Social barriers have been shown to reduce gene flow and contribute to genetic structure among populations in species with high cognitive capacity and complex societies, such as cetaceans, apes and humans. In birds, high dispersal capacity is thought to prevent population divergence unless major geographic or habitat barriers induce isolation patterns by dispersal, colonization or adaptation limitation. We report that Iberian populations of the red-billed chough, a social, gregarious corvid with high dispersal capacity, show a striking degree of genetic structure composed of at least 15 distinct genetic units. Monitoring of marked individuals over 30 years revealed that long-distance movements over hundreds of kilometres are common, yet recruitment into breeding populations is infrequent and highly philopatric. Genetic differentiation is weakly related to geographic distance and habitat types used are overall qualitatively similar among regions and regularly shared by individuals of different populations, so that genetic structure is unlikely to be due solely to isolation by distance or isolation by adaptation. Moreover, most population nuclei showed relatively high levels of genetic diversity, suggesting a limited role for genetic drift in significantly differentiating populations. We propose that social mechanisms may underlie this unprecedented level of genetic structure in birds through a pattern of isolation by social barriers not yet described, which may have driven this remarkable population divergence in the absence of geographic and environmental barriers.

Understanding how gene flow shapes contemporary population structure requires the explicit consideration of landscape composition and configuration. New landscape genetic approaches allow us to link such heterogeneity to gene flow within and among populations. However, the attribution of cause is difficult when landscape features are spatially correlated, or when genetic patterns reflect past events. We use spatial Bayesian clustering and landscape resistance analysis to identify the landscape features that influence gene flow across two regional populations of the eastern massasauga rattlesnake, Sistrurus c. catenatus. Based on spatially explicit simulations, we inferred how habitat distribution modulates gene flow and attempted to disentangle the effects of spatially confounded landscape features. We found genetic clustering across one regional landscape but not the other, and also local differences in the effect of landscape on gene flow. Beyond the effects of isolation-by-distance, water bodies appear to underlie genetic differentiation among individuals in one regional population. Significant effects of roads were additionally detected locally, but these effects are possibly confounded with the signal of water bodies. In contrast, we found no signal of isolation-by-distance or landscape effects on genetic structure in the other regional population. Our simulations imply that these local differences have arisen as a result of differences in population density or tendencies for juvenile rather than adult dispersal. Importantly, our simulations also demonstrate that the ability to detect the consequences of contemporary anthropogenic landscape features (e.g. roads) on gene flow may be compromised when long-standing natural features (e.g. water bodies) co-exist on the landscape.