Talking Back

PAGE 1

TALKING BACK: EFFECTS OF TRAINING SESSION OCCURRENCE ON CAPTIVE DOLPHIN VOCAL BEHAVIOR BY JENNIE CASKEY A Thesis Submitted to the Division of Social Sciences New College of Florida in partial fulfillment of the requirements for the degree Bachelor of Arts Under the sponsorship of Heidi Harley Sarasota, Florida May, 2012

PAGE 2

i Dedication "The spinner is a considerably different animal. It is much more completely a group animal. Held alone it is lost, frightened, and totally dependent on its handlers." (Norris, 1994, p. 303) For Harley It was always about you.

PAGE 3

ii Acknowledgments I would like to thank My thesis sponsor and advisor, Professor Heidi Harley I have been a devout follower of what friends have affectionately dubbed the "Cult of HEH" since wandering into Bon House my first semester. Your guidance, support, and resounding insight have not onl y inspired my academic pursuits but pervaded every aspect of my life. You have profoundly affected my life in a way that I will forever be grateful for. My committee members, Professors Gordon Bauer and Al Beulig, for your participation and patience as t he deadlines loomed closer and then waved as they flew by. None of this would have been possible without the wonderful staff of a "public facility in central Florida" and "a marine hospital and research laboratory in southwest Florida," but most especial ly: Wendi, for invaluable advice and feedback that guided me in the design of this study and unwavering support and assistance throughout the process, and Adrienne, for going above and beyond to accommodate my many demands, and making it happen, thr ough thick and thin. ~ Katie, because when I say "we," I mean you, Chrissy, for let ting me take advantage of your ability to make everything just that. much. better, Nicole, now undeniably the most dolphin savvy chef I know, whether you like it or not, for withstanding the many, many, many hours of talking, crying, talking, talking, writing, and talking about my thesis, Morgan, for reminding me of what I work for, Adam and Kyle, for putting up with me, ever, ~ Mum and Dad, just remember: it cou ld have been philosophy.

PAGE 4

iii Table of Contents Dedication i Acknowledgment s ii Table of Contents .. iii List of Tables and Figures i v Abstract vi Introduction. .. 1 Method .. 44 Results 53 Discussion 73 Appendix A ... 90 References 92

PAGE 5

iv List of Tables and Figures Table 1. Total Number of Phonations Emitted by Each Subject Pair Across Conditions .. 54 Figure 1. Bottlenose Facility Tank System 46 Figure 2. Pantropical Spotted Dolphin Facility: Lagoon ... 47 Figure 3. Pantropical Spotted Dolphin Facility: Med Tank .. 48 Figure 4. Spectrogram Examples of Phonation Types .. 52 Figure 5. Total Number of Phonations Emitted by Subjects in Training and No Training Sessions ... 54 Figure 6. Distribution of Moonshine's Phonation Behavior in Lagoon Sessions 55 Figure 7 Distribution of Moonshine's Phonation Behavior in Med Tank Sessions 56 Figure 8. Comparison of Phonation Emission Distribut ion Between Med Tank and Lagoon 57 Figure 9 Total Phonations Emitted Over Time (Phase) of Session s 58 Figure 10 Total Number of Compound Vocalizations Emitted by Moonshine Across Phases of Sessions ... 59 Figure 11. Total Phonations Emitted by Moonshine Across Phases of Med Tank Session s... 60 Figure 12. Distribution of Phonation Types Between Phases of TS and NTS Sessions in Lagoon 61 Figure 13. Distribution of Phonation Types Between Phases of TS and NTS Sessions in Med Tank 63 Figure 14. Distrib ution of Whistle Contours. 65

PAGE 6

v Figure 15 Distribution of Whistle Contours Between Phases of TS and NTS Sessions in Lagoon 66 Figure 16 Distribution of Whistle Contours in the During Session Phase of TS and NTS Sessions in the Med Tank... 68 Figure 17 Distribution of Khyber's and Ranier's Phonation Behaviors .. 69 Figure 18 Distribution of Whistle Contours Emitted by Khyber and Ranier in TS and NTS Sessions ... 70 Figure 19 Distribution of Phonation Types Emitted by Calvin and Malabar Between Conditions 71 Figure 20 Total Phonations Emitted by Calvin and Malabar Across Phases of Training and No Training Sessions ............ 72 Figure 21 Distribution of Whistle Contours Emitted by Calvin and Malabar in TS and NTS Sessions ... 73 Figure 22. Moonshine's Signature Whistle Contour. 84

PAGE 7

vi TALKING BACK: EFFECTS OF TRAINING SESSION OCCURRENCE ON CAPTIVE DOLPHIN VOCAL BEHAVIOR Jennie Caskey New College of Florida, 2012 ABSTRACT In 1961, John Lilly's claim that dolphins were "kind, cooperative, intelligent" beings that were likely candidates for inter species communication was met with incredulity in the scientific and popular communities of the time. However, in the decades sinc e, both inter species communication and dolphin cognitive and behavioral research have progressed significantly, and evidence suggests that dolphins do, in fact, possess the traits and mechanisms that have so far been shown to correlate with the most succe ssful inter species communication. Dolphins, as a family, are highly gregarious, and utilize contextually flexible vocal behaviors to coordinate, communicate, and perceive their environment. Historically, they have had a uniquely sympatric relationship wit h humans, and in captivity have demonstrated inter species communication mechanisms such as point gesture comprehension and modified production, response to human attentional states, and vocal labeling. However, before more complex forms of inter species c ommunication between humans and dolphins can be explored, it is necessary to demonstrate bidirectionality of communication from the direction of dolphin to human. To do this, changes in the phonation behaviors of five captive dolphins ( Tursiops truncatus a nd Stenella attenuata ) were observed through a within subject design comparing their phonation behavior when no training session occurred (and therefore no

PAGE 8

vii trainer interaction) to rates before, during, and after a training session occurred. The results sup port the previous research that suggests that there is individual variability within a species' capacities for inter species communication with humans based on levels of enculturation. One subject, Moonshine, showed the greatest variability in his overall phonati on behavior with specific pattern differences that might be expla ined by the occurrence, or lack thereof of a training session. The two bottlenose subject pairs demonstrated less variation between conditions, but Khyber and Ranier did exhibit a dif ference overall, whereas Calvin and Malabar showed no significant difference between conditions. More robust data must be collected from all five subjects to support strongly any conclusions that might be drawn from this study, but these early results suggest that there is more to investigate, and that dolphins' vocal behaviors may provide a quantifiable measure of bidirectionality in individual dolphins' communication with hu mans. ______________________ Heidi E. Harley Division of Social Sciences

PAGE 9

Running Head: CAPTIVE DOLPHIN VOCAL BEHAVIOR 1 Talking Back: Effect of Training Session Occurrence on Captive Dolphin Vocal Behavior Our own spot in the universe, our own view of ourselves, will be tremendously modified if such a [dolphin human] communication is established. Our own views of one an other will change radically under the influence of interspecies communication. The very fact that we try to communicate with them is an important indication of our own stage of evolutionary maturity. (Lilly, 1961, p.223) In 1961, the neurophysiologist Dr. John Lilly published in his book, Man and Dolphin this controversial statement: "Within the next decade or two the human species will establish communication with another species: nonhuman, alien, possibly extraterrestrial, more probably marine" (p. 11). He had come to this conclusion after hearing the sounds emitted by a dolphin while the "positive" and "negative" areas of its brain were stimulated with electrodes (Lilly, 1961), and made them first known at the annual meeting of the American Psychiatric Association in 1958 (Wood, 1973). His assertions shook both the scientific and popular communities; the rapid proliferation of aquatic theme parks around the world over the previous decade had piqued global curiosity and shed light on some of the mystery o f marine mammals, but many of their capabilities remained largely unknown. The image of "kind, cooperative, intelligent, trainable, fun loving, romantic, never hostile or vicious but at times, like some persons, exasperating to deal with" (Lilly, 1958, p. 500) dolphins that Lilly described enhanced the emerging popular opinion of dolphins as anthropomorphized, intelligent beings. Popular media such as Flipper a movie and later successful television series produced in 1963, and Day of the Dolphin a book b y Robert Merle published in 1969 and later reproduced as a movie in 1973,

PAGE 10

CAPTIVE DOLPHIN VOCAL BEHAVIOR 2 continued to boost previously overlooked marine mammals, particularly dolphins, into the positive attention of humans (Fraser, Reiss, Boyle, Lemcke, Sickler, Elliott, Newman, & Grube r, 2006). With Lilly's claims, other issues in human and comparative cognition were brought back into the spotlight. In the early 20 th century, Wilhelm von Osten reported that his horse, der Kluge Hans" ("Clever Hans"), was able to read and do arithmetic (Pfungst, 1911). An informal commission sought to prove these claims empirically, and, for a moment, the human idea of animal intelligence was turned upside down before Oscar Pfungst discovered the solution to the riddle. Oscar Pfungst's (1911) exposŽ of Clever Hans' use of the "unintentional minimal movements of the questioner" (p.1) to cue correct responses was one of the first publications to suggest that non human species may attend to human gestural communication devices during human interactions. S ince that time, humans have struggled with the dividing line between man and beast from every perspective, and the distinctions have become increasingly less clear as research has continued to bridge the animal human gap. Noam Chomsky (1972) stated that "h uman language appears to be a unique phenomenon, without significant analogue in the animal world" (p.67). Prior to the 1960's, this view was virtually irrefutable. Lilly's implication that both communication and ultimately "human" language might be possib le outside of the human race jump started a movement towards the study of the ontogeny of human communication and with it, inter species communication. Communication, according to Stanley Smith Stevens (1950) should be defined as "the discriminatory respo nse of an organism to a stimulus." Accordingly, if a stimulus

PAGE 11

CAPTIVE DOLPHIN VOCAL BEHAVIOR 3 (the "message") elicits no response, true communication does not occur. Human communication has evolved to accommodate the complexities of human cognition on a minute level (Lieberman, 1968), re sulting in an elaborate syntactic structural system of vocal "language." Chomsky (1957) defines human language as "a set (finite or infinite) of sentences, each finite in length, and constructed out of a finite set of elements" (p.13). It is this multi fac eted and complex system that, Chomsky (1972) argues, sets humans apart from other animals. It has been five decades since Lilly's assertions. In that time, researchers have made leaps and bounds towards better understanding human language, animal communica tion, and the complexities and capacities of marine mammal biology and cognition. However, progress in inter species communication between humans and other animals is still limited. If the hypothesis that non human animals have the capacity for cooperative communication is true, what is the missing link that prevents this form of communication from being conclusively proven? Cooperative and dynamic communication as defined by Stevens (1950) requires an adaptable (bi directional) exchange between two or more parties. Current research in inter species communication has focused on distilling the components and mechanisms necessary for such communication to occur, but few studies have been done that demonstrate the necessary bidirectionality of communication in appropriate candidate species. If "inter species communication capable" subjects are uninterested in potential channels of interaction, progress may be impossible. This thesis seeks to establish whether it is possible to quantitatively measure the communic ative behaviors of non human subjects (dolphins) while they either are, or are not, directly interacting with

PAGE 12

CAPTIVE DOLPHIN VOCAL BEHAVIOR 4 humans via the occurrence of training sessions. If significant differences are observed, this may serve as evidence of bidirectionality in intenti onal communication from dolphins to humans. Inter Species Communication Tomasello (2008) asserts that "the road to human cooperative communication begins with great ape intentional communication, especially as it manifests in gestures" (p. 320). Many res earchers therefore theorized that by examining the communication mechanisms of great apes, they might better understand the ontogeny of human cooperative communication and language. This led them to pursue language development research with non human prima tes including chimpanzees ( Pan troglodytes ), bonobos ( Pan paniscus ), gorillas ( Gorilla gorilla ), and orangutans ( Pongo sp.), among others. Human Ape Gestural Interaction In 1966, Allen and Beatrice Gardner began training Washoe, a young female chimpanze e, to use American sign language (ASL) gestures to communicate with her human caretakers. In order to generate this interaction, the Gardners took advantage of the chimpanzee's natural tendency to imitate behaviors. Washoe was exposed repeatedly to ASL com municative gestures, and her imitations were met with appropriate responses, encouraging her pursuit of the communication. By 1969, Washoe had acquired 30 ASL gestures, becoming the first non human to utilize standardized gestural communication in appropri ate contexts (Gardner & Gardner, 1969), and to engage regularly in cooperative communication with humans. The relative success of the Gardners' gestural comm unication approach with Washoe, especially when compared to the previous failure to elicit langua ge through

PAGE 13

CAPTIVE DOLPHIN VOCAL BEHAVIOR 5 vocal com munication in the same species ( see Hayes & Hayes, 1952) led to repetition of the process with more primate subjects, including other chimpanzees such as Moja, Pili, Tatu, Loulis (Gardner & Gardner, 1989), and Nim Chimpsky, (Terrace, P etitto, Sanders, & Bever, 1979), among others, as well as a gorilla named Koko (Patterson, 1978). Despite the progress made by the Gardners and their colleagues with gestural communication in chimpanzees, there was still a disconnection between the subje cts and their human companions, and true "language" had not been confirmed (Terrace, 1982). The sign language studies were riddled with anecdotal evidence of the chimpanzees' abilities to create and comprehend novel sentences and, apparently, use ASL as a form of language; however, when these studies were critically reviewed, they lacked solid proof that the subjects had fully grasped the language concepts (Terrace, 1982). These inconsistencies led a group of researchers to pursue inter species communicat ion from a different angle. To bridge the gap between the natural ape gestural communication and human vocal communication, Rumbaugh, Gill, Brown, Glasersfeld, Pisani, Warner, and Bell (1973) developed a "lexigram" system of communication. A computerized d evice contained a graphic array of symbols (lexigrams) meant to represent individual English words, which subjects could press to illuminate their choice as an equivalent of voicing the word. The machine was first used by a young female chimpanzee named La na (Rumbaugh et al., 1973), and later by two male chimpanzees, Sherman and Austen, to communicate both with humans and each other in cooperative tasks (Savage Rumbaugh, Rumbaugh, & Boysen, 1978). By 1980, two new subjects were introduced to the device: m embers of a different branch of the chimpanzee family, bonobos ( Pan paniscus ) Matata and her son, Kanzi.

PAGE 14

CAPTIVE DOLPHIN VOCAL BEHAVIOR 6 Although Matata's success with the lexigram was dubious at best (Savage Rumbaugh, 1998), Kanzi spontaneously began to utilize the device as an infant t o communicate with his mother, other chimpanzees at the facility, and his human caretakers (Savage Rumbaugh, McDonald, Sevcik, Hopkins, & Rubert, 1986). By the time Kanzi was nine years old, he was able to properly carry out 72% of unique, unusual, and nov el requests (e.g., "Put the collar in the water," or "Stab the ball with the sparklers," Savage Rumbaugh, 1998, pp.68 69). Although Kanzi's use of the lexigram and his apparent comprehension of spoken English may be the closest that studies have come to e stablishing language based inter species communication, the system is still limited by species specific and physiological restrictions. Great apes utilize innate, inflexible vocal behaviors that are species wide and do not adapt to communicative situations (Tomasello & ZuberbŸhler, 2002). Great apes also lack the anatomical structures necessary to form the complex phonations that allow "human level" language to develop (Chomsky, 1957; Lieberman, 1968). Therefore, the strength of Savage Rumbaugh's (1998) arg ument that Kanzi was able to use language to communicate with humans is restricted by the limited scope of responses Kanzi could provide that irrefutably support the communication hypothesis (Savage Rumbaugh, 1998). The unidirectionality of this interactio n makes the argument of cooperative, dynamic communication very difficult to support. Effects of Humans on Vocal Mechanisms and Behavior Great apes may have inflexible vocal behaviors (Tomasello & ZuberbŸhler, 2002), but this is not true for all socially vocal animals. Domestic dogs (perhaps the most highly enculturated animals known to man) utilize a variety of barks, whines, growls, and other

PAGE 15

CAPTIVE DOLPHIN VOCAL BEHAVIOR 7 vocal signals to convey signals not only to dog conspecifics but to humans they interact with as well. Dog owner s report perceiving differences in the acoustic qualities of their pet dogs' barks that indicate the context of the vocalization (e.g., stress, warning, or excitement). Moin‡r, Pongr‡cz, and Mikl—si (2010) tested the truth of these claims by asking sightle ss pet owners to identify the contextual information they perceived while listening to their dogs' barks recorded in a variety of contexts (approach of a stranger, an instigated "fight," preparing for a walk, left alone, and when viewing a ball). Participa nts correctly identified the context of the recordings at significantly above chance levels, suggesting that humans are able to perceive, on some level, differences in the signal intent of their pets vocalizations (Moin‡r et al., 2010). Although primitive, the ability for dog owners to identify and appropriately respond to different context clues in the vocal signals of their pets suggests a level of bidirectionality in dog human communication. Although domestic dogs appear to have developed a mechanism fo r bidirectional communication with humans, it is difficult to differentiate between the effects of centuries of domestication and the more immediate effects of dynamic situations and human interactions on domestic pets' vocalizations. In dogs, the bark beh avior that they are best known for appears to be a species wide behavior that likely developed over time as they diverged from their wolf cousins (Yin, 2000). Dogs bark often, and in a wide variety of contexts, yet barks make up only approximately 2.3% of all wolf vocalizations (Schassburger, 1987) and seem limited to only aggressive and territorial contexts (Joslin, as cited in Mech, 1970). Differences in barking behavior do seem related to domestication, as can be seen in Balyaev's domestic silver foxes ( Gogoleva, Volodin, Volodina, Kharlamova, & Trut, 2008, 2009, 2010, 2011).

