Hard edges, soft edges, and species range evolution: A genomic analysis of the Cumberland Plateau salamander

Abstract

Aim: Gene flow from central to edge populations is thought to limit population growth at range edges by constraining local adaptation. In this study, we explore the thesis that range edges can differ in their dynamics and be either ‘hard’ (e.g. a river) or ‘soft’ (e.g. ecological gradients). We hypothesize that soft edge populations will have smaller effective population sizes than central populations and that gene flow will be greater from the centre to the edge than vice versa. Conversely, we hypothesize that hard edge populations should have similar effective population sizes to central populations and that gene flow will be equal between the two. Location: Kentucky, West Virginia, and Virginia, USA. Taxon: Plethodon kentucki (Caudata: Plethodontidae). Methods: We evaluated landscape suitability using an ecological niche model, then we compared gene flow and effective population sizes between edge and central populations and quantified gene flow between populations. Finally, we characterized landscape genetic variation, testing for isolation by distance and isolation by environment. Results: We found continuously decreasing habitat quality along soft edges, with hard edges more variable. Additionally, we found that soft edges had lower effective population sizes than central populations and that gene flow was greater from the centre of the range to the soft edges than the reverse. In hard edges, by contrast, we found effective population sizes in edge populations were similar to central populations, with relatively equal gene flow in both directions. Main Conclusions: Understanding why species have range limits is central to investigations of the structure of biodiversity, yet the evolutionary dynamics of range edges remain poorly understood. We show that within a single species with a small range, the evolutionary dynamics operating at range boundaries may depend on the nature of the boundary.