Studies of whistle development and perception
By Laela Sayigh, PhD, Woods Hole Oceanographic Institution and University of North Carolina, Wilmington, and Vincent Janik, PhD, Sea Mammal Research Unit, St. Andrews, Scotland

      We continue to work toward creating a fully digital whistle database that will be accessible to other researchers. During each capture-release session, high-quality recordings are obtained, making the Sarasota Whistle Database the most comprehensive of its kind in the world. Many individuals have been recorded on multiple occasions, over periods of many years, enabling us to examine questions about dolphin communication that are not possible elsewhere. This year, we used this database to address a claim made in a recent paper that signature whistles do not exist (McCowan & Reiss 2001; Animal Behaviour 62:1151-1162). In our study, we randomly selected 20 whistles from each of 20 randomly selected dolphins in the Sarasota Whistle Database, and asked judges to group whistles according to their overall contour, or pattern of frequency changes over time (Figure 1). Although the judges had no knowledge about the number of dolphins present in the sample, they consistently grouped together whistles produced by the same dolphin (an average of 18.9 out of 20 whistles of a given dolphin were grouped together, and an average of only 0.5 out of 380 possible whistles of other dolphins were included in these groups). Our paper, entitled “Facts about signature whistles of bottlenose dolphins” was published in the journal Animal Behavior, and provides unequivocal evidence for the production of individually distinctive signature whistles by bottlenose dolphins.

Currently, we have several avenues of research underway:

1) We are carrying out playback experiments during capture-release to find out whether dolphins can recognize each other by means of voice cues. Experiments that have been described in previous issues of Nicks’n’Notches demonstrated that dolphins can recognize the frequency modulation patterns of the signature whistles of other individuals, but we do not know if they also can recognize animals by the tone of their voice. We are playing back non-signature whistles to dolphins to see whether they recognize who produced them.

2) We are also using playback experiments to look at how dolphins react when their own signature whistle is copied by someone else. Wild dolphins sometimes mimic the signatures of others, presumably to address or find a specific animal they are looking for. In this study we are looking at how dolphins react to a playback of a copy of their own whistle. This may help us understand whether dolphin signature whistles are used like human names to address individuals.

Figure 1. Examples of signature whistles of ten different individual bottlenose dolphins recorded during temporary capture in Sarasota, Florida. Spectrograms were made with AVISOFT SASLab Pro, at a sample rate of 80,000 Hz and a 256 point FFT.

3) We are studying the factors influencing signature whistle development in bottlenose dolphin calves. Bottlenose dolphins are unusual among mammals in that they learn their individually distinctive signature whistles. We have recorded signature whistles of 111 bottlenose dolphin calves during brief capture-release events in Sarasota Bay, Florida since 1975. Social association data, consisting of at least 10 sightings during the first year of life, are available for 74 of these calves. We are assessing similarity of the calves’ whistles to their mothers, siblings, close associates, and infrequent associates using quantitative comparisons. Visual assessment of whistles from 99 mother-calf pairs has found that 34% of calves had signature whistles similar to those of their mothers. Of the remaining calves, some still shared acoustic features with the mother, such as non-linear elements or constant frequency portions. Several calves produced whistles similar to those of older siblings. We are examining a variety of factors, such as sex, birth order, number of associates, and social association patterns of the mother as possible influences on vocal development. Our results show that more than half of all calves develop whistles that show little resemblance to the mother’s whistle, but that mothers and siblings are often important vocal models to their calves.
      This work was funded by a grant to L. Sayigh and R. Wells from Harbor Branch Oceanographic Institution’s Protect Wild Dolphins Program, and a Royal Society University Research Fellowship from the UK to V. M. Janik.