William L. Brown

The Evolution of the Dacetine Ants

A preliminary synthesis of evolutionary studies of certain morphological and ethological characters in the ant tribe Dacetini is given. A primary trend inferred from the study of living species is the shift from epigaeic-subarboreal to cryptic-terrestrial foraging. Possibly associated in the early stages of dacetine evolution with this shift was a trend toward oligophagous predation, resulting in extreme cases among modern forms in specialized feeding on collembolans. Secondary changes have occurred in hunting behavior and in a diversity of morphological features; these are described briefly. Reversed or, more accurately, countercurrent evolution has occurred in various phyletic lines in most of the characters of this morphological-ethological coadaptive system. Such changes are relatively short-range and usually involve only a few characters at a time. In at least some instances they have produced a local reversal in the overall dacetine trend to specialization, as with secondarily increased polyphagy in some species of Strumigenys. Convergence to the higher dacetine morphological type has occurred independently in several other ant tribes; in at least one case (Rhopalothrix) ethological convergence is also evident.

The First Mesozoic Ants

Two worker ants preserved in amber of Upper Cretaceous age have been found in New Jersey. They are the first undisputed remains of social insects of Mesozoic age, extending the existence of social life in insects back to approximately 100 million years. They are also the earliest known fossils that can be assigned with certainty to aculeate Hymenoptera. The species, Sphecomyrma freyi, is considered to represent a new subfamily (Sphecomyrminae), more primitive than any previously known ant group. It forms a near-perfect link between certain nonsocial tiphiid wasps and the most primitive myrmecioid ants.

General Adaptation and Evolution

W ITHIN single species of animals 11 7(and plants) many characters under genetic control are known to vary in correlation with geographically varying features of the environment. On a local scale, larks (Niethammer, 1940) or mammals on lava flows (Benson, 1933) may vary in color according to the color of the soil on which they live. In such cases, adjacent populations on contrasting soil types may show relatively sharp differences in color. This variation appears to be adaptive in that it presumably renders the animals less conspicuous to predators. Over broader areas, species show heritable variation corresponding to climate, soil, vegetation, the presence of competitors and predators, and so on. In Rana pipiens Schreber, the North American leopard frog, developmental rates, and perhaps also the proportions of the eggs and arrangement of the egg masses, are correlated with temperatures found during the season of reproduction and growth (Moore, 1949). The so-called Ecological Rules of Bergmann, Allen, Gloger and others (brief summary by Mayr, 1942) cover many cases of broad-scale variation, most of them probably directly adaptive, though needing closer study. Where variation is more or less continuously graded to form a cline, the term ecocline has been used (Huxley, 1943; Auffenberg, 1955) to characterize variations that appear to be adaptively correlated with the environment. Geographical variation can be strongly influenced by the kinds of competitors present in parts of a species range. The honeycreeper Loxops virens (Gmelin) ranges through most of the main Hawaiian Islands, and shows some variation from island to island, but its most divergent population occurs on Kauai, the only island where a second closely related species of Loxops is also found (Amadon, 1950). This phenomenon, called character displacement, has been summarized and exemplified in a separate paper (Brown and Wilson, 1956). It seems obvious that prey species should vary according to the kind and density of predators attacking them, to the kinds of distasteful models available for them to mimic, and to the kind and density of other species that are potential prey for the same predators (L. Brower, 1958). The examples and even the categories of variation correlated with definite, limited aspects of the physical and biotic environment could be expanded virtually without end. It will be appreciated, I think, that the demonstration of these correlations is often no easy matter. Frequently, the apparent environmental correlate of a varying character will really bear only a second-orthird-order relationship to that character, so that puzzling exceptions to the correlation occur. Where characteristics of a species vary in correlation with definite attributes of

The Ant Genus Simopelta (Hymenoptera: Formicidae)

