The Social Group and Its Living Space
To the extent that the social organization of a species is a consequence of interaction between its members, the communication system is a basic component in social design. Studies of animal communication should eventually help to explain how the diversity of animal social systems is maintained. Conversely, the social system of a species provides the background against which one may hope to interpret the function that its communication signals serve. Information that field studies have provided on the behavior of the great apes is now sufficient for us to begin appraising the differences in organization that they exhibit and speculating about their evolvement.
The first problem is to define the basic social units, the spatial distribution of individuals, their patterns of contact and separation, and the ways in which they interact with one another—either competitively, thus limiting one another's access to resources, or cooperatively, thus aiding one another in reproduction, resource exploitation, and the avoidance of threats to life and health. Defining social groupings and distinguishing competitive from cooperative social units is less easy in practice than it might first appear. The difficulties are well illustrated by the great apes.
In one cluster of species, the gibbons and siamang, the social units are readily discernible. So far as is known, all species exhibit the same reproductive groupings. Monogamous pairs with long-lasting pair bonds live with their offspring in tropical rain forest, occupying territories from which adjacent families and mature offspring are excluded.
Of the other types of social organization represented in the great apes, the gorilla's pattern is most like the gibbons'. While there are some solitary adult males, most animals live in a social group.
These gorilla groups are perhaps five times larger than those of gibbons, averaging fifteen members in the nine troops studied by Fossey (1972), as compared with an average of 3.4 members in Kloss's gibbon (Tenaza, 1975), with one to three fully mature silverbacked males per group. The balance is made up of blackbacked males, adult females, and their offspring. The overlap between the extensive home ranges of gorilla groups is considerable, but each has exclusive use of a part, apparently as a consequence of an admixture of mutual avoidance and aggressive repulsion of neighboring groups (Schaller, 1963; Fossey, 1972, 1974).
The patterns of social organization in the chimpanzee and the orang-utan are harder to discern. Rather than forming durable groupings that move synchronously as coherent units, individuals and smaller groups tend to separate and coalesce in combinations that vary in composition from day to day. The periods of individual separation seem to be much longer for the orang-utan than for the chimpanzee. Although little is known about social groupings of the orang-utan, the most solitary of the ape species, studies by Mackinnon (1974,1975) suggest a pattern of social organization in at least one of its populations (in Sumatra) that is common in many other mammals (Eisenberg, 1966; Fisler, 1969) but is most unusual for higher primates. The majority of sightings are either of single animals or of adult females with one or two offspring. Single adult males occupy a large home range that encompasses the ranges of several adult females and their offspring. If the inhabitants of this shared living space are regarded as a group, it is not very different in size and composition from groups of the mountain gorilla. One well-counted group of seventeen animals included two adult males, two subadult males, six adult females, two adolescents, two juveniles, and three infants (Mackinnon, 1974). The home range of one group overlaps relatively little with ranges of adjacent groups. There is evidence of active exclusion and avoidance by aggression and vocal signaling. Although all members may be seen occasionally in close proximity, they never seem to move around as a coherent durable group, as the gorilla does. Instead the members are dispersed most of the time, having only transient contacts with one another, except for relationships of mothers and offspring and consortships between an adult male and an estrous female. In other orang-utan populations group members were more dispersed, and the definition of a grouping depends on a distinction between the brief, peaceful encounters occurring between members as compared with the aggressive encounters with members of other groupings. As Mackinnon (1974) summarizes a complex situation, "It would seem that links of familiarity and general tolerance exist between animals that range over the same area.”
Something similar may be true of the more social chimpanzee. This species was most commonly seen by van Lawick-Goodall (1968b, 1971) in groups of two to six animals, but more recent data suggest even smaller subgroupings. Wrangham (pers. comm.) found mean group sizes of 1.1 in one season and 2.0 in another, sampling those in which at least one member was an adult male. The subgroupings of chimpanzees in the Gombe Stream population are constantly changing in composition, apart from the durable subgroups that consist of mothers and their immature offspring. In striking contrast with the orang-utan and all of the other apes, peaceful subgroupings of chimpanzees often include more than one adult male at a time, and more than one adult female. If it is indeed possible to define a larger group on the basis of the predominantly peaceful and cooperative interactions of subgroupings when they meet (there is in fact a good deal of aggression when reunions occur), this grouping is probably larger than in the orang-utan and may amount to as many as fifty animals. Early evidence suggested that such a "community" at the Gombe National Park has north and south neighboring "communities" with which there is both peaceful intermingling, from which high-ranking adult males may perhaps exclude themselves (van Lawick-Goodall, 1971), and also a degree of mutual avoidance (Nishida and Kawanaka, 1972). More recent data have confirmed that adult male members of the community defend a territory against adjacent group males. This territory encompasses many overlapping adult female home ranges, and contacts between communities may involve mainly females with home ranges straddling the boundary (Wrangham, 1975).
Thus, the apes offer an unusually broad spectrum of organizational patterns with durable heterosexual pairs and dependent young, represented by the gibbons at one extreme, and at the other the chimpanzee, with a more labile social organization and several adult males living peaceably together. Having one-male groupings (i.e., one fully mature male) with several adult females and their offspring, the orang-utan and the gorilla fall somewhere between.
In all species the evidence reported suggests some degree of exclusive use of living space by resident groups. It is not always easy to discern whether adjacent communities simply avoid invading parts of one another's home ranges, perhaps aided by long-range vocal and nonvocal sounds, or whether there is active aggressive exclusion. At present there is evidence for territorial behavior in male and female gibbons, and for male group territoriality in the chimpanzee, with suggestions of territoriality in adult male orangutans and, perhaps, of a less well defined form of group territoriality in the gorilla.
Communication Systems and Social Organization
Basic to the complexities of the frequency and quality of different kinds of social interaction that may be mediated by communication signals is the elementary question of signaling distance. The dimensions of the space and the nature of the terrain over which signaling must occur to function normally are important in understanding the structure of the signal system that is used. The social systems of the great apes are diverse enough that we may be prepared for a variety of special requirements for distance signaling among species. In the gibbons for example the requirements of territoriality loom large in both sexes. Having signals that can be received over a distance approximating the normal spacing of neighboring troops permits the maintenance of spacing patterns with a significant economy of effort, as compared with a signal system that only permits location and identification of neighboring groups at close range (Marler, 1968). Vocal signals lend themselves well to such a function, and they are convenient for distance signaling in animals that dwell in dense rain forest. Thus, in the gibbon repertoire we should be prepared for loud vocalizations that are designed for the maintenance of intergroup spacing. Given the small size and relatively coherent nature of the social group in gibbons and given a social network that is presumably simple, at least with regard to number of participants, the requirements for complex intragroup signaling may not be great (cf. Chivers 1974). Thus, the dictates of intergroup signaling might be expected to dominate the gibbon vocal repertoire.
The emphasis on distance signaling may be even greater in the orang-utan, where the requirements for territoriality are added to those of a highly dispersed group organization, though there will still be demands for close-range signaling within subgroups and consortships. By contrast, coherent social groups of the gorilla, which are larger than the gibbons' and have more participants in the network of social interactions, should lead to emphasis on close-range signaling, though in adult males there are still requirements for the maintenance of spacing between groups. The strongest emphasis of all on complex, relatively close-range, within-group signaling might be expected in the labile social system of the chimpanzee, though here too there are requirements for long-range communication between separated group members and between different social groups.
The apes are well provided with the classical sensory modalities for the detection of stimuli from sources at a distance from the body. Following sections will review the use of sounds in ape communication and, in less detail, visual signals (see reviews in Andrew, 1963; van Hooff, 1962, 1967, 1970, 1971; van Lawick-Goodall, 1968a; Wickler, 1967).
