Phylogeny of Formicidae and position of Ponerinae

Bayesian analysis of the combined data set (ALL_DATA) yielded a phylogeny that is broadly congruent with the results of Moreau et al. (2006) and Brady et al. (2006) for subfamily-level relationships within Formicidae (Fig. 2), with strong support for a sister relationship between Leptanillinae and monophyletic poneroid and formicoid groups. The phylogeny differs from these earlier studies primarily in its inclusion of Martialinae, which is inferred as sister to the rest of Formicidae as in Rabeling et al. (2008).

A thorough investigation of the sister group of Ponerinae is outside the scope of this study, but Ponerinae was inferred to be sister to a clade consisting of Paraponera (Paraponerinae) + Tatuidris (Agroecomyrmecinae), with strong support in the MrBayes analysis (BPP = 1.00 for entire clade, BPP = 0.99 for Paraponera + Tatuidris) but much lower support in the RAxML bootstrap analysis (BS = 0.42 for the entire clade, BS = 0.43 for Paraponera + Tatuidris). This relationship was inferred by Moreau et al. (2006), and also by Brady et al. (2006) when their analyses included non-ant outgroups. In contrast, the analysis of Rabeling et al. (2008) reconstructed Ponerinae as sister to the formicoid subfamilies, as did the Formicidae-only analysis of Brady et al. (2006) under certain rootings. Given that these latter studies were unable to reject several alternate rootings of the phylogeny of Formicidae, the sister group of Ponerinae remains unresolved pending additional data. Future sampling of additional non-ant vespoid outgroups and leptanillines, and inclusion of additional genes, may provide better resolution of the basal relationships in Formicidae, including a firmer elucidation of the sister group of Ponerinae .