Integrating paleoecology and genetics of bird populations in two sky island archipelagos.

BACKGROUND: Genetic tests of paleoecological hypotheses have been rare, partly because recent genetic divergence is difficult to detect and time. According to fossil plant data, continuous woodland in the southwestern USA and northern Mexico became fragmented during the last 10,000 years, as warming caused cool-adapted species to retreat to high elevations. Most genetic studies of resulting 'sky islands' have either failed to detect recent divergence or have found discordant evidence for ancient divergence. We test this paleoecological hypothesis for the region with intraspecific mitochondrial DNA and microsatellite data from sky-island populations of a sedentary bird, the Mexican jay (Aphelocoma ultramarina). We predicted that populations on different sky islands would share common, ancestral alleles that existed during the last glaciation, but that populations on each sky island, owing to their isolation, would contain unique variants of postglacial origin. We also predicted that divergence times estimated from corrected genetic distance and a coalescence model would post-date the last glacial maximum. RESULTS: Our results provide multiple independent lines of support for postglacial divergence, with the predicted pattern of shared and unique mitochondrial DNA haplotypes appearing in two independent sky-island archipelagos, and most estimates of divergence time based on corrected genetic distance post-dating the last glacial maximum. Likewise, an isolation model based on multilocus gene coalescence indicated postglacial divergence of five pairs of sky islands. In contrast to their similar recent histories, the two archipelagos had dissimilar historical patterns in that sky islands in Arizona showed evidence for older divergence, suggesting different responses to the last glaciation. CONCLUSION: This study is one of the first to provide explicit support from genetic data for a postglacial divergence scenario predicted by one of the best paleoecological records in the world. Our results demonstrate that sky islands act as generators of genetic diversity at both recent and historical timescales and underscore the importance of thorough sampling and the use of loci with fast mutation rates to studies that test hypotheses concerning recent genetic divergence.

Geographic variation in Mexican jays (Aphelocoma ultramarina): local differentiation, polyphyly or hybridization?

Studies of genetic variation within highly variable taxa can provide valuable insight into the factors influencing biological diversification. We examined six microsatellite loci, a nuclear intron and the mitochondrial control region to determine if the Mexican jay subspecies Aphelocoma ultramarina couchii and A. u. potosina have hybridized with western scrub-jays (A. californica). We suspected hybridization because these Mexican Jay populations resemble scrub-jays in several traits. We sampled six Mexican jay (N = 105) and four scrub-jay (N = 78) populations. Suspected hybrid Mexican Jay populations did not share any mitochondrial types or intron alleles with scrub-jays. All microsatellite alleles found in the suspected hybrid Mexican jay populations are also found in the control Mexican jay populations. Genetic distance-based trees from microsatellites supported reciprocal monophyly of Mexican jays and scrub-jays with bootstrap support > 80%. We randomized genotypes among populations to test scenarios consistent with hybridization. In some areas where Mexican jays and scrub-jays occur in sympatry or geographical proximity, randomization tests yield results expected under hybridization (100% support for seven of nine scenarios). However, these populations were not the primary candidates for hybridization based on phenotype. Even if low-level hybridization did occur, hybridization does not appear to be the main reason some Mexican jay populations resemble scrub-jays more than others. The scrub-jay-like traits in these populations may be due to drift, adaptation or plasticity. Alternatively, ancient hybridization, followed by selection for scrub-jay like traits in some Mexican jay populations, might have given rise to the observed variation.