James Raymond Peter Worth

Warm‐edge populations, that is, those usually small and isolated populations at the boundary of species maximum temperature range limit, are of particular scientific and conservation interest. This is especially the case for those that have long been geographically isolated, possibly through multiple glaciations, as they can contribute significantly to species' overall genetic diversity. However, distinguishing between glacial relict and more recently established warm‐edge populations is difficult especially in the absence of a continuous fossil record. In this study, we investigate the origin of anomalously low‐elevation warm‐edge populations of the dominant cool temperate Japanese tree Fagus crenata in the northeast Kanto (NEK) region. We sought to determine whether these populations are glacial relicts or, rather, were established via dispersal in the Holocene using chloroplast and nuclear genetic data, species distribution modeling and comparison with populations in a well‐established glacial refugia (south‐central Kyushu [SCK]) and of mid‐late Holocene age (southern Hokkaido [SH]). Whole chloroplast genome “mining” revealed a similar amount of haplotype diversity in NEK and SCK while there was no chloroplast variation in SH. Populations in NEK, particularly the lowest elevation sites, harbored unique haplotypes providing evidence for their persistence since the Last Glacial Maximum as supported by species distribution models under two of three general circulation models. NEK populations also lacked the genetic signatures of recent expansion that characterize the most northern‐edge populations in SH. We conclude that warm‐edge populations in NEK are most likely relicts of glacial refugia and should be given high priority for conservation.