PAGE 16

CAPTIVE DOLPHIN VOCAL BEHAVIOR 8 Gogoleva et al. (2008, 2009, 2010, 2011) have demonstrated significant differences in the vocal behaviors of tamed compared to aggressive and "unselected" silver foxes ( Vulpes vulpe s ). The subjects of these studies were selected from a population of foxes that have been selectively bred over several generations to elicit specific features. Due to the effects of early exposure to humans on later responses, petting and other contact be tween humans and foxes on this farm was forbidden. Tame foxes were selected for "friendliness to people, display[ing] emotionally positive responses to any human" (Gogoleva et al., 2011, p. 216). Aggressive foxes were "not afraid of humans, do not try to i ncrease their distance from an approaching human, and instead tend to attack him" (Gogoleva et al., 2009, p. 369). Unselected foxes were bred without regard to these features and thus represented a "control" group most similar to their wild counterparts. U nlike many of the species studied in the past with regard to the effects of domestication, Belyaev's silver foxes provide a unique opportunity to compare the behaviors and abilities of domesticated and wild members of the same species. In Gogoleva et al.' s (2011) study of the vocal activity of tame, aggressive, and unselected foxes when humans approach, the researchers sought to examine the differences in the vocal behavior across the three strains, and the effects of human attention on each strain's behav ior. In order to test this, 45 foxes (15 tame, 15 aggressive, and 15 unselected) were recorded for 5 minutes each as an unfamiliar human approached their enclosures. The recordings were analyzed to find the types of vocalizations made by each fox, and the frequency of their vocalizations over that period of time. The foxes had previously been described as making eight call types toward people: five vocal calls (whine, moo, cackle, growl, and bark) and three non vocal (pant, snort, and cough)

PAGE 17

CAPTIVE DOLPHIN VOCAL BEHAVIOR 9 (Gogoleva et al ., 2008). Of these vocalizations, the analyses showed that only the whine, moo, and growl were found in all three strains when humans approached. The tame foxes also produced cackle and pant, but did not cough or snort. Aggressive and unselected foxes, in contrast, produced cough and snort but not cackle and pant. Barks were the least common vocalizations across all three strains, found only rarely in aggressive foxes (Gogoleva et al., 2011). These results showed a clear effect of domestication and selecti ve breeding on the natural vocal responses of foxes to humans. Even more interesting, perhaps, was the difference in the frequency of the foxes' vocalizations over time throughout the recordings. Recording began when the unfamiliar human approached the fo cal foxes' enclosures, and continued for five minutes while the human remained at a distance of 50 cm from the enclosure but did not engage in any interaction with the focal fox. During this time, the approach of the human proved to elicit "explosive" voca l and physical activity in each of the foxes, but when analyzed, tame foxes showed a decrease in panting, considered an excitatory response, and an increase in whines and cackles, thought to be "attention getting" vocalizations, over the five minutes. Aggr essive foxes showed only an increase in moos over the time of the recording, and unselected foxes showed an increase in moos and a decrease in coughs and snorts. These results suggest that in tame foxes, the approach of the human caused significant immedia te excitement (especially when coupled with the affiliative movements of the foxes, such as wagging tails, mouth ajar, moving on half bent paws, etc.), followed by increased attention getting behavior, perhaps to attempt to counter the apparent lack of att ention of the human and lengthen the interaction. The aggressive and unselected foxes, on the other hand, did not demonstrate an apparent difference in their

PAGE 18

CAPTIVE DOLPHIN VOCAL BEHAVIOR 10 excitatory response or show any affiliative vocalizations throughout the recordings (Gogoleva et a l., 2011). Domestic cat vocalizations have apparent structural differences than their wild cat cousins (Nicastro, 2004) that may be both the result of their species' divergence and levels of human interaction (Yeon, Kim, Park, Lee, Lee, Suh, Houpt, Chan g, Lee, Yang, & Lee, 2011). Nicastro (2004) found that the acoustic features of African wild cats ( Felis silvestris lybica ) were significantly different from those of domestic cats ( Felis catus ). Wild cat vocalizations were generally longer in length and m ore complex, with more harmonics (resulting in "raspier" qualities), whereas domestic cat vocalizations were shorter, simpler (fewer compound, or combined, vocalizations), and had fewer harmonics (Nicastro, 2004). However, Nicastro's (2004) study only comp ared the wild cats' vocalizations to the vocalizations of domestic cats that were owned and cared for by humans, and could not illustrate whether humans had a direct influence over this degree of dissimilarity. Yeon et al. (2011) were able to address thi s effect more directly in a study comparing the vocal behaviors of domestic house (human owned and cared for) and feral cats. Thirty eight domestic cats (25 feral, 13 house) were recorded in a variety of situations involving the approach of familiar and un familiar humans, unfamiliar dogs and cats, and a doll. The cats were placed in a cage and recorded with both video cameras and audio recorders as the various stimuli approached in order to stimulate vocalizations. Five difference tests were conducted: frie ndly approach of a caretaker, threatening approach of a stranger, approach of a doll, stranger approaching with a strange dog, and stranger approaching with a strange cat. The recordings showed a significant difference in the

PAGE 19

CAPTIVE DOLPHIN VOCAL BEHAVIOR 11 vocal responses to the differe nt situations by the two groups of cats. Both groups had a higher occurrence of one type calls (growls, hisses, and meows) than complex calls (combinations of these vocalization types), but feral cats not only demonstrated a higher occurrence of vocalizati ons overall, but significantly more complex calls than house cats. Feral cats produced a significantly higher rate of growls and hisses during both the affiliative and the agonistic test situations than house cats, though both showed significantly higher c all rates during the agonistic situation than the affiliative. The feral cats also demonstrated significant differences in the structure of their vocalizations than those of the house cats, with higher overall frequency of each of their vocalizations (Yeon et al., 2011). These frequency differences indicated a closer similarity between the feral cats' vocalizations and those previously recorded of their wild cat relatives than the house cats' vocalizations (Nicastro, 2004). These results show a clear differ ence in the vocal responses of domestic cats with high enculturation and levels of human interaction to the approach of humans compared to those of less socialized cats, suggesting that the voca l behavior of domestic cats is a ffected by the levels of expos ure to humans. Point Gesture Comprehension and Production In order to overcome the weaknesses of current inter species communication research, the mechanics of bi directional communication can be approached from a different angle. Referential pointing is a ubiquitous human behavior (see Kita, 2003), used as both a complimentary gesture and a "complete communicative act" (Tomasello, 2008) in human to human communication. Furthermore, referential point gesture behavior is not documented in the natural reper toire of any non human species (Leavens

PAGE 20

CAPTIVE DOLPHIN VOCAL BEHAVIOR 12 & Hopkins, 1999), providing a discrete mechanism of communication for comparison in human animal communicative interactions. Comprehension of human referential pointing may be correlated with the amount of human int eraction to which an animal (or species) is exposed (Call & Tomasello, 1994; Mikl—si & Soproni, 2005). The prevalence of pointing in humans makes this an advantageous ability to species who live closely with humans. It is therefore not surprising to find t hat domestic animals including dogs (Mikl—si, Polg‡rdi, Top‡l, & Cs‡nyi, 1998), cats (Mikl—si, Pongr‡cz, Lakatos, Top‡l, & Cs‡nyi, 2005), horses (Maros, G‡csi, & Mikl—si, 2007), and goats (Kaminski, Riedel, Call, & Tomasello, 2005) are all capable of utili zing human point cues to make a correct choice in a choice task. However, these species have undergone centuries of domestication, complex interactions of evolution, adaptation, and selective breeding that have shaped them into the social and service compa nions they represent today. Therefore, to ask questions about the relatively immediate effects of human interaction on mechanisms and comprehension of communication in non human species, non domestic species such as wolves, seals, dolphins, and primates ha ve been studied. Acquisition of point gesture comprehension with exposure and training. A longitudinal study conducted by Vir‡nyi, G‡csi, Kubinyi, Top‡l, BelŽnyi, Ujfalussy, and Mikl—si (2008) found that although wild wolves have significantly lower rate s of human point gesture comprehension than domestic dogs, wolves raised with a comparable level of sociability and enculturation with humans will perform better at a choice task with human point cues. Unlike dogs, wolves do not acquire the ability without repeated exposure and training, but they do demonstrate a capacity to perform this task correctly in

PAGE 21

CAPTIVE DOLPHIN VOCAL BEHAVIOR 13 a form of simple communication that is comparable to their distant cousins' (Vir‡nyi et al., 2008). Shapiro, Janik, and Slater (2003) trained a gray sea l pup to select a target in a choice task based on the gestural indication of his trainer. When they later tested his ability to comprehend various referential point gestures (for a review, see Mikl—si & Soproni, 2005), the seal performed well above chance (Shapiro et al., 2003). Although it is a positive step in studying communication, many studies of point gesture comprehension in non domestic species, including those cited above, can be interpreted via operant conditioning of their subjects (Mikl—si & Soproni, 2005; Shapiro et al., 2003). Hence, the strongest conclusion that can be drawn from these studies is that subjects are able to generalize the rules of trained referential gestures to dynamic pointing in humans (Shapiro et al., 2003). The human "in terference," or active pursuit of such interactions may decrease the bi directional aspects of this form of communication. Acquisition of referential point gesture behavior. Although many studies of primate comprehension of referential pointing have been conducted (e.g., Anderson, Sallaberry, & Barbier, 1995; Anderson, Montant, & Schmitt, 1996; Call & Tomasello, 1994; Itakura & Tanaka, 1998; Neiworth, Burman, Basile, & Lickteig, 2002; et al.), primates demonstrate surprising variability in their ability t o comprehend the human point gesture. Capuchin monkeys appear to understand the referential meaning of human pointing when it is linked to the attentional gaze of the human pointer, but not when the point gesture is performed alone (Anderson et al., 1995). Rhesus monkeys do not appear to comprehend the referential nature of human pointing at all (Anderson et al., 1996). On the other hand, Itakura and Tenaka (1998) found that chimpanzees and an orangutan were

PAGE 22

CAPTIVE DOLPHIN VOCAL BEHAVIOR 14 able to comprehend each of five referential gestu res (tapping, gaze plus point, close gaze, gaze alone, and glancing) used by human collaborators to indicate a correct choice. The results of Anderson et al.'s (1995) study of Capuchin monkeys' abilities would suggest that the method of Itakura and Tenaka 's (1998) study may have influenced the results. The primate subjects may not have demonstrated the same level of comprehension if pointing alone had been tested rather than pointing with gaze cues. Despite the inconsistent results of point comprehension studies in primates, captive primates have been observed to imitate and utilize the human point gesture during referential interactions (Call & Tomasello, 1994). Call and Tomasello (1994) observed the referential pointing behavior of two orangutans ( Pongo pygmaeus ) from different backgrounds of enculturation. Chantek, raised in a highly enculturated environment, consistently showed higher overall understanding of referential pointing than Puti, who was raised by humans in a nursery as an infant before bein g placed with conspecifics. However, the two orangutans had varying levels of success across the three experiments in this study. In the first experiment of Call and Tomasello's (1994) study, Chantek and Puti were asked to point to the location of a reques ted tool. Chantek began to reliably point to the correct location after only two trials, but Puti never spontaneously pointed to a tool location. In the second study, the orangutans selected a baited bucket after an experimenter pointed to the correct choi ce. Although Chantek did significantly better than Puti, he only made the correct choice during 33 out of 63 trials. During the final experiment Chantek and Puti were observed while they selected a reward by pointing at it as the experimenter either turned his back and walked away, left the room, kept his eyes open, or closed them. Chantek reliably attended to the experimenter's attentional state,

PAGE 23

CAPTIVE DOLPHIN VOCAL BEHAVIOR 15 and pointed more often when the experimenter faced him with his eyes open. Chantek never pointed when the exper imenter left the room or turned his back. Puti did not appear to differentiate between open and closed eye conditions (Call & Tomasello, 1994). The significant differences between Puti and Chantek support the theory that primate point gesture comprehens ion and production is related to subjects' levels of enculturation and human interaction during their upbringing (Mikl—si & Soproni, 2005). Unlike primates, captive dolphins have proven to demonstrate the ability to comprehend and correctly utilize the hu man referential pointing gesture reliably to make selections in a choice task (Herman, Abichandani, Elhajj, Herman, Sanchez, & Pack, 1999; Pack & Herman, 2007). Herman et al. (1999) illustrated this by training two dolphins, Akeakamai and Phoenix, to selec t an object placed to the left, right, or behind them based on the point of a human trainer. The dolphins were able to respond correctly to direct points, cross body points, and familiar symbolic gestures, suggesting that the dolphins did understand the re ferential character of human points. A later study conducted by Herman and Pack (2007) further supported this contention, when the dolphins were able to identify an object cued by human pointing by selecting it from a matching set or acting upon it. The do lphins were never formally trained to respond to the human point gesture, but both subjects had been unintentionally and casually exposed to the behavior over many years prior to the studies (Herman et al., 1999). A study conducted by Xitco, Gory, and Kuc zaj (2001) resulted in the spontaneous discovery of a "pointing" behavior in captive dolphins that appears to be a direct adaptation to the needs of humans. In this study, dolphins were paired with humans in an object seeking task that required the human p artner to retrieve the object after the dolphin

PAGE 24

CAPTIVE DOLPHIN VOCAL BEHAVIOR 16 located it. Because the human partners were significantly slower than the dolphins in arriving at the object location, the dolphins were forced to wait for the humans. During this time, dolphins would orient and align their bodies towards the object, occasionally turning their heads to monitor their human partners' progress. This behavior has not been observed in the wild, and is unlikely to be found there due to the echoic eavesdropping abilities of dolphins; dolphins working with other dolphins to locate an object could usually utilize the echolocation cues of conspecifics to gather the same information that would be unavailable to human partners (Xitco & Roitblat, 1996). Just as Call and Tomasello (1994) o bserved in the orangutans Chantek and, to a lesser extent, Puti, dolphins are also aware of the attentional state of the humans with whom they interact (Xitco, Gory, & Kuczaj, 2004). During a later study of the point and gaze behaviors of captive dolphins with human partners, Xitco et al. (2004) found that the point and gaze behaviors were significantly more likely to occur if the human partner was apparently attending to the dolphin (via orientation). A study conducted by Tomonaga, Uwano, Ogura, and Saito (2010) further supported this discovery when dolphins given instruction through gestural signs responded more frequently when the humans delivering the signals oriented their bodies towards the dolphins. This study did not find, however, that the head orie ntation of the trainers significantly influenced the performance of the dolphins. Dolphins as Candidates for Inter Species Communication Primates have been central to inter species communication research due, in part, to the insights they may provide on the ontogeny of human language as well as their relative similarity to humans on both genetic and socio behavioral levels. These similarities make

PAGE 25

CAPTIVE DOLPHIN VOCAL BEHAVIOR 17 them a logical choice as likely candidates for human animal communication. Enculturation theories suggest th at domestic animals may be more likely to share mechanisms of communication with their human companions or owners due to the close relationship and sympatry between them. The research on these animals suggests that Lilly's (1961) claim that true inter spec ies communication is possible may be not be as far fetched as it appeared. Lilly's (1961) claim caused a stir for another reason: prior to this decade, humans knew very little about marine mammals, and at the time of his proposal, dolphin research was in its primitive stages. Lilly's support of dolphins as intelligent and plausible candidates for human animal communication was radically separate from the popular perception of animals at the time. Despite this, dolphins have a rich history of interaction with humans that may have set the groundwork for the fascination they have inspired in comparative cognition research today. History of Man and Dolphin Throughout history, dolphins and humans have had a unique relationship. Dolphins are prominent in the r ich mythology, folklore, and legends of the ancient Greeks, Chinese, Japanese, Indians, Brazilians, Africans, and Native Americans, as well as cultures in the Australasian and Mediterranean regions (Burgoyne, 2000, as cited in Hampson, 2005; Cravalho, 1999 ; Hall & Vanderhoop, 2004; Higham, 1960; MacKenzie, 2005; Murray, 1897; Pliny the Younger, 1915; Slater, 1994; Taylor, 2003). In some legends, gods appear as dolphins, such as the incarnation of Vishnu (Taylor, 2003), the Buddhist and Hindu makara (Mackenz ie, 2005), and Apollo as dolphin (Sax, 2001;

PAGE 26

CAPTIVE DOLPHIN VOCAL BEHAVIOR 18 Taylor, 2003). In others, dolphins are transformed humans, such as in the Wampanoag American Indian legend of Katama and her lover, Mattakesett, who escape a war between their tribes by diving into the ocean, wh ere they became dolphins to be eternally together (Hall & Vanderhoop, 2004). In the story of Dionysus' revenge on the Tyrrhenian pirates, Dionysus transforms the pirates into dolphins as a form of punishment (Homeric Hymn 7 to Dionysus, trans. 1914). The B razilian boto encantado are believed to be human dolphin shape shifters, causing mischief and harm to land bound humans (Cravalho, 1999; Slater, 1994). Dolphins also appear in legends as friends or assistants to humans, playing with or even protecting huma ns (e.g., the Boy on a Dolphin, Taylor, 2003; the Burmese guardian dolphins, Busnel, 1973; the Brazilian Amazonia tucuxi Slater, 1994). Anthropological accounts describe dolphin human interactions ranging from pet like to apparent alliances. Pliny the El der (A.D. 32 79) describes in his book, The Natural History, (vol. IX, verse 8, translated to English by John Bostock and H.T. Riley in 1855 from the original Latin, retrieved from Perseus), the story of Simo, a dolphin in Lake Lucrinus, in southern Ital y, with apparent affection for a young boy that passed by each day, as well as of a boy in Hippo, Africa, and his companionship with a dolphin in the waters off the shore, among others. In response to requests by scholars of Pliny for support of such stori es, Higham (1960) describes accounts of Opo Joe, a dolphin that allowed people to pet her and for children to ride on her back near Opononi beach, New Zealand, as well as Pelorus Jack, a Risso's dolphin that frequently socialized with sailors and boaters i n Cook Strait near New Zealand. Later in The Natural History (vol. IX, verses 9 and 10), Pliny describes fishermen of a region in France who, during a migration of mullet through the area, call to dolphins

PAGE 27

CAPTIVE DOLPHIN VOCAL BEHAVIOR 19 for assistance. The dolphins are reported to resp ond, arranging themselves "in a battle line" to "block the access to the deep waters and drive the disturbed fish toward the shallower waters" (Busnel, 1973, p.112) before the fishermen surround the fish with nets and use pitchforks to spear them (Pliny th e Elder, Natural History vol. IX, as it appears in Busnel, 1973, translated to English from E. Saint Denis' French translation from the original Latin). Accounts of "co fishing" strategies like the one described by Pliny are found in the histories of Abori ginal Australians (Neil, 2002), fisherman in the Ayeyawady region of India (Tun, 2004), Africans (Busnel, 1973), and rural populations of Brazilian Amazonia (Cravalho, 1999; Pryor, et al., 1990; Slater, 1994). Some of these co fishing alliances continue to this day. Most literature detailing these interactions describes the fishermen and dolphins as cooperating through a system of rudimentary communication to produce a "symbiotic" relationship (Busnel, 1973). The fishermen use splashing, rhythmic beating, and vocal sounds possibly reminiscent of a school of potential prey to attract the attention of the dolphins, who in turn drive actual schools of fish towards the fisherman (as described by Busnel, 1973; Cravalho, 1999; Neil, 2002; Pryor, et al., 1990; Sla ter, 1994; and Tun, 2004). The prey, caught between two predators, become readily available to both (Busnel, 1973). There are some variations to this system: in India, fishermen believe that a specific "guardian" dolphin is particular to certain fishermen, and therefore any prey caught with the assistance of that dolphin should rightfully belong to the fishermen associated with it (Busnel, 1973). In Brazil, similar familiarity and individualized associations form between fishermen and "their" boto (Busnel, 1973; Slater, 1994). These individualized relationships have, in some cases, led to more complex communication

PAGE 28

CAPTIVE DOLPHIN VOCAL BEHAVIOR 20 systems, with varied calls and splash or rhythm patterns between fishermen to identify the caller or recipient (Tun, 2004). The basic system of a uditory communication, however, remains constant. The emergence of this behavior may be a result of an extension of the natural hunting behaviors of these dolphins, reinforced by the enhanced hunting success due to the partnership (Busnel, 1973). Interacti ons between humans and dolphins in both a current and historic perspective lay the groundwork for the pursuit of cooperative communication between the two species. Though primates are evolutionarily closest to humans, they lack the frequent sympatric inter actions that domestic animals, and to a lesser extent, dolphins, have historically exhibited with humans. The Natural History of Dolphins In order to understand the qualities tha t make dolphins remarkably well suited for inter species communication rese arch, it is important to understand the natural history of their species. Cetaceans as an order include a vast number of species, which despite sharing, at a family level, common natural histories, environments, and behaviors, also have many species speci fic differences (Rendell & Whitehead, 2001). The current study includes subjects of both the bottlenose dolphin ( Tursiops truncatus ) and the pantropical spotted dolphin ( Stenella attenuata ) species. Therefore, species related information will be limited to those two species. Social Structure The social and ecological pressures of the vast environment of dolphins (including predation, resource distribution, and reproduction, among others) have

PAGE 29

CAPTIVE DOLPHIN VOCAL BEHAVIOR 21 contributed to a highly gregarious social strategy among most dolphin species (May Collado, Agnarsson, & Wartzok, 2007; Pryor & Shallenberger, 1991; Wells, 1991). Population structure appears to be most affected by habitat structure and activity patterns (Shane, Wells, & WŸrsig, 1986), as well as time of day (though this is likely related to daily activity cycles) (Pryor & Shallenberger, 1991; Shane, 1977, as cited in Shane et al., 1986; Wells, 1978, as cited in Shane et al., 1986) or season (Garc’a & Dawson, 2003; Scott, Wells, & Irvine, 1990; Shane et al., 1986). Norris and Dohl (1980) note that across species and environments, a pattern emerges; shallower, more restricted environments are correlated with smaller group size, whereas vaster environments are correlated with larger groups. They note that river dolphin s are usually found in small groups or solitary, whereas coastal dolphins form larger but still relatively limited groups, and pelagic dolphins are often observed in groups of several thousand (Norris & Dohl, 1980). This correlation has been observed in bo th bottlenose and spotted dolphin groups. In coastal regions, bottlenose dolphin groups may include up to approximately 30 members, usually ranging between 2 15 members (Jefferson, Leatherwood, & Webber, 1994), and spotted dolphins usually remain in groups of under 100, usually ranging between 1 50 members (Garc’a & Dawson, 2003; Jefferson et al., 1994). In pelagic regions, groups may consist of over 200 bottlenose dolphins and upwards of thousands of spotted dolphins (Jefferson et al., 1994). Fission fus ion society In contrast to the popular idea of "pods," described by Reynolds, Wells, and Eide (2000) as "a group of animals that remains together for a long time and whose membership is virtually unchanged during that time" (p.112), dolphins live in fluid groups that can change in size and composition from moment to moment.

