Despite more than half a century of research, the evolutionary origin of termites remains unresolved [1] [2] [3]. A clear picture of termite ancestry is crucial for understanding how these insects evolved eusociality, particularly because they lack the haplodiploid genetic system associated with eusocial evolution in bees, ants, wasps and thrips [4] [5]. Termites, together with cockroaches and praying mantids, constitute the order Dictyoptera, which has been the focus of numerous conflicting phylogenetic studies in recent decades [6] [7] [8] [9] [10] [11] [12]. With the aim of settling the debate over the sister-group of termites, we have determined the sequences of genes encoding 18S ribosomal RNA, mitochondrial cytochrome oxidase subunit II (COII) and endogenous endo-beta-1, 4-glucanase (EG) from a diverse range of dictyopterans. Maximum parsimony and likelihood analyses of these sequences revealed strong support for a clade consisting of termites and subsocial, wood-feeding cockroaches of the genus Cryptocercus. This clade is nested within a larger cockroach clade, implicating wood-feeding cockroaches as an evolutionary intermediate between primitive non-social taxa and eusocial termites.

Parallel evolution is the independent appearance of similar derived phenotypes from similar ancestral forms. It is of key importance in the debate over whether evolution is stochastic and unpredictable, or subject to constraints that limit available phenotypic options. Nevertheless, its occurrence has rarely been demonstrated above the species level. Climate change on the Australian landmass over the last approximately 20 Myr has provided conditions conducive to parallel evolution, as taxa at the edges of shrinking mesic habitats adapted to drier biomes. Here, we investigate the phylogeny and evolution of Australian soil-burrowing and wood-feeding blaberid cockroaches. Soil burrowers (subfamily Geoscapheinae) are found in relatively dry sclerophyllous and scrubland habits, whereas wood feeders (subfamily Panesthiinae) are found in rainforest and wet sclerophyll. We sequenced and analysed mitochondrial and nuclear markers from 142 specimens, and estimated the evolutionary time scale of the two subfamilies. We found evidence for the parallel evolution of soil-burrowing taxa from wood-feeding ancestors on up to nine occasions. These transitions appear to have been driven by periods of aridification during the Miocene and Pliocene across eastern Australia. Our results provide an illuminating example of climate-driven parallel evolution among species.

Morphologically similar cockroaches in the subfamilies Panesthiinae and Geoscapheinae (Blattaria: Blaberidae) display contrasting feeding habits, behaviour and biogeographical distributions. Panesthiinae, found throughout Asia and Australia, all live in and feed on decaying wood that they burrow into. Geoscapheinae are restricted to Australia and construct and live in burrows in the soil, where they feed on dry leaves taken from the surface. A lack of knowledge about phylogenetic relationships among these cockroaches hinders an understanding of the factors that have shaped the evolution of their diverse lifestyles and biogeography. To address this issue, we sequenced three genes from representatives of nine of the 10 genera in the two subfamilies, and performed phylogenetic analyses. The well-supported topology revealed the Panesthiinae to be paraphyletic with respect to the Geoscapheinae. Soil-burrowing cockroaches appear to have evolved from a lineage of wood burrowers that invaded Australia from the north some time after the merging of the Asian and Australian tectonic plates ca. 20 Myr ago. The main factor promoting the evolution of soil burrowing is likely to have been one of the periods of strong aridity that Australia has experienced since the Miocene period.

The genus Thaumastocoris is revised. Nine new species are described (T. busso, T. freomooreae, T. kalaako, T. majeri, T. nadeli, T. ohallorani, T. roy, T. safordi, and T. slateri) and the five previously described species are redescribed. A diagnostic key to species is provided, supported with illustrations of key character systems and maps depicting their distributional range. Host plants are tabulated, and biology and host plant associations are discussed.

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.