Digging deep: a revised phylogeny of Australian burrowing cockroaches (Blaberidae: Panesthiinae, Geoscapheinae) confirms extensive nonmonophyly and provides insights into biogeography and evolution of burrowing

ABSTRACT Soil‐burrowing cockroaches (Blaberidae: Geoscapheinae) are large insects endemic to Australia. Originally thought to represent a monophyletic group, these enigmatic species have in fact evolved burrowing behaviour, associated fossorial morphological modifications, and dietary transitions to dry leaf litter feeding multiple times from the wood‐feeding Panesthiinae in a striking example of parallel evolution. However, various relationships within these two subfamilies remain unresolved or poorly understood, notably the apparent paraphyly of Panesthiinae with respect to Geoscapheinae, the position and diversification of certain species within major clades, and several aspects of the overall group's biogeography and morphological evolution. Here, we investigate the phylogeny of Australian members of these two subfamilies using whole mitochondrial genomes paired with nuclear ribosomal markers and highly conserved genes from the bacterial endosymbiont Blattabacterium . Using the resulting robust, fossil‐calibrated phylogeny from these three sources we confirm the nonmonophyly of both subfamilies and recover Geoscapheinae as polyphyletic within a paraphyletic Panesthiinae. The nonmonophyly of natural groups, at all levels from subfamily to species, has been driven by repeated, independent acquisitions of burrowing forms in Geoscapheinae from panesthiine ancestors that colonized the continent on two separate occasions during the Miocene. We additionally find morphological variation within Geoscapheinae itself is correlated with species distributions. Older soil‐burrowing clades living in comparatively arid environments have additional morphological reductions beyond obvious fossorial adaptations compared to those in younger burrowing clades from more temperate habitats. Ultimately, the results presented here demonstrate connections among phylogeny, biogeography and morphology throughout Australian representatives of these two subfamilies, factors that could not be previously consolidated using existing phylogenetic frameworks. Given the discordance between molecular data implemented here and the existing taxonomic classification, we find no support for retaining Geoscapheinae as a discrete taxonomic grouping. In closing, we discuss the taxonomic implications of these results and present a roadmap for future research on Geoscapheinae and their panesthiine relatives.