PAGE 30

CAPTIVE DOLPHIN VOCAL BEHAVIOR 22 Dolphin groups and subgroups may be better described as temporary gatherings of individuals engaged in similar activities in the same area (Reynolds et al., 2000). Gender and age spec ific social patterns. Within the grand scheme of the complex social dynamics of dolphins, there are some notable patterns of relationships between dolphins as well as stable individual bonds (Wells, Scott, & Irvine, 1987). In bottlenose dolphins, a dult ma les may form coalitions (pair bonds) of two, sometimes three dolphins (Connor, Helthaus, & Barre, 1999; Watwood, Tyack, & Wells, 2004). Wells (1991) observed that pair bonds between males in Sarasota Bay appear to solidify at sexual maturity and last as lo ng as twenty years, or until a pair member dies. Males that lose their partner may form a pair bond with another single male (Wells, 1991). In Shark Bay, male alliances have two levels: pair bonds between two or even three males, as well as "superalliances (Connor et al., 1999) of several bonded pairs that cooperate to take females from other alliances and defend against such attacks (Connor et al., 1999). Female associations in both species vary from relatively solitary to sociable (Gibson & Mann, 2007; W ells, Irvine, & Scott, 1980), the strongest bond being between a mother and her calf (Wells et al., 1980). Some mother and calf pairs are relatively solitary, possibly to avoid close competition for resources (Wells, 1991). Other pregnant females and moth er calf pairs may form loose alliances (Wells et al., 1987; Wells, 1991), potentially to increase protection and therefore reproductive success (Wells et al., 1987), decrease energy costs in foraging (Gibson & Mann, 2007), or to allow social learning (e.g. female dolphins without offspring may engage in allo parental care of infants,

PAGE 31

CAPTIVE DOLPHIN VOCAL BEHAVIOR 23 possibly to learn and practice how to raise infants of their own) (Gibson & Mann, 2007; Mann & Smuts, 1998). In both bottlenose dolphins and pantropical spotted dolphins, i nf ant dolphins tend to stay with their mothers for three to five years (Gibson & Mann, 2007; Herzing, 1997), though infant dolphins begin to "emulate, in the first few weeks of life, the fission fusion nature of the society at large" (Mann & Smuts, 1998, p.1 110) by separating and reuniting with their mothers frequently (though briefly) by the second week of their lives (Mann & Smuts, 1998). Once juveniles separate from their mothers, they often join to form subadult groups composed of both males and females until they reach maturity, usually between eight and twelve years old (Wells, 1991). Males tend to associate with subadult groups longer than females, likely because females reach sexual maturity and quickly become impregnated, and leave the subadult group s to join female bands (Reynolds, et al., 2000; Wells, 1991). Connor, Mann, and Watson Capps (2005) observed significant affiliative behavior (contact swimming) between female bottlenose dolphins ( Tursiops aduncus ) (pregnant, nursing, and single no calf, most commonly in non pregnant single females) in Shark Bay, Australia, that may be a signal of cooperation among females. This cooperation may benefit the females, who were most often observed in male dominated groups, by providing support against the com mon herding and harassment to which they were subjected by male members (Connor et al., 2005). There is some evidence that female dolphins and even some male dolphins of both species have some level of preferential kinship associations (Mann & Smuts, 1998;

PAGE 32

CAPTIVE DOLPHIN VOCAL BEHAVIOR 24 Mšller, Beheregaray, Allen, & Harcourt, 2006; Welsh & Herzing, 2008), which may explain observations of bottlenose dolphin female associative groups consisting of multiple generations of daughters (Reynolds, et al., 2000). Mann and Smuts (1998) noted tha t the few experienced female associates of mother calf pairs tended to be kin. Mixed species groups. The common overlap of different species' ranges and the potential benefits of association have resulted in dolphins often socializing and functioning in large mixed species groups (Connor, Wells, Mann, & Read, 2000; Reynolds et al., 2000). These associations can be agonistic or sympatric in nature, dependent on the species and/or context (Herzing & Johnson, 1997; WŸrsig, 1986). Herzing and Johnson's (1997 ) study of the interaction between groups of Atlantic spotted dolphins and bottlenose dolphins off of the coast of the Bahamas found that 15% of all encounters with either species involved inter species interactions. These interactions included both affili ative (including inter species sexual behavior and allo parental care) and aggressive behaviors, as well as apparent cooperative foraging. Melillo, Dudzinski, and Cornick (2009) observed further interactions between Atlantic spotted dolphins and bottlenose dolphins off the coast of the Bahamas. Over four years, the researchers conducted focal animal sampling of the two groups during inter species interactions, and found that 50% of these interactions were sexual in nature, usually initiated by male bottleno se dolphins. Bottlenose dolphins are well known to associate with many species found within their range (Scott & Chivers, 1990; Reynolds et al., 2000). In Scott and Chivers's (1990) assessment of the distribution of bottlenose dolphins in the Eastern Trop ical Pacific Ocean, bottlenose dolphins were sighted with at least 13 other species of cetaceans,

PAGE 33

CAPTIVE DOLPHIN VOCAL BEHAVIOR 25 including the short finned pilot whale (present in 40% of all mixed species herds), the spotted dolphin (36%), and the spinner dolphin (19%, though only 7% of the mixed species herds that contained both bottlenose and spinner dolphins did not contain spotted dolphins). QuŽrouil, Silva, Casc‹o, Magalh‹es, Seabra, Machete, and Santos (2008) observed the interactions of common dolphins ( Delphinus delphis ), bottlen ose dolphins, striped dolphins, and spotted dolphins in the Azores archipelago. Each of these species were sighted in mixed species groups with each other, with the exception of striped and bottlenose dolphins (QuŽrouil et al., 2008). Bottlenose dolphins h ave been associated with humpbacked dolphins in the Moreton Bay area in Australia, though bottlenose dolphins seem most dominant and prevalent in these groups (Corkeron, 1990). Possible reasons for these mixed species associations may include protection, f oraging efficiency, or simply social behavior. Pantropical spotted dolphins found in mixed species groups with Hawaiian spinner dolphins ( Stenella longirostris ) may have an association based on mutual protection (Norris & Dohl, 1980; WŸrsig, 1986). The two species, though closely related, have different activity patterns; most notably, spinner dolphins are considered nocturnal feeders and rest during the afternoon and evening (Benoit Bird & Au, 2008; WŸrsig, Wells, & Norris, 1994), whereas spotted dolphins forage primarily during the day (Jefferson et al., 1994). Foraging spotted dolphins provide some level of cover or protection while spinner dolphins rest at deeper levels of the open water (Fitch & Brownell, 1968, as cited in Norris & Dohl, 1980), and vice versa at night, when spotted dolphins are resting and spinner dolphins become more active (WŸrsig, 1986). Bottlenose dolphins observed in Mellilo et al.'s (2009) study were often sexually immature or apparently of low social rank (based on rake marks and behavioral

PAGE 34

CAPTIVE DOLPHIN VOCAL BEHAVIOR 26 observation), which may have limited their access to females of their own species. This may have led them to seek copulation with spotted dolphin females (Melillo et al., 2009). QuŽrouil et al. (2008) frequently found mixed species groups engage d in foraging behaviors, suggesting potential cooperative forging tactics that may increase hunting efficiency among the different species, but it is also possible that the mixed groups were a result of separate groups passively associating while pursuing the same food sources. However, species of tuna have often been observed in apparently sympatric groups with spinner and spotted dolphins (who share their diet) (Scott & Chivers, 1990), presumably to benefit from the combined foraging efforts (Scott & Chiv ers, 1990; WŸrsig, 1986). Range Dolphins are both freshwater and saltwater mammals, though the majority of species live in an oceanic environment (Reynolds, Wells, & Eide, 2000). Species of the family Delphinidae (including Tursiops truncatus and Stenel la attenuata ) range throughout temperate, sub tropical, and tropical oceans worldwide, in both coastal to pelagic regions (Jefferson, Leatherwood, & Webber, 1994; Rice, 1977). Bottlenose dolphins ( Tursiops truncatus ) range as far north as the Sea of Okhots k in the North Pacific and as far south as the coastal waters of Tierra del Fuego in the South Pacific/Atlantic oceans, from approximately 60 N to 50 S (Jefferson et al., 1994; Wells & Scott, 1999). Pantropical spotted dolphins ( Stenella attenuata ), tho ugh less prevalent, are also found within this region, ranging as far north as the Sea of Japan and as far south as the Argentine Basin of the South Atlantic, from approximately 40 N to 40 S (Jefferson et al., 1994; Perrin, 1975; Perrin & Hohn, 1994).

PAGE 35

CAPTIVE DOLPHIN VOCAL BEHAVIOR 27 Although the two species' ranges overlap, bottlenose dolphins are mostly associated with coastal regions (Schmidly, 1981; Wells & Scott, 1999), whereas pantropical spotted dolphins are generally more pelagic (Jefferson, et al., 1993; Perrin, 1975; Perrin & Hohn, 1994). Visual and Acoustic Environment The immediate attenuation of light as it penetrates the ocean's depths creates an environment of limited visual scope. Light is scattered and absorbed by the ocean waters at an exponential rate with increas ing depth, leading to drastic reduction of light intensity. At one meter of depth, only 45% of the light energy that falls on the surface of the ocean remains. By 10 meters, long wavelengths have been absorbed, and by 100 meters, only one percent of light energy (short wavelength, or blue green light) remains (Thurman & Trujillo, 1999), leaving a great majority of the dolphin's natural environment visually imperceptible. However, the ocean is a highly acoustic environment, with sound transmitting through sa ltwater at an average velocity of 1,450 meters per second, approximately 4.5 times faster than in air (Thurman & Trujillo, 1999). Dolphin Communication Dolphins have many potential mechanisms for communication. Touch appears to play a significant role in social signaling between conspecifics (e.g., Connor et al., 2005; Paulos, 2004; Herman & Tavolga, 1980; Mann & Smuts, 1999; Tamaki, Morisaka, & Taki, 2006). Tactile behavior with pectorals, flukes, dorsal fins, or trunks between affiliated dolphins is w ell documented (e.g., Herman & Tavolga, 1980; Mann & Smuts, 1998; Tavolga & Essapian, 1957; Norris & Dohl, 1980). Affiliative behavior between dolphins can be characterized, in part, by contact swimming, in which one dolphin rests

PAGE 36

CAPTIVE DOLPHIN VOCAL BEHAVIOR 28 its pectoral fin against the flank of the other, either behind the pectoral or dorsal fin (Connor et al., 2005). This behavior may serve as a touch based communication tool signaling levels of affiliation both to participating and observing parties. Contact swimming is observed al most exclusively in female dolphins, between female female associates (Connor et al., 2005), mother calf pairs (Connor et al., 2005; Mann & Smuts, 1998; Mann & Smuts, 1999; Tavolga & Essapian, 1957, as cited in Connor et al., 2005 and Mann & Smuts, 1999), and even allo mothers and calves (Mann & Smuts, 1998). Connor et al. (2005) noted "petting" (defined as touch between conspecifics lasting under five minutes, p.635) between male conspecifics, but observed no contact swimming. Tamaki, Morisaka, and Taki (2006) studied the tactile behaviors between three captive bottlenose dolphins (two female adults and one male infant) following aggressive interactions, and found that longer spans of time passed between aggressive events between conspecifics that engaged in "flipper rubbing" (p.210) after an aggressive interaction. This correlation suggests that flipper rubbing and possibly other tactile behaviors between dolphins may contribute to the reparation of relationships (Tamaki et al., 2006). Dolphins also hav e a highly developed sense of sight, with adaptations for both in air and underwater vision (Herman, Peacock, Yunker, & Madsen, 1975 ). Bottlenose dolphins are able to see approximately 8.2' of arc underwater and approximately 12.5' of arc in air (Herman et al., 1975). Although their in air acuity appears inferior to their underwater acuity, if the magnification effects of the underwater environment are controlled for, bottlenose dolphins have comparable vision in both environments (Herman et al., 1975). The relative strength of dolphins' sight suggests that it may serve

PAGE 37

CAPTIVE DOLPHIN VOCAL BEHAVIOR 29 an important function in their behavior and, therefore, communication, though when compared to their other strengths, may play on ly a minimal or restricted role Although touch and vision a re both useful tools in communication between conspecifics, they are restricted by the limited visual field of the dolphins' environment and the highly variable degree of distance between members of a group. On the other hand, the ocean provides a favorabl e environment for acoustic signaling, and dolphins have a highly developed sense of hearing which reflects this advantage (Reynolds et al., 2000). Bottlenose d olphins have a hearing range from 0 160 kHz, with fine discrimination between sounds ranging in frequency between 2 kHz and at least 140 kHz (Thompson & Herman, 1975). Hearing is most sensitive between 40 100 kHz (Reynolds, et al., 2000). Humans, in comparison, hear a range from 2 0 Hz to 20 kHz in childhood (Schneider, Trehub, Morongiello & Thor pe, 1986), but high frequency hearing degrades with age in both humans and other species. Fine discrimination in humans is only possible at frequencies below 8 kHz (Thompson & Herman, 1975). Vocal behavior. Given dolphins' uniquely suited environment an d remarkable auditory capabilities, it is not surprising that they have developed an elaborate vocal orientation and communication system. Dolphins are able to emit both broadband sounds that span multiple frequencies simultaneously, and narrowband sounds that are frequency modulated and omnidirectional (Evans, 1967, as cited in Caldwell & Caldwell, 1968). These phonations consist of three major categories: whistles, echolocation clicks, and burst pulses (Lilly & Miller, 1961; McBride & Hebb, 1948; Wood, 19 53). Although there are various constraints to sound travel through the ocean, dolphin sounds can be heard

PAGE 38

CAPTIVE DOLPHIN VOCAL BEHAVIOR 30 over at least 20 km in channels, and up to 2 km in shallow, muddy areas (Quintana Rizzo, Mann, & Wells, 2006). As such, they are an ideal medium for the organization of dolphin society; they are used to maintain contact within widespread groups; to promote cohesion; to identify separate species, groups, or individuals; to coordinate cooperative behavior; as a means of perceiving the environment; and to convey intent (e.g., aggression). Whistles. Whistles are omnidirectional, narrowband, and frequency modulated signals that usually consist of a fundamental whistle below 20 kHz, with harmonics up to 100 kHz (Lammers, et al., 2003), and can last between 0.05 and 3.2 s (Bazœa Dur‡n, 2004). Whistles can have many contours, varying from individual to individual, group to group, and species to species, though many are shared amongst all dolphins (McCowan & Reiss, 1995a). The most prominent whistle contours produced by individual dolphins are individually distinctive, called "signature" whistles that may be utilized as a form of individual identification (Caldwell & Caldwell, 1965), contact and cohesion call (Herzing, 2000; Janik & Slater, 1998; Sayigh, Esch Wells, & Janik, 1997; Tyack, 1997), and even affiliative signal (Watwood, Tyack, & Wells, 2004). Caldwell and Caldwell (1968) recorded four common dolphins' ( Delphinus delphis bairdi ) whistles over 34 days and found that each dolphin produced a unique c ontour specific to each individual, though one subject produced two apparently unique whistles. These signature whistles were not used by any other members of the group, though all four subjects were captured at the same time from the same area, suggesting that they were affiliated in the wild (Caldwell & Caldwell, 1968).

PAGE 39

CAPTIVE DOLPHIN VOCAL BEHAVIOR 31 The signature whistle hypothesis has undergone scrutiny by some researchers; McCowan and Reiss (1995a) utilized a new quantitative technique for categorizing whistle contours, which they termed the "contour similarity technique (CS technique)" (McCowan, 1995) to analyze whistles from three different social groups of bottlenose dolphins. The results suggested that many whistle contours are shared, and one predominant whistle is shared by ev ery individual across every social group (McCowan & Reiss, 1995a). McCowan and Reiss (2001) found that in contexts of isolation, the predominant whistle contour produced by subjects is, in fact, shared, and subtler, individual variation in the production o f the whistle may account for individual recognition. Their results were in contrast to the signature whistle hypothesis suggested by the Caldwells (1965, 1968), which suggested that the predominant whistle of each individual is unique to that individual. However, McCowan's and Reiss's (2001) argument has been refuted by o the r research supporting the existence of signature whistles in dolphins (i.e., Janik & Slater, 1998) With the signature whistle hypothesis in mind, Janik and Slater (1998) studied cap tive bottlenose dolphins in a variety of contexts to understand more about the use of their whistles. To do this, the researchers recorded 4 bottlenose dolphins living together in a two tank system, and noted when the group was together and when one or mor e separated from the group and swam to the other tank. The researchers were able to identify four individualized stereotyped whistles that, due to their prominence in the recordings, were considered signature whistles and were correlated with each of the f our subjects. They found that individual specific stereotyped whistling was virtually non existent when the group was together, but predominant in situations in which one or more

PAGE 40

CAPTIVE DOLPHIN VOCAL BEHAVIOR 32 dolphins were in a visually separate but accessible (both physically and acou stically) location. The use of signature whistles in such a specific context appears to support the hypothesis that these whistles contain identification information and are used as a form of "contact call" or cohesion call, helping the dolphins to keep in touch while separate. The calls were not found to cause any particular behavioral reaction, such as pursuit or rejoining, which suggests that it is simply a tracking type mechanism (Janik & Slater, 1998). The signature whistle hypothesis proposed by the Caldwells (1965) as well as their use as contact calls as suggested by Janik and Slater (1998) has been supported in both captive and free ranging contexts. Over a period of 17 years, Watwood, Owen, Tyack, and Wells (2003) observed and recorded 29 free ra nging dolphins while both free swimming and gently restrained to obtain data on their whistling behavior in both contexts. Of these, 13 males were selected for focal animal sampling to identify individual whistle repertoires. Eleven of the 13 focal animals produced at least one matching whistle type in both conditions. The researchers found that 76% of matched whistles were individually specific signature contours. Free swimming subjects were most likely to produce their signature whistles when voluntarily separated from their allied partners, and least likely to produce them when together with their partners. These findings support the use of signature whistles as contact calls, used to keep partners in touch with each other while apart (Watwood et al., 200 3). This has also been found between mother calf pairs, the other strong pair bond in dolphin society (Sayigh, Tyack, Wells, & Scott 1990). Mello and Amundin (2005) found

PAGE 41

CAPTIVE DOLPHIN VOCAL BEHAVIOR 33 that mother calf pairs whistled more frequently when separated than when together, and more when calves reunited with their mothers than when mothers retrieved their calves. Signature whistles may also be a social signal of affiliation. Watwood et al. (2004) compared the signature whistles produced by allied male dolphins to those of t heir partners and non partners. Both human observers and the CS technique (McCowan, 1995) rated signature whistles of paired males as more similar than the whistles of non paired males (Watwood et al., 2004). Smolker (1993, as cited in Tyack, 1997) found t hat a group of three adult males exhibited signature whistle convergence over two years as they formed a coalition, adopting one dolphins' signature whistle as the most common whistle for all three. The convergence observed in paired male signature whis tles reflects the social learning necessary for dolphins' open and dynamic vocal repertoires. Janik (2000) suggests that the similar variable whistles produced by many dolphins are a product of vocal imitation and whistle "matching" (p. 1355) utilized for individual identification, i.e., one member emits a specific contour, and as a means of acknowledging and/or addressing that member, another repeats the contour back (Janik, 2000). Like humans, dolphins are born into a highly gregarious and vocal society and therefore must begin development of their vocal repertoire immediately upon birth. Mello and Amundin (2005) recorded captive pregnant females for the last seven months of their gestation period and found significant changes in their vocal behavior. T he expectant mothers increased whistling behavior over time, with an accelerated increase in the days prior to their offsprings' births. This behavior may function not only as priming for the newborn infants to respond to their mothers' vocalizations (simi lar to the phenomenon of