There is also evidence that specialized chemical signals are emitted in social situations. The most obvious circumstance, well studied in recent years in macaques by Michael and his colleagues, is found in sexual relationships between estrous females and adult males (Michael, 1969). Sexual pheromones may be as important in apes as in macaques, though they are yet to be studied. Certainly there is every sign of keen olfactory interest among males when a female is in estrus. A finger is often touched to the genital area of the female chimpanzee and then brought to the nose and sniffed. In chimpanzees the olfactory investigation of the female genital area is most frequent at the first sign of sexual swelling and during her detumescence, and is less common when the female has a large sexual swelling (van Lawick-Goodall, 1968b). Similar behavior by captive female gorillas toward infants has been recorded (Hess, 1973). Apart from genital secretions, there seems to be no evidence of the discrete glands specialized to produce chemical signals that are commonly found in prosimians and are also present in both platyrrhine and catarrhine monkeys.
Production and Perception of Sounds
Apes have ample means for producing a variety of well-structured sounds. All of them make some use of nonvocal sounds, notably the drumming of male chimpanzees on the ground and on buttresses of large trees, and the chest beating by the gorilla, augmented in silverbacked and large blackbacked males by the resonant properties of inflated air sacs on the chest (Schaller, 1963). Vocalizations nevertheless provide the main medium for auditory communication. The basic structure of the larynx of the great apes is similar to that of man, and it probably operates on the same fundamental principles (Negus, 1949). However, the post-laryngeal tract tends to lack a pharyngeal region. Together with differences in the range of tongue movements that are possible, this lack would restrict the capacity of apes to produce certain of the sounds of human speech (Lieberman, 1968; Lieberman, Crelin, and Klatt, 1972).
Of greater potential biological significance is the presence in all of the apes, other than some gibbons, of laryngeal air sacs, especially marked in adult males. In the adult male chimpanzee, for example, small paired air sacs communicate directly with the laryngeal ventricle. After the lungs are filled, the sacs can be inflated by closing the mouth and nose or the ventricular bands of the larynx. In the gorilla these sacs are larger still, even in adult females, and the sacs of the adult male orang-utan are huge, with a capacity as great as seven liters (Huber, 1931) and extend into the armpit as far as the shoulder blades. Although Negus (1949) attributes a respiratory function to these sacs, which he thinks would permit the recirculation of breath without the inhalation of new air, a function in sound production seems more likely. The sacs are inflated before and during the production of certain loud vocalizations. Possibly the sacs serve as resonators. Another possibility is that flow of air between the sacs and the pharynx over the ventricular ligaments sets them into vibration, thus providing a secondary sound source. Such sounds have been recorded (Brandes, in Huber, 1931), but it is likely that in the apes they are no more than by-products of the process of inflation and deflation of the sacs. Gautier (1971) has analyzed the role of the laryngeal sacs of Cercopithecus monkeys in sound production, and since the morphology is similar to that in the great apes, we may take his analysis as a provisional model.
His studies of the "double boom" of the adult male Cercopithecus neglectus show that air expelled from the sac does not excite the laryngeal cords, and probably not the false cords either. Inflation of the sac immediately before the first vocalization is silent, though if the sac is punctured, then a sound is generated at the time when inflation would normally begin. Before the second boom there is an audible sound during inflation of the sac. Although air is forcefully expelled from the sac during sound production, the fundamental frequency of the sound produced is unchanged from that resulting without sac involvement. By puncturing and resealing the sac of an adult male, Gautier demonstrated that it serves as a resonator and amplifier, selectively emphasizing the fundamental frequency at the expense of higher harmonics, which become more evident when the sac is experimentally opened.
Probably the laryngeal sacs possessed by apes serve a similar function of selectively emphasizing certain frequencies that are produced in the larynx by the vocal cords in the normal way. Thus, the relatively narrow range of frequencies characterizing the loud pant-hooting of chimpanzees, as compared with their other vocalizations, may be attributable to involvement of the laryngeal sacs, which are visibly inflated during this vocalization. To judge by the sound spectograms that Mackinnon (1974) presents of the long calls of adult male orang-utans, again with a narrower range of frequencies than some of the other vocalizations, the same may apply.
Data on hearing are available only for the chimpanzee (Stebbins, 1971). On the basis of what is known about this species, we can assume provisionally that they hear their vocalizations in much the same way we do. The auditory threshold curve for the chimpanzee resembles our own, with the greatest sensitivity between about 1 kHz and 8 kHz, though the range does extend into the ultrasonic, to about 30 kHz (see Stebbins, 1971, for details). Their capacity to discriminate between different frequencies is also much like our own, so that our ears are probably a good basis for at least a preliminary analysis of chimpanzee sounds and probably of the sounds of the other species of great ape as well.
The apes commonly use nonvocal sounds in their displays, the most elementary being branch shaking, breaking, and dropping. The noise of gibbon brachiation displays is audible for 100 m or more. The hand may be used to thump on the ground or on the bole of a tree. This action has become specialized in the chimpanzee as stereotyped drumming behavior in which a chimpanzee will
leap up at a tree and drum on the trunk or buttresses with his feet—usually the two feet pound down one after the other in quick succession making a double beat; there is then a slight pause before the next double beat. From one to three double beats are normal. . . . Frequently, when a number of chimpanzees travelling together come upon some favored 'drumming tree' along the track (a tree with wide buttresses) each male in turn drums in this manner. This results in a whole series of one to three double beats with irregular intervals between each, [van Lawick-Goodall, 1968b:273].
The gorilla creates a nonvocal sound with rather similar structural and temporal patterning by a quite different means, namely chest beating, documented in detail by Schaller (1963). Chest beating occurs as one component in a sequence that typically includes eight other distinct acts.
The hands are almost invariably held flat while beating the chest; that is, with fingers extended and the palm often slightly cupped. The animal tends to hold its hand within six inches or less of the chest and the alternate beats are rapid and direct. The fingers are often spread on females and young animals, but those of adult males tend to touch each other. The sound produced when a silverbacked or large blackbacked male beats his chest may be described as a hollow "pok-pok-pok" somewhat resembling the noise produced by rapping an empty gourd with sticks. Under favorable conditions the sound carries for as much as a mile. In silverbacked males the prominent air sacs presumably act as resonators, for their sudden inflation on each side of the throat is sometimes readily apparent before the chest beats. Small blackbacked males, females and youngsters produce a mere slapping sound when beating the chest. [Schaller, 1963:225]
Although as many as twenty slaps may be given in gorilla chest beating, the modal value is between three and four, not too different from the number of chimpanzee drumming beats that is typical (from one to three double beats are normal; van Lawick-Goodall, 1968b:273). Both are effective long-distance signals, audible in the forest at considerable distances.
Orang-utans and gibbons spend most of their time in the forest canopy, where sound attenuation is less than on the ground. They have no nonvocal signals equivalent in carrying power to drumming and chest beating. However, the noisy branch breaking by disturbed orang-utans (Schaller, 1963; Mackinnon, 1974) may be communicative. Female Kloss's gibbons brachiate among branches of tall trees while singing (Tenaza, 1976), but the noises of rustling vegetation thus produced are audible for much shorter distances than the songs occurring with them. This form of branch shaking occurs in all gibbons, and is probably more a visual than an auditory signal.
VOCALIZATIONS OF THE CHIMPANZEE
Although the vocal behavior of the chimpanzee has been studied more than that of any other ape (Reynolds and Reynolds, 1965; van Hooff, 1962, 1967; van Lawick-Goodall, 1968a, 1968b), including some sound spectrographic analyses (van Hooff, 1971; Marler, 1965, 1969), a comprehensive acoustical description has yet to be published. Based on long and intimate study, the verbal descriptions of vocalizations of chimpanzees in the Gombe National Park in Tanzania by van Lawick-Goodall provide the basis for the present account, augmented with data from a ten-week study of behavior of the Gombe population (Marler, 1969, 1976; Marler and Hobbet, 1975).
