Demographic inference
We compared four different models to infer the evolutionary history of this group of S. corvina subspecies in Costa Rica and Panama. We found one model was preferred over the other three (Table 3). The best model included the default migration bouts between S. c. hoffmanni and S. c. hicksii, and between S. c. corvinaand S. c. hicksii only (Model B; Figure 2). This model suggests a scenario of rapid divergence followed by isolation of the black and pied lineages until recent gene flow facilitated by secondary contact and is consistent with reduced hybridization between S. c. corvina andS. c. hoffmanni . Genomic homogenization due to gene flow in addition to rapid divergence inhibits accurate reconstruction of the phylogenetic relationship between subspecies, estimates of times of divergence, and rates of gene flow.
DISCUSSION
We compiled a dataset of 14,839 genome-wide SNPs from 255 individuals from Costa Rica and Panama to test whether phenotypic traits covary with population structure and hybridization patterns between S. corvina subspecies that differ in plumage coloration across three independent contact zones (Figure 1A). We found three genetic clusters of moderate genetic differentiation, consistent with subspecies distributions (Figures 1C and S4), that hybridized extensively across two of the three contact zones (Figures 3 and 4). Hybridization occurs despite their differences in plumage coloration, suggesting a limited role of plumage divergence in reproductive isolation. Across these two contact zones, subspecies differ in the plumage brightness of a few male color patches. For these color patches, we found coincident centers with the respective hybrid index/genetic clines but a sharper transition (i.e., narrower clines), suggesting limited introgression of these plumage traits (Figure 5; Table 2). We found no other clear morphometric variation among subspecies. Finally, our demographic inference supports our observations of no hybridization between S. c. corvina andS. c. hoffmanni, which recently established secondary contact in the Central Valley of Costa Rica (Table 3).