Coevolutionary maintenance of forked tails and song in hirundines (Aves: Hirundininae)

Both conspicuous plumage ornamentation and song are well‐known examples of sexually selected traits, but their interrelationship is not well known, perhaps in part because of confounding factors, including interspecific variation in ecology, habitat, morphology, and type of ornamentation. Here, using a phylogenetic comparative approach and the 69 species with available information (i.e. 96% of all 72 species), I examined the evolutionary relationship between forked tails and the presence/absence of song in hirundines (Aves: Hirundininae). Hirundines have similar ecology (e.g. aerial insectivores, social monogamy, and biparental provisioning), morphology (e.g. syrinx with nearly complete bronchial rings), and plumage ornamentation (i.e. a sexually selected forked tail), which provides a unique opportunity to examine the evolutionary associations between plumage ornamentation and song. In particular, hirundines have repeatedly lost their ornamentation, forked tails, setting up a condition to test their association with the evolutionary gain/loss of their simple song. After controlling for phylogeny and covariates, I demonstrated that song was less likely to be found in species with forkless tails than in species with forked tails. Two correlates of tail shape, sexual dimorphism in the overall plumage characteristics as a well‐known proxy of sexual selection and incubation type as a measure of extrapair mating opportunity, were not positively related to the presence or absence of song, indicating the importance of forked tails, rather than their correlates. The analysis of the correlated evolution of discrete characters further supported the correlated evolution of the two traits, in which forked tails and song are maintained together and less likely to be lost under the presence of each other. The current study provided macroevolutionary support for the integrated use of visual and acoustic courtship traits.