Speciation in savanna birds in South America: The case of the Least Nighthawk Chordeiles pusillus (Aves: Caprimulgidae) in and out of the Amazon.

The Least Nighthawk Chordeiles pusillus is widespread wherever there are savannas in the South American tropics, often in isolated patches, such as white-sands savannas in the Amazon rainforest realm. Here, we investigate genetic relationships between populations of the Least Nighthawk to understand historical processes leading to its diversification and to determine dispersal routes between northern and southern savannas by way of three hypothesized dispersal corridors by comparing samples from white-sand savannas to samples from other savannas outside of the Amazon rainforest region. We use 32 mtDNA samples from the range of C. pusillus to infer a dated phylogeny. In a subset of 17 samples, we use shotgun sequences to infer a distance-based phylogeny and to estimate individual admixture proportions. We calculate gene flow and shared alleles between white-sand and non-Amazonian populations using the ABBA-BABA test (D statistics), and Principal Component Analysis (PCA) to examine genetic structure within and between lineages. Finally, we use species distribution modelling (SDM) of conditions during the Last Glacial Maximum (LGM), currently, and in the future (2050-2080) to predict potential species occurrence under a climate change scenario. Two main clades (estimated to have diverged around 1.07 million years ago) were recovered with mtDNA sequences and Single Nucleotide Polymorphism (SNPs) and were supported by NGSadmix and PCA: one in the Amazon basin white-sand savannas, the other in the non-Amazonian savannas. Possible allele sharing between these clades was indicated by the D-statistics between northern non-Amazonian populations and the white-sand savanna population, but this was not corroborated by the admixture analyses. Dispersal among northern non-Amazonian populations may have occurred in a dry corridor between the Guianan and the Brazilian Shield, which has since moved eastward. Our data suggest that the lineages separated well before the Last Glacial Maximum, consequently dispersal could have happened at any earlier time during similar climatic conditions. Subsequently, non-Amazonian lineages became more divergent among themselves, possibly connecting and dispersing across the mouth of the Amazon River across Marajó island during favourable climatic conditions in the Pleistocene.