The catalog of vocalizations prepared by van Lawick-Goodall (1968a) included twenty-four classes of vocal sounds. For purposes of comparison with other ape species the list was reduced to thirteen. Chimpanzee vocalizations are highly graded and the list of twenty-four includes several obvious variations on a common theme (e.g., pant-hoots and pant-shrieks; barks and shrieks), which were merged into single categories. Sounds distinguished primarily by context were merged wherever acoustical analysis failed to reveal reliable characters for distinguishing them (e.g., hoo and hoo-whimper and whimper, scream calls and screaming, panting and copulatory panting). Sounds distinguished by the age and sex of the vocalizer were also grouped, even though acoustical differences were detectable (e.g., screaming and infant screaming). Finally, two sounds occurred so rarely that their acoustical identity was not established (soft grunt, groan). Since each of the two sounds is similar to another category (grunting, rough grunting), they are not treated separately here. In addition to these changes some new names have been coined either where acoustical analysis suggests a more appropriate term or where sounds had been defined by their accompanying posture or facial expression (e.g., bobbing pants: changed to pant-grunting; cf. also mixed names used by van Hooff, 1971).
Signaling Behavior of Apes
A catalog of chimpanzee vocalizations.
The thirteen remaining basic vocal categories are listed in Table 1 together with four nonvocal sounds ranging in loudness from lip smacking and teeth clacking, which are audible only up to a distance of two or three meters, branch shaking, which is somewhat louder, ground thumping, which we hear at many meters distance, and drumming, audible several hundred meters away. The frequency with which each type of vocalization and one nonvocal sound—lip smacking—were used in one study (Marier, 1976) by each of ten classes of individuals identified by sex and age is shown in Table 2 (center figure in each cell). Also included are the percentage that each makes up of the total recorded for that class of vocalization (top figure in each cell) and the percentage of all vocalizations recorded for that class of individuals (bottom figure in each cell). At the top of each column the number in parentheses indicates the number of individuals representing each class in the study population at Gombe at the time. These numbers range from one (juvenile male) to nine (mother with infant), some individuals making a disproportionally large contribution to the total number of vocalizations recorded. To correct for this Table 3 was prepared by dividing the frequencies with which each vocalization was recorded by the number of individuals in each class, providing a clearer picture of differences in frequency of usage between sex and age classes.
The following brief account of thirteen basic chimpanzee sounds illustrates the kind of communicative sounds that an ape may use. Vocal sounds are considered in order of the frequency of use by the Gombe study population as a whole (Marler, 1976).
(1) Pant-hoot is by far the most frequent vocalization heard at Gombe Stream. Every entry in Tables 1 and 2 denotes a pant-hooting sequence lasting from seven to eleven seconds, each with many parts. Pant-hooting is one of three vocalizations that are typically voiced on both exhalation and inhalation, the others being laughter and pant-grunting. Fig. 1 shows a portion of a typical pant-hooting sequence by an adult female. Aligned with it are single frames from synchronized movie film, illustrating the configurations of the mouth during pant-hooting. One can see that the frequency spectrum is broader in exhalation sounds than in sounds produced on inhalation and that this in turn relates to the degree of mouth opening. One exhalation sound in Fig. 1 is given with the mouth opened less widely than usual and the frequency spectrum is correspondingly narrowed.
Although both sexes and all ages were recorded pant-hooting, the vocalization is given most often by males (males, 453; females, 195). It is by far the most common sound of an adult male and constitutes 50 percent of all sounds that an older male typically made in the Gombe study (Table 3). In both males and females it is used more frequently with increasing age and makes up an increasing proportion of the sounds uttered as a chimpanzee matures. There are consistent differences in the structure of male and female pant-hooting, readily audible to the unaided human ear. There are also individual differences that permit an experienced human listener to identify a distant pant-hooting individual (van Lawick-Goodall, 1968a; Marler and Hobbet, 1975). Fig. 2 illustrates the kind of differences by which the pant-hooting of three adult males can be distinguished.
Pant-hooting is given in many situations. It frequently occurs in response to pant-hooting of distant individuals, often after the chimpanzee has listened with obvious close attention. On many occasions it seems to occur spontaneously as animals are feeding, or during the night, a feature it shares with other primate calls that serve a long-distance spacing function (Marler, 1968). It may also accompany special occasions such as the eating of prey after capture (Teleki, 1973b) or the rejoining of groups after separation (van Lawick-Goodall, 1968a). It is a loud sound that carries far, and its functions, perhaps multiple, include the long-range announcement of an individual's presence and sex, which permits the continued separation of some individuals and groups and the reunion of others.
(2) Pant-grunt is another vocalization with sound on both inhalation and exhalation. It is quieter and faster than pant-hooting, with breathing cycles averaging about five per second, as compared with one per second in pant-hooting. It is a low-pitched sound, with little energy above 1.5 kHz. As can be seen in Fig. 3, exhalation sounds are lower-pitched than inhalation sounds, the former usually below 500 Hz. The latter often have breathy overtones. Rare or absent at younger ages, pant-grunting first becomes evident in adolescence and its frequency increases with maturity. It was recorded in roughly equal numbers in males and females (males, 131; females, 154).
Pant-grunting is typically given by a subordinate approaching or being approached by a dominant animal. As proximity is reduced, the pant-grunting individual may crouch in the bobbing movement described by van Lawick-Goodall (1968a), or it may lose its nerve and flee, and the vocalization may grade into pant-screaming. Though not always successful, it seems to function as a submissive signal, at least on occasion facilitating the establishment of peaceful proximity with the dominant partner.
(3) Laughter was so named for its resemblance to human laughter, though it is by no means identical(Fig. 4). Van Hooff (1971) refers to it as ah-grunting. It is a highly variable sound with at least three distinct forms: steady exhaled laughter (A), pulsed exhaled laughter which has the quality of a chuckle (B), and wheezing laughter with phonation on both exhalation and inhalation (C). Type A elements may be given singly in the early stages of a laughter sequence or in rhythmical succession, each element lasting a quarter of a second or so. As a play bout proceeds and its vigor increases, type A may grade into type B, as a single exhalation becomes broken into a series of bursts followed by a pause for silent inhalation and then another burst of pulses or chuckles. This in turn may grade into the C form as tickling and biting reach their highest intensity. At this stage phonation on inhalation begins and the laughter assumes a labored wheezing quality, and it may be accompanied by efforts to ward off the tickler. When given rhythmically, the rate varies from about two to five breath cycles per second. The frequency characteristics vary widely. The maximum emphasized frequency is higher in young animals than in adults though a high-frequency, breathy component is often present irrespective of age.
Though recorded from all age classes other than nulliparous adult females, laughter is much more common in young animals than in adults. Males laughed more than females in the Gombe sample, and male infants may well be more outgoing and prone to engage in the vigorous play that generates laughter. However, this sample is probably biased by overrepresentation of laughter by one male infant (Flint) that spent much time in the study area with three siblings and the mother. All were prone to spend much time in play and many sound recordings were made. Not only did young animals make the biggest contribution to laughter, but it was also the most frequent vocalization recorded from infants of both sexes and from juvenile females, though it made a lesser contribution to the vocal output of juvenile males. The context for laughter was always some variety of play, usually with physical contact in the forms of tickling by the hand and biting with the teeth. On a significant number of occasions it was recorded in the absence of physical contact, such as when two animals chased each other around a bush in play. The closing scene in the film "Vocalizations of Wild Chimpanzees" (Marier and van Lawick-Goodall, 1971) illustrates such a chase in play between two infant males competing for possession of a palm flower. There is a close correlation between play and laughter, and laughter terminates without transitional forms when play is interrupted.
(4) Squeak is a short shrill call, one to two tenths of a second in duration, with a fundamental that is frequency-modulated, peaking in midcall at 0.5 to 1 kHz and then falling (Fig. 5). There is a rich array of four to eight or more harmonics often with significant energy up to 6 to 8 kHz. Squeaks are usually given in series at two to five per second, often grading into other sounds in the course of a sequence. Although squeaking was recorded from all age and sex classes, it was more frequent in females than in males in the main sample (female, 170; male, 94). It is used especially often by adolescent animals of both sexes, and in the Gombe study made up a high proportion of the average output of both (Table 3). There was a general tendency for frequent use by animals of intermediate age, which perhaps indicates a correlation with lability of dominance relations. Among males it is notably less common in older adults than in any other age class. It is given by animals showing other signs of distress, for example after an aggressive attack or during the approach of a higher-ranking animal. Van Lawick-Goodall (1968a) notes that the touching or embracing of the subordinate evoked by squeaking may reduce its distress.
