) A New General Catalogue of the Entomol. (London), 133. Naturwissenschaften 84, 22–23. Ants of the World (Harvard Univ. Press, Cam- 17. Carpenter, F. M. (1930) Bull. Mus. Comp. Zool. 31. Rust, J. & Anderson, N M. (1999) Zool. J. Linn. bridge, MA). 70, 1–66. Soc. 125, 331–348.

Schultz PNAS ͉ December 19, 2000 ͉ vol. 97 ͉ no. 26 ͉ 14029 Downloaded by guest on October 2, 2021