Nightjars (Aves: Caprimulgidae) are a species-rich family of birds, with the “eared nightjars” (Eurostopodinae) being an early-branching group endemic to the Indo-Pacific. While much research has focused on species-rich nightjar genera and their higher-level relationships, the evolutionary history of Eurostopodinae (Eurostopodus, Lyncornis) remains understudied. We generated a genome-scale dataset to produce the first fully sampled phylogeny of all Eurostopodus and one Lyncornis species, including sequencing two type specimens of critically endangered and extinct species. Tree-building methods inferred concordant, well-resolved topologies that reveal intriguing biogeographic patterns within Eurostopodus. Our results show Eurostopodus as sister to all other nightjars, while Lyncornis, previously considered related, is more closely allied with other caprimulgids. We propose that the term "eared nightjars" should apply only to the two Lyncornis species, which should be classified within the subfamily Caprimulginae. Accordingly, since only Eurostopodus species remain in Eurostopodinae, we recommend renaming this subfamily "Indo-Pacific nightjars" to reflect their geographic distribution in this significant region.

The goal of this study is to assess the evolutionary history of this understudied group of nightjars using a genome-scale dataset. We sampled all Eurostopodus species, including three species that have never been part of a molecular analysis before—two of which (E. exul and E. diabolicus) were sourced from toepad clippings of type specimens housed at the American Museum of Natural History.

We collected a comprehensive genome-scale dataset of ultraconserved elements (UCEs), utilizing an updated probe set specifically refined to better capture UCEs from degraded, toepad-sourced samples. For wider taxonomic sampling, we downloaded whole genomes from GenBank to harvest UCE data in silico. We used both maximum likelihood and species tree methods to infer species relationships. For the maximum likelihood tree, we used IQtree 2. We determined the optimal substitution model for each locus (-m TESTONLY) first and then analyzed the partitioned, concatenated alignment with 1000 ultrafast bootstrap replicates. We generated a species tree with SVDQuartets in Paup v4.0a166. We analyzed all quartet possibilities (n = 16,356 quartets) and assessed node support with 100 bootstrap replicates.