(5) Scream is a high-pitched, far-carrying sound that represents 9 percent of the total sample. There are intergrades with squeaks, from which it can be distinguished by the longer duration—more than a quarter or even half a second —and the tendency for energy to be concentrated in the lower harmonics (Fig. 6). The fundamental is usually between 0.5 kHz and 1 kHz or above, but it is often weaker than the second harmonic and even the third. After an upward inflection at the start the pitch is usually steady for the major part of the sound, descending at the end. The tone may be fairly pure, or frequency modulated at a rapid rate (e.g., 100 warbles per second), or there may be an overlay of noise of varying density. As a result the tonal quality of screams can vary widely in the same individual.
Screams were recorded in the Gombe study from all classes of animals, with females contributing rather more to the total sample than males (122:94). The overall frequency of screaming by adult females with young and by adolescent females was especially high. Among males the overall peak was shown by younger adult males and adolescent males. When the figures are converted to frequencies per class of individual, the usage by females is fairly consistent across age classes, apart from a peak in infants that is probably an artifact of the small sample. The distribution among males is more uneven, with the oldest adults screaming considerably less than the adolescent and juvenile males. Like squeaking, screaming is given by victims of social strife associated with dominance relations— hence the infrequent use by high-ranking adult males. It may evoke reassurance behavior in others or, according to van Lawick-Goodall (1968a), recruitment of help against an aggressor.
(6) Whimper is a soft, low-pitched sound about one tenth of a second in duration with a fundamental frequency usually between 200 and 350 Hz (Figs. 7 and 8). This is either a pure tone or a harmonic series, sometimes with only two or three overtones, sometimes a dozen or more. The spectral structure seems to correlate somewhat with the pattern of delivery, the purer form being given singly or spaced at least one or two seconds apart, the more complex forms being more often delivered in series at a rate of two to four per second. It was recorded for both sexes (female, 66; male, 109). All age classes used it at a roughly similar level with the exception of a higher rate in juvenile and infant males, for whom it also makes up a larger proportion of the utterances of a typical individual. The circumstances of whimpering are varied, and van Lawick-Goodall (1968a) lists three separate calls for this category occurring in rather different situations.
One call is given on hearing a strange sound or seeing something strange (Fig. 7). Van Lawick-Goodall described this as the "hoo" call, but sound spectrographic analysis reveals no consistent difference between it and other types of whimpering. However, we confirmed the distinctive pattern of delivery in this context of either single or spaced calls. The other forms of call are given when an animal begs for food, when a frightened infant clings to its mother (Fig. 8) or searches for a nipple, and when an individual is being threatened or is having a temper tantrum, which produces the most rapid sequences. Responses by others to the call include donating food, retrieving or suckling the infant, and offering reassurance gestures; or the call may be ignored.
(7) Bark is a loud sharp call that varies considerably in pitch (Fig. 9). The duration ranges from .1 to .25 seconds with a fundamental between 300 Hz and 1 kHz. So far the description is very like that of a squeak. Barks differ from squeaks in their abrupt onset and their tendency to be noisy. This may take the form of a noise overlay or sudden breaks in pitch with a step change of .1 kHz to .5 kHz. The modulation of frequencies in the bark is often steeper than in the squeak. Finally barks differ from squeaks in that, although there may be a broad array of up to eight or more harmonics, the main energy is usually concentrated in the first one or two so that the call tends to be rather low-pitched. Barks grade into squeaks at the higher-pitched end and into grunts at the lower end.
While barks were not recorded for juvenile males, the numbers for other classes are so small that this lack is hardly significant. Most notable is the high frequency of barking by adolescent females. In the Gombe study barking makes up nearly 20 percent of the vocal output of a typical adolescent female. Although some of this excess results from the indulgence in frequent barking by a single animal (Gigi), adolescent females do seem to bark more than other classes. Barking is associated with annoyance or mild aggressiveness toward another individual, though it may also occur in generalized excitement or while threatening others at a distance. The response of others is not clear although, as van Lawick-Goodall (1968a) notes for the waa bark, others may be prone to look toward the caller and to bark themselves.
(8) Waa bark. Van Lawick-Goodall (1968a) does not distinguish between barks and waa barks, lumping both under the latter heading. However, with practice a distinction can be made both by sound spectrograph and by ear between the shorter bark and the waa-barks, which are more drawn out, often lasting half a second or longer. The mouth tends to be open more widely and conspicuous frequency breaks are consistently present (Fig. 10). There is an overall tendency for the pitch to rise and fall audibly during the course of the sound, but often with a heavy overlay of noise, usually restricted to the early or middle section of the call. Given singly or in series, the call often grades into another call such as a bark or a scream.
The waa bark was recorded in similar numbers from males and females (43:41) with no striking preponderance in any age class other than a rather frequent use by adolescents of both sexes (Tables 2 and 3). Some complementarity between the use of barks and waa barks by a typical individual adolescent male and female (Table 3) suggests that adolescent males are prone to use waa barks in situations in which adolescent females use barks instead. The use of the waa bark when animals are both aggressive and apprehensive, as when a subordinate threatens a dominant animal at a distance, is perhaps consistent with frequent use by an age class whose dominance relations are not yet fully crystallized. Van Hooff (1971) also interprets barking (shrill bark = bark + waa bark) as associated with balanced aggressive and submissive motivation.
(9) Rough grunt. A wide array of sounds ranging from squeaks to barks, with a pulsed grunt as the most typical form, is given by animals eating a favored food. By comparison with other gruntlike sounds of chimpanzees, the tempo of rough grunting is distinctively slow: a typical rate of delivery would be two to four calls per second. However, in a group, with the attendant social excitement commonly generated, the rate may be higher and the structure of the call may also vary. In adult animals that are relaxed and feeding, each call consists of up to about ten glottal pulses with a wide frequency range, given in rapid succession, producing a sound something like a groan (Fig. 11). In younger animals the higher frequency emphasis and pulse rate result in the "tonal grunt" of van Hooff (1971) and an intermediate form that is somewhat barklike (the barking sounds of van Lawick-Goodall, 1968a). With increased excitement the pulse rate may increase still further and the result is a squeak or a "shriek" (ibid.). When food is present in a complex social situation, many different forms are often intermingled. However, in the spaced calls of a relaxed individual the pulsatile structure of a call is usually evident; hence the name for the category as a whole.
Rough grunting was heard more from males than from females in the Gombe study (69:14). Within each sex it was used more frequently by adults, and in males the more frequent use of rough grunting by older adults was striking (Table 3). This call also made up a significant portion of the vocal output of a typical older adult male—almost 10 percent. The close association with preferred food suggests that it functions as a food call. The transfer of attention to the calling animal and the often hurried approach of others are consistent with the notion that rough grunting does in fact convey information about the presence of food. The call is soft, and only animals close by can hear it. The tendency of high-ranking adult males to assume control of limited food sources may explain their frequent use of this call.
(10) Pant. There are several situations in which animals of either sex, typically adults, give rapid panting sounds. There is phonation on both inhalation and exhalation, cycling rapidly at a rate comparable to that of pant-grunting, at about five pairs of sounds per second (Fig. 12). Voicing is rare, though it is occasionally heard on the expiratory phase. The sounds are soft, with most energy concentrated in low frequencies, though some noise may be spread across a wide spectrum. It is given during grooming, upon the reunion of two animals after separation, and as a component in peaceful greeting behavior, when it is often associated with placing the mouth on the face of another, as van Hooff (1971) has indicated. It is also given during copulation by the thrusting male. Its intensity varies in these contexts, and copulatory panting tends to be the loudest (van Lawick-Goodall, 1968a). These are all "peaceful" occasions, and although its communicative significance is obscure, panting may indicate a low probability of aggression.