The taxonomic history of Simopelta (subfamily Ponerinae, tribe Ponerini) has been discussed in detail by W. M. Wheeler (I935) and by Borgmeier (I95O). Borgmeier was the first to describe the queen of any species in the genus that of S. peryandei-which he showed deserved to be called dichthadiiform, or belonging to a particular form of queen caste characterized by extreme reduction or loss of eyes, loss of wings, hypertrophy of petiole and gaster, and other characters. He explained its great similarity to certain females of Eciton by convergence in its hypogaeic way o,f life, a statement that is puzzling because, as Father Borgmeier well knows, .Eciton is. not really hypogaeic in its habits, at least as compared to the majority of ants that spend most of their time on or below the ground level. _At any rate, as we shall show in this paper, the. convergence between the queens of at least one Simopelta species and certain army ants, so discerningly noted by Father Borgmeier, is only one aspect of the army-ant or legionary lifeform that two and perhaps all Simopelta species share with the true army ants of subfamily Dorylinae. It is the purpose of this contribution to list and key the workers of the knmvn species of Simopelta, to describe two new species of the genus, and to set forth on the behavior of one species some o.bservations, however fragmentary, that will establish that it follows the army-ant way of life in important respects.

New Species of the Ant Genus Myopias (Hymenoptera: Formicidae:Ponerinae)

The work reported upon here began in the early 1950’s as a revision of genus Myopias, including as a synonym Trapeziopelta. For a year or more it served as the trial focus of RBW’s doctoral thesis research, until his interests shifted into other channels, and he laid the revisionary work aside. Meanwhile, WLB’s interest in the revision continued, but he had no opportunity at that time to do much more than supervise the drafting of a set of illustrations by artist Nancy Bufflermmany of which are now offered heremand to make some of the dissections of mouthparts, etc. As WLB’s work on the reclassification progressed for over 25 years through the tribes of subfamily Ponerinae, much new material was added to what had been available for the original Myopias study, and additional new synonymies and new species were discovered, as well as valuable information on the larvae, males, distribution and bionomics of species new and old. Even the status of Myopias as a genus apart from Pachycondyla came into question. Although in some ways it would be best if the old findings to which we both contributed could simply be incorporated in the reclassification part dealing with tribe Ponerini s. str., there seemed in this course no convenient way to recognize the legitimate claim of RBW to authorship based on the considerable amount of work he had done on Myopias in 1955. The compromise reached sees the larger Myopias review, with keys to species and discussions of synonymy, biology, etc. to be included in Brown’s forthcoming Part VII of Contributions toward a Reclassification of the Formicidae, while descriptions of the new species included in various drafts of our joint manuscript of the

Numerical Taxonomy, Convergence, and Evolutionary Reduction

Let us be clear from the beginning-I am writing to criticize numerical taxonomy, not quantitative systematics. I am sure that measurement and mathematical methods will and should be much more widely employed in systematic studies. It should also be understood that numerical taxonomy, by which I mean specifically the doctrine and methods outlined by Sokal and Sneath (1963) in their book of this title, is to me hot without interest, import, and potential usefulness for systematic biology. But I intend to show that in at least one important respect, numerical taxonomy is headed in the wrong direction. I am not going to attempt to deal with mathematical questions, and I am going to skip over the obvious objections based on the difficulty of properly analyzing characters, so that I can get to the subject that interests me most. While considering this fault in numerical taxonomy, I hope to bring the criticism around to a more constructive position by hinting at an approach to systematics that the numerical taxonomists might find worth trying. The aspect I have chosen to discuss deals with phyletic convergence, a term that I here interpret broadly to include the finergrained phenomenon, parallelism. Michener and Sokal (1957) make the following statement:

A New Pheidole With Reversed Phragmosis (Hymenoptera: Formicidae)

