Aim We aimed to elucidate how the current geographic distribution of alpine plants in the Japanese archipelago was shaped during Quaternary climatic oscillations, using Potentilla matsumurae as a case study. According to previous phylogeographic studies, post-glacial range fragmentation (vicariance scenario) and stepwise migration (dispersal scenario) are both possible. We thus aimed to assess which scenario is more probable for the distribution changes of alpine plants in the Japanese archipelago.
Location The alpine zone in the Japanese archipelago.
Methods Using amplified fragment length polymorphism we determined the genotype of 161 individuals of P. matsumurae from 22 populations. Relationships among individuals and populations were examined using principal coordinates analysis and a neighbour-joining (NJ) tree, respectively. To examine the genetic population structure, we performed analysis of molecular variance (amova) and structure analysis.
Results Differentiation between central Honshu and northern Japan was not very strong based on the principal coordinates analysis among individuals, the NJ tree of populations (59% bootstrap support), or amova (12% of genetic variation). Moreover, structure analysis did not detect clear geographic differentiation across populations. Although the populations in central Honshu were structured geographically (Mantel test: r = 0.45, P < 0.005; NJ tree), those in northern Japan did not exhibit geographic structure regardless of geographic distance (Mantel test: r = 0.26, P = 0.03; NJ tree). Population relationships in the NJ tree did not always reflect the geographic location.
Main conclusions The current geographic structure of P. matsumurae could not be explained by stepwise migration. This suggests that a single continuous distribution during the last glacial period was later fragmented, perhaps by recovering forest, during the post-glacial period, resulting in the current distribution and phylogeographic structure of P. matsumurae. Our data support the vicariance scenario.