Heterospecific song quality as social information for settlement decisions: an experimental approach in a wild bird

Assessing local habitat quality via social cues provided by conspecific or heterospecific individuals sharing the same needs is a widespread strategy of social information use for breeding habitat selection. However, gathering information about putative competitors may involve agonistic costs. The use of social cues reflecting local habitat quality acquired from a distance, such as acoustic cues, could therefore be favoured. Bird songs are conspicuous signals commonly assumed to reliably reflect producer quality, and thereby local site quality. Birds of various species have been shown to be attracted to breeding sites by conspecific and heterospecific songs, and to use conspecific song features as information on producer (and by extension habitat) quality. Whether they can do the same with heterospecific song features, and whether this depends on the individual's own phenotype and especially its competitive ability, remains unknown. We used a playback experiment in a wild population of collared flycatchers, Ficedula albicollis, a species known to eavesdrop on the presence and performance of dominant great tits, Parus major. We tested whether flycatchers, whose aggressiveness was experimentally assessed, preferred to settle near playback of a high-quality great tit song (i.e. song with large repertoire size, long strophes, high song rate), a low-quality great tit song or a chaffinch song (control). Among old females, aggressive ones preferred to settle near playback of high-quality tit song and avoided playback of low-quality tit song, while less aggressive females preferred to settle near playback of low-quality tit song. Male personality or age did not influence settlement decisions. This shows that collared flycatcher females use great tit song quality features as information for settlement decisions, although this depended on their own competitive ability and/or previous experience with great tit songs. Our study therefore further illustrates the complex condition-dependent use of heterospecific social information for breeding habitat selection.