Whistled Vocal Behavior in Dolphins A research paper for Prof. Heidi Harley's Marine Mammal Cognition By Julia Orth, New College, 10 May 2000.

Ever since the time of the ancient Greeks, intriguing anecdotes have fueled interest in dolphins' use of sounds. Pliny the Elder wrote, in his Natural History (IX, 8, 24), "The dolphin is an animal that is not only friendly to mankind but is also a lover of music, and it can be charmed by singing in harmony, but particularly by the sound of the water organ." This belief in dolphins' interest in human music is made clearer by another story that tells of a musician named Arion. I shall retell the most common variant of this story here. Arion of Methymna was known throughout the land for his skill in music. He lived at the court of King Periander (or "Pyranthus," ca 625-585 BC) in Corinth, but one day decided to sail across the sea to Italy and Sicily. Once there he became rich by his skills, and at last desired to return home. He chose a crew of Corinthians and set out from Tarentum. When the sailors saw what treasures he had, they conspired to dispose of him and take them for their own. Arion begged that they grant him a last request; he wanted to sing and play one more song before throwing himself to the waves. The pirates quicky consented, pleased at the chance to hear this musician who was second to none. Arion plucked sweet notes from his lyre, and then hurled himself, in full costume, into the sea. It is said that this music attracted a dolphin, who saved him from the depths and carried him to Taenarum, in Laconia. There he went ashore and made his way to Corinth and Periander's court. Periander doubted his story, and hid him away, sending for the sailors. They told that Arion had chosen to remain in Italy, and he was doing well. At that, Arion came before them, and the astonished sailors could no longer deny their guilt.

This story was said to be true, and the coins of Methymna, in Lesbos, depicted him astride a dolphin. (Lilly & Montagu, 1963)

This seems like no more than a pretty story, but tales from more recent years suggest that there may be more to it than meets the eye. Modern interest in the acoustic abilities of dolphins was rekindled in part by the writings of a Dr. John Lilly. Dr. Lilly was a neurologist who became interested in dolphins, and acquired some of his own for experimentation. He became fascinated by what he claimed was dolphin mimicry of human speech (1965). After a few years had passed he'd become confident that dolphins have a language of their own, which was termed "delphinese." (Lilly, 1961, 1978) His books were widely read, and fueled the current public interest in dolphin communication.

Though "delphinese" as Lilly claimed it is far-fetched and nearly impossible to verify, sound does fulfill an important social role for dolphins. Tales more true to that of Arion are still told, of dolphins that are not only attracted by human music, but that seek out musical interactions with humans. Orca researcher Paul Spong, working with captive orcas (Orcinus orca) at Vancouver Aquarium, claims to have successfully used music to reward a lethargic orca named Tung Jen for activity. Tung Jen was, he reported, sufficiently motivated by novel sounds that he began to move and jump around the tank instead of remaining still at the bottom (Weyler, 1986).

In another such anecdotal case, a pod of wild orcas returned night after night at 10:30 PM to interact with a musician on a boat equipped with an audio set-up through which sounds from the boat were broadcast into the water and sounds from the water broadcast into the boat. (Nollman, 1987) The more interesting of these interactions include one in which one of the orcas appeared to break a whistle-type-sound into pieces, repeating them seperately and then putting them back together (Nollman, 1991). Similar curiousity and manipulation of sound was said to occur in interactions with a group of wild Atlantic spotted dolphins (Stenella attenuata). A crew of individuals who were aquainted with this pod, and who had been interacting with it for over a year, designed a small machine that made sounds similar to dolphins' whistles. They brought this underwater and pressed the various buttons to produce the different sounds. The dolphins were attracted by this and came over to investigate, and quickly, by the crews' account, began not only repeating the computer-generated sounds, but also modifying them and adding things to them. (Jones, 1984)

Anecdotes such as these suggest a number of interesting things: that at least some species of dolphins are motivated by novel sounds even after initial investigation of them, that they are capable of mimicry of some sounds or some aspects of sounds, and that they are capable of holding some features of a sound constant while manipulating others. Verification of these possibilities is in the province of scientific research, and fortunately, a body of research addressing these questions has begun to accumulate. I shall now proceed to review some of the preliminary answers these studies have turned up with respect to the question of dolphins' vocal abilities.

