Possible LDD-driven divergence process of migratory species
Previous studies have illustrated a divergence scenario for migratory species that assumed vicariance in generating isolated populations (Hewitt, 2004; Weir & Schluter, 2004). While this scenario has provided suitable explanations for many cases, some divergence events may not adequately be explained by it. For example, the traditional vicariance scenario suggests that the 100 km wide strait between the Korean Peninsula and the Japanese archipelago caused deep genetic divergence of archipelagic populations of mammals from the continent (McKay, 2012; Sato, 2017). From an ecological perspective, such a small barrier would ineffectively restrict the movement of a highly volant bird species (e.g. Gotelli & Graves, 1990), and establishment of a new population from a more distant area as a consequence of LDD is likely to be its alternative explanation. However, traditional frameworks based on the continental distribution for most migratory species have probably hindered the opportunities to test it (Warren et al., 2015).
Our integrative approach using a Japanese migratory bird species provided a suitable system within which to discriminate the mode of genetic divergence under a hypothesis testing framework. We successfully provided a hypothetical scenario for how LDD and paleogeography have contributed to divergence of a migratory species. We also showed that discriminating LDD from vicariance in our framework benefited a more comprehensive understanding of diversification mechanisms of migratory species in this region. We believe that consideration of LDD is also important for continental regions where most migratory species have diversified. Attention has recently been paid to LDD and colonisation as an important phenomenon for distribution and speciation of continental migratory species (Rushing, Dudash, Studds, & Marra, 2015; Winkler et al., 2017). Like the ECS in our system, some geographic barriers (such as seas, deserts, and mountain ranges) in continental areas have also changed size and shape during their geological history (Hewitt, 2004; Sun et al., 2018). Some of the genetic divergence in migratory species over these barriers should be revisited taking the LDD hypothesis into account (e.g. Reeves, Drovetski, & Fadeev, 2008). Further verification of the usability of our integrative approach among several species will facilitate our understanding of the relative importance of LDD to vicariance in the speciation of migratory birds and other animals.
ix. Figures embedded in the text:
Figure 1 Predicted colonisation route (black arrows) and expected results from integrative approaches in (a) long distance dispersal (LDD) and (b) vicariance hypotheses in the Brown Shrike. Predicted gene networks (upper left panes), migratory routes (lines with open circles), and suitable glacial regions (coloured circles) under the two hypotheses are shown. Colours correspond to subspecies in Figure 1c. We predicted that L. c. lucionensis represents the basal lineage of the Brown Shrike since close relatives of this species are distributed around South-east Asia (Zhang et al., 2007). Because there are various possibilities for the genetic position of L. c. cristatus in the LDD hypothesis, we did not present it in Figure 1a. Grey coloured regions represent terrestrial areas during the Early to the Middle Pleistocene, and dotted lines indicate the present coastlines. (c) DNA sampling localities. Circles and triangles represent samples from breeding and non-breeding localities, respectively, and their colours indicate haplogroups inferred in the median-joining haplotype network of COI (Figure 2a). Land colours indicate breeding ranges of presumed subspecies following Worfolk (2000). The striped region indicates a presumed hybrid zone between subspecies.
Figure 2 (a) A median joining haplotype network constructed using 521 bp of the partial COI gene. Each circle indicates a unique haplotype and its size is proportional to the number of samples. Colours of circles indicate regions where samples were collected. One mutational gap was indicated by one bar, and a black node indicates a median vector. Clusters analogous to clades found in the Bayesian inference tree (Figure S1.2) are indicated. (b) The multi-locus network constructed for the Brown Shrike using the four loci under the NeighborNet algorithm. Coloured shades correspond to clades to which their mitochondrial haplotypes belong.
Figure 3 Estimated suitable distribution for L. c.superciliosus for different time periods; a constructed model was projected to (a) the present climate, and (b-d) three different climate models for the Last Glacial Maximum (LGM, 20 kya). Suitability for a 2.5 min grid increases from yellow, green to indigo, whereas the grey area was determined as absence. The black area is that of strict extrapolation indicated by multivariate environmental similarity surface analysis. Black lines indicate present coastlines. Note that the terrestrial landmass expanded during the LGM due to the lowered sea level.
Figure 4 Migratory routes of three individual L. c. superciliosus around the East China Sea. Mean posterior probabilities of location estimates of the three birds for each raster cell are indicated by colour gradation from yellow (high) to indigo (low). Dots with line segments indicate median tracks of their inferred migratory routes (white and red dots indicate migrating and stationary sites, respectively). Autumn and spring migrations are indicated by solid and hatched lines, respectively.
Figure 5 Schematic presentation of the hypothetical phylogeographic scenario of divergence in the Brown Shrike. (a) The ancestral population colonised the Japanese archipelago via LDD across the East China Sea (ECS) in a glacial period around the late Early Pleistocene. (b) Increased isolation of the archipelagic population through expansion of the ECS facilitated its divergence. Its migratory route was shaped to retrace the colonisation route. (c) In a later glacial period, part of the archipelagic population colonised the Korean Peninsula over a land bridge. (d) Northward expansions caused vicariance of both the archipelagic and northern clades and resulted in their present geographical distributions. The hatched lines in each pane indicate the predicted coastlines in the previous inter/glacial periods, following Ota (1998). Coloured circles indicate ancestral populations of each lineage. Lines with open circles indicate the migration route and arrows indicate movements of populations.