Characterization of hybrids and contact zones
Based on our test using only informative markers, we found that individuals of hybrid ancestry had generally low levels of heterozygosity across the three contact zones (Figure 4). Moreover, we also found a continuous distribution of hybrid indices between the two parental subspecies in the corvina-hicksii andhoffmanni-hicksii contact zones (Figure 4A,B). Interestingly, we found virtually no corvina-hoffmanni hybrids (Figure 4C), in this contact zone; all individuals with mixed ancestry can be confidently assigned to one of the parental subspecies based on the genetic makeup, contrary to the other two contact zones.
We fitted geographic clines along both contact zones in which subspecies hybridize, we fitted clines for the hybrid index, plumage traits, and significantly different morphometric traits. First, we identified morphometric differences between distant populations, testing for the effect of subspecies and sex, and found corvina andhicksii differed in beak size, wing length, and tail length, while hoffmanni and hicksii differed in beak size, tail length, and tarsus length (Table 1). For color patches, we focused on those patches that differed between subspecies. In both contact zones, the best fit model for the hybrid index had a fixed scaling from 0 to 1 and no tails fitted (Table 2). Plumage brightness of the different patches modeled showed a clinal distribution in which the brightness of the rump and belly in males for the corvina-hicksii contact zone, and the brightness of the throat for the hoffmanni-hicksiicontact zone had cline centers that were coincident with their respective hybrid index clines. However, the center of the throat’s brightness for the corvina-hicksii contact zone is displaced ~60 km to the East, with no overlap of the 95% CI (Figure 5A,C; Table 2). Plumage brightness and genetic clines differ in cline widths, with plumage clines showing narrower widths than the hybrid index clines (Figure 5A,C; Table 2).
When the most distant populations differed in morphometric traits, we found a general pattern in which hicksii is smaller than the other two subspecies, showing on average smaller beaks and shorter tails than corvina and hoffmanni , and shorter wings thancorvina . However, hoffmanni had, on average, shorter tarsi than hicksii (Table 1). Morphometric traits along thecorvina-hicksii contact zones (tail length and beak size) were not coincident with respect to the hybrid index, with tail length and beak size having displaced centers (Figure 5B; Table 2). Along thehoffmanni-hicksii contact zone, cline centers of beak size and tail length are coincident with the correspondent hybrid index based on the overlapping of the 95% CI (Figure 5D; Table 2). However, we found important intra-population variation, with none of the morphometric traits exhibiting a clear clinal distribution when modeling average trait values per population and, instead, showed a more linear, smoother transition from one parental population to the other, which result in low confidence of cline parameters (Figure 5B,D; Table 2). For example, the CI of the cline width of the tarsus length betweenhoffmanni-hicksii and wing length between corvina-hicksiiwere estimated to be between ~20 km to 1200 km wide (not graphed, Table 2).