(11) Grunt. A soft, brief, low-frequency sound, given singly or in trains with a variable rate, occurs in a variety of somewhat ill-defined situations (Fig. 12). The difficulty of identification may well have led to its underrepresentation in the Gombe sample, which suggests rarity or absence in young animals and only occasional use by older ones. It occurs during feeding behavior and during occasions of social excitement, but the contexts for this call are not yet well defined. It often grades into other calls as the situation changes, with transitions into rough grunting, barking, pant-grunting, and other calls. It is an acoustically simple sound produced on a single exhalation through a mouth that is closed or only slightly open.
(12) Cough is similar to grunting but unvoiced, produced by a rush of air out through a more open mouth, giving it a breathy tone (Fig. 13). It is usually given only once, in association with a brief threatening hand gesture to a companion. Adult males and adult and adolescent females account for almost all utterances of this call, which generally accompanies a mild confident threat toward a subordinate animal.
(13) Wraaa. Most rarely heard of all chimpanzee sounds in our study is this unusual call, probably a variant of the waa bark (Fig. 14). It begins in the same way, but the last tonal section is drawn out into a howl that may last a second or more, to judge by ear. No completely satisfactory recordings of this call were made, and a full physical description awaits further study. It was noted only from adult females and adult and adolescent males, with the older adults responsible for the majority of utterances. It is given in response to man and other predators or to strange animals (once to a distant buffalo, for example) and seems to combine elements of threat and alarm. It has been noted in response to discovery of a dead companion (Teleki, 1973), again with evidence of ambivalent aggressive and fearful responses. It probably functions as the only distance alarm call of the chimpanzee.
VOCALIZATIONS OF THE GORILLA AND CHIMPANZEE COMPARED
Fossey (1972) has assembled a list for the mountain gorilla of sixteen vocalizations, indicating where she concurs or differs with Schaller's (1963) classification. A summary of her final list is presented in Table 4. Brackets around three groups of calls suggest where categories might be further lumped, because the calls "could be grouped together on the bases of similarities in their physical structure, a subjective impression of the sounds, the context in which they occurred and the responses they elicited."However, the sixteen types present a compromise between lumping and splitting similar to that used in chimpanzee classification, making it most useful for comparative purposes. Fossey Hdivides the vocalizations into seven functional groupings: aggressive calls (3), mild alarm calls (2), fear and alarm calls (2), distress calls (2), group coordination vocalizations (3), calls for intergroup communication (1), and finally miscellaneous calls (3). A comparison between Tables 1 and 4 indicates a similar array of general functions inferred for gorilla and chimpanzee vocalizations. Fossey also presents statistics on the response evoked in other group members by a large number of examples of gorilla calls, and she documents the usage of calls by sex and age class as well. The latter permits a number of illuminating comparisons with the chimpanzee.
A catalog of gorilla vocalizations.
Table 5 presents a digest of the numbers of vocalizations recorded by Fossey from four classes of individuals (Fossey, 1972, Tables 2 and 3, excluding data from unidentified individuals). The data on chimpanzee vocalizations have been rearranged to fit the same categories. The table also shows the number of individuals in each sex and age class in the study population, permitting calculation of the average number of vocalizations uttered by each class member, given directly (D) and as a percentage (E). A striking difference is apparent.
Whereas vocal behavior of the chimpanzee is evenly distributed throughout all classes of individuals, there is a huge preponderance in the gorilla of vocalizations by adult males, which contribute more than 90 percent of all vocal behavior recorded. This is true both as overall figures and as data reduced to the output of an average class member. Fossey's data also show that there is a further asymmetry in the vocal behavior of adult males, silverbacked males making a much greater contribution than blackbacked males (1490:73).
Using the physical descriptions of gorilla calls presented by Fossey (1972), we have attempted to compare them with chimpanzee vocalizations. Table 4 indicates that all thirteen chimpanzee calls have plausible acoustical parallels in the gorilla. The circumstances in which they are given also seem to be similar, indicating a surprising degree of correspondence in vocal behavior of the two species.
A comparison of the frequency of gorilla and chimpanzee vocal behavior
Retabulation of Fossey's data in a format corresponding to that used for the chimpanzee data permits further comparisons (Tables 6 and 7). Having arranged the calls in order of the frequency of use in the general sample, we can see that the hoot series, which corresponds to the most frequently used chimpanzee vocalization, pant-hooting, ranks only third in the gorilla. However, the circumstances for the two species were not strictly comparable. Whereas the chimpanzee study population at Gombe National Park is fully habituated to human observers, Fossey's study population was less well habituated, especially in the early phases of study. She noted that while "alarm calls" (wraaghs) were the most frequent vocalization, it was probable that if more data had been collected from the same groups later, the alarm calls would have been relatively less frequent and the belch or other group coordination vocalizations would be the most frequent for nearly all age and sex classes (Fossey, 1972:40).
Working with the same study population as Fossey, but now better habituated, Harcourt (pers. comm.) suggests that the rank order of usage is now the following: (1) belch, (2) chuckles, (3) pig grunt, (4) hoot bark or hiccup bark (not distinguished), (5) hoot series, (6) whine and cries, (7) question barks, (8) screams, (9) pant series, (10) copulatory pants (not quantified in Fossey's sample), (11) growl, (12) wraagh, (13) roar, and (14) whinny. Harcourt expressed less confidence in the relative rankings of 4-14 than in that of the first three, and notes that chuckles might rank first on a duration measure though third on an onset measure.
The change in estimated rank of the gorilla wraagh from 1 to 12 is as Fossey had predicted. The frequent calling in the early study was presumably triggered by the observer. If Harcourt's data are confirmed quantitatively, then this becomes another case of close resemblance between chimpanzee and gorilla. The estimated decline in rank of gorilla cries from 8 to 6 in frequency of use is perhaps another manifestation of the habituation process, as is the decline in ranking of screams from 5 to 8. The increased ranking of chuckles as estimated by Harcourt to 2 from 15 perhaps reflects the other side of the habituation coin, namely that certain activities would have been inhibited during observation of nervous animals, notably play behavior. This presumably became relatively more common as habituation proceeded.
Frequency of use of gorilla vocalizations. I.
Perhaps the most intriguing of these changes in relative frequency of use of gorilla calls is the estimated rank increase of the belch from 6 to 1, bringing the gorilla data into even more marked contrast with that for the chimpanzee, where the equivalent call, "rough grunting," ranked 9 in the overall Gombe sample. The situations that Fossey (1972) describes for gorilla belching include not only feeding, but also sunning, grooming, and play, with calling of one individual often evoking belching in several others—a broader array of contexts than those in which chimpanzee rough grunting occurs. It seems likely that this is a significant difference between the two species.
One consequence of the provisioning of the chimpanzee population during gathering of the Gombe sample is the presence of satiated animals, possibly favoring the occurrence of play (Wrangham, pers. comm.). Gathering of animals in the provisioning area may have made agemates more readily available for activities such as play. Furthermore, the presence in the camp area of the Flo family—an unusually large and coherent family group—may also have favored the occurrence of play around the camp. Thus chimpanzee laughter, a subject of special study, is surely overrepresented in the Gombe sample. It now seems conceivable that it may actually be less frequent than in the gorilla in normally dispersed chimpanzee populations.
There was probably an increase in aggressive activities in the Gombe Stream study population as a result of provisioning (Wrangham, 1974), and it is possible that the high ranking in frequency of use of pant-hooting is to some extent attributable to the high level of general arousal maintained in the camp area.
Frequency of use of gorilla vocalizations. II.
Although there is a clear and intriguing contrast in the ranking of the chimpanzee rough grunt and the gorilla belch in frequency of use, perhaps the most striking conclusion to be drawn from the data is again the surprising degree of correspondence between the two species in the rank order of use of corresponding calls.