PAGE 42

CAPTIVE DOLPHIN VOCAL BEHAVIOR 34 speech recognition found in human newborns, e.g., DeCasper & Fifer, 1980), but also to begin the process of their vocal learning. McCowan and Reiss (1995b) recorded the whistles of captive born infant bottlenose dolphins over the first year of their development and found evidence for vocal plasticity and social learning. Over one year, five out of eleven whistle contours shared between infants and adults were found in the new infant repertoire, suggesting acquisition from the surro unding members of the dolphin group (McCowan & Reiss, 1995b). Fripp, Owen, Quintana Rizzo, Shapiro, Buckstaff, Jankowski, Wells, and Tyack (2005) compared the signature whistles of five dolphin calves to the whistles of dolphins that were either the calves own associates, other calves' associates, captive dolphins, or free ranging dolphins from another area. The calves' whistles were most similar to the whistles of dolphins from the same area (associates of themselves or the other calves), and more likely to be similar to the whistles of dolphins that were rarely in close association with them. These results support the theory that calves may model their signature whistles from the whistles of community members, though not necessarily their closest associat es (Fripp et al., 2005). The tendency for signature whistles to be modeled on those of distant associates rather than the closest members to an individual may be a mechanism of mate selection, reflected by the heritability of signature whistle contours ( Tyack, 1997). Tyack (1997) observed that daughters in Sarasota Bay, Florida, have signature whistles that are more different from their mothers' than sons' are, suggesting a possible mechanism for indications of relatedness, allowing dolphins to better avo id inbreeding. Echolocation clicks. Echolocation clicks are unidirectional trains of broadband pulses with interclick intervals exceeding 10 ms (Lammers, et al., 2002). They have a

PAGE 43

CAPTIVE DOLPHIN VOCAL BEHAVIOR 35 frequency range of 0.2 150 kHz (Reynolds, et al., 2000) with peak freque ncies between approximately 110 and 130 kHz (Au, 1993; Au & Herzing, 2002). Echolocation clicks are vital to the foraging and hunting strategy of dolphins and provide cues about the spatial orientation and environment (Au & Herzing, 2003; Benoit Bird & A u, 2009; Herzing, 1996; Herzing, 2000; Lammers et al., 2003; Wood, 1953; Xitco & Roitblat, 1996). Dolphins emit pulses and wait to receive the "target echo" before sending out another click in order to collect information about the target object (Evans & P owell, 1967; Johnson, 1967). The resulting lag time can vary from at least 15.4 ms to a minimum of 2.5 ms as dolphins near the target (Evans & Powell, 1967). Cooperative behavior. Dolphins are able to take advantage of the information derived from conspec ifics' echolocations, as shown by a study conducted by Xitco and Roitblat (1996). In this study, two dolphins performed an echolocation task, in which one dolphin was allowed to produce echolocation signals towards an object that could not be perceived vis ually, while the non echolocating dolphin listened to the echolocation emissions and echoes from the object. The listener then identified the object ensonified by the echolocating dolphin. This sort of cooperative ability suggests possible use of echolocat ion clicks as indirect cues for group movement, as well as higher hunting efficiency due to less acoustic "traffic" and lower chance of eavesdropping by competing predators. Benoit Bird and Au (2008) found evidence of cooperative behavior and cohesion thro ugh echolocation clicks in foraging groups of Hawaiian spinner dolphins. In contrast to the researchers' hypothesis, whistling behavior (usually associated with cohesion, see Watwood et al., 2004) was significantly less common than echolocation,

PAGE 44

CAPTIVE DOLPHIN VOCAL BEHAVIOR 36 and echolo cation clicks were not emitted from all members of foraging parties (Benoit Bird & Au, 2008). Burst pulses. Burst pulses are similar to echolocation clicks, consisting of broadband pulses in denser groups, with interclick intervals of less than 10 ms (L ammers, et al., 2002). The frequency range of burst pulses has been debated, with reports ranging from between 0.2 16 kHz (Reynolds, et al., 2000) to having little to no energy below 20 kHz (Lammers, et al., 2002). It is possible that this debate has dev eloped from the highly variable nomenclature that may encompass burst pulse vocalizations (Lammers, et al., 2002). Vocalizations described in terms of "rasps," "grates," "mews," "barks," "yelps" (Reynolds, et al., 2000), "squawks," "squeaks," "creaks" (Lam mers, et al., 2002), "bleats," "barks" (Herzing, 2000), etc., represent only a few of the wide variety of terms used to explain the variable sounds emitted by dolphins as they are perceived by humans. This variation of terminology or equipment constraints across many studies may be reflected in the variation in reporting parameters of these sounds (Lammers, et al., 2002). The broad nomenclature used to describe burst pulses has led to a wide variety of sounds and behaviors associated with the classificatio n (Lammers, et al., 2003). Despite this, little is fully understood about the role of burst pulses in dolphin acoustic signaling (Herzing, 2000). It is generally believed that burst pulses play an important social role, especially when non whistling specie s of cetacea are considered (Herzing, 1996; Herzing, 2000; Lammers, et al., 2003). In bottlenose and spotted dolphins, burst pulses are often associated with aggression (Herzing, 1996; Herzing 2000; Lammers, et al., 2003; Overstrom, 1983) as well as other situations of excitement, distress, or alarm (Herzing, 1996; 2000).

PAGE 45

CAPTIVE DOLPHIN VOCAL BEHAVIOR 37 Overstrom (1983) observed the vocal behavior of 5 captive bottlenose dolphins during agonistic interactions. Open mouthed interactions occurred more often between dolphins separated by a fence that limited physical contact but not visual or acoustic contact than between dolphins swimming together "peacefully" (p. 99), suggesting an aggressive association with this behavior. More burst pulses were emitted during mouth to mouth interactions than mouth to body interactions and jaw clap behavior was positively correlated with duration of burst pulse bouts (Overstrom, 1983). These results suggest an association between burst pulse phonations and aggressive behaviors. Social Learning and Dolphin Vocal Behavior Strong evidence of social learning and flexibility in dolphins' vocal repertoires can be seen in the structure and development of signature whistles (Fripp et al., 2005; Janik, 2000; McCowan & Reiss, 1995b; Watwood et al., 2004) as well as in the dynamic and wide ranging contours of non signature (termed "variable") whistles used by dolphins in a variety of other contexts. Dolphins' abilities to learn and incorporate novel sounds into their vocal behaviors are necessary to accommodate the variable composition of cooperative groups of individuals with highly varied vocal repertoires. Sound Imitation and Vocal Learning Richards, Wolz, and Herman (1984) found that captive dolphins can imitate computer generated sounds and incorporate them in to their whistle repertoire. Prior to the study, the dolphin subject had a limited repertoire of specific whistle contours that did not have any similarities to the model sounds produced by the computer. After exposure, the dolphin was able to reproduce im mediately a few of the novel model sounds, and could successfully repeat the sounds (then familiar) on subsequent trials. The dolphin then was

PAGE 46

CAPTIVE DOLPHIN VOCAL BEHAVIOR 38 trained to use these sounds to, in effect, label objects, and could reliably utilize the unique labels to identi fy objects during 91% of trials (Richards et al., 1984). Captive dolphin signature whistles. Miksis, Tyack, and Buck (2002) compared the signature whistles of dolphins in captivity to those of wild conspecifics, and found that captive dolphins developed s ignature whistles that are more similar to the bridge whistles used by their trainers. These whistles are flatter, with fewer "turns" or drastic frequency contours in comparison to wild dolphins, much like the flat whistles of their trainers. Context Dep endent Vocalization by Captive Dolphins Towards Their Trainers In Janik and Slater's (1998) study, an interesting event occurred which suggests the potential use of signature whistles as contact calls between dolphins and their trainers: an occurrence of a late feeding session resulted in a high number of signature whistles emitted by all four dolphins, despite all four members of the dolphin group being within visual contact with one another. Further observation of this event showed that the dolphins ofte n emitted these signature whistles while oriented towards an observation window through which the trainers could be seen. It is possible that the expectation of the trainers' presence at the dolphin tank in order to conduct the feeding session caused the d olphins to attempt to contact the trainers in a similar manner as they would another conspecific that was separated from the group. Of course, this is a speculation that requires significantly more research before such an assertion can be made. However, this anecdote highlights an important issue and piece to the inter species communication puzzle: if captive dolphins do, in fact, use their natural vocal communication mechanisms to attempt to communicate a signal to human trainers, this

PAGE 47

CAPTIVE DOLPHIN VOCAL BEHAVIOR 39 could be used as a measure for bi directionality in inter species communication, an important element to "true" communication (Stevens, 1950). On the other hand, it is possible that Janik and Slater's (1998) observation was a coincidence, a unique situation that held no com municative meaning, but rather was the result of a combination of other variables. Similar to studies of attentional state dependent gestural communication (e.g., Tomasello & Call 1994; Xitco et al., 2004), dolphins must also exhibit attentional state depe ndent vocal communication before any leaps can be made regarding communicative intent. In order to establish this possibility, a within subjects design must be utilized to examine the situational dependence (whether humans are present or not) of dolphin vo cal behavior. Current Study Beginning with the case of Clever Hans (Pfungst, 1911), animal cognition and mechanisms of animal human interaction have undergone close scrutiny. Ideas of what makes humans unique have been challenged repeatedly and undergon e many evolutions as the line between man and beast has blurred with new findings in comparative cognition. Chomsky's claim that complex language is the uniquely human trait (1972) has not been solidly refuted, but research into mechanisms of human animal interaction, communication, and forms of gestural and symbolic "language" is rapidly progressing. Up to this point, most inter species communication research has fallen into two predominant categories: examination of the ontogeny of human language and th e mechanisms necessary for its evolution (e.g., Gardner & Gardner, 1969; Rumbaugh et al., 1973; Savage Rumbaugh et al., 1986; Savage Rumbaugh, 1998; Terrace et al., 1979; Tomasello, 2008), or the role of human interaction on animal communication, either as a

PAGE 48

CAPTIVE DOLPHIN VOCAL BEHAVIOR 40 result of enculturation (e.g., Call & Tomasello, 1994; Kaminski et al., 2005; Maros et al., 2007; Mikl—si et al., 1998; Mikl—si et al., 2005; Mikl—si & Soproni, 2005; Vir‡nyi et al., 2008) or domestication (e.g., Gogoleva et al., 2008, 2009, 2010, & 20 11; Nicastro, 2004; Moin‡r et al., 2010; Schassburger, 1987; Yeon et al., 2011; Yin, 2000). As such, primates and domestic animals have naturally been the most prevalent subjects of human animal communication research. The results of these studies suggest that inter species communication occurs between humans and animals, varying from primitive forms (i.e., gestural communication) (e.g., Call & Tomasello, 1994; Kaminski et al., 2005; Maros et al., 2007; Mikl—si et al., 1998; Mikl—si et al., 2005; Mikl—si & Soproni, 2005; Shapiro et al., 2003; Tomasello, 2008; Vir‡nyi et al., 2008) to complex interactions (i.e., symbolic communication) (e.g., Gardner & Gardner, 1969; Patterson, 1978; Savage Rumbaugh, 1998; Savage Rumbaugh et al., 1986; Savage Rumbaugh et al., 1978; Terrace et al., 1979). Although the progression of human animal communication research resulted in emphasis on primate or domestic animal subjects, John Lilly (1961) may have been ahead of his time with his suggestion that dolphins were superior c andidates for inter species communication. The historic relationship of man and dolphin in mythology, folklore, and reality is evidence of long standing inclusion of dolphins in human social awareness (e.g., Burgoyne, 2001; Busnel, 1973; Cravalho, 2003; Ha mpson,2005; Hall, 2004; Higham, 1960; MacKenzie, 2005; Murray, 1897; Neil, 2002; Pliny the Elder, 1848; Pliny the Younger, 1915; Pryor, 1990; Sax, 2001; Slater, 1994; Taylor, 2003; Tun, 2004). Cooperative fishing exhibited between some dolphin groups and h uman societies around the world suggests a historical awareness, at least, between dolphins and humans (e.g.,

PAGE 49

CAPTIVE DOLPHIN VOCAL BEHAVIOR 41 Busnel, 1973; Lockley, 1979; Neil, 2002; Pliny the Elder, pub. 1848; Pryor, 1990; Tun, 2004), though the extent of intentional cooperation on the behalf of the dolphin may be debatable (WŸrsig, 1986). Like humans and great apes, dolphins are highly gregarious (Connor et al., 2000; May Collado et al., 2007; Pryor & Shallenberger, 1991; Wells, 1991), as reflected in their fluid fission fusion socie ty. Many species of dolphins readily accept sympatric relationships with other dolphin species (Connor, et al., 2000; Herzing, 1996; Herzing, 2000; Johnson, 1986; Lammers, et al., 2003; McCowan & Reiss, 1995; Pryor & Norris, 1998; Reynolds, et al., 2000; T yack, 1997), and regularly associate with non dolphin species (i.e., tuna) in hunting and foraging contexts (Connor et al., 2000). It is possible that the relationship between dolphins and fishermen in cooperative fishing arrangements is an extension of th ese sympatric tendencies. The fluid social structure of dolphins in the wild suggests highly adaptive and inclusive social behavior in dolphins, qualities that encourage the development of cooperative communication. Unlike the inflexible vocal behavior of great apes (Tomasello & ZuberbŸhler, 2002), dolphins utilize an expansive and flexible vocal repertoire (Caldwell & Caldwell, 1968; Lilly & Miller, 1961; McBride & Hebb, 1948; Wood, 1953). The particular demands of dolphins as gregarious mammals necessita te contact, coordination, and recognition of members across broad distances. To accommodate these needs, dolphins utilize both narrowband and broadband sounds (Caldwell & Caldwell, 1968) that can be classified as whistles, burst pulses, and echolocation cl icks (Lilly & Miller, 1961; McBride & Hebb, 1948; Wood, 1953), each with context specific uses (Au & Herzing, 2003; Benoit Bird & Au, 2009; Caldwell & Caldwell, 1968; Herzing, 1996; Herzing,

PAGE 50

CAPTIVE DOLPHIN VOCAL BEHAVIOR 42 2000; Janik & Slater, 1998; Lammers et al., 2003; Sayigh et al., 1997; Tyack, 1997; Watwood et al., 2004; Xitco & Roitblat, 1996). These complex vocal behaviors are vital to nearly all elements of dolphin living; environmental perception, orientation, foraging and hunting, coordination, cohesion, identification, and soc ial interaction all rely on the acoustic cues of individuals and groups (Au & Herzing, 2003; Benoit Bird & Au, 2009; Herzing, 1996; Herzing, 2000; Janik & Slater, 1998; Lammers, et al., 2003; McCowan & Reiss, 2001; Overstrom, 1983; Sayigh, et al., 2007; Ty ack, 1997; Watwood, et al, 2005; Xitco & Roitblat, 1996). The combined results of previous human animal interaction studies illustrate a pattern of salient and important mechanisms for communication: gestural communication comprehension and production (C all & Tomasello, 1994; Kaminski et al., 2005; Maros et al., 2007; Mikl—si et al., 1998; Mikl—si et al., 2005; Mikl—si & Soproni, 2005; Shapiro et al., 2003; Tomasello, 2008; Vir‡nyi et al., 2008), contextually flexible vocal behavior (Tomasello & ZuberbŸhl er, 2002; Yin, 2002) and awareness of the attentional state of a communicative partner (Call & Tomasello, 1994; Xitco et al., 2001, 2002). Captive dolphins have demonstrated all of these traits: They can comprehend the intent of the human point gesture (He rman, et al., 1999; Herman & Pack, 2007), and even develop a modified "point" gesture during cooperative tasks with humans (Xitco et al., 2001). As residents of a highly acoustic environment, dolphins utilize a wide variety of contextually specific complex phonations to communicate between individuals and groups (Au & Herzing, 2003; Benoit Bird & Au, 2009; Herzing, 1996; Herzing, 2000; Janik & Slater, 1998; Lammers, et al., 2003; McCowan & Reiss, 2001; Overstrom, 1983; Sayigh, et al., 2007; Tyack, 1997; Wat wood, et al, 2005; Xitco & Roitblat, 1996), and can add to their

PAGE 51

CAPTIVE DOLPHIN VOCAL BEHAVIOR 43 vocal repertoire through social learning (Fripp et al., 2005; Janik, 2000; McCowan & Reiss, 1995b; Miksis et al., 2002; Richards et al., 1984; Watwood et al., 2004) and other acoustic influen ces throughout their lives In cooperative tasks, captive dolphins visually monitor their human partners and attend to their partner's attentional state before initiating communication (Xitco, et al., 2004). These qualities combined with dolphins' other s imilarities to humans (e.g., gregarious society, vocal communication) suggest that dolphins are, in fact, likely candidates for inter species communication. Dolphins also present a previously unavailable opportunity to investigate bi directional communic ation from a new perspective. Vocal behaviors are a quantifiable measure of interaction between dolphin conspecifics (Watwood et al., 2004), and this may be true between humans and dolphins as well. Janik & Slater's (1998) study included an anecdote that s uggested that captive dolphins might use vocal behaviors to attempt communication with human trainers. However, current literature contains no information regarding the overall effect of human presence over the vocal behaviors of captive dolphins, and with out this integral step, it is difficult to move forward with further investigation of Janik & Slater's (1998) suggestion. The current study seeks to set the groundwork for further investigation of inter species communication between humans and dolphins. In order to establish whether human interaction has a significant, quantifiable effect on the vocal behavior of captive dolphins, we designed a within subjects measure of vocal behavior when training sessions (thus, trainer interaction) did and did not occ ur. If human interactions do impact the vocal behavior of dolphins, we would expect to see a difference between the number of phonations and type of vocalizations produced in each condition. Based on the

PAGE 52

CAPTIVE DOLPHIN VOCAL BEHAVIOR 44 prevalence of vocal behavior as a social signal in d olphins, we would expect that trainer interactions during training sessions would result in higher numbers of vocalizations. The circumstances surrounding this study offered a unique opportunity for comparison. Shortly before the beginning of the study, the tank mate of one of the study's subjects, Moonshine, passed away. Since that time, he has been living alone, a rare situation for captive dolphins (see United Nations Environmental Programme, 2006). His social environment is a stark contrast to that of the other four subjects of the study, who live together, forming an environment that is closer to what they would experience in the wild. These contrasts suggest that there will be significant differences between the vocal behaviors of the two "groups," such that Moonshine will vocalize less overall than the social group, but that traine r interactions via training session occurrence will have a more significant effect on his vocal behaviors. Method Subjects Five male dolphins were observed in the study. Ranier, age (approximately) 31, Khyber, 20, Calvin, 18, and Malabar, 11, were bottlenose dolphins ( Tursiops truncatus ) housed in a public facility in central Florida (hereafter called "bottlenose facility"). Moonshine, 12, was a pantropical spotted dolphi n ( Stenella attenuata ), housed in a marine hospital and research laboratory in southwest Florida (hereafter called "spotted facility"). As Call and Tomasello (1994) and Mikl—si and Soproni (2005) demonstrated, it is important to understand the individual history of each subject in studies of human animal

PAGE 53

CAPTIVE DOLPHIN VOCAL BEHAVIOR 45 interactions. Differences in enculturation and overall human interaction in the subjects' upbringings can influence their current interactive behaviors (Call & Tomasello, 1994). Ranier was wild caught fr om the Gulf of Mississippi in 1988 and had lived at three other facilities before his current residence. Calvin's history is unique: he was born in captivity in 1994, but was orphaned at four months of age, at which time he was living with four female dolp hins with nursing calves of their own. Calvin nursed from all four females in addition to being fed fish by his human caretakers. He lived at his birth facility before moving to his current residence in 2003. In November of 2005, Calvin and Ranier were joi ned by Khyber and Malabar. Khyber was captive born in 1992 at a facility in the Florida Keys and resided in one other facility before arriving at his current residence. Malabar was captive born in 2000 at a facility in Bermuda. All four dolphins had been i nvolved in previous cognitive and behavioral studies. The four dolphins live together in a large tank system at their current facility but are usually separated into pairs (Ranier and Khyber, Calvin and Malabar). Moonshine stranded in the Florida Keys in the summer of 2003 and was found with severe sunburns on his dorsal surface. He was brought to his current home for rehabilitation but was later deemed unreleasable by the National Marine Fisheries Service due to health problems. Moonshine was very young at the time of his stranding and was, in effect, raised by his trainers from that point on. In 2005, Moonshine was joined by an infant Hawaiian spinner dolphin and shared a tank with her until her death in September of 2011. He has lived alone since. In or der to counter the effects of Moonshine's solitude, his trainers interact with him regularly, conducting up to eight formal sessions (sometimes more) per day as well as many more informal interactions.

