A peninsula as an island: multiple forms of evidence for overwater colonization of Baja California by the gartersnake Thamnophis validus

The recent shift toward dispersal rather than vicariant explanations of disjunct distributions has been driven by the use of molecular data to estimate divergence dates between lineages. However, other kinds of evidence can also be critical in evaluating such biogeographic hypotheses. In the present study, we used a multifaceted approach employing diverse analyses of mitochondrial DNA sequences to assess explanations for the disjunct distribution of the gartersnake Thamnophis validus. The occurrence of this species in the Cape Region of the Baja California peninsula, isolated from mainland populations that occur along the west coast of Mexico, might be explained by: (1) separation of the peninsula from mainland Mexico through rifting 4–8 Mya (tectonic vicariance); (2) fragmentation of the range of this semi‐aquatic species because of post‐Pleistocene aridification (vicariance by aridification); (3) natural overwater dispersal across the Gulf of California; or (4) human introduction. Divergence dating indicates that peninsular and mainland T. validus separated from each other within the last 0.5 Myr, thus rejecting tectonic vicariance. In addition, the estimated closest mainland relatives of peninsular snakes are found farther north than expected under this hypothesis. Three findings argue against vicariance by aridification: (1) peninsular snakes and their closest mainland relatives are more genetically similar than predicted; (2) the location of closest mainland relatives is farther south than predicted; and (3) the species is absent from areas where one might expect to find relict populations. Taken together, refutation of the vicariance hypotheses and the fact that the estimated closest mainland relatives are found almost directly across the Gulf from the Cape Region supports some form of overwater colonization. Various additional arguments suggest that natural dispersal is more likely than human introduction. The present study emphasizes the need for multiple kinds of evidence, beyond divergence dates, to discriminate among hypotheses and to provide independent sources of corroboration or refutation in historical biogeography. © 2008 The Linnean Society of London, Biological Journal of the Linnean Society, 2008, 95 , 409–424.