Evolution of rattlesnakes (Viperidae; Crotalus) in the warm deserts of western North America shaped by Neogene vicariance and Quaternary climate change.

During Pleistocene, the Laurentide ice sheet rearranged and diversified biotic distributions in eastern North America, yet had minimal physical impact in western North America where lineage diversification is instead hypothesized to result from climatic changes. If Pleistocene climatic fluctuations impacted desert species, the latter would reflect patterns of restricted gene flow concomitant with indications of demographic bottlenecks. Accordingly, molecular evidence for refugia should be present within these distributions and for subsequent range expansions as conditions improved. We sought answers to these questions by evaluating mitochondrial DNA (mtDNA) sequences from four species of rattlesnakes [Crotalus mitchellii (speckled rattlesnake), Crotalus cerastes (sidewinder), Crotalus tigris (tiger rattlesnake), Crotalus ruber (red diamond rattlesnake)] with distributions restricted to desert regions of southwestern North America. We inferred relationships using parsimony and maximum likelihood, tested intraspecific clades for population expansions, applied an isolation-with-migration model to determine bi-directional migration rates (m) among regions, and inferred divergence times for species and clades by applying a semiparametric penalized likelihood approach to our molecular data. Evidence for significant range expansion was present in two of eight regions in two species (Crotalus mitchellii pyrrhus, C. tigris region north). Two species (C. cerastes, C. mitchellii) showed a distribution concomitant with northward displacement of Baja California from mainland México, followed by vicariant separation into subclades. Effects of Pleistocene climate fluctuations were found in the distributions of all four species. Three regional diversification patterns were identified: (i) shallow genetic diversity that resulted from Pleistocene climatic events (C. tigris, C. ruber); (ii) deep Pleistocene divisions indicating allopatric segregation of subclades within refugia (C. mitchellii, C. cerastes); and (iii) lineage diversifications that extended to Pliocene or Late Miocene (C. mitchellii, C. cerastes). Clade-diversifying and clade-constraining effects impacted the four species of rattlesnakes unequally. We found relatively high levels of molecular diversification in the two most broadly distributed species (C. mitchellii, C. cerastes), and lower levels of genetic diversification in the two species (C. tigris, C. ruber) whose ranges are relatively more restricted. Furthermore, in several cases, the distributions of subspecies were not congruent with our molecular information. We suggest regional conservation perspectives for southwestern deserts cannot rely upon subspecies as biodiversity surrogates, but must instead employ a molecular and deep historical perspective as a primary mechanism to frame biodiversity reserves within this region.