ABSTRACTThe golden-crowned (Zonotrichia atricapilla) and white-crowned (Z. leucophrys) sparrows have been presented as a compelling case for rapid speciation. They display divergence in song and plumage with overlap in their breeding ranges implying reproductive isolation, but have almost identical mitochondrial genomes. Previous research proposed hybridization and subsequent mitochondrial introgression as an alternate explanation, but lacked robust nuclear gene trees to distinguish between introgression and incomplete lineage sorting. We test for signatures of these processes between Z. atricapilla and Z. leucophrys, and investigate the relationships among Z. leucophrys subspecies, using mitochondrial sequencing and a reduced representation nuclear genomic dataset. Contrary to the paraphyly evident in mitochondrial gene trees, we confirmed the reciprocal monophyly of Z. atricapilla and Z. leucophrys using large panels of single nucleotide polymorphism (SNPs). The pattern of cytonuclear discordance is consistent with limited, historical hybridization and mitochondrial introgression, rather than a recent origin and incomplete lineage sorting between recent sister species. We found evidence of nuclear phylogeographic structure within Z. leucophrys with two distinct clades. Altogether, our results support the true species status of Z. atricapilla and Z. leucophrys, and indicate deeper divergences between the two species than inferred using mitochondrial markers. Our results demonstrate the limitations of relying solely on mitochondrial DNA for taxonomy, and raise questions about the possibility of selection on the mitochondrial genome during temperature oscillations (e.g. during the Pleistocene). Historical mitochondrial introgression facilitated by past environmental changes could cause erroneous dating of lineage splitting in other taxa when based on mitochondrial DNA alone.
INFERRING PHYLOGENIES FROM mtDNA VARIATION: MITOCHONDRIAL-GENE TREES VERSUS NUCLEAR-GENE TREES REVISITED