Neutral processes and reduced dispersal across Amazonian rivers may explain how rivers maintain species diversity after secondary contact

Riverine Barriers as Obstacles to Dispersal in Amazonian Birds

Amazonian rivers represent known barriers for avian dispersal, reducing gene flow and enhancing differentiation. Despite the importance of rivers in the avian evolutionary process, we have made only minor advances in understanding the limitations imposed by rivers on flying birds. To fill that gap, we conducted dispersal-challenge experiments over water, assessing the flying capabilities of 84 tropical bird species of 22 different avian families. We mist-netted and released 484 birds from a stationary boat on the Rio Branco, northern Amazonia, at increasing distances from the shore, including 249 individuals at 100; 219 at 200; 8 at 300; and 5 at 400 m. A successful trial was represented by a bird reaching the riverbank, whereas a failure would refer to birds not reaching the shore and landing on the water, when they were rescued by our team. Our main goal was to understand if the outcome in the experiments could be predicted by (i) phylogenetic constraints, (ii) morphology (body mass and wing shape), (iii) flight speed, (iv) ecological preferences (stratum, habitat, and river-island specialization), and (v) psychological reluctance to fly. Nearly two thirds of the individuals (332) were successful in reaching the riverbank, whereas 152 failed. We found significant differences among lineages. Whereas seven avian families succeeded in all of the trials, two families (antbirds and wrens) were particularly bad dispersers (<40% success). The hand-wing index (HWI) was the single most powerful predictor of trial success. Flying speed was also a significant predictor of success. Overall, ecological attributes had a low explanatory power. Only forest stratum preference had a significant, although weak, effect on dispersal ability: canopy- and ground-dwellers performed better than understory birds. However, we found no effect of habitat preference or river-island specialization on dispersal ability. Our speed estimates for 64 bird species are among the first produced for the tropics and suggest slower flying speeds than those reported from temperate migratory birds. Although birds showed behavioral differences when presented with the opportunity to fly away from the boat, we found no evidence that their reluctance to fly could predict the outcome in the experiments. This represents the first experimental study evaluating the riverine effect through dispersal ability of Amazonian birds, providing important insights to better understand dispersal limitations provided by riverine barriers.