Saux, C., B. L. Fisher & G. S. Spicer. 2004. Dracula ant phylogeny as inferred by nuclear 28S rDNA sequences and implications for ant systematics (Hymenoptera: Formicidae: Amblyoponinae).
Molecular Phylogenetics and Evolution 33: 457-468.
Ants are one of the most ecologically and numerically dominant families of organisms in almost every terrestrial habitat throughout the world, though they include only about 1% of all described insect species. The development of eusociality is thought to have been a driving force in the striking diversification and dominance of this group, yet we know little about the evolution of the major lineages of ants and have been unable to clearly determine their primitive characteristics. Ants within the subfamily Amblyoponinae are specialized arthropod predators, possess many anatomically and behaviorally primitive characters and have been proposed as a possible basal lineage within the ants. We investigate the phylogenetic relationships among the members of the subfamily, using nuclear 28S rDNA sequence data. Outgroups for the analysis include members of the poneromorph and leptanillomorph (Apomyrma, Leptanilla) ant subfamilies, as well as three wasp families. Parsimony, maximum likelihood, and Bayesian analyses provide strong support for the monophyly of a clade containing the two genera Apomyrma + Mystrium (100% bpp; 97% ML bs; and 97% MP bs), and moderate support for the monophyly of the Amblyoponinae as long as Apomyrma (Apomyrminae) is included (87% bpp; 57% ML bs; and 76% MP bs). Analyses did not recover evidence of monophyly of the Amblyopone genus,
while the monophyly of the other genera in the subfamily is supported. Based on
these results we provide a morphological diagnosis of the Amblyoponinae that includes Apomyrma. Among the outgroup taxa, Typhlomyrmex grouped consistently with Ectatomma, supporting the recent placement of Typhlomyrmex in the Ectatomminae. The results of this present study place the included ant subfamilies into roughly two clades with the basal placement of Leptanilla unclear. One clade contains all the Amblyoponinae (including Apomyrma), Ponerinae, and Proceratiinae (Poneroid clade). The other clade contains members from subfamilies Cerapachyinae, Dolichoderinae, Ectatomminae, Formicinae, Myrmeciinae, and Myrmicinae (Formicoid clade).
Saux et al's paper is real food for myrmecological thought. I'm not speaking of the Amblyoponines, even though they are the ostensible subject of the study. Rather, the taxon sampling of ants outside the Amblyoponines is complete enough to give us a tantalizing glimpse of the larger formicid history: three major lineages of ants, and a suggestion (albeit one without much statistical support) that the highly specialized subterranean Leptanilla
could be sister to the rest of all ants. This is not a possibility that has received much consideration, but the implications for reconstruction of the common ant ancestor are considerable. Of course, usual caveats about single-gene phylogenies should apply...