Resolving taxonomic turbulence and uncovering cryptic diversity in the musk turtles (Sternotherus) using robust demographic modeling.

Accurate and consistent delimitation of species and their relationships provides a necessary framework for comparative studies, understanding evolutionary relationships, and informing conservation management. Despite the ever-increasing availability of genomic data, evolutionary dynamics can still render some relationships exceedingly difficult to resolve, including underlying speciation events that are rapid, recent, or confounded by post-speciation introgression. Here we present an empirical study of musk turtles (Sternotherus), which illustrates approaches to resolve difficult nodes in the Tree of Life that robust species-tree methods fail to resolve. We sequence 4430 RAD-loci from 205 individuals. Independent coalescent-based analyses, corroborated with morphology and geography, strongly support the recognition of cryptic species within Sternotherus, but with conflicting or weak support for some intraspecific relationships. To resolve species-tree conflict, we use a likelihood-based approach to test support for alternative demographic models behind alternative speciation scenarios and argue that demographic model testing has an important role for resolving systematic relationships in recent, rapid radiations. Species-tree and demographic modeling strongly support the elevation of two nominal subspecies in Sternotherus to species and the recognition of a previously cryptic species (S. intermedius sp. nov.) described within. The evolutionary and taxonomic history of Sternotherus is discussed in the context of these new species and novel and well-supported systematic hypotheses.