ABSTRACT
Theory suggests speciation without geographic isolation is plausible if divergent natural selection is strong enough to counteract gene flow. Tropical mountains provide strong and temporally stable environmental gradients that can promote local adaptation and population genetic structure. Pairs of closely related species with adjacent but divergent elevational ranges are common in these environments, a pattern that suggests parapatric speciation (i.e., speciation with moderate gene flow), but evidence for this process is scarce. Here we use genomic data from modern and historical museum specimens to investigate speciation in a pair of New Guinea kingfishers that segregate ranges by elevation. We find that the lowland species Syma torotoro and montane species S. megarhyncha form discrete genotypic clusters with bimodal variance in phenotypic traits. Phylogenetic relationships among lineages are discordant between mitochondrial and nuclear genomes, which D-tests and demographic inference indicate is partly driven by interspecific gene flow over long time periods. Summary statistics reveal differentiation is concentrated in a handful of small regions of the genome. These data are consistent with ecological speciation driven by adaptation to abiotic or biotic factors varying with elevation. More broadly, they suggest selection across elevational gradients can maintain species boundaries without intrinsic reproductive isolation, a mechanism contributing to high tropical biodiversity.
Footnotes
We have altered our approach to phylogenetic inference, using improved mitochondrial DNA assemblies, a method to infer an ultrametric tree using a strict molecular clock, and estimated a coalescent-based species tree. We now clarify which dataset accompanied each analysis, and also include this information in a new table of sampling data in the main text. We include numerous new details to allow readers to understand and reproduce our methods, and justify choices (e.g. with SNP filters) whenever possible. Throughout the manuscript, we shift from a hypothesis-testing framework to a more descriptive approach, including in a revised demographic inference analysis. This analysis implements a different method of estimating parameter uncertainty, correcting erroneous values from the previous draft. We temper our conclusions and highlight sources of uncertainty in the title, abstract, introduction, and discussion sections. Finally, we have removed material from the previous version we believe was redundant or otherwise unnecessary.