PAGE 54

CAPTIVE DOLPHIN VOCAL BEHAVIOR 46 Facilities Bottlenose dolphin facility. The bottlenos e dolphins live in a large tank comprising three pools connected by gates and a canal: Main Tank, A pool, and B pool. The Main Tank is a portion of a larger tank that holds a variety of aquatic life on public exhibit. The entire system holds 22 million lit ers of salt water. A and B pools are not visible to the public and only accessible to facility staff. A pool measures 7.6m by 7m by 2.1m deep, and B pool measures 8.2m by 7m by 2.1m deep. The recording station was located near B pool, visible from the back pools (see Figure 1 for a diagram of the tank system). Figure 1. Bottlenose Facility Tank System.

PAGE 55

CAPTIVE DOLPHIN VOCAL BEHAVIOR 47 Pantropical spotted dolphin facility. Due to reconstruction of Moonshine's main habitat during the observation period, recording occurred in two loca tions, the Lagoon, and the Med Tank. The Lagoon is a publicly viewable area; the Med Tank is only accessible to staff and researchers. Lagoon. Moonshine lives primarily in an oval tank that holds 757,000 liters of salt water, measuring 38.1m by 15.25m. A dock and gate split the tank into two secti ons: the shallower east end, 1.3 m de ep, and the deeper west end, 3.7 m deep. The recording station was located on the north side of the Lagoon, visible from both the tank and the observation areas surrounding the tank (see Figure 2 for a diagram of the Lagoon). Figure 2. Pantropical Spotted Dolphin Facility: Lagoon.

PAGE 56

CAPTIVE DOLPHIN VOCAL BEHAVIOR 48 Med Tank. Moonshine was intermittently housed in a round holding tank from January 24 th March 2 nd The med tank was 9.14 m in diameter, 2.74 m deep, and held 189,271 liters of salt water. The recording station was located beside the ladder area used to enter and exit the tank, partially obscured by a shade cloth (see Figure 3 for a diagram of the Med Tank). Materials Hydrophones. Both facili ties were equipped with High Tech, Inc. HTI 96MIN hydrophones (flat frequency response of 2 Hz to 30 kHz, although the actual recording range appeared to be 0 Hz to 50 kHz). Computers. We recorded on a Lenovo T410 laptop computer at the bottlenose facili ty and a Dell Inspiron 1520 laptop computer at the spotted facility. Figure 3. Pantropical Spotted Dolphin Facility: Med Tank.

PAGE 57

CAPTIVE DOLPHIN VOCAL BEHAVIOR 49 Recording/Analysis Programs. We used Avisoft RECORDER USG version 4.2.8 for all recordings at the bottlenose facility, a nd Avisoft SASlab Pro version 5.1.22 for all recordings at the spotted facility. We used Avisoft SASlab Pro Sound Analysis and Synthesis Laboratory version 5.2.01 to convert all recordings (recorded as .wav files) into spectrograms. Video Recording Devic es. The bottlenose facility used a PC Osprey 4 channel video card with H.264 Webcam software to simultaneously video record the surface of A pool and B pool, as well as both above and under water in the Main Tank. A Submertec Spyball underwater camera (mo del SB CZ) was used to view the Main Tank at the bottlenose facility. A Canon Vixia HF R20 camcorder was used to record Moonshine's behavior during recording sessions at the spotted facility. Equipment configuration. Bottlenose dolphin facility. The bottl enose facility was equipped with six hydrophones placed approximately one meter deep in locations chosen to maximize triangulation to identify vocalizing dolphins. We used hydrophones 1, 2, 3, and 6 to record A pool, B pool, and the Main Tank simultaneousl y (see Figure 1 for an illustration of the hydrophone array). Pantropical spotted dolphin facility. The spotted dolphin facility was equipped with one hydrophone. In the Lagoon, we placed the hydrophone approximately one meter deep near Moonshine's usual target station (see Figure 2 for hydrophone location), an area he apparently preferred. In the Med Tank, we placed the hydrophone approximately one meter deep near the entrance area of the tank (see Figure 3 for hydrophone location).

PAGE 58

CAPTIVE DOLPHIN VOCAL BEHAVIOR 50 Procedure Recording sessions. Data collection occurred over 73 days from January 10, 2012 to March 22, 2012. We recorded one session per day of observation at each facility (n=8). Recording sessions were haphazardly assigned to either the "training" (TS) (n=4) or "no training (NTS) (n=4) conditions. Bottlenose facility recording sessions. To balance all variables, each bottlenose subject pair participated in two recording sessions, one training session and one no training session. All recording began at approximately 1400 h ours, when the dolphins were routinely separated into pairs for training. One pair moved to the main tank for a "window" session (a training session that audiences could view), while the other pair remained in A and B pools. Only vocalizations emitted by t he pair in A and B pools were included in analyses. TS condition. In the TS condition, the trainer entered the tank area and began a session 15 minutes after recording had started. During the session, pairs were intermittently separated by a grate and wer e restricted to either A or B pool. Training sessions consisted of informal "play" behaviors, such as object manipulation (e.g., tossing balls, beaching on floating mats). The dolphins received food rewards during the session. The trainer did not cue vocal izations. The training session lasted 15 minutes after which the trainer left the area. Recording continued until the window session in the main tank ended, and the second pair returned to the back pools (M = 775.5 sec, SD = 312.9). NTS condition. Recordi ng began at the same time, 1400 hours, but no training session or trainer dolphin interactions occurred. Recording continued for 45 minutes, or

PAGE 59

CAPTIVE DOLPHIN VOCAL BEHAVIOR 51 until the window session in the main tank ended. Subjects were able to access both A and B pools throughout the recording session. Spotted facility recording sessions Moonshine was recorded during four sessions, balanced between the two conditions. To counter balance variables, one training session and one no training session occurred in each tank. All recording se ssions began at approximately 1500 hours. TS condition. In the TS condition, the trainer entered and began a session 15 minutes after recording had started. In the Lagoon, the training session included husbandry behaviors, object retrieval/delivery, and a swim session with the trainer, which included free swimming (synchrony), imitation tasks, object retrieval, and husbandry behaviors. In the Med Tank, the training session consisted of a swim session with the trainer, including the same behaviors. The train er did not cue vocalizations during the sessions. Training sessions lasted 15 minutes after which the trainer left the area. Recording continued for 15 minutes after the session. NTS condition Recordings began at the same time, 1500 hours, but trainers re mained out of view of the dolphin and no training session or trainer dolphin interactions occurred. Recording continued for 45 minutes. Behavioral Observation. We noted the initial location (i.e., A pool, B pool, Main Tank; East or West end) and behaviora l states of each subject at the start of each recording session. Any changes in behavior and location of the subjects were noted by a researcher at the recording stations as they occurred. All recording sessions were video recorded to provide visual feedba ck of the behaviors and locations of the subjects.

PAGE 60

CAPTIVE DOLPHIN VOCAL BEHAVIOR 52 Analysis/Playback settings. We recorded all sessions at a rate of 100,000 Hz/second and listened to audio playback both in real time and reduced by 48%. Due to the quietness of the recordings obtained at the spotted facility, all of Moonshine's recordings were later modulated in amplitude at a constant rate of six dB/10 ms to increase volume and, therefore, clarity on the spectrogram. We also converted the sampling frequency from 100,000 Hz to 48,000 Hz o n those recordings to make the spectrogram clearer at lower frequencies, since the fundamental frequencies of Moonshine's vocalizations did not range above 30 kHz Spectrogram Review and Phonation Extraction. We visually reviewed spectrograms on Avisoft S ASlab Pro while listening to the audio. Phonations were numbered and then sorted into the following categories: burst pulse, echolocation click, single click, jaw clap, or whistle (See Figure 4 for spectrog ram examples of each phonation type). Compound voc alizations, in which a single dolphin emitted more than one phonation type simultaneously (see Figure 4), were categorized as both the combination of phonations and the individual phonation elements. Whistles were further separated into subcategories b y co ntour (see Appendix A for descriptions and examples of each whistle contour). We reviewed video files in real time to cross reference above water vocalizations and confirm jaw claps. We omitted any above water vocalizations that did not appear in the spe ctrograms from final analyses. We reviewed videos as necessary to confirm locations of bottlenose subjects.

PAGE 61

CAPTIVE DOLPHIN VOCAL BEHAVIOR 53 Figure 4. Spectrogram examples of vocalization types. Horizontal axes measure time, and vertical axes measure frequency (30 50 kHz). Results A total of 2158 phonations were extracted from eight 45 minute recordings of five dolphins. A total of 1461 phonations were extracted from Moonshine's four sessions, 1384 from the two Lagoon sessions, and 77 from the two Med Tank sessions. A total of 698 phonations were extracted from the bottlenose subjects' sessions, 486 from Khyber and Ranier's two sessions, 211 from Calvin and Malabar's two sessions (see Table 1). The data obtained show that the TS (training session) condition elicited more phonation behavior than the NTS (no training session) condition in some subjects, but not in others. Moonshine emitted a much higher number of phonations during the TS sessions than the NTS sessions in both the Lagoon (TS = 805, NTS = 579) and the Med

PAGE 62

CAPTIVE DOLPHIN VOCAL BEHAVIOR 54 Tank (TS = 61 NTS = 16). Ranier and Khyber did not show as much variability but did emit more phonations during the TS session (N = 318) than the NTS session (N = 168). In contrast, Calvin and Malabar did not show any difference in overall number of phonations betwee n the two conditions (TS = 108, NTS = 103) (see Figure 5). Total Number of Phonations Emitted by Subjects in Training and No Training Sessions Subject Figure 5. Total number of phonations emitted by subjects between conditions.

PAGE 63

CAPTIVE DOLPHIN VOCAL BEHAVIOR 55 Distribution of M oonshine's Phonation Behavior in Lagoon Sessions Figure 6. Total number of emissions of each phonation type in both TS and NTS sessions (left), TS session alone (top right), and NTS session alone (bottom right). Moonshine Lagoon Phonat ions included whistles, echolocation clicks, jaw claps, single clicks, and burst pulses. Whistles made up 42.85% of overall phonations (TS = 22.36%, NTS = 71.33%), echolocation clicks made up 34.47% (TS = 44.72%, NTS = 20.21%), jaw claps made up 17.11% (TS = 25.95%, NTS = 7.08%), single clicks made up 3.63% (TS = 2.48%, NTS = 1.38%), and burst pulses made up 3.29% (TS = 4.47%) (see Figure 6 for distributions).

PAGE 64

CAPTIVE DOLP HIN VOCAL BEHAVIOR 56 Distribution of Moonshine's Phonation Behavior in Med Tank Sessions Figure 7. Tot al number of emissions of each phonation type in both TS and NTS sessions (left), TS session alone (top right), and NTS session alone (bottom right). Med Tank Phonations included whistles, echolocation clicks, single clicks, and burst pulses. Echolocat ion clicks made up 36.36% of overall phonations (TS = 37.70%, NTS = 31.35%), single clicks made up 32.47% (TS = 40.98%), whistles made up 15.58%) (TS = 1.64%, NTS = 68.75%), and burst pulses made up 15.58% (TS = 19.67%) (see Figure 7 for distributions).

PAGE 65

CAPTIVE DOLP HIN VOCAL BEHAVIOR 57 Comparison of Phonation Emission Distribution Between Med Tank and Lagoon Figure 8. Total number of phonation emissions across phases of each condition in both the Lagoon and Med Tank.

PAGE 66

CAPTIVE DOLP HIN VOCAL BEHAVIOR 58 Distribution of phonation emissions within conditions in lagoon. Pre In the pre session phase, Moonshine emitted 15.03% of overall TS session phonations, and 45.42% of overall NTS session phonations. During During the training session phase, 11.55% of overall TS session phonations were produced, and 27. 29% of overall NTS session phonations were produced. Post In the post session phase, 73.42% of TS phonations were produced, and 27.29% of NTS phonations were produced (see Figure 9 for graphed distribution of phonations across conditions). Total Phonat ions Emitted Over Time (Phase) of Sessions Phase of Session Figure 9. Total number of phonations emitted by Moonshine in the pre session, during session, and post session phases of the Training and No Training Session Lagoon recording sessions.

PAGE 67

CAPTIVE DOLP HIN VOCAL BEHAVIOR 59 Compo und vocalizations Compound vocalizations were only observed during the Lagoon sessions. In the pre session phase, Moonshine emitted 17.22% of overall TS session compound vocalizations, and 56.92% of overall NTS session compound vocalizations. In the train ing session phase, .66% of overall TS session compound vocalizations were produced, and 24.62% of overall NTS session compound vocalizations were produced. In the post session phase, 82.12% of overall TS session compound vocalizations were produced, and 18 .46% of overall NTS compound vocalizations were produced (see Figure 10 for graphed distribution of compound vocalizations). Total Number of Compound Vocalizations Emitted by Moonshine Across Phases of Sessions Phase of Session Figure 10. Distributio n of compound vocalizations emitted by Moonshine in the Lagoon in Training and No Training Session sessions.

PAGE 68

CAPTIVE DOLP HIN VOCAL BEHAVIOR 60 Distribution of phonation emissions within conditions in med tank. Pre In the pre session phase, Moonshine emitted 3.29% of overall TS sessio n phonations, and 25% of overall NTS session phonations. During During the training session phase, 85.25% of overall TS session phonations were produced, and 18.75% of overall NTS session phonations were produced. Post In the post session phase, 11.4 8% of TS phonations were produced, and 56.25% of NTS phonations were produced (see Figure 11 for graphed distribution of phonations across conditions). Total Phonations Emitted by Moonshine Across Phases of Med Tank Sessions Phase of Session Figure 11. Total number of phonations emitted by Moonshine across phases of the Training and No Training Session sessions in the Med Tank.

PAGE 69

CAPTIVE DOLP HIN VOCAL BEHAVIOR 61 Distribution of Phonation Types Emitted by Moonshine Between Phases of TS and NTS Sessions in Lagoon Pre Session Durin g Session Post Session Figure 12. Total number of each phonation type emitted during the three phases (pre during, and post ) of each TS session ( left column) and NTS session ( right column).

PAGE 70

CAPTIVE DOLP HIN VOCAL BEHAVIOR 62 Distribution of Moonshine's phonation types within conditions. Lagoon Pre Within the pre session phase, whistles, echolocation clicks, single clicks, and jaw claps were found. Whistles made up 55% of all phonations (TS = 43%, NTS = 61%), echolocation clicks made up 33% (TS = 55%, NTS = 23%), single c licks made up 1% (TS = 2%, NTS = 1%), and jaw claps made up 10% (NTS = 15%) (see Figure 12 for charted distributions). During During the training session phase, whistles, echolocation clicks, burst pulses, and single clicks were found. Whistles made up 56% of all phonations (TS = 17%, NTS = 79%), echolocation clicks made up 25% (TS = 38%, NTS = 18%), and single clicks made up 5% (TS = 8%, NTS = 3%). Burst pulses were only found in the TS session (38%) (see Figure 12). Post Within the post session pha se, whistles, echolocation clicks, burst pulses, single clicks, and jaw claps were found. Whistles made up 32% of all phonations (TS = 44%, NTS = 81%), echolocation clicks made up 38% (TS = 44%, NTS = 18%), burst pulses were only found in the TS session (1 7%), single clicks made up 2% (TS = 2%, NTS = 1%), and jaw claps made up 28% (TS = 35%, NTS = 1%) (see Figure 12).

PAGE 71

CAPTIVE DOLP HIN VOCAL BEHAVIOR 63 Distribution of Phonation Types Emitted by Moonshine Between Phases of TS and NTS Sessions in Med Tank Pre Session During Sessi on Post Session Figure 13. Total number of each phonation type emitted during the three phases (pre during, and post ) of each TS session ( left column) and NTS session ( right column).

PAGE 72

CAPTIVE DOLP HIN VOCAL BEHAVIOR 64 Med Tank Pre Within the pre session phase, Moon shine only emitted echolocation clicks and a burst pulse. Echolocation clicks made up 83.33% (TS = 50%, NTS = 100%), burst pulses made up 16.67% (TS = 50%) (see Figure 13 to compare distributions). During During the training session phase, whistles, echo location clicks, burst pulses, and single clicks were found. Whistles made up 7.4% of all phonations (TS = 1.85%, NTS = 100%). Echolocation clicks (34.54%), burst pulses (20.37%), and single clicks (41.18%) were only found in the TS session (see Figure 13) Post Within the post session phase, echolocation clicks, single clicks, and whistles were found. Echolocation clicks made up 25% of all phonations (TS = 11.11%, NTS = 42.86%). Single clicks were only found in the NTS session (57.14%), and whistles wer e only found in the TS session (88.89%) (see Figure 13).

PAGE 73

CAPTIVE DOLP HIN VOCAL BEHAVIOR 65 Distribution of Whistle Contours Figure 14. Total whistle contour distribution of both TS and NTS sessions (left), with TS session distribution (top right) and NT S distribution (bottom right). See Appendix A for definitions of contours. Distribution of whistle contours across conditions. In the Lagoon sessions, 12 contours between 593 total whistles were identified. The contour "Moonshine" (unique to Moonshine, not found in the bottlenose dolphin subjects' repertoire) was most prevalent overall (74.37%), followed by "flat" (12.31%). During the TS condition, "Moonshine" was the most prevalent (41.67%), followed by "flat" (29.44%) and "complex" (10.56%). During the NTS condition, "Moonshine" was most prevalent (88.62%), followed by "flat" (4.84%) (See Figure 14 for complete distributions). In the Med Tank sessions, three contours were identified between 12 total whistles. In the TS session, only one whistle, "up," was emitted. In the NTS session, "Moonshine" was most prevalent (81.81%).

PAGE 74

CAPTIVE DOLP HIN VOCAL BEHAVIOR 66 Distribution of Whistle Contours Between Phases of TS and NTS Sessions in Lagoon Pre Session During Session Post Session Figure 15. Total number of each whistle c ontour emitted across the three phases (pre during, and post ) of TS sessions (left column) and NTS sessions (right column) in the lagoon.

PAGE 75

CAPTIVE DOLP HIN VOCAL BEHAVIOR 67 Distribution of Moonshine's whistle contours within conditions. Lagoon Pre Whistles made up 55% of all ph onations (TS = 43%, NTS = 61%). Within those, seven contours were identified: "complex" (TS = 6%, NTS = 6%), "flat" (TS = 4%, NTS =10%), "up" (TS = 4%, NTS =3%), "rug" (NTS = 1%) and "Moonshine" (TS = 85%, NTS =79%). The "boat" contour was only identified in the TS session (2%), and the "gur" contour was only identified in the NTS session (1%) (see Figure 15 for charted distributions). During Whistles made up 56% of all phonations (TS = 17%, NTS = 79%). Within those, eight contours were identified: "up" ( TS = 25%, NTS = 4%) and "Moonshine" (TS = 56%, NTS = 88.62%). The "complex" (.8%), "flat" (1.6%), "gur" (.8%), and "rug" (.8%) contours were only identified in the NTS session, and the "blip" (2%) and "reverse tilde" (6%) contours were only identified in t he TS session (see Figure 15). Post Whistles made up 32% of all phonations (TS = 44%, NTS = 81%). Within those, 9 contours were identified: "boat" (TS = 2%, NTS =.78%), "Moonshine" (TS = 20%, NTS = 96.88%), and "flat" (TS = 46%; NTS = 1.56%). The "gur" contour was only identified in the NTS session (.78%), and the "bump" (2%), "down" (2%), "reverse tilde" (2%), "up" (4%), "crinkle" (10%), and "complex" (14%) contours were only found in the TS session (see Figure 15).

PAGE 76

CAPTIVE DOLP HIN VOCAL BEHAVIOR 68 Distribution of Whistle Cont ours in the During Session Phase of TS and NTS Sessions in Med Tank Figure 16. Total number of each whistle contour emitted during the during session phase of the TS session (left) and NTS session (right) in the med tank. Med Tank Pre No whis tle phonations were emitted in the pre session phase in either condition. During Whistles made up 7.27% of all phonations (TS = 1.92%, NTS = 100%). In the TS session, only one "up" whistle was produced. In the NTS session, one "up" whistle, one "reverse tilde" whistle, and one "Moonshine" whistle was produced (see Figure 16). Post Whistles were only found in the post session phase of the NTS session. All of the whistles (N = 8) were "Moonshine" contours.