_A_ number of ants have plug-shaped heads, known or assumed to be used to stopper the nest entrance. This phenomenon, called phragmosis, is best understoo.d in certain groups o.f Camponotus and Cephalotini (e.g. Wheeler I9IO: 209-212 Szabd-Patay I928 Creighton and Gregg 954), but Wheeler (I9OI: 534; 1927) also, described species with phragmotic-like heads in Pheidole, Crematogaster and the dacetine genus Colobostr,uma, while Patrizi (I948) published the problematical Solenopsis (Crateropsis) elmenteitae (placed by Ettersha.nk, 1966, in Oligomyrmex). In most of these forms, either the queens or the soldiers, or both, are the phragmotic castes; in Colobostruma leae, the assumed phragrnotics are the queen and the monomorphic workers. In the new species of Pheidole described below, phragmotic behavior at some stage of the life cycle is indicated only or t’he queen caste, and even then is inferred from her aberrant body form. But in this case, the plug is tormed, not by the head, but by the highly mo.dified gaster (Figs. 6 and 7). Two queens showing this plug-like modification of the gaster were taken separately trom rotten wood in rain to.rest in the. general vicinity of Manaus during my collecting trip of I962 in the Brazilian Amazon. In one case, a.nd possibly in both, the queens belonged to definite colonies with workers, soldiers and brood. (Notes on collection M-6o are ambiguous because two Pheidole queens, one of embolopx and one o.f a to.tally different species, were in the vial with this number, but the notes state that the queen was taken apart from the soldiers, workers and brood. Probably one of the two queens was taken up in the aspirator along with bits of rotten wood without my knowing it.) The queen’s thickened scape base (with a gelatinous sheath) and the largely smooth and shining alitrunk with overhanging scutal margins, are protective characters suggesting social parasitism as a way of nest-founding for this caste. The strong similarities in color, sculpture and pilosity between the queens and the accompanying

A Supplement to the World Revision of Odontomachus (Hymenoptera: Formicidae)

In a recent publication (Brown, 1976, Studia Entomologica 19: 67-171) I reviewed the world fauna of the ponerine ant genus Odontomachus. While that contribution was in press, and since its appearance, some significant new Odontomachus material, and information about material, has come to my notice. It seems appropriate to supplement the revision now, while it is fresh, by offering the new information for incorporation. The first and most important addition is the description, of a new (twentieth) species in the haematodus group.

The Neotropical Species of the Ant Genus Strumigenys Fr. Smith: Group of Gundlachi (Roger)

This paper is a continuation of my series on the New World fauna of the dacetine ant genus Strumigenys Fr. Smith. Earlier parts, containing keys to the abbreviations for measurements and proportions, may be found in Jour. New York Ent. Soc. 61: 53-59, 1o1-11o (1953). In addition to these, other parts have been published or are being prepared. At the end of the series, an illustrated key to the New- World members of the genus will be forthcoming.

The Neotropical Species of the Ant Genus Strumigenys Fr. Smith: Miscellaneous Concluding Studies

very close, and seem, from the limited material available, to replace one another in a chain extending from Mexico to Panama, and perhaps beyond. So far as I can see now, the differences, are complex enough and strong enough to indicate that each form is a distinct species; perhaps together [they constitute] one superspecies. However, it is not beyond possibility that one or more o these forms intergrades with a neighbor. More material is needed. Since that writing, material has turned up which, though small in amount, tends to bridge the gap between S. micretes Brown and S. ,.acacoca Brown, indicating perhaps that they belong to. a single variable species. Nevertheless, the new material poses certain problems itself, and the discussion next offered is intended to give details that should help in eventually straightening this complex out. A sample consisting of parts of four nest series from Boquete, Chiriqui Province, Panama (F. Y[. Gaige leg., see below) contains 25 workers with highly variable preapical mandibular dentition, the denticles varying in number from to 4 in the two mandibles taken together, and also varying markedly in size, acuteness and position, so as to bridge virtually completely the chief diagnostic character-gap between micretes (each mandible with a small but acute preapical tooth, and a little arther up a minute denticle) and lacacoca (man-