The whimpers and squeaks of the chimpanzee probably correspond with the "cries" of the gorilla. Whereas cries are confined to infant gorillas, the corresponding chimpanzee calls are used by all ages and both sexes, even though infants still whimper more. This may be taken as a sign of the greater emotionality and expressiveness of the chimpanzee at all ages as compared with the gorilla. However, it may be noted that screams, which rank fifth in frequency of general usage for both species, are used by gorillas of all ages and both sexes, most frequently by silverbacked males. There is a contrast in the chimpanzee in that although all classes of individuals engage in screaming, it is less frequent in males than females and is notably infrequent in the older, higher-ranking males. By contrast silverbacked male gorillas scream the most frequently.
Thus, comparisons of temperament in the two species, which tend to emphasize the phlegmatic nature of the gorilla, should not overlook the remarkable vocal range of silverbacked males. If vocal output is indeed to be used as an index of temperament, then silverbacked males can hardly be viewed as any less expressive and emotional than chimpanzees, however phlegmatic other classes of individuals may appear. One wonders whether their restraint might not be correlated with the relative exuberance and assertiveness of the silverbacked male. Certainly the fact that silverbacks assume prime responsibility for intergroup spacing correlates with their high vocal output and with the existence of at least one vocalization that is unique to adult males (in addition to the roar, Fossey also mentions the whinny as restricted to adult males but notes that this may be an anomalous call assodated with sickness). A territorial organization places heavy responsibilities on the class of individuals that must maintain it. More complex within-group organization in the chimpanzee perhaps helps to explain the lack of any vocalizations that are unique to adult males (see below, p.1027).
VOCALIZATIONS OF THE ORANG-UTAN
The social life of the orang-utan, long considered the most mysterious of the apes, is becoming known from the research of Harrisson (1960), Horr (1972), Rodman (1973), Mackinnon (1974, 1975), and Galdikas-Brindamour (1975). In a preliminary account of vocalizations Mackinnon lists sixteen different types and illustrates some of them with sound spectrograms. While the data permit no detailed comparison with the chimpanzee and gorilla, some comments can be made. It can be seen from Table 8, which lists the vocalizations and a summary of the circumstances in which they are given, that vocalization occurs in much the same general situations as it does in the other two apes. According to Mackinnon, some sounds are given on exhalation, others on inhalation, the latter perhaps occurring more frequently than in other species. One interesting sound is the kiss squeak, associated with a sharp intake of air through pursed trumpet-lips. In one population in Sumatra "animals frequently held the knuckles or back of the hand to the lips while making this noise and this had the affect of deepening the tone/' Most distinctive of all is the "long call" of adult, aggressively dominant males. This call consists of a long train of low-pitched calls which the inflated laryngeal pouch imbues with a rich deep tone, the series lasting from one to three minutes. There is a bubbly introduction building up to a climax of roars, then trailing off into a series of sighs. The number of call units varies from twenty-five to fifty, with differences between Sumatran and Bornean populations in length of the units and duration of the series.
A catalog of orang-utan vocalizations.
Mackinnon says that adult males give long calls irregularly, sometimes several in one day, sometimes none over several days. Evidently this call is given less frequently than chimpanzee pant-hooting or gorilla pant series. In Sumatra the calling peaks sharply in the early morning, and Mackinnon points out an intriguing complementarity with the loud calling of other sympatric species, each with species-specific timing.
According to Mackinnon, adult male long calls are a major vehicle for long-range interaction between orang-utans, sometimes occurring spontaneously, other times triggered by sudden noises such as a tree falling, a branch breaking, or most often by the calling of other orang-utans. Behavior accompanying calling such as hair erection, laryngeal pouch inflation, rocking, and branch shaking—also associated with aggressive display—suggest a corresponding motivation for calling. Males may react similarly to calling by rushing and branch shaking or calling in return, though at times males seem to become silent or to withdraw from another's calling. Mackinnon concludes that long calls are important for spacing out adult males; as a consequence face to face encounters between them are rare.
Although males are highly competitive, the resource at issue is not clear—food, space, or females. There are suggestions that the long call attracts sexually receptive females (Mackinnon, 1974; Horr, 1972), though females may withdraw and hide in response to calling and may sometimes ignore it Tn any case, these observations suggest that the male long call functions to coordinate both intragroup and intergroup interactions. The initiative taken by adult male orang-utans in this regard is reflected in the extreme sexual dimorphism of this species, both in size and external morphology, and in the vocal repertoire, the long call being unique to the adult male. In this respect the orang-utan is closer to the gorilla than to the chimpanzee or the gibbons.
VOCALIZATION OF GIBBONS
Despite the contrasts in social organization between gibbons and other apes, gibbon vocal repertoires are similar in size to those of the other apes. The most complete descriptions of vocalizations by wild gibbons are for the Kloss's gibbon, and white-handed gibbon (Tables 9 and 10), and the siamang (Chivers, in press). Ellefson's list of twelve vocalizations compiled during sixteen months of detailed observation of four white-handed gibbon families may be more complete than our list of ten Kloss's gibbon vocalizations, which is based on a fourteen-week study of thirteen family groups. Chivers (in press) describes eight siamang vocalizations after seventeen months' study of two groups. Boutan (1913) describes fifteen vocalizations used by a whitecheeked gibbon that he kept as a pet for more than five years, and he notes that he heard no vocalizations by other wild or captive members of this species that were not used by his captive specimen. We can conclude for the present that the vocal repertoires of gibbons include something in the vicinity of eight to fifteen vocalizations, possibly with minor differences among species. The most frequently occurring vocalizations in the repertoire of Kloss's gibbon are illustrated in Figs. 15, 16, and 17.
All classes of gibbon vocalizations are used by both sexes and, as might be predicted from the intrasexual nature of gibbon territorial aggression, sexual dimorphism occurs only in the songs.
Similarly, most vocalizations are not confined to particular age classes, though in the siamang (Chivers, in press), Kloss's gibbon, and the white-handed gibbon one vocalization is given only by juveniles (Tables 9 and 10). Although primarily the prerogative of adults, spacing calls are used even by young gibbons still with their parent (Chivers,1972; Ellefson, 1974). However, such singing by young gibbons might in some cases be inhibited in the presence of an adult of the same sex (Tenaza, 1976).
A catalog of Kloss's gibbon vocalizations.
If long-distance vocalizations are to assist in maintaining of spatial organizations based on established social relationships among individuals and groups, as apparently they do, they will be more efficient if they identify vocalizing individuals to others. Countersinging interactions of male Kloss's gibbons occur between males separated by distances of 150-500 m (Tenaza, 1976) and songs of neighboring males predictably show a high degree of individuality (Fig. 18).
Evidence from field studies shows that gibbon songs serve primarily as intergroup rather than intragroup signals, inducing close neighbors of the same sex to sing (Fig. 19). Furthermore, about half of all singing bouts for both sexes of Kloss's gibbons are by individuals engaged in dyadic countersinging with adjacent neighbors (Tenaza, 1976). White-handed gibbons (Carpenter, 1964; Ellefson, 1974) and siamangs (Chivers, in press) behave in a basically similar fashion, though in these species, unlike H. klossii, bisexual choruses also occur.
A catalog of white-handed gibbon vocalizations.
When an individual or a pair of gibbons breaks silence with its song or duet it usually stimulates others to begin singing, and a chorus of temporally overlapping bouts of singing by two or more individuals or pairs of gibbons develops. In Kloss's gibbon 96-97 percent of song bouts by both sexes occur in such choruses, while only the remaining 3-4 percent are produced by gibbons singing alone (Tenaza, 1976).
With information presently available from wild populations of gibbons we can distinguish three kinds of consisting entirely of males singing; choruses consisting entirely of females singing; and choruses consisting of duets sung by mated pairs of gibbons. Interspecific differences in the occurrence and co-occurrence of these chorus types are summarized in Table 11.
All-female choruses occur only in Kloss's gibbon and are confined to the four hours following sunrise after the gibbons have left their sleeping trees for the day. Choruses consisting entirely of males singing occur in the white-handed gibbon (Carpenter, 1964; Ellefson, 1974; Chivers, 1972), the Sumatran dark-handed gibbon (W. Wilson and C. Wilson, pers. comm.), the Bornean gibbon (N. Fittinghoff, P. Rodman, and D. Lindburg, pers. comm.), and Kloss's gibbon (Tenaza, 1976). In all but the white-handed gibbon, all-male choruses typically begin before dawn while the gibbons still are in their sleeping trees. In the white-handed gibbon they typically begin later in the morning, after a period of foraging (Carpenter, 1964; Chivers, 1972; Ellefson, 1974), though males occasionally sing from their sleeping trees just before dawn.
