Geographic and climatic differences between islands and continents may affect the evolution of their biota, and promote divergent selection in species distributed in both landscapes. To assess spatial-genetic structure, we genotyped 18 expressed sequence tag-simple sequence repeat (EST-SSR) loci and sequenced two mtDNA markers (ND5 and COI) and one nuclear marker (EF1α) in two subspecies of the butterfly Parantica sita. Compared with nuclear markers, mtDNA had a stronger signal of population structure. Approximate Bayesian computation (ABC) suggested that a continuousgene-flow model best described the data. According to this model, the two subspecies diverged approximately 23.1 kya, with 10 times more introgression from the continental (ssp. sita) to the insular subspecies (ssp. niphonica) than vice versa. Ecological niche modelling was performed to predict the paleo-and current potential distributions and elucidate the geohistorical process, which revealed a northeastern, insular origin.Winter precipitation and annual temperature range were the main determinants of the subspecies distributions. Maximum-likelihood population-effects models showed that the population differentiation of the insular and continental subspecies was primarily affected by environmental resistance and local climate. Sex-biased migration capacity and long-term precipitation-driven divergence between the continental and insular lineages shaped the current genetic structure of P. sita. Evidence from the nuclear markers confirmed intersubspecific gene flow despite adaptive divergence between the subspecies. These results imply that the continental subspecies is still capable of returning to the island and introgressing with the insular subspecies.

Aim: Rapid global warming is threatening global biodiversity, and it will likely lead to varying degrees of local adaptation, particularly amongst plant species. Besides, rising temperatures frequently result in upslope distribution shifts towards climatic optima (i.e. the escalator effect) within a limited dispersal space, such as in insular environments. Here, we integrated ecological and genetic approaches to investigate how climate change will impact the genetic compositions and spatial distributions of Taiwan's endemic maple Acer caudatifolium.Location: Taiwan. Methods:We estimate the distribution range shifts of A. caudatifolium under climate change through species distribution modelling (SDM). We also use 368 genotyped samples to infer dispersal and genetic hotspots and quantify the contributions of geography/environments to genetic variations. We further assess the potential risk to A. caudatifolium under different climate warming scenarios. Results:We detected three genetic diversity hotspots near mountainous glacial refugia and two dispersal hotspots in northern Taiwan and the central-to-southern Central Mountain Range. Overall range reductions and an altitudinal upslope shift were observed in SDM. Using both linear and nonlinear regression approaches, we found that genetic variation was significantly associated with geographic distance and elevation-related climatic variables. The potential risk analysis revealed that the northernmost summit-dwelling populations were the most vulnerable. Furthermore, the major risk factor differed amongst populations: for central populations, temperature and precipitation jointly determined the potential risk, whereas precipitation was the only risk factor for northern and southern populations.Main conclusions: This case study demonstrates how various climate factors, mountain height and the availability of corridors jointly determine the demographic fates and sustainability of island maples in the face of climate change. This study also provided estimates of the implications of global warming, which can be conducive to developing appropriate conservation strategies.

No abstract