PAGE 77

CAPTIVE DOLP HIN VOCAL BEHAVIOR 69 Distribution of Khyber's and Ranier's P honation Behaviors Figure 17. Total number of emissions of each phonation type in both TS and NTS sessions (left), in TS session alone (top right) and NTS session alone (bottom right). Khyber and Ranier Phonations included echolocation clicks, burst pulses, whistles, and single clicks. Echolocation clicks made up 45.68% of overall phonations (TS = 30.18%, NTS = 75%), burst pulses made up 29.63% (TS = 42.14%, NTS = 5.95%), whistles made up 21.39% (TS = 26.42%, NTS = 11.90%), and single clicks mad e up 3.29% (TS = 1.26%, NTS = 7.14%) (see Figure 17 for distribution).

PAGE 78

CAPTIVE DOLP HIN VOCAL BEHAVIOR 70 Distribution of Whistle Contours Emitted by Khyber and Ranier in TS and NTS Sessions Figure 18. Total number of each whistle contour emitted by Khyber and Ranier in TS and NTS Sessions (left), TS session alone (top right), and NTS session alone (bottom right). Distribution of whistle contours emitted by Khyber and Ranier. Between 104 total whistles, 12 contours were identified. The contour "up" was most prevalent overall (24.04%), followed by "flat" (22.12%), and "Khyber" (20.19%). During the TS condition, "Khyber" was most prevalent (25.00%), followed by "up" (21.43%) and "flat" (17.86%). During the NTS condition, "flat" was most prevalent (40%), followed by "up" (35%) (See Figure 18 for complete distribution).

PAGE 79

CAPTIVE DOLP HIN VOCAL BEHAVIOR 71 Distribution of Phonation Types Emitted by Calvin and Malabar Between Conditions Figure 19. Total number of each phonation type emitted by Calvin and Malabar in TS and NTS Sessions (l eft), TS session alone (top right), and NTS session alone (bottom right). Calvin and Malabar Phonations included echolocation clicks, burst pulses, whistles, and single clicks. Echolocation made up 54.98% of overall vocalizations (TS = 38.89%, NTS = 71. 84%), burst pulses made up 25.59% (TS = 41.67%, NTS = 8.73%), whistles made up 9.95% (TS = 16.67%, NTS = 2.91%), and single clicks made up 9.48% (TS = 2.77%, NTS = 16.50%) (See Figure 19 for distribution).

PAGE 80

CAPTIVE DOLP HIN VOCAL BEHAVIOR 72 Pre In the pre session phase, Calvin and Mal abar emitted .93% of overall TS session phonations and 1.94% of overall NTS session phonations. During During the training session phase, 93.52% of overall TS session phonations were produced, and no phonations were produced in the NTS session. Post In the post session phase, 5.95% of TS phonations were produced, and 98.06% of NTS phonations were produced (See Figure 20 for graphed distribution of phonations across conditions). Total Phonations Emitted by Calvin and Malabar Across Phases of Training and No Training Sessions Phase of Session Figure 20. Total number of phonations emitted by Calvin and Malabar across phases of the Training and No Training Sessions.

PAGE 81

CAPTIVE DOLP HIN VOCAL BEHAVIOR 73 Distribution of Whistle Contours Emitted by Calvin and Malabar in TS and NTS Se ssions Figure 21. Distributions of whistle contours emitted in both TS and NTS sessions (left), TS session alone (top right), and NTS session alone (bottom right). Distribution of whistle contours. Within 21 total whistles, six contours were identified. The contour "flat" was most prevalent overall (38.09%), followed by "up" (23.81%). During the TS condition, "flat" was most prevalent (33.33%), followed by "up" (27.78%). During the NTS condition, "flat" and "bump" made up 100% of all whistles (see Figure 21 for complete distributions). Discussion Effect of Conditions on Phonation Rates The first hypothesis, that captive dolphins will produce more phonations during sessions in which training sessions occur, was partially supported by this s tudy. There

PAGE 82

CAPTIVE DOLP HIN VOCAL BEHAVIOR 74 was a clear difference in the number of phonations emitted across conditions by some, but not all, of the subjects. However, the limited sample size and highly variable data suggest that these results may not necessarily indicate an absolute ef fect of training sessions (and thus trainer interaction) on phonations, but rather may be the result of coincidence or other underlying factors. The second hypothesis, that Moonshine will vocalize less overall than the bottlenose group but will show a g reater variability in the number of vocalizations between training and no training session sessions, was also only partially supported. The effect of TS sessions on increased vocal behavior was, in fact, strongest in Moonshine's phonation behavior in the L agoon, but also affected Khyber's and Ranier's phonation behavior, though to a lesser extent. Calvin and Malabar showed no changes in their phonation behavior between TS and NTS sessions. On the other hand, Moonshine emitted a high number of phonations ( N = 1461) overall, more than twice the number of phonations emitted in all conditions by all four bottlenose dolphins combined (N = 697). This directly refutes the first half of the second hypothesis but is not entirely surprising when other variables are considered. The high variability found in the results across the three subjects (or subject pairs) makes it impossible to conduct a between subject comparison of Moonshine, Khyber and Ranier, and Calvin and Malabar. In general, the sample size of this s tudy was not robust enough to generalize these results as the typical behavior of any of its subjects. However, if these results do reflect the typical responses of these subjects to the occurrence of training sessions, they may be explained by exploring t he interaction between species

PAGE 83

CAPTIVE DOLP HIN VOCAL BEHAVIOR 75 differences, levels of sociability (used here in terms of dolphin to dolphin relations), individual levels of enculturation, and stages of social development. Differences in Effect of Conditions Between Subjects Perhaps the first explanation that should be addressed is the difference in species between Moonshine and Khyber, Ranier, Calvin, and Malabar. Though both pantropical spotted dolphins and bottlenose dolphins are gregarious by nature, bottlenose dolphins are found in smaller groups, on average, than are pantropical spotted dolphins (Jefferson et al., 1994). The prevalence of vocal communication in the organization, coordination, and cohesion of dolphins as a whole (Au & Herzing, 2003; Herzing, 1996; Herzing, 2000; Jani k & Slater, 1998) may make pantropical spotted dolphins, as a species, more likely to exhibit higher rates of phonation overall. Unfortunately, little is known about the relative phonation rates between bottlenose dolphins and pantropical spotted dolphins, mostly due to the limited scope of captive pantropical spotted dolphin research. Moonshine is only one of two of his species in captivity in the world, so although the data obtained from this and other studies he participates in are invaluable, there are few data to contextualize it. Effects of solitude on inter species socialization Moonshine's considerable uniqueness is not limited to his rarity as a captive pantropical spotted dolphin; it is possible that his high rates of phonation were also a resul t of his solitude. With little social interaction outside of his interactions with his trainers, it is likely that he is more responsive to these occasions. This theory is supported when Moonshine's phonation behavior in the Lagoon is compared to his pho nation behavior in the Med Tank. Moonshine emitted nearly twenty

PAGE 84

CAPTIVE DOLP HIN VOCAL BEHAVIOR 76 times as many phonations while in the Lagoon (N=1384) than while in the Med Tank (N=77). This is potentially due to the substantially reduced opportunity for human interaction in the Med Tank as compared to the Lagoon. In order to fully understand the implications of this comparison, it is important to understand a few significant details about these two environments: The Lagoon, Moonshine's usual habitat, is a pool that is located in the center of an outdoor aquarium compound, viewable only from the surface but entirely visible. A stream of guests regularly pass es through the area to view Moonshine, resulting in nearly constant human proximity and attention. The Med Tank, in contrast, is b ehind a walled barrier that is off limits to guests. It has a raised wall that ends approximately 1 meter above ground level, and the water level is kept low enough that there is a steep drop from the top of the tank to the pool itself. This resulted not o nly in reducing the possibility for, and availability of, humans in Moonshine's environment, but also significantly reduced Moonshine's ability to see outside of his own tank. While in the Lagoon, Moonshine had ample opportunity to pursue interactions wi th both passing guests and his trainers. In the Med Tank, however, Moonshine's social interaction was entirely restricted to the time his trainer spent inside the tank with him, since he could not so much as see humans outside of that time. The great dif ference between the numbers of phonations Moonshine emitted while in the Lagoon compared to the Med Tank suggests that his vocal behaviors may be affected by the presence of humans in his environment on a broad level. However, it is also possible that Moon shine was responding to stress from the relocation or other outside variables Aside from the stress of relocation, Moonshine's phonation behavior

PAGE 85

CAPTIVE DOLP HIN VOCAL BEHAVIOR 77 may have been affected by the physics of the Med Tank itself. However, baseline recordings of the tank before Moonshine resided in it show that the level of ambient sound was below 5 kHz, comparable to that of the Lagoon. After his relocation, there were no apparent differences in the acoustics of the tank visible on the spectrograms; that is, no apparent echoes or increased volume were observed that might have discouraged Moonshine's phonations due to discomfort. More data would be necessary from both environments to truly establish any significance of this difference. However, the distribution of Moonshine's ( admittedly very limited) phonation behavior while in the Med Tank supports the theory that human presence (or lack thereof) affects his overall vocal behavior. During the NTS session in the Med Tank, Moonshine only emitted 16 total phonations, mostly durin g the terminal (post session) period (56.25%) of the session. In contrast, Moonshine emitted 61 total phonations during the TS session, and nearly all of these phonations (85.25%) occurred during the training session itself, while his trainer was in the ta nk with him. In addition, although compound vocalizations were relatively common in the Lagoon sessions (N = 216), Moonshine did not emit any compound vocalizations while in the Med Tank. Effects of enculturation Differences in enculturation resulting from the individual histories of each subject may be another explanation of the difference of effect found between the subjects. Call and Tomasello (1996) suggest that enculturation may have a significant impact over the extent of interaction and communic ation between humans and dolphins. Although enculturation has not been quantified, it is possible to infer relative differences.

PAGE 86

CAPTIVE DOLP HIN VOCAL BEHAVIOR 78 Moonshine was stranded at a very young age, when he would still have been relatively dependent on his mother (Gibson & Mann, 2007; Herzing, 1997). Human trainers raised him, and even while he lived with another dolphin (a young female of a sympatric species, Stenella longirostris ), human interactions were a near constant part of his daily routine. These interactions only increas ed during his time in isolation, in an effort to keep him socialized and stimulated. Khyber, Ranier, Calvin, and Malabar may represent comparatively less enculturated dolphins. Although all but Ranier were born and raised in captivity, they have not been as dependent as Moonshine on humans for general care taking and social interactions integral to the health and development of dolphins. In fact, the differences between Khyber and Ranier compared to Calvin and Malabar across conditions in this study may r elate to their own relative levels of enculturation and even perhaps their stage of social development. Ranier and Khyber are the eldest pair in the study (31 and 20 years old, respectively), and have been living in captivity the longest. Ranier, althoug h wild caught, has been living in aquariums for over 20 years, and during that time has regularly interacted with human trainers not only through general husbandry and training interactions, but also through his participation in many cognitive and behavior al studies. Likewise, Khyber has lived in a captive environment for all of his 20 years, and has participated in many studies in addition to his daily interactions with humans. This suggests that of the four bottlenose subjects of this study, Ranier and Kh yber were the most enculturated, and this may explain why a stronger effect of training sessions on their phonation behavior was observed than was observed in Calvin and Malabar.

PAGE 87

CAPTIVE DOLP HIN VOCAL BEHAVIOR 79 Calvin and Malabar had relatively parallel histories to Khyber and Ranier: both were captive born and also frequent participants in cognitive and behavioral studies. However, they may have shown little change in their vocal behavior in part due to their "less enculturated" status as younger individuals, though this effect may be less true for Calvin (who is 18, only 2 years younger than Khyber) than for Malabar (the youngest subject of the study at 11 years old). However, Calvin has unique differences in his personal history that may set him apart, in terms of sociability (in a do lphin to dolphin sense) and enculturation. Malabar's youth may also have an impact on the pair's interactions, both with each other and regarding their relationship to their human trainers. In the wild, Malabar would be at a stage of his life in which the formation of social relationships, including possibly a pair bond with another male (Connor et al., 1999; Watwood et al, 2004, Wells, 1991) would be integral to his survival (Wells, 1991). It is possible that his level of social maturity interferes with hi s (and by association, possibly Calvin's) attention to human interactions. In direct contrast to the human interaction inducing environment that Moonshine lives in, the constant dolphin dolphin social interaction that pervades the bottlenose dolphins' li ving environment may have an over arching effect on their attention to human trainers. Distribution of Phonations Within Conditions To examine the more immediate effects of the occurrence of training sessions on the subjects' vocal behavior, we compared t he overall rate of phonation distributed across both conditions for each subject. If the occurrence of a training session during the recording period had no effect on the vocal behavior of the subjects, the distribution of

PAGE 88

CAPTIVE DOLP HIN VOCAL BEHAVIOR 80 phonations is expected to be para llel in both conditions, and percentages should be roughly the same. As it was in the first analyses, this comparison highlighted individual differences in the behaviors of each subject (or subject pair): The distribution of Khyber's and Ranier's phonati ons over time did not show high variability (M = 4.4, SD = 1.045) between sessions, suggesting that their overall patterns of phonation emissions were not affected by the occurrence of a training session. Calvin and Malabar had significantly different di stributions (M = 62.21, SD = 53.01). The majority of TS session phonations occurred during the training session phase (93.52%), compared to the complete lack of phonation behavior during that time in the NTS session. In contrast, the majority of NTS sessio n phonations occurred during the terminal (post session) phase (56.25%) and numbered substantially higher than the TS session (11.47%). Without more robust data to support this finding, however, it is more likely that the stark contrast of these distributi ons is a result of outside variables and chance. The low number of phonations outside of the two most prominent phases of both conditions limits the ability to investigate the contextual differences of the two sessions. Moonshine demonstrated a significan tly different distribution (M = 30.75, SD = 15.2) of phonation behavior between the Lagoon TS and NTS sessions: In the Lagoon TS session, Moonshine emitted the majority (73.42%) of his phonations during the terminal (post session) phase of the session. In contrast, only 27% of his overall phonations in the NTS session occurred during the same time span. This may be explained, in part, by Moonshine's common pattern of jaw clapping, particularly following training sessions. This frequent behavioral pattern has been observed by both the researcher and Moonshine's trainers (A. Cardwell, personal

PAGE 89

CAPTIVE DOLP HIN VOCAL BEHAVIOR 81 communication, November 2011). This observation is supported by the data presented here: jaw claps occurred in the TS session only during the terminal phase and made up 35% of the overall phonations that occurred during that time. The jaw clap behavior likely caused a "cushion" effect on the distribution of Moonshine's phonations. However, even if jaw claps are removed from the analysis, the difference in the distribut ion is still substantial (M = 23.28, SD = 10.73). There was a difference in the distribution of compound vocalizations both across and between sessions, as well. Overall, Moonshine emitted more compound vocalizations (69.91% of total) during the TS sessi on than the NTS session. Within the TS session, Moonshine emitted most compound vocalizations (82.12%) during the terminal (post session) phase of the session, significantly more than during that phase of the NTS session (M = 50.29, SD = 45.01). It is po ssible that Moonshine's increased phonation is a result of the departure of his trainer. Janik & Slater's (1998) anecdotal evidence of signature whistle use by captive dolphins towards their trainers suggests that dolphins might utilize a contact call with humans in a similar manner as they would with conspecifics. It would make sense, if this were true, for the departure of his human trainer to cause Moonshine to elevate his phonation rate and perhaps utilize a higher number of whistles, particularly signa ture whistles. In order to investigate this, it is necessary to examine the types of phonations produced during these periods of high vocal behavior. Contextual Differences in Moonshine's Phonation Behavior Burst pulses were only found during the TS sess ion (4.47% of all phonations), and only during the training session (38%) and post session phases (17%). The disparity

PAGE 90

CAPTIVE DOLP HIN VOCAL BEHAVIOR 82 between the two conditions suggests a social mechanism behind burst pulse behavior, as suggested by Herzing (1996, 2000). Although burst pulses have been associated with aggression (Herzing, 1996; Herzing 2000; Lammers, et al., 2003; Overstrom, 1983), in Moonshine's case, the behavior did not seem particularly aggressive. Although we did not code for behavioral associations, no obvious aggr essive interactions were observed either during the recording session or in the reviewed video footage. The overall prevalence of echolocation behavior in the TS sessions (and to a lesser extent, the NTS sessions) is likely due to the role of echolocati on in dolphins' spatial orientation and perception of their environments (Au & Herzing, 2003; Benoit Bird & Au, 2009; Herzing, 1996; Herzing, 2000; Lammers et al., 2003; Xitco & Roitblat, 1996). Because training sessions involved more active movement and, in some cases, searching for objects and other orientation and identification tasks, the significantly higher number of echolocation clicks during the TS session (over twice as much as during the NTS session, M = 238.5, SD = 171.82) is understandable. Howe ver, it is therefore surprising that the majority of these phonations (44%) were emitted during the terminal phase of the session, when no training session was occurring. It is possible that the umbrella effect of Moonshine's increased overall phonations d uring that phase resulted in this discrepancy. Single clicks, echolocation click trains of 6 pulses or less, were relatively rare (3.63% of all phonation behaviors), and did not vary significantly between conditions (M = 1.88, SD = .71) nor across phases of sessions (M = 2.33, SD = 2.31). It is possible that these phonations were under represented in the data due to noise; ambient noise often masks or has a similar appearance to isolated echolocation clicks.

PAGE 91

CAPTIVE DOLP HIN VOCAL BEHAVIOR 83 Moonshine's whistling behavior, both between sessions and across phases of the TS session, supports the findings of other researchers regarding the production of signature whistles during separations (e.g., Janik & Slater, 1998; Watwood et al., 2003). In Moonshine's case, his solitude can be seen as a more extreme and permanent form of isolation, and the coming and going of his trainer approximates, after a fashion, the temporary separations that have been proven to elicit high rates of signature whistle production. Whistles were the most common pho nation type produced by Moonshine overall (42.85%), though there was a significant difference between the TS and NTS sessions (M = 46.85, SD = 34.63), such that he produced significantly more whistles during the NTS session than during the TS session. In b oth the TS and NTS sessions, the "flat" and "Moonshine" contours were most prevalent. The "flat" contour is a whistle which begins and ends at the same frequency and does not significantly deviate from that frequency throughout its duration. The frequency "contour" of this whistle is reminiscent of that of the bridge whistle used by Moonshine's trainer, and its production is likely a product of imitation. Dolphins are well known for their imitative abilities (e.g., Miksis et al., 2002; Richards et al., 1984 ), and commonly incorporate sounds prevalent in their environment into their own repertoires (Fripp et al., 2005; Janik, 2000; McCowan & Reiss, 1995b; Watwood et al., 2004). Miksis et al. (2002) even found that captive dolphins' signature whistles reflect incorporation of their trainers' bridge whistles. The relative prevalence of "flat" contours (12.31%) in Moonshine's whistle productions may also be a product of the positive association of the flat bridge whistle contour, used as a secondary reinforcer on a regular basis during trainer interactions.

PAGE 92

CAPTIVE DOLP HIN VOCAL BEHAVIOR 84 The "Moonshine" contour represented 74.37% of all whistles emitted in both sessions (see Figure 22 for spectrogram examples of the contour) The contour was so named due to its overall prevalence in Moonshi ne's repertoire and unique parameters (when compared to the already existing whistle dictionary synthesized from the bottlenose subjects' repertoires), suggesting that it is Moonshine's signature whistle (Caldwell & Caldwell, 1968). The substantially highe r (M = 65.15, SD = 33.2) production of this whistle in the NTS session (88.62%) when compared to the TS session (41.67%) suggests that when no training session occurs, Moonshine behaves similarly to both other captive dolphins and wild dolphins when separ ated from their social group. In contrast to the relatively even distribution of signature whistles throughout the NTS session (M = 122, SD = 6.25), there was more significant variance in the number of signature whistles produced across the different pha ses of the TS condition (M = 25, SD = 17.69). Signature whistles occurred most frequently in the pre session phase, and least frequently during the training session phase. It is possible that this is indicative of predictive behavior; since the training se ssions occurred at the same time every day, Moonshine may have been expecting the trainer's arrival and emitting signature whistles as cohesion calls, similar to the increased whistling emitted by mothers when calves reunited with them (Mello & Amundin, 2 005). A B Figure 22. Moonshine's signature whistle contour.

