The ecology of vocal mimicry in the superb lyrebird, Menura novaehollandiae

Abstract

The ability of some species of bird to accurately mimic the songs and calls of other species is one of the most spectacular but puzzling forms of animal communication. In this thesis I review existing work on evolution of avian vocal mimicry before presenting a detailed study of the ecology of mimicry in an oscine passerine: the superb lyrebird, Menura novaehollandiae. Few studies investigate avian vocal mimicry, and research has been hindered by conceptual confusion. However, adopting a modified version of Vane-Wright's (1980) functional definition of mimicry will provide a more practical and coherent framework for future enquiry. Encouragingly, there is increasing evidence that vocal mimicry can function deceptively, although some of the most elaborate mimetic forms remain unexplained. Clarifying the ultimate and proximate causes of vocal mimicry requires closer engagement with signal theory and further empirical work on the ecology of avian vocal mimics. My study shows that male superb lyrebirds partition their large repertoire of mimetic song types between two contrasting sexual displays. Mimicry produced during 'recital' displays, when males were perched and visually inconspicuous, was highly varied and included imitations of many other species' songs. However, only seven sounds were imitated during 'dance' displays, six of which were alarm calls. Such context-dependent mimicry indicates that mimicry in lyrebirds has multiple functions. Male lyrebirds are highly accurate but imperfect mimics of the complex songs of the grey shrike-thrush, Colluricincla harmonica. During playback experiments, shrike-thrushes rarely differentiated between model and mimetic song but could integrate contextual information with differences in signal structure to distinguish between the two. Acoustic analyses showed that lyrebirds sang fewer repetitions of individual element types, suggesting a trade-off between demonstrating both mimetic accuracy and versatility. Thus, evaluating the similarity between model and mimetic sounds illuminates signal discrimination by models and the evolutionary forces shaping mimetic 'recital' song in lyrebirds. Mimicry is integrated within an unusually complex display. During dances, lyrebirds coordinated a song type repertoire containing mimetic and lyrebird-specific songs, with a repertoire of display movements unnecessary for vocal production, so that specific dance movements were associated closely with specific song types. Thus, lyrebirds produced a display of a level of complexity previously only associated with humans. Dance mimicry consists of a remarkable acoustic illusion of a mixed-species mobbing flock. This behaviour may have evolved to: 1) reduce the risk of predation during terrestrial displays; or 2) manipulate an anti-predator response in the female in order to prolong copulation. Consistent with both these hypotheses, a playback experiment showed that mimicry of a mixed-species mobbing flock attracted small passerines just as often as a recording of an actual mixed species mobbing flock. Hence, dance mimicry is highly accurate and can deceive heterospecific passerines. This study provides evidence that intense competition for mates among male lyrebirds has selected for both deceptive and non-deceptive vocal mimicry in dance and recital displays respectively. These results highlight the sophistication of mimetic forms and function. Despite centuries of observing nature, the diversity and complexity of animal signals still continues to surprise. - provided by Candidate.