Imitation of computer-produced whistle sounds has been described under controlled circumstances, as well as the informal settings described above. One group of researchers in Haawai'i trained a female bottlenose dolphin (Tursiops truncatus), known as Akeakamai or Ake, to mimic computer generated whistles using operant conditioning. After it was demonstrated that she could faithfully mimic computer-generated whistles, and had a concept of mimicry such that she mimicked auditory features that were novel or that she was not specifically being rewarded to mimic, an attempt was made to teach her computer-generated whistle "labels" for five familiar objects. The first method used was to present an object and then present the sound associated with it. As she learned the sounds, they were gradually faded out, the idea being that eventually the sounds would be phased out entirely and Ake would be producing the labels on her own. Unfortunately for this method, Akeakamai mimicked the fading of the sounds as well. When this problem was recognized, the method was changed so that the sound was always presented at the same intensity, but that as time went on it was paired less and less frequently with its object. Akeakamai learned the five sounds this way. Having learned them, she was presented with 167 trials where she was presented with an object and given a command for labeling it; a blind observer categorized her responses. 91% of her responses were correct, and all of her errors but one involved confusion of two particular sounds, those for "person" and "frisbee" (Richards, Wolz, & Herman 1984).

A study at Marineland of Florida utilized a contrasting method, involving an underwater "keyboard" with keys composed of various plastic shapes. Instead of being trained, the dolphins were presented with the keyboard and allowed to push the keys at will. The activation of any given key resulted in a distinct computer-generated whistle and then a particular object or event (a ball, a ring, a fish, &c.) provided by one of the human experimenters (the keyboard was only placed in the water when experimenters were present to fulfill this task!). The computer generated whistles were designed to be similar to, yet distinct from, the dolphins' normal whistles, so that they could be imitated if the dolphins so chose, and imitation could be recognized by the experimenters if it occurred. Two young, male Atlantic bottlenose dolphins were the primary users of this system. The dolphins were found to quickly begin imitating these sounds--the first mimicry occurred after only 13 exposures to the target sound, another after only 9. A later sound was mimicked after only 2! The dolphins began by mimicking the end of the whistle, then the beginning, and then the whistle as a whole, bringing to mind the "breaking up" of whistle-phrases by the wild Vancouver orcas interacting with the musician. These dolphins also appeared to combine whistles, running the whistles associated with "ring" and "ball" into each other. A majority (ranging from 73% - 100%) of the dolphins' productions of associated whistles were emitted during appropriate activities. That is, the "ball" whistle when playing with the ball, the "rub" whistle during instances of physical contact between the dolphin and an experimenter, and the novel "ring-ball" combination during simultaneous play with the ring and the ball (Reiss & McCowan, 1993).

Research on both captive and wild bottlenose dolphins has shown that individuals have unique, stereotyped vocalizations, known as "signature whistles." This phenomenon was first noticed and described by scientists Melba and David Caldwell, who claimed that they made 90% of dolphins' vocalizations (1965). Signature whistles have been shown to remain stable for up to at least 12 years (Sayigh et al., 1990). Mimicry of these whistles has been known to occur (Tyack, 1986), but many studies have found imitation of signature whistles to be rare (Janik & Slater, 1998; Smolker, Mann, & Smutts, 1993).

Bottlenose dolphin calves in captivity have been observed to develop signature whistles by around 6 months of age (Caldwell & Caldwell, 1979). Wild studies have refrained from capturing dolphins this young out of concern for their safety, but all wild dolphins investigated had developed a stable signature whistle by the age of one year (Sayigh et. al., 1990).

