It is often expected that temperate plants have expanded their geographic ranges northward from primarily southern refugia. Evidence for this hypothesis is mixed in eastern North American species, and there is increasing support for colonization from middle latitudes. We studied genome-wide patterns of variation in RADseq loci to test hypotheses concerning range expansion in a North American forest herb (Campanula americana). First, spatial patterns of genetic differentiation were determined. Then phylogenetic relationships and divergence times were estimated. Spatial signatures of genetic drift were also studied to identify the directionality of recent range expansion and its geographic origins. Finally, spatially explicit scenarios for the spread of plants across the landscape were compared, using variation in the population mutation parameter and Tajima's D. We found strong longitudinal subdivision, with populations clustering into groups west and east of the Mississippi River. While the southeastern region was likely part of a diverse Pleistocene refugium, there is little evidence that range expansion involved founders from these southern locales. Instead, declines in genetic diversity and the loss of rare alleles support a westward colonization wave from a middle latitude refugium near the southern Appalachian Mountains, with subsequent expansion from a Pleistocene staging ground in the Mississippi River Valley (0.51 - 1.27 Mya). These analyses implicate stepping stone colonization from middle latitudes as an important mechanism of species range expansion in eastern North America. This study further demonstrates the utility of population genetics as a tool to infer the routes traveled by organisms during geographic range expansion.

This dataset contains R scripts necessary to generate the results in our paper. All scripts are custom.

Users will need sequence data (BioProject ID: PRJNA560272) and some knowledge of iPyrad parameterization (see Supporting Table 1).