Geographic variation in the whistle vocalisations of dolphins has previously been reported. However, most studies have focused on the whole whistle repertoire, with little attempt to classify sounds into biologically relevant categories. Common bottlenose dolphins (Tursiops truncatus) use individually distinctive signature whistles which are thought to help maintain contact between conspecifics at sea. These whistles may show a different kind of variation between populations than non-signature whistles. Here I investigate signature whistle use and variation in the two recognised species of bottlenose dolphins (T. truncatus and T. aduncus) from populations inhabiting the coastal waters of the North America, Scotland, South Africa, Tanzania, Japan, Australia and New Zealand, and one captive colony. I identified likely signature whistles (signature whistles types, SWTs) from acoustic recordings by combining two novel techniques: automated contour categorisation in ARTwarp (Deecke and Janik, 2006) and a specific bout analysis based on the timing of signature whistle production in T. truncatus termed SIGID (Janik et al. in press). Three ways of categorising the contours were tested and between 87 and 111 SWTs were identified in total. Repeated emissions of stereotyped contours were apparent in the repertoire of all T. aduncus populations using both automated and human observer categorisation, providing good evidence for signature whistle use in this species. There was significant inter-specific variation in the frequency parameters, looping patterns and duration of SWTs. Inflection points, duration and measures of SWT complexity showed high variation within populations, suggesting inter- and intra-individual modification of these parameters, perhaps to enhance identity encoding or convey motivational information. Using 328 bases of the mtDNA control region, I found high levels of population differentiation (FST and φST) within the genus Tursiops. These data do not support a link between mtDNA population differentiation and variability in call type. Instead, morphological variations at the species level, and learned differences at the population level, better explain the variation found.

Bottlenose dolprlin (Tursiops trunratus) whistles are currently studied by subjective visual comparison of whistle spectrograms. This thesis describes the novel use of stochastic modeling to automate the comparison of dolphin whistles and to yield an objective, quantitative measure of whistle similarity. The relationship of bottlenose dolphin whistle production to a model of human speech production is discussed, providing a basis for the use of human speech recognition techniques for creating whistle models. Discrete hidden Markov models based on vector quantization of linear prediction coefficients are used to create whistle models based on statistical information derived from a sample set of dolphin whistles. Whistle model comparison results are presented indicating that evaluation of bottlenose dolphin whistles via hidden Markov modeling provides an objective measure of similarity between whistles. The results also demonstrate that hidden Markov models provide robustness against the effects of temporal and frequency variance in the comparison of whistles. The extensibility of stochastic modeling techniques to other animal vocalizations is discussed and possibilities for further work in areas such as the determination of possible structural components, similar to phonemes in human speech, is provided. / Master of Science

Thesis (Ph. D.)--Joint Program in Biological Oceanography (Massachusetts Institute of Technology, Dept. of Biology; and the Woods Hole Oceanographic Institution), 1999. / Includes bibliographical references. / Odontocetes are assumed to use echolocation for navigation and foraging, but neither of these uses of biosonar has been conclusively demonstrated in free-ranging animals. Many bats are known to use echolocation throughout foraging sequences, changing the structure and timing of clicks as they progress towards prey capture. For odontocetes, however, we do not know enough about their foraging behavior to describe such sequences. To conduct detailed behavioral observations of any subject animal, the observer must be able to maintain continuous visual contact with the subject for a period commensurate with the duration of the behavior(s) of interest. Behavioral studies of cetaceans, which spend approximately 95% of their time below the water's surface, have been limited to sampling surface behavior except in special circumstances, e.g. clear-water environments, or with the use of technological tools. I addressed this limitation through development of an observation platform consisting of a remote controlled video camera suspended from a tethered airship with boat-based monitoring, adjustment, and recording of video. The system was used successfully to conduct continuous behavioral observations of bottlenose dolphins in the Sarasota Bay, FL area. This system allowed me to describe previously unreported foraging behaviors and elucidate functions for behaviors already defined but poorly understood. Dolphin foraging was modeled as a stage-structured sequence of behaviors, with the goal-directed feeding event occurring at the end of a series of search, encounter, and pursuit behaviors. The behaviors preceding a feeding event do not occur in a deterministic sequence, but are adaptive and plastic. A single-step transition analysis beginning with prey capture and receding in time has identified significant links between observed behaviors and demonstrated the stage-structured nature of dolphin foraging. Factors affecting the occurrence of specific behaviors and behavioral transitions include mesoscale habitat variation and individual preferences. The role of sound in foraging, especially echolocation, is less well understood than the behavioral component. Recent studies have explored the use of echolocation in captive odontocete foraging and presumed feeding in wild animals, but simultaneous, detailed behavioral and acoustic observations have eluded researchers. The current study used two methods to obtain acoustic data. The overhead video system includes two towed hydrophones used to record 'ambient' sounds of dolphin foraging. The recordings are of the 'ambient' sounds because the source of the sounds, i.e. animal, could not be localized. Many focal follows, however, were conducted with single animals, and from these records the timing of echolocation and other sounds relative to the foraging sequence could be examined. The 'ambient' recordings revealed that single animals are much more vocal than animals in groups, both overall and during foraging. When not foraging, single animals vocalized at a rate similar to the per animal rate in groups of>=2 animals. For single foraging animals, the use of different sound types varies significantly by the habitat in which the animal is foraging. These patterns of use coupled with the characteristics of the different sound types suggest specific functions for each. The presence of multiple animals in a foraging group apparently reduces the need to vocalize, and potential reasons for this pattern are discussed. In addition, the increased vocal activity of single foraging animals lends support to specific hypotheses of sound use in bottlenose dolphins and odontocetes in general. The second acoustic data collection method records sounds known to be from a specific animal. An acoustic recording tag was developed that records all sounds produced by an animal including every echolocation click. The tag also includes an acoustic sampling interval controller and a sensor suite that measures pitch, roll, heading, and surfacing events. While no foraging events occurred while an animal was wearing an acoustic data logger, the rates of echolocation and whistling during different activities, e.g. traveling, were measured. / by Douglas Paul Nowacek. / Ph.D.