As part of a long-term study of the free-ranging bottlenose dolphin population around Sarasota, Florida, the signature whistles of several mothers and offspring were recorded. The signature whistles of male offspring were much more likely to resemble their mother's whistle than the whistles of female offspring were; the signature whistles of female offspring were more likely to be different. The researchers in this study hypothesize that this sex difference is related to the fact that female dolphins tend to continue to associate frequently in a common group with their mothers after reaching adulthood, where male dolphins join separate groups. The females then would need whistles that could not easily be confused with those of their mothers, since both will remain in proximity throughout their lifespans, whereas the males do not face such a problem (Sayigh, et al, 1990, 1995).

There is clear potential for the use of signature whistles as a method of identifying individuals. Dolphins must address an interesting dilemma in being highly social mammals in a dangerous ocean environment (Norris, 1994). Schooling type behaviours and lack of differentiation are beneficial, yet the complex social structure and plasticity of dolphin behavior necessitate individuality as well. Signature whistles may provide a way of marking individuals to others in the group without necissitating physical differences. They may provide further information on top of that as well.

To address the question of whether dolphins do in fact recognize signature whistles as demarking a particular individual, a playback study was conducted with wild dolphins. Wild bottlenose dolphins from a well-known group were temporarily caught, and their reactions to recordings of signature whistles of offspring or parents vs. similar, familiar non-kin were analyzed. The dolphin subject was held in place and presented to one side with an audio recording. The recordings presented were of independent offspring and of a familiar individual of similar characteristics as the offspring for mother dolphins, and of the mother and a familiar individual of similar characteristics as the mother for the independent offspring. Individuals' whistles were carefully chosen to control for the fact that the subject may simply respond to a whistle s/he had heard more often. The subjects' responses to the recording were judged by how often they attempted to turn their heads towards the recording apparatus, and this was measured against a baseline. The dolphnis did in fact respond significantly more strongly to the kin whistles than to the familiar non-kin whistles (Sayigh et al. 1998).

In an attempt to determine the context in which signature whistles are actually used, a study was conducted wherein recordings were made of a group of four captive bottlenose dolphins. The dolphins were housed in two connected pools. Recordings from occasions when an individual had spontaneously separated from the group in an undisturbed context and entered the other pool were compared with recordings from occasions when the group was intact. It was found that signature whistles were most commonly produced when individuals were seperated from the group, both by the seperated individual, and by the remaining three members of the group. Signature whistles are otherwise rare in undisturbed contexts. (Janik, 1998) These findings suggest that signature whistles may serve to maintain group cohesion. It is interesting to not that signature whistle copying was noted very rarely in this study, though it could only be recorded when the copy was produced by a seperated individual, due to the difficulties in determining which individual was producing which whistle. In the five cases in which copying occurred, the signature whistle copied had been produced by a member of the group within a minute prior to the copy, and was produced again within a minute afterwards.

A different study suggests that signature whistles are used to initiate reunions between individuals, as well as to simply maintain cohesion. Wild bottlenose mothers and calves living around Shark Bay, Australia, were observed and recordings were made during seperations of the pair. Whistling occurred most often by calves during seperations, often beginning when the distance between the two was at its greatest, preceeding a reunion. Anecdotal evidence suggests that in some cases a signature whistle may prompt waiting behavior by the reciever, who will stop or slow down until the whistler catches up (Smolker, Mann, & Smuts, 1993). Yet another potential communicative use of signature whistles involves the imposition of context-specific information on them. Variations of their basic form may provide clues about the state of the individuals producing them. To test this hypothesis, a study was conducted comparing the signature whistle vocalizations of captive bottlenose dolphins under different circumstances. Vocalizations were recorded when the dolphins were isolated, when they were performing a trained discrimination task and had performed correctly, and when they were performing the discrimination and had just performed incorrectly. Frequency and timing were found to vary by these contexts, suggesting that some form of context-specific information is imbedded in signature whistles (Janik, Dehnhardt, & Todt, 1994).

