Phylogeographic analysis and environmental niche modeling of the plain-bellied watersnake (Nerodia erythrogaster) reveals low levels of genetic and ecological differentiation

Abstract

Species that exhibit geographically defined phenotypic variation traditionally have been divided into subspecies. Subspecies based on phenotypic features may not comprise monophyletic groups due to selection, gene flow, and/or convergent evolution. In many taxonomic groups the number of species once designated as widespread is dwindling rapidly, and many workers reject the concept of subspecies altogether. We tested whether currently recognized subspecies in the plain-bellied watersnake Nerodia erythrogaster are concordant with relationships based on mitochondrial markers, and whether it represents a single widespread species. The range of this taxon spans multiple potential biogeographic barriers (especially the Mississippi and Apalachicola Rivers) that correspond with lineage breaks in many species, including other snakes. We sequenced three mitochondrial genes (NADH-II, Cyt-b, Cox-I) from 156 geo-referenced specimens and developed ecological niche models using Maxent and spatially explicit climate data to examine historical and ecological factors affecting variation in N. erythrogaster across its range. Overall, we found little support for the recognized subspecies as either independent evolutionary lineages or geographically circumscribed units and conclude that although some genetic and niche differentiation has occurred, most populations assigned to N. erythrogaster appear to represent a single, widespread species. However, additional sampling and application of nuclear markers are necessary to clarify the status of the easternmost populations.

Figure 1

Map depicting species range, subspecies breaks, and collecting localities. Due to the spatial scale of the map, not all localities are discernable, although detailed information is available in Appendix I.

Figure 2

Bayesian phylogram produced using a total evidence approach with Cyt-b, NADH-II, and COX-I. Numbers above nodes correspond to the posterior probability / maximum likelihood bootstrap proportion (BP) / parsimony BP. Outgroups have been collapsed and their branch length shortened.

Figure 3

Map depicting species range, subspecies breaks, and major lineage for each sample. See Fig. 2 for symbol definitions.

Figure 4

PCoA results with geographically defined subspecies as the grouping factor.

Figure 5

Scatterplot of first two principal components from the bioclimatic data with grouping based on a) major lineage and b) taxonomy.

Figure 6

Ecological niche models for the a) Eastern lineage and b) non-eastern lineage using 19 WorldClim data layers. Locality points are detailed with major lineage identification symbols, defined in Fig. 2.