PAGE 93

CAPTIVE DOLP HIN VOCAL BEHAVIOR 85 Individuality in Inter species Communication Research on more complex levels of inter species communication (i.e., symbolic communication) emphasizes that individual qua lities of its subjects make them more or less successful (e.g., Gardner & Gardner, 1969; Savage Rumbaugh, 1998; Tomasello & Call, 1994). Moonshine, Khyber and Ranier, and Calvin and Malabar demonstrate differences between themselves that may predict their individual success in future communication research, though the results obtained in this study must be supported with more extensive data before this theory can be tested. Study Limitations and Suggestions for Future Research In order to support the impl ications of these tentative results more solidly, several more sessions must be run with each subject. However, the results suggest that there is more to be gained from further investigation of this effect. The variability between cases may be indicative o f levels of individual candidacy for further inter species communication research: if Moonshine does, in fact, consistently demonstrate the differences found in this study, this may be evidence that humans do have an impact on his vocal behavior, providing solid support for the potential for bidirectional communication. With more data for comparison, it would be possible to investigate the more specific differences further in Khyber and Ranier's behaviors, if there are any to be found. Likewise, it would es tablish a stronger baseline of phonation behavior for each subject, including Calvin and Malabar, which would allow stronger comparisons to be made between the results of the subjects as a whole. It is also possible that the data presented here could be inaccurately represented due to errors made during the spectrogram analysis, resulting in potential inclusion of non

PAGE 94

CAPTIVE DOLP HIN VOCAL BEHAVIOR 86 dolphin sounds, mis labeling of phonations, or omission of phonations altogether. Additional coders must review the data to provide inter c oder reliability and reduce the chance for these errors. This step was planned for this project but did not occur due to time constraints. In the current climate of technological advancement, it may also be possible to develop a computer program that cou ld extract and label dolphin vocalizations with the same level of accuracy, or possibly greater, as humans. It appears that dolphins discriminate between whistles by the frequency contour (Harley, 2007), and so far, no computer program has been able to mat ch the human ability to sort and categorize these contour differences (for a review, see Janik, 1998). However, these approaches have all involved quantitative measures that are not viable systems of categorization. It may be possible to develop a qualitat ive A.I. (Artificial Intelligence) program that can be "trained" by humans to discriminate between whistle contours, and even other phonations, and learn with experience to sort phonations and contours accurately. The method of this study had several othe r limitations that may have resulted in inaccurate reporting of the total phonation behaviors of the subjects: Many above water phonations could be heard at the time of the recordings, and through the video playback, that were not visible on the analyzed spectrograms. Because acoustic recording was only done underwater, any in air phonations that did not appear on the spectrograms were omitted from the analyses entirely. This may have caused the omission of integral data regarding human dolphin acoustic in teractions, since, of the potential receivers of this signal, humans are the most likely beneficiaries of the behavior.

PAGE 95

CAPTIVE DOLP HIN VOCAL BEHAVIOR 87 In future studies of dolphin human acoustic interactions, both underwater and in air recordings should be obtained. Because Khyber, Ra nier, Calvin, and Malabar live in units of two that are rarely separated, it was necessary to conduct sessions with the subjects as pairs instead of individuals. Isolating any subjects during the study would likely have caused an effect on their vocal beha viors (Caldwell & Caldwell, 1968) that would skew any data that might have been collected. This made it not only impossible to identify any individual specific differences, but also presented challenges in excluding vocalizations from the non participating pair. Although the two pairs were separated by water tight gates during recordings, the hydrophones in the tanks were able to pick up sounds from both sides of the gates. A deductive system was used to identify the vocalizing dolphin (whether participatin g or not) of every phonation, but it is possible that mistakes were made (i.e., a non participating dolphin echolocated directly towards a hydrophone on the other side of the water tight gate, making the phonation appear strongest in a channel that usually picks up participating dolphin's phonations). To combat this, video was simultaneously recorded of both tanks used in the study, as well as both above and under water in the tank where the non participating dolphins were held. Unfortunately, it is extrem ely difficult to synchronize the exact timing of both the acoustic and video recordings, making it difficult to state with certainty whether a specific dolphin was definitively vocalizing at any given time. It would require either complete acoustic isolati on of the subjects or better time synchronization to conclusively differentiate between vocalizers. The subjects' vocal behaviors may also have been influenced by outside variables that could not be controlled:

PAGE 96

CAPTIVE DOLP HIN VOCAL BEHAVIOR 88 Although trainers did not interact with t he dolphins during NTS sessions, other people were in the area of the recording sessions: at Moonshine's location, guests were able to pass through and observe the sessions throughout the process; at both locations, other employees and guests occasionally passed through the study area, and the researchers conducting the recordings remained at the recording station throughout all sessions. Dolphins' awareness and response to human attentional states (Xitco et al., 2001, 2004), coupled with the stark contrast of Moonshine's behavior between two environments, suggests that there is an effect of human presence, whether direct interaction occurs or not. Therefore, future studies should ideally control for this effect by removing human presence entirely outside of the human interaction condition. During no training session recordings, another possible outside variable may have influenced the phonation behaviors of the subjects. The recording sessions occurred at times that the dolphins were accustomed to particip ating in training sessions, and the training session still occurred at some point during the broader time frame before or after the recording session. In Moonshine's case, it may have occurred before or after the recording session time. If it occurred befo re, it was ended at least 45 minutes prior to the recording session. The bottlenose subjects participated in their training session immediately following the recording session. It is therefore possible that the post session phonation behaviors of the dolph ins were affected by anticipation of the upcoming session. Although the data presented here are not enough to make any conclusive statements about the effect of training sessions and trainer interaction on the phonations of captive dolphins, they do prov ide the groundwork for possible new methods of quantitative

PAGE 97

CAPTIVE DOLP HIN VOCAL BEHAVIOR 89 analysis of dolphin human interaction via vocal behavior. With some methodological refinements and more robust data, future studies may fill the integral gap in inter species communication literat ure regarding bidirectional communication by demonstrating communicative behaviors of dolphins directed to humans.

PAGE 98

CAPTIVE DOLP HIN VOCAL BEHAVIOR 90 Appendix A Whistle Dictionary Blip Any whistle so short or messy that its contour cannot be determined clearly. Boat A roughly symmet rical whistle that goes down first and then up, whether it be curved or straight. The inverse of bump. Bump A roughly symmetrical whistle that goes up first and then down, whether it be curved or straight. The inverse of boat. Calvin Calvin's signat ure whistle. An upsweep that starts around 5 kHz, rises to around 10 kHz slowly, then ends with a sharp uptick (resembling a hockey stick). Complex A whistle that does not fit the previous categories. Frequent forms of complex involve combinations of u p/flat, down/flat, bump/flat. Crinkle A whistle that is relatively flat in overall contour, deviating no more than 5kHz, but frequently modulates up and down within that span. Down Any whistle that primarily goes down, with no major flat/upward po rtions. Flat Any whistle that is primarily flat, with only slight deviations. Gur A whistle that is roughly flat, dips into a bump, and then is roughly flat again. Khyber Khyber's signature whistle. A sharp bump, followed by a lower flat, a sec ond taller bump. Often strung together with a large dip between iterations. Resembles a castle. Malabar A Whistle of 2+ looped bumps with the right side of each bump culminating in a higher peak Moonshine Moonshine's signature whistle. A whistle of one or more looped tildes, usually begins around 10 kHz, quickly rises to around 15 kHz, then curves back down to around 8 kHz before resolving at around the original 10 kHz. Ranier A series of peaks, usually sharp but sometimes more round, with upsweep s connecting them. Typically ends in a tail similar to Calvin's hockey stick. Reverse Tilde Looks like Tilde, but beginning with a curve downward, followed by a curve upward, and the beginning of a curve downward. Rug A whistle that is roughly flat, rises into a boat, and then is roughly flat again. Tilde Looks like ~; a curve upward followed by a curve downward and the beginning of a curve upward. Up Any whistle that primarily goes up, with no major flat/downward portions.

PAGE 99

CAPTIVE DOLP HIN VOCAL BEHAVIOR 91 B lip Boat Bump Bumpbump Calvin Complex Crinkle Down Flat Gur Khyber Moonshine Reverse Tilde Ranier Up Rug Tilde

PAGE 100

CAPTIVE DOLP HIN VOCAL BEHAVIOR 92 References Anderson, J. R., Montant, M., & Schmitt, D. (1996). Rhesus monkeys fail to use gaze direction a s an experimenter given cue in an object choice task. Behavioural Processes, 37 (1), 47 55. doi:10.1016/0376 6357(95)00074 7 Anderson, J. R., Sallaberry, P., & Barbier, H. (1995). Use of experimenter given cues during object choice tasks by capuchin monkey s. Animal Behaviour, 49 (1), 201 208. doi:10.1016/0003 3472(95)80168 5 Au, W. W. L. (1993). Characteristics of dolphin sonar signals. In Au, W.W.L., The sonar of dolphins (pp. 115 139) New York: Springer Verlag. Au, W. W. L., & Herzing, D. L. (2003). Ech olocation signals o f wild Atlantic spotted dolphin (S tenella frontalis). The Journal of the Acoustical Society of America, 113 (1), 598 604. doi:10.1121/1.1518980 Bazœa Dur‡n, C. (2004). Differences in the whistle characteristics and repertoire of bottlenos e and spinner dolphins. Anais Da Academia Brasileira De Cincias, 76 (2), 386 392. Benoit Bird, K. J., & Au, W. W. L. (2009). Phonation behavior of cooperatively foraging spinner dolphins. The Journal of the Acoustical Society of America, 125 (1), 539 546. d oi:10.1121/1.2967477 Busnel, R. G. (1973). Symbiotic relationship between man and dolphins. New York Academy of Sciences Transactions Series, 35 (2), 112 131. Caldwell, M. C., & Caldwell, D. K. (1965 ). Individualized whistle contours in bottle nosed dol phin s (T ursiops truncatus). Nature, 207 434 435. Caldwell, M. C., & Caldwell, D. K. (1968). Vocalization of naive dolphins in small

PAGE 101

CAPTIVE DOLP HIN VOCAL BEHAVIOR 93 groups. Science, 1 59 (3819), pp. 1121 1123. doi: 10.1126/science.159.3819.1121 Call, J., & Tomasello, M. (1994). Production and comprehension of referential pointi ng by orangutans (P ongo pygmaeus). Journal of Comparative Psychology, 108 (4), 307 317. doi:10.1037/0735 7036.108.4.307 Chomsky, N. (1957). Syntactic structures (2 nd ed.) New York: Mouton. Chomsky, N. (1972). Language and mind. New York: Harcourt Brace Jovanovich. Connor, R., Wells, R. S., Mann, J., & Read, A. J. (2000). The bottlenose dolphin: Social relationships in a fission fusion society. In J. Mann (Ed.), Cetacean societies: Field studies of dolphins and whales (pp. 91 126) Chicago: University of Chicago Press. Connor, R. C., Heithaus, M. R., & Barre, L. M. (1999). Superalliance of bottlenose dolphins. Nature, 397 (6720), 571 572. doi:10.1038/17501 Connor, R., Mann, J., & Watson Capps, J. (2006). A sex specific a ffi liative contact behavior in I nd ian ocean bottlenose dolphins, T ursiops sp. Ethology, 112 (7), 631 638. doi:10.1111/j.1439 0310.2006.01203.x Corkeron, P.J. (1990) Aspects of the behavioural ecology of inshore dolphins Tursiops tr uncatus and Sousa chinensis in Moreton Bay, Australi a. In S. Leatherwood & R.R. Reeves (Eds.), The Bottlenose Dolphin, pp. 285 293. London: Academic Press. Cravalho, M. A. (1999). Shameless creature s: An ethnozoology of the A mazon river dolphin. Ethnology, 38 (1), pp. 47 58. DeCasper A., & Fifer, W. (1980). Of human bonding: Newborns prefer their mothers' voices. Science, 208 (4448), 1174 1176. doi:10.1126/science.7375928

PAGE 102

CAPTIVE DOLP HIN VOCAL BEHAVIOR 94 Evans, W. E., Powell, B. A., & Naval Undersea Warfare Center (U.S.) (1967). Discrimination of different metallic p lates by an echolocating delphinid China Lake : Naval Undersea Warfare Center. Fraser, J., Reiss, D., Boyle, P., Lemcke, K., Sickler, J., Elliott, E., Gruber, S. (2006). Dolphins in popular literature and media. Society & Animals: Journal of Human An imal Studies, 14 (4), 321 349. doi:10.1163/156853006778882402 Fripp, D., Owen, C., Quintana Rizzo, E., Shapiro, A., Buckstaff, K., Jankowski, K., Tyack, P. (2005). Bottlenose dolphin (T ursiops truncatus) calves appear to model their signature whistles on the signature whistles of community members. Animal Cognition, 8 (1), 17 26. doi:10.1007/s10071 004 0225 z Garcia, C., & Dawson, S. (2003). Distribution of pantropical spotted dolphin s in pacific coastal waters of P anama. Latin American Journal of Aquat ic Mammals, 2 (1), 29 38. Gardner, R. A. & Gardner, B.T. (1969). Teaching sign language to a chimpanzee. Science, 165 (3894), 664 672 Gibson, Q. A., & Mann, J. (2008). The size, composition and functi on of wild bottlenose dolphin (T ursi ops sp.) mother calf groups in Shark Bay, A ustralia. Animal Behaviour, 76 (2), 389 405. doi:10.1016/j.anbehav.2008.01.022 Gogoleva, S. S., Volodin, I. A., Volodina, E. V., Kharlamova, A. V., & Trut, L. N. (2010). Vocalization toward conspecifics in silver foxes (V ulpes vul pes) selected for tame or aggressive behavior toward humans. Behavioural Processes, 84 (2), 547 554. doi:10.1016/j.beproc.2010.01.021

PAGE 103

CAPTIVE DOLP HIN VOCAL BEHAVIOR 95 Gogoleva, S. S., Volodin, I. A., Volodina, E. V., Kharlamova, A. V., & Trut, L. N. (2009). Kind granddaughters of angry gra ndmothers: The effect of domestication on vocalization in cross bred silver foxes. Behavioural Processes, 81 (3), 369 375. doi:10.1016/j.beproc.2009.03.012 Gogoleva, S. S., Volodin, I. A., Volodina, E. V., Kharlamova, A. V., & Trut, L. N. (2011). Explosive vocal activity for attracting human attention is related to domestication in silver fox. Behavioural Processes, 86 (2), 216 221. doi:10.1016/j.beproc.2010.12.001 Hall, S. E., & Vanderhoop, J. (2004). The legend of K atama: The creation story of dolphins, a W ampanoag legend of Martha's V ineyard Pennsylvania: Island Moon Press. Hampson, G.A. (2005). Human: Machine, ape or dolphin? Journal of Futures Studies, 9 (4), 29 44. Hayes, K. J., & Hayes, C. (1952). Imitation in a home raised chimpanzee. Journal of compar ative and physiological psychology 45 (5), 450 459. Herman, L. M., Abichandani, S. L., Elhajj, A. N., Herman, E. Y. K., Sanchez, J. L., & Pack, A. A. (1999). Dolphins (T ursiops truncatus) comprehend the referential character of the human pointing gesture. Journal of Co mparative Psychology, 113 (4), 347 364. doi:10.1037/0735 7036.113.4.347 Herman, L. M., Peacock, M. F., Yunker, M. P., & Madsen, C. J. (1975). Bottlenosed dolphin: Double slit pupil yields equivalent aerial and underwater diurnal acuity, Science 189, 650 652. Herman, L. M. and Tavolga, W. N. (1980). The communication systems of cetaceans. In

PAGE 104

CAPTIVE DOLP HIN VOCAL BEHAVIOR 96 L. M. Herman (Ed.), Cetacean behavior: Mechanisms and functions 149 209. New York: Wiley Interscience. Herzing, D. L. (1996). Vocalizations and associat ed underwater behav ior of free ranging A tlantic spotted dolphins, S tenella frontalis and bottlenose dolphins, T ursiops truncatus Aquatic Mammals, 22 (2), 61 79. Herzing D L (1997) The life history of free ranging Atlantic spotted dolphins ( Stenella fron talis ): Age classes, color phases, and female reproduction. Mar ine Mamm al Science, 13 (4), 576 595. Herzing, D. L. (2000). Acoustics and social behaviors of wild dolphins: Implicat ions for a sound society. In W. L. Au A. N. Popper & R. R. Fay (Eds.), He aring by whales and dolphins (pp. 225 227). New York: Springer. Herzing, D. L., & Johnson, C. M. (1997). Interspecific interactions between Atlantic spotted dolphins (Stenella frontalis) and bottlenose dolphins (Tursiops truncatus) in the Bahamas, 1985 1 995. Aquatic Mammals 23 (2), 85 99. Higham, T. F. (1960) Nature note: Dolphin riders. A ncient stories vindicated. Greece & Rome, 7 (1), pp. 82 86. Hopkins, W. D., & Leavens, D. A. (1998). Hand use and gestural communication in chimpanzees (P an troglodytes ). Journal of Comparative Psychology, 112 (1), 95 99. doi:10.1037/0735 7036.112.1.95 Itakura, S., & Tanaka, M. (1998). Use of experimenter given cues during objec t choice tasks by chimpanzees (Pan troglodytes), an orangutan (P ongo pygmaeus), and human infan ts (H omo sapiens). Journal of Comparative Psychology, 112 (2), 119 126. doi:10.1037/0735 7036.112.2.119

PAGE 105

CAPTIVE DOLP HIN VOCAL BEHAVIOR 97 Janik, V. M. (2000). Whistle matchin g in wild bottlenose dolphins (T ursiops truncatus). Science, 289 (5483), 1355 1357. doi:10.1126/science.289.5483.1355 Janik, V. M., & Slater, P. J. B. (1998). Context specific use suggests that bottlenose dolphin signature whistles are cohesion calls. Animal Behaviour, 56 (4), 829 838. doi:10.1006/anbe.1998.0881 Jefferson, T. A., Leatherwood, S., & Webber, M. A. (1993). Ma rine mammals of the world United Nations Environment Programme. Johnson, C. S. (1968). Masked tonal thresholds in the bottlenosed porpoise. Journal of the Acoustical Society of America, 44 (4), 965 967. doi:10.1121/1.1911236 Kaminski, J., Riedel, J., Call, J. & Tomasello, M. (2005) Domestic goats (Capra hircus) follow gaze direction and use social cues in an object choice task. Animal Behaviour, 69 (1), 11 18. Kita, S. (2003). Pointing: A foundational building block in human communication. In S. Kita (Ed.), Pointing: Where language, culture, and cognition meet (pp. 1 8). Mahwah, NJ: Erlbaum. Lammers, M. O., Au, W. W. L., & Herzing, D. L. (2003). The broadband social acoustic signaling behavior of spinner and spotted dolphins. The Journal of the Acoustical Soc iety of America, 114 (3), 1629 1639. doi:10.1121/1.1596173 Leavens, D. A., & Hopkins, W. D. (1999). The whole hand point: The structure and function of pointing from a comparative perspective. Journal of Comparative Psychology, 113 (4), 417 425. doi:10.1037/ 0735 7036.113.4.417 Lieberman, P. (1968). Intonation, perception, and language. Cambridge, MA: M.I.T Press.