Predawn chorusing by male Kloss's gibbons occurred on 90 percent of all observation days on Siberut Island (Tenaza, 1976), with choruses beginning as early as five hours before sunrise. The adaptive advantage of predawn chorusing in Kloss's gibbon may be related to the demonstrably short supply of safe sleeping trees in Siberut, and to consequent pressures upon males controlling desirable sleeping trees to warn others away by singing in them (Tenaza, 1975; Tenaza and Tilson, in prep.). Whether this hypothesis might be extended to other gibbons with predawn male choruses cannot be determined until the nature and abundance of their sleeping sites have been evaluated.
Nonvocal interactions between adult gibbons of neighboring groups are, with the exception of courtship activities, concerned totally with excluding one another from their respective territories (Carpenter, 1964; Ellefson, 1974; Tenaza, 1976). Hence it seems safe to assume that song serves a similar function. The fact that gibbons without adjacent neighbors sing less often than those holding territories contiguous with other territories (Chivers, 1971; Berkson et al., 1971) provides further evidence for the spacing function of gibbon songs.
The vocal repertoire of adult Kloss's gibbons includes two calls, sirening and alarm trills (Table 9 and Fig. 17), that occur only when a predator is detected. These alarm calls are loud enough to be heard by a man on the ground 800 m away, whereas members of a Kloss's gibbon family rarely are more than 10 m from one another (Tenaza, 1976). It therefore seems likely that these calls function to warn gibbons in adjacent territories of the presence and location of predators, as well as to warn immediate group members, for which the softer hoo and howl alarm calls (Tables 9 and 10 and Fig. 17) that are audible to about 100 m also suffice.
Distribution among species of three kinds of gibbon choruses.
The frequency-modulated syllables of sirening and the rapidly repeated sound elements in the alarm trill(Figs. 17 and 20) make both of these calls easy to locate by binaural comparison of time, phase and intensity cues (cf. Marler, 1955). Hence they accurately reveal the location of a calling animal to predators as well as to neighboring gibbons. Evolution of such behavior is understandable in theory if animals benefiting from one another's alarm calls, i.e., those occupying adjacent territories, have a high probability of being closely related (cf. Andrew, 1963; Smith, 1965).
There are suggestions that adult gibbons occupying adjacent territories might often be related as parents and offspring or perhaps as siblings. Aldrich-Blake and Chivers (1973) observed a male siamang establish a home range contiguous to that of his parents. Tenaza and Tilson (in prep.) witnessed four instances of settling by young adult Kloss's gibbons. In each case the gibbon (two females, two males) left its parents at maturity and settled in a territory contiguous to that of its parents. Three of the four definitely mated in those territories subsequently. Thus proximity of close kin can plausibly be invoked as a factor in the evolution of loud alarm calls.
Unique among vocalizations of higher primates are the duets in which paired males and females of several gibbon species engage. Songs and duets of captive siamangs (Fig. 21C) have been analyzed in detail by Lamprecht (1970). In all species of gibbons for which information is available, other than Kloss's gibbon, males and females sing duets with their mates (Table 11). Male and female songs overlap during duetting in the pileated gibbon (Marshall et al., 1972), as they do in the siamang and hoolock (Figs. 21 and 22). In other gibbons male and female alternate songs with one another.
In the white-handed gibbon the female may emit a few short notes but for the most part is silent while her mate is singing. She signals her readiness to give a full song by uttering a series of short, monotonous notes. When she begins her full song, the male normally stops singing until she has completed it, whereupon he adds a short coda (Fig. 2IB), then pauses for several seconds before starting to sing again. Occasionally a male does not cease singing when the female begins her song; in these rare instances the female does not complete her song but stops and makes another series of presong notes be¬fore starting another song (Tenaza, unpubl.). That the female presong notes serve to signal her mate is supported by our observations of a mated pair of white-handed gibbons at the San Francisco Zoo, made from September through November of 1970. About halfway through the observation period the male of this pair died. Although the female continued to sing normal songs after her mate's death, she stopped pro¬ducing the presong notes. Similarly, an adult female pileated gibbon caged with a mangabey monkey at the San Francisco Zoo did not pre¬cede her songs with series of short notes (Fig. 23E), although mated pileated gibbon females in the wild do produce presong notes (Marshall, et al., 1972). In Kloss's gibbon, which does not duet, mated females do not produce long series of short notes before each song but make only three or four short whups before uttering the first long syllable of song (Fig. 23F).
Duetting by wild dark-handed gibbons, recorded in Malaya by D.J. Chivers and in Sumatra by W. Wilson and C. Wilson, is similar to the duetting of white-handed gibbons described above. Duetting by Bornean gibbons has been recorded in eastern Kalimantan by P. Rodman and N. Fittinghoff. It differs from duetting in other typical gibbons in that the male does not sing between female songs but simply adds a brief coda to the end of her song (Fig. 23A). In a continuous ten-minute recording made by J. T. Marshall of singing by a wild female Javan gibbon, (Fig. 23C), duetting did not occur.
In duetting by siamangs (Fig. 21C) the male sings stereotyped phrases that overlap particular, predictable portions of the female song (Lamprecht, 1970). Thus the sequence of events in a siamang duet is nearly as predictable as it is in the typical gibbons described above.
Singing by hoolocks contrasts with that of the siamang and with that of the typical gibbons in two respects. First, there is no striking sexual dimorphism in the structure of sound elements contributed to a duet. Although in the pair we tape recorded at the Vancouver Zoo (British Columbia) the male sounds tended to be lower in pitch than those of the female, both sexes produced basically the same sounds. Second, the course of events in a duet is not readily predictable. Thus the hoolock duet selected for illustration (Fig. 22) is not necessarily typical for the species or even for the pair that produced it. Duets of this pair varied in duration, in relative contributions of male and female, in the ways basic sounds were combined, and in the order of singing by male and female. Recordings of wild hoolocks in Assam by R. Tilson and in Burma by J. T. Marshall support this general picture.
Captivity seems to have no effect upon the song structure or the nature of duetting in gibbons. Six white-handed gibbon pairs that we recorded in various zoos all have the same basic song and duet patterns, with slight variations within and among individuals and pairs; they do not differ from those we recorded on the Malay Peninsula and heard in Thailand. Similarly, siamang pairs we have recorded in zoos and in the wild are basically alike and their patterns do not differ from those described by Lamprecht (1970).
Several species of gibbons have hybridized in captivity (see records in The International Zoo Yearbook). Duetting between mates of different species and the inheritance of songs by their hybrid progeny remain unstudied. In 1970 we tape recorded duets between a white-handed gibbon male and a Bornean gibbon female that had been caged together for nine years and had produced two viable offspring, a male and a female, at the Micke Grove Zoo (Lodi, California). Both individuals sang their sex- and species-typical songs. The male responded to his mate's songs in the manner typical of his species, although his mate was of another species. He would stop singing when the female began a song and would add the typical coda to the end of her songs. After death of the parents the hybrid offspring were kept together and began duetting late in 1975. They behaved like a mated pair. Songs of the hybrid female are structurally intermediate between songs of the parental species, demonstrating clearly the inheritance of these vocal patterns. Both song structure and duetting behavior of the male hybrid resemble those of the male whitehanded gibbon.
