Evolutionary and ecological drivers of local adaptation and speciation in a North American avian species complex.

Throughout the speciation process, genomic divergence can be differentially impacted by selective pressures, as well as gene flow and genetic drift. Disentangling the effects of these evolutionary mechanisms remains challenging, especially for nonmodel organisms. Accounting for complex evolutionary histories and contemporary population structure often requires sufficient sample sizes, for which the expense of full genomes remains prohibitive. Here, we demonstrate the utility of partial-genome sequence data for range-wide samples to shed light into the divergence process of two closely related ducks, the Mexican duck (Anas diazi) and mallard (A. platyrhynchos). We determine the role of selective and neutral processes during speciation of Mexican ducks by integrating evolutionary and demographic modelling with genotype-environment and genotype-phenotype association testing. First, evolutionary models and demographic analyses support the hypothesis that Mexican ducks originally diverged ~300,000 years ago in climate refugia arising during a glacial period in southwest North America, and that subsequent environmental selective pressures played a key role in divergence. Mexican ducks then showed cyclical demographic patterns that probably reflected repeated range expansions and contractions, along with bouts of gene flow with mallards during glacial cycles. Finally, we provide evidence that sexual selection acted on several phenotypic traits as a co-evolutionary process, facilitating the development of reproductive barriers that initially arose due to strong ecological selection. More broadly, this work reveals that the genomic and phenotypic patterns observed across species complexes are the result of myriad factors that contribute in dynamic ways to the evolutionary trajectories of a lineage.