Agroecomyrmecinae is a subfamily of ants containing two extant and two fossil genera.[1] The subfamily was originally classified in 1930 by Carpenter as Agroecomyrmecini, a Myrmicinae tribe.[2] Bolton raised the tribe to subfamily status in 2003, suggesting that Agroecomyrmecinae might be the sister taxon to Myrmicinae. It has since been discovered to be one of the earliest lineages of ants, a clade from the basal polytomy for all ants.[3][4] In 2014, the subfamily was expanded to two tribes. The tribe Ankylomyrmini was moved from the subfamily Myrmicinae to Agroemyrmecinae.[5]
Since the original description, the systematic status of the Agroecomyrmecini tribe has been the focus of intense debate. Bolton (2003) was the first to suggest the taxonomic instability of Tatuidris within Myrmicinae and raised the genus to the level of a new subfamily, the Agroecomyrmecinae, suggesting the Agroecomyrmecinae might be the sister taxon to Myrmicinae. This assessment was based on these diagnostic characters:[7][8]
The subfamily rank of the armadillo ants was reassessed by Baroni Urbani & de Andrade (2007) in their last systematic assessment of the dacetines. They analyzed a morphological dataset that included former dacetines, basicerotines, phalacromyrmecines, and Tatuidris, as well as other non-Myrmicinae taxa such as the Australian genus Myrmecia and the Neotropical genus Pseudomyrmex. This work was the first attempt to include Tatuidris as a terminal taxon in a morphological cladistic analysis. In their study, Baroni Urbani & de Andrade (2007) identified six morphological synapomorphies shared between Tatuidris and the dacetines, justifying the inclusion of the genus within Myrmicinae. These characters included:[9][10]
In addition, two autapomorphies (a differently shaped petiolar tergum and sternum, and the eyes at or close to the apex of the antennal scrobe) separated Tatuidris from all other extant ant genera included in their study.[9][11]
Unlike phylogenetic studies based on morphological traits, molecular analyses of the internal phylogeny of the ants have given strong evidence that the armadillo ants are neither closely related to nor nested within the Myrmicinae. Brady et al. (2006), Moreau et al. (2006) and Rabeling et al. (2008) reconstructed phylogenetic trees with the agroecomyrmecines inside the 'poneroid' group of subfamilies, close to the Paraponerinae, and gave support for the exclusion of the genus from the Myrmicinae, a subfamily located inside the 'formicoid' clade.[12] Given the early appearance of the Agroecomyrmecinae in the geologic record, the similarities of armadillo ants to Myrmicinae were hypothesized to represent convergence and/or retention of plesiomorphic forms.[13][14]
Recently, Keller (2011) challenged the phylogenetic relationships of the poneromorph subfamilies (including Tatuidris).[14][15]