Signature whistles have been a major focus of research on dolphin communication, and indeed, while they are clearly of great social importance to the dolphins, some researchers suggest that their centrality has been overemphasized. Drs. McCowan and Reiss categorized the whistles of three groups of captive bottlenose dolphins, and found that they had diverse whistle repertoires, and shared many whistles among them. There are several problems in researching dolphins' vocalizations. It is very difficult to tell which individual is producing a given sound. Sound travels 4.5 times faster in water than it does in air, making it impossible for humans to determine the direction of a sound underwater unaided. Dolphins do not make any obvious movements during sound production. Technological efforts are being made to alleviate this problem. One such method involves a telemtery device, termed a "vocalight", that is attached to a dolphin with a suction cup and lights a series of LEDs when that individual vocalizes (Tyack, 1991). Any observation of dolphins in the wild is difficult for a number of reasons, including the speed with which dolphins move over great distances, the difficulty of supporting humans in an aquatic environment, the necessity of not harassing the dolphins, the fact that most dolphin interactions take place below the surface, and the fact that individuals are often difficult to differentiate visually. Given these restrictions in study, one approach to attempting to learn about the potential content of dolphin vocalizations is to statistically analyze the vocalizations of a small group of captive cetaceans. A preliminary project using this technique was recently conducted. Vocalizations from captive bottlenose dolphins were collected for four developmental stages: less than a month old, 2 - 8 mo, 9 - 12 mo, and adult. These samples were analyzed using Zipf's relation, which measures the distribution of signals. If signals are repetitive or overly limited, the complexity of the information that can be conveyed is also limited. If the signals are too diverse, they have less potential organization, and the same message would be represented by different signals; a randomly generated sample of signals represents the furthest extreme of this diversity. Zipf's relation cannot, of course, measure the content of signals or the information contained in structural relationships of the signals, but it can provide an idea of the potential for a given system's ability to carry information. Analysis of vocalizations from all four groups of dolphins had regression coefficients ranging from -0.82 to -1.07, with older juvenile and adult vocalization regression factors at -0.95. Regression coefficients between -1.00 and 0.00signify higher diversity and randonimity, where coefficients lower than -1.00 signify higher repetition. A system with a regression coefficient of -1.00 is considered to have the most potential for communication. Analysis that examines higher entropic orders by measuring the probability that a given signal will follow another given signal may also be of use in attempting to learn more about the possible use of dolphin vocalizations. Such analysis can determine the potential internal complexity of a communication system; as when measuring the occurance of signals themselves, there is less potential for communication in a particular system if the relationships between signals tend to be either random or stereotyped. This type of analysis may be useful in examining dolphin vocalizations and other animal communication systems (McCowan, Hanser, & Doyle, 1999). (The use of Zipf's relation with respect to dolphin vocalizations has also been discussed by a pair of Russian researchers, in the same paper in which they claim to have found a structure to bottlenose vocalizations (Markov & Ostrovskaya, 1990).)

This review does not claim to cover evey piece of work that has been done on dolphin vocal behavior, but much work has been represented. Dolphins live in an auditory world, and appear to express great interest in novel sounds. They have the capacity to imitate unfamiliar tones and whistles, and are intrinsically motivated to do so even when not presented with a reward. Bottlenose dolphins possess distinct individual signature whistles which are used in times of distress or individual seperation from the group. Numerous theories exist as to the specifics of their significance, functioning, and content. In at least one population of wild bottlenose dolphins, males tend to produce signature whistles similar to those of their mothers' and females signature whistles that are different. Dolphins are known to modify aspects of their signature whistles, as they are capable of modifying aspects of other whistles while holding certain features, such as frequency contours, constant. Dolphins in a group seem to be able to recognize one another by their signature whistles, and are capable of imitating one anothers' signature whistles, although they do so only occasionally. Even the variety of ideas about signature whistle use may be too limited a lens through which to study dolphin vocalizations, as some researchers have suggested that a variety of other types of shared vocalizations are frequently produced by undisturbed groups. Statistical methods of analyzing this variety of sounds are being tested, but their usefulness will only be determined for sure when more is known about the actual function of vocalizations in dolphin societies.

