|dc.description.abstract||The role of behavioural ecology in the conservation of species and biodiversity remains poorly understood. Observational and experimental studies of bird song, using a biogeographic approach, provide an opportunity to address this issue. Here I test hypotheses on the emergence and divergence of song cultures related to historical as well as more recent population fragmentation events. The North Island kōkako (Callaeas wilsoni) is a duetting songbird endemic to New Zealand that was once widespread but, due to habitat loss and introduced predators, has become endangered. These birds are sedentary and monogamous with poor flying abilities. Therefore, all of the c. 1500 surviving individuals are marooned in 13 scattered surviving populations and 11 translocated populations. The historical fragmentation of populations, the kōkako’s territorial nature and capacity to learn songs, and the establishment of new populations through translocation make this species a good candidate for studies on the emergence of song cultures, social divergence and landscape effects on conservation from a behavioural perspective.
Song sharing and comparisons of acoustic characteristics were used to reconstruct associations of song traditions and acoustic traits at both macrogeographic and microgeographic levels in the six largest surviving populations of kōkako (Chapter 2 and 3). Distinct song cultures or dialects were discovered in all populations examined. However, patterns of vocal similarity observed were not correlated with historical fragmentation patterns or linear geographical distance (Chapter 2). Instead, rapid vocal drift following fragmentation and isolation as well as founder effects linked to population contraction may explain current macrogeographic song variation (Chapter 3). Nevertheless, a fragmentation effect on song similarity, generating song repertoire divergence, was observed at a microgeographic scale within discontinuous populations (Chapter 2). In addition, juvenile dispersal across portions of tenuously connected habitat, which have strikingly different song cultures, suggests that kōkako learn songs after dispersal (Chapter 2). Post-dispersal learning may explain the lack of consistency between dialect membership and genetic relatedness among surviving populations. Dialect formation may occur as different cultural trajectories are forged by fragmentation and isolation; subsequent song convergence may then occur due to social selection or chance events.
The effects of population size on song attributes, repertoires and sharing, vocal versatility and syntactical structure were examined within the six largest natural populations, as well as two translocated populations of kōkako (Chapter 3). Repertoire size, song diversity and the predictability of syntactical structure were positively correlated with population size. Song traditions and the rate of cultural evolution may depend on social interaction and population size, as the neutral model of song microevolution predicts. Smaller populations of kōkako, with reduced social stability under relaxed selection pressures, may resemble small colonizing groups undergoing founder effects and vocal drift leading to loose syntax, consistent with the ‘withdrawal of learning’ hypothesis. Furthermore, social modulation of vocal behaviour and song traditions are discussed in the context of reproductive success and adaptation to differing social conditions as in the case of translocation events (Chapter 3).
Recently isolated translocated populations exhibit remarkable acoustic divergence and reduced song sharing with the corresponding source population (Chapter 4). Translocated populations showed relatively low song diversity and increased song sharing as well as acoustically distinct song features. To investigate how meaningful this vocal divergence was for individuals in source and translocated populations, a reciprocal playback experiment was performed to examine vocal and approach responses to simulated local and non-local intruders. Responses to simulated intruders differed among populations but there was little evidence of discrimination based on stimulus type. The degree of song divergence (i.e., repertoire and acoustic features) and vocal response to playback relative to the source population was greater in an older translocated population than a more recently translocated population (Chapter 4). Consistent with the consequences of founder effects and vocal drift, these results demonstrate that kōkako populations vocally diverge in isolation following the establishment of a population with a vocal subset and subsequently a different cultural trajectory (in accordance with the ‘withdrawal of learning’ hypothesis). Cultural erosion in small translocated populations may affect their viability as reduced vocal virtuosity can be associated with a reduced capacity for territory defense and mate attraction, and consequently low reproductive success. Song divergence can ultimately have important implications for our understanding of the evolution of animal societies and our ability to conserve animal populations.||