The role of physical and ecological barriers in the diversification process of birds in the Guiana Shield, northern Amazonia

Abstract

Understanding the factors that influence the formation and location of distribution boundaries is important for the study of evolutionary processes. These factors can be studied effectively at suture zones, regions containing disproportionally high numbers of contact zones (CZs) and phylogeographic breaks (PBs). Together, CZs and PBs offer complementary views of current and historical factors that separate or bring together populations of closely related taxa. For my dissertation, I studied a suture zone in northern Amazonia, where ~100 pairs of taxa replace one another geographically. I analyzed the Guiana Shield avifauna, using bird distributions to redefine the boundaries of the Guianan area of endemism. I showed that the Rio Branco is an important biogeographical barrier and a natural western limit for this area, although smaller rivers, savannas, and mountains also play a significant role. A multivariate approach revealed that the Branco/Negro interfluvium represents a transitional zone for birds, suggesting that the longstanding view of Amazonia as a mosaic of parapatric areas of endemism likely represents an oversimplification of current patterns. I investigated the role of rivers in maintaining and generating biodiversity by testing predictions of the ‘riverine barrier hypothesis’. Using a comparative phylogeographic approach, I found that phenotypically differentiated populations across rivers are reciprocally monophyletic. The lower Rio Negro represents a stronger barrier to gene flow than does the upper Rio Negro, but no genetic homogenization occurs towards the headwaters. Most ‘riverine barrier hypothesis’ predictions were not supported, suggesting that rivers are key to maintaining biodiversity, but not for its generation. Finally, I explored the role of physical and ecological factors in the location of CZs and PBs. PBs cluster along physical barriers, whereas CZs aggregate at the headwaters of large rivers. Nearly half of the pairs that come into contact hybridize, and show significantly lower genetic distances than pairs that come into contact and do not interbreed, suggesting that time of isolation as inferred from genetic data may predict their likelihood of hybridization. Ecological niche models showed significant levels of niche divergence between pairs, suggesting that environmental variables cannot be ruled out as factors influencing the location of suture zones.