Differentiated historical demography and ecological niche forming present distribution and genetic structure in coexisting two salamanders (Amphibia, Urodela, Hynobiidae) in a small island, Japan

Abstract

Background: The climatic oscillations in the Quaternary period considerably shaped the distribution and population genetic structure of organisms. Studies on the historical dynamics of distribution and demography not only reflect the current geographic distribution but also allow us to understand the adaption and genetic differentiation of species. However, the process and factors affecting the present distribution and genetic structure of many taxa are still poorly understood, especially for endemic organisms to small islands.

Methods: Here, we integrated population genetic and ecological niche modelling approaches to investigate the historical distribution and demographic dynamics of two co-existing salamanders on Tsushima Island, Japan: the true H. tsuensis (Group A), and Hynobius sp. (Group B). We also examined the hypothesis on the equivalency and similarity of niches of these groups by identity and background tests for ecological niche space.

Results: Our result showed that Group A is considered to have undergone a recent population expansion after the Last Glacial Maximum while it is unlikely to have occurred in Group B. The highest suitability was predicted for Group A in southern Tsushima Island, whereas the northern part of Tsushima Island was the potential distribution of Group B. The results also suggested a restricted range of both salamanders during the Last Interglacial and Last Glacial Maximum, and recent expansion in Mid-Holocene. The genetic landscape-shape interpolation analysis and historical suitable area of ecological niche modelling were consistent, and suggested refugia used during glacial ages in southern part for Group A, and in northern part of Tsushima Island for Group B. Additionally, we found evidence of nonequivalence for the ecological niche of the two groups of the salamanders, although our test could not show either niche divergence or conservatism based on the background tests. The environmental predictors affecting the potential distribution of each group also showed distinctiveness, leading to differences in selecting suitable areas. Finally, the combination of population genetics and ecological modeling has revealed the differential demographic/historical response between coexisting two salamanders on a small island.

Keywords: Biogeography; Cytochrome b; Ecological niche models; Hynobius tsuensis; Population genetics; Quaternary climate; Syntopic; Tsushima Island.

Figure 1. The position of Tsushima Island, Japan (A) and sampling localities on Tsushima Island (B).

(B) Red triangles and blue inverse triangles show the collection localities of Group A and Group B, respectively. The shaded areas indicate elevations of 200 m or more.

Figure 2. Upper: A lateral view of the result of a genetic landscape-shape interpolation analysis for Group A (A) and Group B (B). Lower map: Sampling localities of Group A (red triangles) and Group B (blue inverse triangles).

Right and left tips of the upper figures correspond to the northernmost and southernmost localities of Group A (red triangles) and Group B (blue inverse triangles) in Tsushima Island (lower maps). High peaks showed high genetic diversity in the upper figures. Red triangles and blue triangles in the lower maps correspond to “Locality” in Table S1. The shaded areas indicate elevations of 200 m or more.

Figure 3. Mismatch distribution analyses based on the partial cyt b sequences (413 bp).

Dashed and solid lines indicate the sudden expansion and spatial expansion models, respectively.

Figure 4. The Bayesian skyline plot based on the partial cyt b sequences (413 bp).

Population dynamics of Group A (A) and Group B (B).

Figure 5. The potential distribution of Group A (A), Group B (B) and overlap area between Group A and Group B (C) under current environmental conditions from Maxent model.

Figure 6. Left: The response curves of the annual precipitation (upper) and elevation (lower) of Group A (red solid line) and Group B (blue dashed line). Right: Distribution of the variables on Tsushima Island.

Figure 7. The potential distribution of Group A (upper) and Group B (lower) for Last Interglacial (LIG), Last Glacial Maximum (LGM), Mid-Holocene (Mid-Holo), and present scenarios.

Figure 8. The results of the identity test, and background tests of Group A and Group B by ENMTools version R.

Black dashed line indicates the results of niche overlap representing the true calculated niche overlap. Red columns show the result of 100 replicates. The left plots showed the Schoener’s D index, and the right plots indicated the Hellinger’s-based I.

Figure 9. The results of ecospat identity test, ecospat background tests of Group A and Group B by ecopat function in ENMtools version R.

Black dashed line indicates the results of niche overlap representing the true calculated niche overlap. Red columns show the result of 100 replicates. The left plots showed the Schoener’s D index, and the right plots indicated the Hellinger’s-based I.