Polyploids recurrently emerge in angiosperms, but most polyploids are likely to go extinct before establishment due to minority cytotype exclusion, which may be specifically a constraint for dioecious plants to evolve polyploid populations. Investigations into the frequency and distribution of polyploids in natural populations is thus necessary for understanding polyploid evolution in plants. This study determined the ploidy levels of 28 populations and 351 individuals of Salix polyclona, and identified the type of polyploidy (auto- vs. allo-) using whole genome re-sequencing data. We further investigated the phylogeny, population genetic diversity and species range shifts to explore the origin and spatiotemporal evolution of the polyploid complex. Our analyses revealed a high frequency (52%) of autopolyploids in it with a clear geographic distribution confined to the western part of its range where complex mountain systems create higher levels of environmental heterogeneity. Comparisons of diploid male and female genomes suggested a female heterogametic sex-determining factor on chromosome 15, which likely also acts in the dioecious polyploids. Fossil-calibrated phylogeny showed a more recent diversification of the polyploids (ca. 2.3 Ma) than the diploid (ca. 6.2 Ma), and population demographic histories largely corroborated the geological and climatic history of the region. Our results suggest that climatic oscillations and uplift of eastern Qinghai-Tibetan Plateau and the connecting mountains may have facilitated the preservation and establishment of polyploid populations. This study provides an example of the evolution of a diploid-polyploid complex in a willow species and illustrates a role of polyploidization in mountain biodiversity.

Pine wilt disease (PWD), Bursaphelenchus xylophilus, is an extremely threatening invasion forest disease throughout the world, especially in Asia. B. xylophilus is spread in Asia by vector beetles of Monochamus alternatus, which has long no effective control method. Understanding of landscape effects on the dispersal and outbreaks of forest pests is crucial to establishing effective ecological control strategies. Here, we analyzed the samples of M. alternatus collected at landscapes in order to estimate the effects of landscape types on the genetic structure and dispersal of M. alternatus. The landscapes included the geographical scales, forest types and land uses. The individuals of M. alternatus were genotyped by using whole-genome resequencing. Population genetic structures were clearly differentiated at the intermediate scale, suggesting the intermediate scale is an effective barrier against natural dispersal of M. alternatus. We used the least-coat distances, least-cost transect analysis, and distance-based redundancy analysis to estimate the effects of forest types and land uses within the fine scales. The results showed that the gene flow and genetic diversity were positively correlated with host and mixed forests, whereas negatively with non-host forests. Among land-use landscapes, the roads had the positive effect on gene flow and genetic diversity but farmland and urban uses had negative effects. This highlights that human-mediated transport via roads was likely to be the main factor leading to the long-distance invasion of M. alternatus, whereas non-host landscapes could suppress the spread of this species. These findings may be useful to control the PWD dispersed by M. alternatus.