Fig. 1

Studied Taiwanese Rufous-capped Babbler populations. A A map showing the two low-altitude populations (L1 and L2) and the two high-altitude ones (H and H′). The low- and high-altitude common gardens for reciprocal transplant experiments are also indicated on the map (l and h, respectively). Note that H′ was only used for population genetic analyses, not the transplant experiments. B An asymmetric neighbor joining tree for the evolutionary relationships among studied populations plus an additional one from the mainland China. For tree building, a nonreversible model of genetic drift [55] is used to estimate drift-caused changes in allelic frequencies (the drift parameter) between split events; the use of a nonreversible model renders this method being able to root the tree without using outgroups. C Principal component analysis (PCA) plot, showing projections of individuals on the first two principal components. Analyses of B and C were conducted with respective sets of 10,000 randomly selected autosomal single-nucleotide polymorphisms (SNPs) that segregate over the total samples