The evolutionary dynamics of learned bird song

Learned behaviors are fascinating to study, but it is often difficult to compare the evolution of these behaviors across many species, primarily because it is challenging to measure aspects of complex behaviors in a standardized way for different species. This set of studies proposes to use new computational tools developed in the investigator’s lab to analyze bird songs recorded in nature and extract quantitative features that are comparable across all songbirds, such as syllable and song duration. The first study will explore whether different features of song evolve quickly or slowly, or at different rates in different groups of birds. By producing and analyzing new datasets of song information, this study will test evolutionary hypotheses about the role of selection pressures on song evolution. For example, this study will test the hypothesis that song features that have been linked to female preferences will be the song features that evolve more quickly. The second study will generate new metrics to quantify the temporal patterns of birdsong and test the hypothesis that rhythm, a song trait potentially evolving under natural selection, will show correlated evolution with other song traits and life-history traits. The third study will quantify whether closely related birds can be distinguished by their songs to determine how songs have changed over the course of relatively recent speciation events. If song evolved at a steady rate over evolutionary time, then song distinguishability between species would positively correlate with genetic distance. Alternatively, if song plays a role in reproductive isolation, then sister species with overlapping geographic ranges will evolve distinguishable songs more quickly. This finding would suggest that song differentiation accumulates more quickly when species have the potential to hybridize. The researchers will have ongoing interactions with local birders and community scientists and will launch educational modules aimed at educating middle-schoolers in evolution and conservation using birdsong as a readily observed example. In addition, the investigators will add to the computational training of underrepresented students preparing for graduate school by partnering with the Fisk-Vanderbilt Masters-to-PhD Bridge program. Birdsong is unique in how densely sampled it is through community science efforts, but large-scale analyses have remained challenging due to a lack of flexible tools for processing these recordings. Using new tools developed to analyze a broad taxonomic sample of songbirds, this study will quantify song features from numerous individuals per species and analyze the dynamics of song evolution over long timescales. With new cross-species metrics, this research also aims to quantify rhythm and analyze temporal patterns in birdsong across oscine songbirds. Finally, these studies will assess the potential role of learned song in reproductive isolation between sister species, thus assessing whether learning appears to facilitate speciation. Together, these studies will advance knowledge about the evolution of behavior on multiple timescales, improving our understanding of the role of learning in evolutionary processes. The proposed studies will also further develop computational tools to facilitate the analysis of vocalizations by the broader research community. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.