Independent evolution of song diversity and song motor performance in canaries, goldfinches and allies indicates clade‐specific trade‐offs in birdsong

Abstract

The diversity and the motor performance of birdsongs can both be sexually selected. In wood warblers, most species with high motor performance sing a greater proportion of trills, presumably to advertise performance, and thus have lower syllable diversity. We tested if this trade‐off between motor performance and syllable diversity extends to canaries, goldfinches and allies, a clade with much longer and more varied songs. We assembled a molecular phylogeny and inferred song motor performance based on the speed of frequency modulation either in trills or in within‐song intervals. The two metrics of performance were positively, but only mildly, related across species. While performance evaluated in intervals had high phylogenetic signal, performance evaluated in trills changed independently of phylogeny and was constrained by body size. Species in densely vegetated habitats sang fewer trills, but did not differ in motor performance. Contrary to wood warblers, song motor performance did not predict the proportion of trilled syllables nor within‐song syllable diversity, perhaps because large differences in the song duration of canaries, goldfinches and allies prevent trills from severely compromising syllable diversity. Opposed results in wood warblers and in these finches indicate the existence of clade‐specific trade‐offs in the evolution of birdsong.

Anthropogenic noise, song, and territorial aggression in southern house wrens

“…trill rate) and frequency bandwidth of the trill in many bird species (Podos et al 2009) including the house wren (Cramer 2013b). Here, we measured the double quantile vocal deviation (DQ vocal deviation) to characterize vocal performance (Cardoso 2019, Cardoso et al 2020). The DQ vocal deviation was measured separately for spontaneous trills and trills produced in response to the playback.…”

“…trill rate), and it is robust to small sample sizes (Cardoso 2019). Higher values of DQ vocal deviation indicate higher vocal performance, and lower values of DQ vocal deviationindicate lower vocal performance (Cardoso et al 2020).…”