PAGE 106

CAPTIVE DOLP HIN VOCAL BEHAVIOR 98 Lilly, J. C. (1961). Man and dolphin Oxford England: Doubleday. Lilly, J. C., & Devereux Foundation. (1958). Some considerations regarding basic me chanisms of positive and negative types of motivations. Santa Barbara, Calif.; Devon, Pa.: Devereux Foundation. Lilly, J. C., & Miller, A. M. (1961). Sounds emitted by the bottlenose dolphin. Science, 133 (3465), pp. 1689 1693. MacKenzie, D. A. (2005). Myth s of China and J apan Whitefish, MO: Kessinger Publishing. Mann, J., & Smuts, B. B. (1998). Natal attraction: Allomaternal care and mother infant separations in wild bottlenose dolphins. Animal Behaviour 55 1097 1113. Mann, J., & Smuts, B. (1999). Behav ioral development in wil d bottlenose dolphin newborns (T ursiops sp.) Behaviour, 136 (5), pp. 529 566. Maros, K., G‡csi, M., & Mikl—si, (2008). Comprehension of human pointing gestures in horses ( E quus caballus ). Animal Cognition,11 (3), 457 466. doi:10. 1007/s10071 008 0136 5 May Collado, L., Agnarsson, I., & Wartzok, D. (2007). Phylogenetic review of tonal sound production in whales in relation to sociality. BMC Evolutionary Biology, 7 (1), 136. doi:10.1186/1471 2148 7 136 McBride, A. F., & Hebb, D. O. (1 948). Behavior of th e captive bottle nose dolphin, T ursiops truncatus. Journal of Comparative and Physiological Psychology, 41 (2), 111 123. doi:10.1037/h0057927

PAGE 107

CAPTIVE DOLP HIN VOCAL BEHAVIOR 99 McCowan, B. (1995). A new quantitative technique for categorizing whistles using simulated si gnals and whistles from captive bottlenose dolphins (Delphindae Tursiops truncatus ). Ethology, 100 177 193. McCowan, B., & Reiss, D. (1995). Quantitative comparison of whistle repertoires from capt ive adult bottlenose dolphins (Delphinidae, T ursiops trun catus): A re evaluation of the signature whistle hypothesis. Ethology, 100 (3), 194 209. McCowan, B., & Reiss, D. (1995). Whistle contour development in captive bo rn infant bottlenose dolphins (T ursiops truncatus): Role of learning. Journal of Comparative P sychology, 109 (3), 242 260. doi:10.1037/0735 7036.109.3.242 McCowan, B., & Reiss, D. (2001). The fallacy of 'signature whistles' in bottlenose dolphins: A comparative perspective of 'signature information' in animal vocalizations. Animal Behaviour, 62 (6), 1151 1162. doi:10.1006/anbe.2001.1846 Mech, L. D. (1970). The wolf : The ecology and behavior of an endangered species Minneapolis: University of Minnesota Press. Melillo, K. E., Dudzinski, K. Ms., Cornick, L. A. (2009). Interactions between Atlantic spot ted (Stenella frontalis) and bottlenose (Tursiops truncatus) dolphins off Bimini, the Bahamas, 2003 2007. Aquatic Mammals, 35 281 291. Mello, I., & Amundin, M. (2005). Whistle production pre and post partum in bottlenose d olphins (Tursiops truncatus) in human c are. Aquatic Mammals 31 (2), 169 175. Mikl—si, Pongr‡cz, P., Lakatos, G., Top‡l, J., & Cs‡nyi, V. (2005). A comparative study of the use of visual communicative signals in interactions between dogs (C anis fam iliaris) and humans and cats (F e lis catus) and humans. Journal of Comparative Psychology, 119 (2), 179 186

PAGE 108

CAPTIVE DOLP HIN VOCAL BEHAVIOR 100 Mikl—si, Soproni, K., (2006). A comparative analysis of animals' understanding of the human pointing gesture. Animal Cognition, 9 (2), 81 93 Miklosi A., Polgardi R., Topal J., & Csanyi V. ( 1998 ). Use of experimenter given cues in dogs. Animal Cognition, 1 113 121. Miksis, J. L., Tyack, P. L., & Buck, J R. (2002). Captive dolphins, T ursiops truncatus, develop signature whistles that match acoustic features of human made model sounds. Journal of the Acoustical Society of America, 112 (2), 728 739. doi:10.1121/1.1496079 Mšller, L. M., Beheregaray, L. B., Allen, S. J., & Harcourt, R. G. (2006). Association patterns and kinship in female I nd o Pacific bottlenose dolphins (T ur s iops aduncus) of southeastern A ustralia. Behavioral Ecology and Sociobiology, 61 (1), 109 117. doi:10.1007/s00265 006 0241 x Moln‡r, C., Pongr‡cz, P., & Mikl—si, (2010). Seeing with ears: Sightless humans' perception of dog bark provides a test for struc tural rules in vocal communication. The Quarterly Journal of Experimental Psychology, 63 (5), 1004 1013. doi:10.1080/17470210903168243 Murray, A. S., & Klapp, W. H. (1897). Manual of mythology: Greek and Roman, Norse and old German, Hindoo and E gyptian myth ology New York: Tudor Publishing Company. Neil, D. T. (2002). Cooperative fishing i nteractions between aboriginal A ustr alians and dolphins in eastern A ustralia. Anthrozošs, 15 (1), 3 18. doi:10.2752/089279302786992694

PAGE 109

CAPTIVE DOLP HIN VOCAL BEHAVIOR 101 Neiworth, J. J., Burman, M. A., Basil e, B. M., & Lickteig, M. T. (2002). Use of experimenter given cues in visual co orienting and in an object choice task by a new world monkey species, cotton top tamarins (Saguinus oedipus ). Journal of Comparative Psychology, 116 (1), 3 11. doi:10.1037/073 5 7036.116.1.3 Nicastro, N. (2004). Perceptual and acoustic evidence for species level differences in meow v ocalizations by domestic cats (Felis catus} and African wild cats (F elis silvestris lybica). Journal of Comparative Psychology, 118 (3), 287 296. doi :10.1037/0735 7036.118.3.287 Norris, K. S., & Dohl, T. P. (1980). Behavior of the Hawaiian spinner dolphin Stenella longirostris. Fishery Bulletin, 77 821 849. Overstrom, N. (1983). Association between burst pulse sounds and aggressive behavior in captive A tlantic b ottlenosed dolphins (T ursiops truncatus). Zoo Biology, 2 (2), 93 103. doi:10.1002/zoo.1430020203 Pack, A. A., & Herman L. M. (2007). The dolphin's (T ursiops truncatus) understanding of human gazing and pointing: Knowing what and where. Journal o f Comparative Psychology, 121 (1), 34 45. doi:10.1037/0735 7036.121.1.34 Patterson, F. G. (1978). The gestures of a gorilla: Language acquisition in another pongid. Brain and Language, 5 (1), 72 97. doi:10.1016/0093 934X(78)90008 1 Paulos, R. D. (2004) Non vocal communication in the Atlantic spotted dolphin (Stenella frontalis) and the Indo Pacific bottlenose dolphin (Tursiops aduncus). Retrieved from: www.usm.edu/psy kuczaj/Dissertations_Theses.html

PAGE 110

CAPTIVE DOLP HIN VOCAL BEHAVIOR 102 Perrin, W. F., & Hohn, A. A. (1994). Pantropical spotted dolphin S tenella attenuata In S. H. Ridgway, & F. R. S. Harrison Sir Richard (Eds.), Handbook of marine mammals volume 5: The first book of dolphins (pp. 71 98). San Diego, CA: Academic Press. Perrin, W. F. (1975). Spotted porpoise: Historical review. In W. F. Perrin (Ed.), Variation of spott ed and spinner porpoise (genus S tenella) in the eastern P acific and Hawaii ( pp. 13 21) Los Angeles, California : University of California Press. Pfungst, O., & Rosenthal, R. (1911). Clever Hans, the horse of Mr. von O sten. New York: Holt, Rinehart and Winston. Pliny the Elder, The Wernerian Club (trans. Hol land, P.) (1848). Pliny's natural history in thirty seven books London, Eng land: George Barclay. Pliny the Younger. (1915). XXI: To C aninus. In W. M. L. Hutchinson (Ed.), Letters (W. Melmoth Trans.). ( pp. 499 500 ) W. Heinemann. Retrieved from http://books.google.com/books?id=yXhfAAAAMAAJ Pryor, K., & Shallenberger, I. K. (1991). Social structure in spotted dolphins ( S tenella attenuata ) in the tuna purse seine f ishery in the easte rn tropical P acific. In K. Pryor, & K. S. Norris (Eds.), Dolphin societies: Discoveries and puzzles (pp. 161 196). Los Angeles, California: University of California Press. Pryor K., & Lindbergh, J. (1990). A dolphin human fishing cooperative in Brazil. Ma rine Mammal Science, 6 (1), 77 82. doi:10.1111/j.1748 7692.1990.tb00228.x QuŽrouil, S., Silva, M. A., Casc‹o, I., Magalh‹es, S., Seabra, M. I., Machete, M. A., & Santos, R. S. (2008). Why do dolphins form mix ed species associations in the A zores? Ethology, 114 (12), 1183 1194. doi:10.1111/j.1439 0310.2008.01570.x

PAGE 111

CAPTIVE DOLP HIN VOCAL BEHAVIOR 103 Quintana Rizzo, E., Mann, D. A., & Wells, R. S. (2006). Estimated communication range of social sound s used by bottlenose dolphins (T ursiops truncatus). Journal of the Acoustical Society of America, 120 (3), 1671 1683. doi:10.1121/1.2226559 Rendell, L., & Whitehead, H. (2001). Culture in whales and dolphins. Behavioral and Brain Sciences, 24 (2), 309 382. doi:10.1017/S0140525X0100396X Reynolds, J. E., Wells, R. S., & Eide, S. D. (2000). The bottlenose dolphin: Biology and conservation Florida: University Press of Florida. Rice, D. W. (1977). A list of the marine mammals of the world Seattle : U.S. Dept. of Commerce, N ational Oceanic and Atmospheric Administration, National Marine Fisheries Service. Richards, D. G. Wolz, J. P., & Herman, L. M. (1984). Vocal mimicry of computer generated sounds and vocal labeling of obj ects by a bottlenosed dolphin, T ursiops truncatus. Journal of Comparative Psychology, 98 (1), 10 28 Rumbaugh, D., Gill, T., Brown, J., von Glasersfeld, E., Pisani, P., Warner, H., & Bell, C. (1973). A computer controlled language training system for investigating the language skills of young apes Beha vior Research Methods & Instruments, 5 (5), 385 392 doi:10.3758/BF03200213 Savage Rumbaugh, S., McDonald, K., Sevcik, R. A., Hopkins, W. D., & Rubert, E. (1986). Spontaneous symbol acquisition and communicat ive use by pygmy chimpanzees (P an paniscus). Jo u rnal of Experimental Psychology: General 115 (3), 211 235.

PAGE 112

CAPTIVE DOLP HIN VOCAL BEHAVIOR 104 Savage Rumbaugh, E., Rumbaugh, D. M., & Boysen, S. (1978). Symbolic communic ation between two chimpanzees (P an troglodytes). Science, 201 (4356), 641 644. Savage Rumbaugh, S., Shanker, S. G., & Taylor, T. J. (1998). Apes, language, and the human mind New York, NY : Oxford University Press. Sax, B. (2001). The mythical zoo: An encyclopedia of animals in world myth, legend, and literature Santa Barbara, California: ABC CLIO. Sayigh, L. S., Esch, H. C., Wells, R. S., & Janik, V. M. (2007). Facts about signature wh istles of bottlenose dolphins, T ursiops truncatus. Animal Behaviour, 74 (6), 1631 1642 Sayigh, L. S., Tyack, P. L., Wells, R. S., & Scott, M. D. (1990). Signature whistles of fr ee rangi ng bottlenose dolphins T ursiops truncatus: Stability and mother offspring comparisons. Behavioral Ecology and Sociobiology, 26 (4), pp. 247 260. Schassburger, R. M. (1987). Wolf vocalization: An integrated model of structure, motivation and ontogeny. In H. Frank (Ed.) Man and wolf: Advances, issues, and problems in captive wolf research (pp. 313 347). Dordrecht, Netherlands: Dr W Junk Publishers. Schmidly, D. J. (1981). Marine Mammals of the southeastern United S tates coast and the gulf of M exico Office of Biological Services, Washington, D.C.: U.S. Fish and Wildlife Service. doi:FWS/OBS 80/41 Schneider, B. A, Trehub, S. E., Morongiello, B. A., & Thorpe, L. A. (1986). Auditory sensitivity in preschool children. Journal of the Acoustical Society of America 79 447 452.

PAGE 113

CAPTIVE DOLP HIN VOCAL BEHAVIOR 105 Scott, M. D., & Chivers, S. J. (1990). Distribution and herd structure of bottlenose dolphins in the eastern tropical Pacific Ocean. In S. Leatherwood, & R. R. Reeves (Eds.) The Bottlenose Dolphin (p p. 387 402) San Diego, California: Acade mic Press, Inc. Scott, M. D., Wells, R. S., & Irvine, A. B. (1990). A long term study of bottlenose dolphins on the west coast of F lorida. In S. Leatherwood, & R. R. Reeves (Eds.), The bottlenose dolphin (pp. 235 244). San Diego, California: Academic Pres s, Inc. Shane, S. H., Wells, R. S., & WŸrsig, B. (1986). Ecology, behavior and social organizat ion of the bottlenose dolphin: A review. Marine Mammal Science, 2 (1), 34 63. doi:10.1111/j.1748 7692.1986.tb00026.x Shapiro, A. D., Janik, V. M., & Slater, P. J. B. (2003). A gray seal's (H alichoerus grypus) responses to experimenter given pointing and directional cues. Journal of Comparative Psychology, 117 (4), 355 362 Slater, C. (1994). Dance of the dolphin: Transforma tion and disenchantment in the A mazonian imagination Chicago, Illinois: University of Chicago Press. Stevens, S. S. (1950). Introduction: A definition of communication. The Journal of the Acoustical Society of America, 22 (6), 689 690 Tamaki, N., Morisaka, T., & Taki, M. (2006). Does body con tact contribute to wards repairing relationships? T he association between flipper rubbing and aggressive behavior in captive bottlenose dolphins. Behavioural Processes, 73 (2), 209 215. doi:10.1016/j.beproc.2006.05.010

PAGE 114

CAPTIVE DOLP HIN VOCAL BEHAVIOR 106 Tavolga, C., & Essapian, F.S. (1957). The behavior of bottle nosed dolphins (Tursiops truncatus): M ating, pregnancy, parturition and mother infant behavior. Zoologica 42 (1), 11 31. Taylor, S. (2003). Souls in the sea: Dolphins, whales, and human destiny Berkeley, California: Frog Books. Ter race, H. S. (1982). Evidence for sign language in apes: What the ape signed or how well the ape was loved? PsycCRITIQUES, 27 (1), 67 68. doi:10.1037/020651 Terrace, H., Petitto, L., Sanders, R., & Bever, T. (1979). Can an ape create a sentence? Science, 206 (4421), 891 902. doi:10.1126/science.504995 Thompson, R. K. R., & Herman, L. M. (1975). Underwater frequency discrimination in the bottlenose dolphin (1 140 kHz) and the human (1 8 kHz). Journal of the Acoustic Society of America, 57, 943 948. Thurman, H. V., & Trujillo, A. P. (1999). Water and seawater. In H. V. Thurman, & A. P. Trujillo (Eds.), Essentials of oceanography (6th ed., pp. 132 167). Upper Saddle River, New Jersey: Prentice Hall. Tomasello, M. (2008). Origins of human communication Cambridge, Mass.: MIT Press. Tomasello, M., & ZuberbŸhler, K. (2002). Primate vocal and gestural communication. In M. Bekoff, C. Allen, & G. M. Burghardt (Eds.), The cognitive animal: Empirical and theoretical perspectives on animal cognition (pp. 293 299). Cambridge MA, US: MIT Press. Tomonaga, M., Uwano, Y., Ogura, S., & Saito, T. (2010). Bottlenose dolphins' (T ursiops truncatus) theory of mind as demonstrated by responses to their trainers' attentional states. International Journal of Comparative Psychology, 23 (3) 386 400.

PAGE 115

CAPTIVE DOLP HIN VOCAL BEHAVIOR 107 Tun, T. (2004). Irrawaddy dolphins in Hsithe M andalay segment of the A yeyawady river a nd cooperative fishing between Irrawaddy dolphin, O rcaella brevirost ris, and cast net fishermen in M yanmar. Sanchaung, Yangon, Myanmar: Wildlife Conservation Society Tyack, P. L. (1997). Development and social f unctions of signature whistles in bottlenose dolphins T ursiops truncatus. Bioacoustics, 8 (1 2), 21 46. Vir‡nyi, Z., G‡csi, M., Kubinyi, E., Top‡l, J., BelŽnyi, B., Ujfalussy, D., & Mikl—si, (2008). Comprehension of human pointing gestures in young human reared wolves (Canis lupus) and dogs (C anis familiaris). Animal Cognition, 11 (3), 373 387. doi:10.1007/s10071 007 0127 y Watwood, S. L., Owen, E. C. G., Tyack, P. L., & Wells, R. S. (2005). Signature whistle use by temporarily restrained and free swimming bottlenose dolphins, T ursiops truncatus. Animal Behaviour, 69 (6), 1373 1386. doi:10.1016/j.anbehav.2004.08.019 Watwood, S. L., Tyack, P. L., & Wells, R. S. (2004). Whistle sharing in pa ired male bottl enose dolphins, T ursiops truncatus. Behavioral Ecology and Sociobiology, 55 (6), 531 543. doi:10.1007/s00265 003 0724 y Wells, R. S. (1991). The role of long term study in understanding the social structure of a bottlenose dolphin community. In K. Pryor, & K. S. Norris (Eds.), Dolphin societies: Discoveries and puzzles (pp. 199 225). Los Angeles, California: University of California Press. Wells, R. S., & Scott, M. D. (1999). Bottlenose dolphin T ursiops truncatus (M ontagu, 1821). In S. H. Ridgway, & F. R. S. Harrison Sir Richard (Eds.), Handbook of

PAGE 116

CAPTIVE DOLP HIN VOCAL BEHAVIOR 108 marine mammals volume 6: The second book of dolphins and the porpoises (pp. 137 182). San Diego, CA: Academic Press. Wells, R. S., Scott, M. D., & Irvine, A. B. (1987). The social structure of free ranging bottleno se dolphins. In H. H. Genoways (Ed.), Current mammalogy (Vol. 1, pp. 247 305). New York: Plenum Press. Welsh, L. S., & Herzing, D. L. (2008). Preferential association among ki n exhibited in a population of Atlantic spotted dolphins (St enella frontalis). In ternational Journal of Comparative Psychology, 21 (1), 1 11. Wood, J. (1953). Underwater sound production and concurrent behavior of captive porpoises, Tursiops truncatus and S tenella plagiodon. Bulletin of Marine Science, 3 (2), 120 133 WŸrsig, B. (1986 ). Delphinid foraging strategies. In R. Schusterman, J. Thomas & F. Wood (Eds.), Dolphin cognition and behavior: A comparative approach (pp. 347 359). New Jersey: Lawrence Erlbaum Associates. WŸrsig, B., R.S. Wells, and K.S. N orris (1994). Food and feeding I n K.S. Norris, B. WŸr sig, R.S. Wells, and M. WŸrsig (Eds.), The Hawaiian Spinner Dolphin ( pp. 216 231) Los Angeles, California: University of California Press. Xitco, M. J., Jr., Gory, J. D., & Kuczaj II, S. A (2001). Spontaneous pointing by bottle no se dolphins (T ursiops truncatus). Animal Cognition, 4 (2), 115 123. doi:10.1007/s100710100107 Xitco, M. J., Jr., Gory, J. D., & Kuczaj II S. A. (2004). Dolphin pointing is linked to the attentional behavior of a receiver. Animal Cognition, 7 (4), 231 23 8. doi:10.1007/s10071 004 0217 z

PAGE 117

CAPTIVE DOLP HIN VOCAL BEHAVIOR 109 Xitco, M., & Roitblat, H. (1996). Object recognition through eavesdropping: Passive echolocation in bottlenose dolphins. Learning & Behavior, 24 (4), 355 365. Yeon, S. C., Kim, Y. K., Park, S. J., Lee, S. S., Lee, S. Y., Su h, E. H., Lee, H. J. (2011). Differences between vocalization evoked by social stimuli in feral cats and house cats. Behavioural Processes, 87 (2), 183 189. doi:10.1016/j.beproc.2011.03.003 Yin, S. (2002). A new perspective on b arking in dogs (Canis f amilaris ). Journal of Comparative Psychology, 116 (2), 189 193. doi:10.1037/0735 7036.116.2.189