References:

Caldwell, M. C., & Caldwell, D. K. (1965). Individualized whistle contours in bottlenosed dolphins (Tursiops truncatus). Nature , 207, 434-435.

Janik, V. M., Dehnhardt, G., & Todt, D. (1994). Signature whistle variations in a bottlenosed dolphin, Tursiops truncatus. Behavioural Ecology & Sociobiology, 35, 243-248.

Janik, V. M. & Slater, P. J. B. (1998). Context-specific use suggests that bottlenose dolphin signature whistles are cohesion calls. Animal Behaviour, 56, 829-838.

Jones, H. & Whitty, J. (Producers) (1984). In the Kingdom of the Dolphins. [Video Recording]. (Available from Home Vision, New York).

Lilly, J. C. (1961). Man and Dolphin.

Lilly, J. C. (1965). Vocal mimicry in Tursiops: Ability to match numbers and durations of human vocal bursts. Science, , 300-301.

Lilly, J. C. (1978). Communication Between Man & Dolphin. New York: Crown Publishers, Inc.

Markov, V. L. & Ostravskaya, V. M. (1990). Organization of communication system in Tursiops truncatus Montagu. In J. Thomas & R. Kastelein (Eds.), Sensory Abilities of Cetaceans (pp. 599-622). New York: Plenum Press.

McCowan, B. & Reiss, D. (1995). Quantitativecomparison of whistle repertoires from captive adult bottlenose dolphins (Delphinidae, Tursiops truncatus): A re-evaluation of the signature whistle hypothesis. Ethology, 100, 194-209.

McCowan, B., Hanser, S. F., & Doyle, L. R. (1999). Quantitative tools for comparing animal communication systems: Information theory applied to bottlenose dolphin whistle repertoires. Animal Behaviour, 57, 409-419.

Montagu, A. & Lilly, J. C. (1963). The Dolphin in History. Los Angeles: University of California Press.

Nollman, J. (1987). Dolphin Dreamtime. New York: Bantam Books.

Nollman, J. (1991). Playing Music with Animals. [Audio Recording]. The Smithsonian Institution, Flokways Cassette Series, 06118.

Norris, K. S. (1994). The Life and Times of the Spinner Dolphin. Berkeley: University of California Press.

Reiss, D. & McCowan, B. (1993). Spontaneous vocal mimicry and production by bottlenose dolphins (Tursiops truncatus): Evidence for vocal learning. The Journal of Comparative Psychology, 107(3), 301-312.

Sayigh, L. S., Tyack, P. L., Wells, R. S. & Scott, M. D. (1990). Signature whistles of free-ranging bottlenose dolphins, Tursiops truncatus: mother-offspring comparisons. Behavioral Ecology and Sociobiology, 26(4), 247-260.

Sayigh, L. S., Tyack, P. L., Wells, R. S., Scott, M. D., & Irvine, A. B. (1995). Sex difference in signature whistle production of free-ranging bottlenose dolphins, Tursiops truncatus. Behavioral Ecology & Sociobiology, 36, 171-177.

Sayigh, L.S., Tyack, P. L., Wells, R. S., Solow, A. R., Scott, M. D., & Irvine, A. B. (1998). Individual recognition in wild bottlenose dolphins: A field test using playback experiments. Animal Behaviour, 57, 41-50.

Smolker, R. A., Mann, J., & Smuts, B. B. (1993). Use of signature whistles during separations and reunions by wild bottlenose dolphin mothers and infants. Behavioural Ecology and Sociobiology, 33, 393-402.

Tyack, P. (1986). Whistle repertoires of two bottlenosed dolphins, Tursiops truncatus: mimicry of signature whistles? Behavioral Ecology and Sociobiology, 18, 251-257.

Tyack, P. (1991). Use of a telemetry device to identify which dolphin produces a sound. In K. Pryor & K. S. Norris (Eds.), Dolphin Societies: discoveries and puzzles (pp. 319 - 344). Berkeley: University of California Press.

Weyler, R. (1986). Song of the Whale. New York:Anchor Press/Doubleday.