Changes in signaling repertoires across species allow for insight into the macroevolutionary forces that control signaling systems. Signaling systems are theorized to be affected by both the social and ecological environments of species. With respect to social variables, increased social complexity is thought to lead to increased vocal complexity. Although ecology can affect signaling systems in numerous ways, one potential effect of ecology is that more cluttered habitats should lead to greater reliance on non-visual (e.g. vocal) signals. To test these concepts on a macroevolutionary scale, we compiled a large dataset of avian vocal repertoires. We amassed vocal repertoires for 821 species of birds and for many of these species categorized their vocalizations into usage categories (e.g. alarm, contact). To analyze the social and ecological forces that act on repertoire evolution, we incorporated datasets with several social variables (e.g. cooperative breeding and length of social bond), and included data on the habitat and foraging behaviors of species within the dataset. We used Bayesian phylogenetic analysis to test for potential relationships within the data. We found that cooperative breeding was a significant predictor of larger repertoire size in birds; we also find several, more targeted effects. For instance, foraging strata affected repertoire size and repertoire composition. In sum, we find considerable evidence that social features affect repertoire size while certain ecological variables have more targeted effects on vocal repertoires.

To develop the dataset we relied heavily on data within the Birds of North America (BNA)(2018). We follow Bradbury and Vehrencamp (1998)’s definition of functional repertoire size to compile repertoires. Briefly, a functional repertoire represents sounds associated with distinct behavioral contexts. The functional repertoires for virtually every species from BNA is included in the current dataset, and paired with previously published data from Handbook of Australian, New Zealand, and Antarctic Birds (HANZAB)(1998) and the primary literature (Leighton, 2017). Both of the secondary sources (BNA and HANZAB) are curated volumes of species information for their respective geographic locations. For species from the primary literature, journal articles focused specifically on defining vocal repertoires were collected, and vocalizations given in specific behavioral contexts were described as a functional repertoire; in addition to the species identified from studies in the primary literature, species from HANZAB and BNA closely related to primary literature were added (see Leighton (2017) for details). Comparative analyses (see below) on overall repertoire size were performed on all species; given the differences between sources in repertoire size, we included data source as a fixed factor in all models to control for differences between the sources. We therefore control for differences in repertoire size between primary and secondary sources.