Content and efficacy-based selection on signal structure: dogs and dingoes as canid models

Animals' signalling solutions come in all shapes and sizes. These designs are the result of selective pressures acting at the content (i.e. 'what' is conveyed) and the efficacy (i.e. 'how best' is it conveyed) level. Understanding why and how these pressures have impacted a signal's structure can help shed light on the evolutionary history of complex signals and thus complex communication systems. In this thesis, I explore the content and efficacy of three canid signals, which vary in terms of their contexts of production, forms and functions. I demonstrate that domestic dogs, Canis familiaris, produce multimodal aggressive signals, composed of at least acoustic and visual components, which convey redundant information. Then, I investigate the content of dingo, C. f. dingo, howls demonstrating that these signals encode identity cues, which allow for class-level recognition of familiarity levels. In addition, dingo howls may convey sex cues, although it is unknown whether this potential information is salient to receivers. Propagation tests confirm that these signals are effective in long-range communication. Finally, I investigate the content and efficacy of dingo bark-howls, which are multicomponent vocal signals, composed of a bark and a howl segment. I show that dingo bark-howls are individually distinctive and that these individual signatures are shared with howl vocalisations. Furthermore, bark howlsare effective medium-range signals, regardless of environmental conditions.Using a playback experimental approach, I test the hypotheses that bark-howls function as alarm signals and that each component has a different role. Results support the hypotheses, and specifically, indicate that barks serve to alert conspecifics while howl segments allow for familiar-unfamiliar discrimination. I discuss these findings in terms of context-specific selective pressures, which can give rise to different, more-or-less 'complex' signals. I then expand the discussion,proposing hypotheses for why we find some signals produced across contexts and discuss future research avenues.