Lack of detectable genetic differentiation between den populations of the Prairie Rattlesnake (Crotalusviridis) in a fragmented landscape

Abstract

Numerous studies have reported genetic fragmentation of species whose habitat has been modified by roads and other anthropogenic features, but it is still not clear how most species respond to roads and whether genetic effects can be detected over a limited number of generations. We used road-crossing models and population genetic analysis (based on microsatellite loci) to make inferences about functional connectivity between populations of the Prairie Rattlesnake (Crotalus viridis (Rafinesque, 1818)) on opposite sides of the Trans-Canada Highway near Medicine Hat (Alberta, Canada). The road-crossing model predicted a high probability of mortality while crossing the Trans-Canada Highway. However, model-based genetic clustering methods (STRUCTURE and BAPS) did not detect structure; a nonmodel-based clustering method (DAPC) found structure, but most groups consisted of individuals captured throughout the study area. Estimates of effective population size were immeasurably large and power to detect genic differentiation was diminished if the effective size exceeded 500; this reduction in power was intensified when the number of loci was reduced (from eight to five to account for null alleles). Our results corroborate accounts of long-distance migration by this species and indicate that genetic fragmentation may not be easily detectable over this spatial and temporal scale.