Sounds beneath the surface : a multiple-approach study of bottlenose dolphin acoustic communication

Abstract

Bottlenose dolphins (Tursiops truncatus) live in complex individualized societies that combine stable social units with fluid groups, and they largely rely on sound to interact with one another. This thesis applies a suite of approaches to study their acoustic communication. I studied the wild population of Sarasota, which offers long-term behavioural data and opportunities to deploy sound-and-movement recording tags. Bottlenose dolphins display a wide call repertoire, including graded sounds challenging to classify. I use dolphin social interactions to present a novel tag-based approach for studying communicative roles of call parameters changes. Applying continuously-sampled parameters, this approach examines how individuals change movements as a function of signal features. I focus then on signature whistles, sounds encoding identity information that in Sarasota are documented for most individuals. Signature whistles function as individually-distinctive contact calls, for instance between closely-bonded animals during separation, but their function between groups is poorly documented. Analysing Dtag data from instances of group encounters, I show that dolphins did not regularly produce signature whistles upon detecting another group, nor systematically before joining; instead, signatures appeared to be used strategically depending on the encountered individuals. Dolphins sometimes imitate the signature whistle of others, which is thought to function for addressing known individuals. Performing first-ever playbacks of natural signature whistle copies, I show that free-ranging subjects turned more frequently towards the playback upon hearing copies of their signature vs signature whistles of others, supporting the addressing function of copies. Signature whistle copies are often produced right after the subject’s signature, in so-called vocal matching interactions. Using playbacks with temporarily-captured subjects, I test whether signature whistle matching, or vocal matching per se, has an addressing signalling role. Subjects turned more times and produced their signature more frequently in response to signature whistle matching vs matching of other calls, supporting the former hypothesis.