The origin and evolution of Cynodontia (Synapsida, Therapsida): Reassessment of the phylogeny and systematics of the earliest members of this clade using 3D‐imaging technologies

Abstract

The origin of cynodonts, the group ancestral to and including mammals, is one of the major outstanding problems in therapsid evolution. One of the most troubling aspects of the cynodont fossil record is the lengthy Permian ghost lineage between the latest possible divergence from its sister group Therocephalia and the first appearance of definitive cynodonts in the late Permian. The absence of cynodonts and dominance of therocephalians in middle Permian strata has led some workers to argue that cynodonts evolved from within therocephalians, rendering the latter paraphyletic, but more recent analyses support the reciprocal monophyly of Cynodontia and Therocephalia. Furthermore, although a fundamental dichotomy in the derived subclade Eucynodontia is well‐supported in cynodont phylogeny, the relationships of more stemward cynodonts from the late Permian and Early Triassic are unresolved. Here, we provide a re‐evaluation of the phylogeny of Eutheriodontia (Cynodontia + Therocephalia) and an assessment of character evolution within the group. Using computed tomographic data derived from extensive sampling of the earliest known (late Permian and Early Triassic) cynodonts and selected exemplars of therocephalians and later (Middle Triassic onwards) cynodonts, we describe novel aspects of the endocranial anatomy of these animals. These data were incorporated into a new phylogenetic data set including a comprehensive sample of early cynodonts. Our phylogenetic analyses support some results previously recovered by other authors, but recover therocephalians as paraphyletic with regards to cynodonts, with cynodonts and eutherocephalians forming a clade to the exclusion of the “basal therocephalian” families Lycosuchidae and Scylacosauridae. Though both conservatism and homoplasy mark the endocranial anatomy of early non‐mammalian cynodonts, we were able to identify several new endocranial synapomorphies for eutheriodont subclades and recovered generally better‐supported topologies than previous analyses using primarily external craniodental characters.