The pelage of three species of gibbons—the hoolock, the pileated, and the white-cheeked—is sexually dimorphic: adult males are black and adult females are generally yellowish (cf. Fooden, 1969). Brockelman and his colleagues (1974) have raised the question of whether color dimorphism might substitute for song dimorphism to facilitate sexual recognition in these species. Sonagraphic analyses demonstrate a high degree of sexual dimorphism in pileated gibbon songs but little in hoolock songs (Figs. 22 and 23). Adult white-cheeked gibbons show a degree of sexual song dimorphism comparable to that in pileated gibbons (Marshall et al., 1972). Hence among the sexually dichromic gibbons we find two species with pronounced sexual dimorphism in songs and one with apparently slight dimorphism. Thus sexual dimorphism in color does not replace sexual dimorphism in song except perhaps in the hoolock. Instead it may complement song by facilitating instant visual recognition of an adult's sex in the same way that song allows rapid auditory recognition of sex in all species but the hoolock. None of the sexually dichromic gibbon species has yet been studied more than casually in the field.
Production and Perception of Visual Signals
The visual system of the great apes is very similar to our own, with extensive overlap of the visual fields, a mixed foveal retina with color vision, and high visual acuity. The correspondence between apes and ourselves in visual signaling equipment is also notable. The lack of a tail is compensated by extensive use of the hands in visual signaling. The facial musculature is elaborate and is undoubtedly specialized for the production of a variety of facial expressions (Huber, 1931; Andrew, 1963) made more visible by the lack of hair on the face. In gibbons conspicuousness of the dark, expressive area of the face is enhanced by a band of white hairs above the face (H. hoolock), beside it (H. concolor), or surrounding it (H. lar, H. agilis, H. pileatus, and H. muelleri). Piloerection provides another way of changing the contour of the trunk and limbs, and slower physiological changes are conspicuous in the visible genital swelling of female chimpanzees during estrus and in some cases during pregnancy (van Lawick-Goodall, 1968b, pers. comm.).
There is thus ample morphological equipment for generating a wide array of visual as well as auditory signals, with a somewhat lesser emphasis on olfactory signaling. There are parallel trends on the sensory side. Thus, we should prepare for the likelihood that vision and audition will be the most important sensory modalities for social communication, though, as indicated earlier, olfaction is probably important in sexual behavior. Finally, mention should be made of the tactile sense. Although sensory and motor specializations are hard to detect, there is evidence that certain parts of the body are well provided with tactile nerve endings, especially the face, hands, and genitalia. There is ample evidence that tactile signals are especially important in the kinds of social communication that emphasize contact with those parts of the body that are richly innervated (e.g., Simpson, 1973).
Of all parts of the ape body none is more concerned with the production of visual signals than the face. Some facial configurations occur only with particular vocalizations, which they structure by altering the size and shape of the mouth aperture and resonating cavities. Whatever information these vocalization-bound facial modifications might communicate to recipients about the performer's motivational state therefore appears to depend mainly on their accompanying vocalizations. Here we shall disregard these expressions and focus on others which, while they can be accompanied by vocalizations, apparently are independent enough in some if not all species of apes to convey information about what the performer is likely to do next even when given silently. Andrew (1963) considers most such facial expressions to have been freed or "facilitated" during their evolution from a former association with vocalizations. Ignoring unfacilitated facial configurations and lumping together those distinguished only by minor differences of context, intensity, or accompanying vocalizations, the repertoire of chimpanzee facial expressions can be reduced to six (Table 12).
Because chimpanzee facial expressions have been intensively studied under nearly ideal observation conditions in the field (van Lawick-Goodall, 1968a, 1968b, 1971) and in a captive colony (van Hooff, 1971), they provide the best baseline for comparison with other apes. Thus, we find that every general category of chimpanzee facial expression has been described in at least two of the other three apes (Table 13). Furthermore, the facial expressions of gibbon, gorilla, and orang-utan function within the same range of social circumstances in which comparable chimpanzee expressions occur (Table 12) (Baldwin and Teleki, 1976; Schaller, 1963; Mackinnon, 1974). Similar facial expressions apparently serve comparable social functions among the apes.
Ultimately the similarity of facial expressions among apes and between apes and monkeys (van Hooff, 1967; Chevalier-Skolnikoff, 1974) can be traced to similarities in facial anatomy. The simplest facial musculature among apes is that of the gibbon, its structural complexity somewhere between the simpler cercopithecoid monkeys and the more complex great apes. This primitive "ground plan" of ape facial musculature seen in the gibbon has increased in complexity along divergent lines leading to the orang-utan on one hand and to the chimpanzee, gorilla, and human on the other (Huber, 1931). Chimpanzee facial expressions are shown in relation to their effector muscles in Fig. 24.
Modifications of orang-utan facial musculature are most conspicuous in the platysma muscles over the cranium, and in the labio-buccal musculature. They are related to support and control of the greatly enlarged laryngeal air sacs, to support of the male cheek pads (Huber, 1931), and perhaps, as Chevalier-Skolnikoff (1974) has suggested for the chimpanzee, to elaboration of the rather prehensile lips for plucking and manipulating food items.
In the chimpanzee/gorilla/human line of facial development homologs of gibbon facial muscles, particularly in the midfacial region, are structurally differentiated into increasingly more and finer subunits proceeding from gibbon to chimpanzee to gorilla to human. The advantage of this increasing complexity might, as Huber (1931) suggested for humans, be related to an increased ability to produce subtle degrees of expression with nuances of meaning, but there is little evidence on the matter.
A classification of facial expressions of the chimpanzee.
Lip smacking is presumed to have originated from jaw, tongue, and lip movements performed while eating ectoparasites and other foreign particles removed from body surfaces during auto- and allo-grooming (van Hooff, 1967). In the white-handed gibbon, orang-utan, and chimpanzee these movements, which have not been observed in gorillas, still occur in their original context of ingesting particles removed with the lips, teeth, and fingers from another's body during allo-grooming(Baldwin and Teleki, 1976; Carpenter, 1964; van Lawick-Goodall, 1968b; Mackinnon, 1974). In the orang-utan they have been described only in this functional context, whereas white-handed gibbons (Tenaza, pers. obs.) and chimpanzees (van Hooff, 1967, 1971; van Lawick-Goodall, 1968b) often perform lip smacking during grooming without taking any foreign particle into the mouth. Noningestive lip smacking in the chimpanzee "does not resemble the original functional smacking in that the protrusion of the tongue is barely noticeable," (van Hooff, 1967:41). In adult white-handed gibbons full downward protrusion of the tongue through parted lips occurs during non-ingestive lipsmacking, but the tongue is only slightly protruded when the behavior is ingestive (Tenaza, pers. obs.).
Probable homologs of chimpanzee facial expressions in other apes.
The Design and Function of Communication Signals in Apes
Selective pressure for species-specificity in communication signals, such a powerful force in the evolution of avian vocalizations, has less effect on the signals of primates, though in certain groups such as the Cercopithecus monkeys(Marler, 1973) and gibbons (Figs. 21-23), its influence can be discerned. Although the chimpanzee and the gorilla are sympatric in the gross sense, they seem to live within earshot of one another only rarely (Gartlan, pers. comm.). They are presumably separated by differences in habitat selection. It is worth considering the possibility that there is active repulsion between them. Comparison of vocalizations of the two species has revealed a surprising degree of correspondence that includes some of those signals thought to be used for maintaining the spacing of conspecific groups. Chimpanzee pant-hooting and gorilla hoot series are somewhat similar acoustically, and one can hardly overlook the convergence by gorillas and chimpanzees on two quite different means of producing far-carrying drumming sounds, chest beating and tree drumming. Both species also use ground thumps and branch breaking. Areas where both species live in close proximity should be studied for possible interspecific reactivity.
Orang-utans and gibbons do live in close contact. While more complete descriptions of their vocal behavior are needed before we can speculate about possible evolutionary interactions, Mackinnon (1974) has an intriguing observation on the timing of adult male orang-utan long-calling in Sumatra in relation to that of loud calls of the white-handed gibbon, the siamang, and the leaf monkey (Presbytis aygula) (Fig. 25). The notion that the nocturnal peak of orangutan calling is timed to avoid temporal overlap with the calling of other species is reinforced by the absence of such a peak in calling of Borneanorang-utans. There primate densities are lower, the siamang is absent, and the only other primate heard calling regularly was the Bornean gibbon (Hylobates muelleri). Irrespective of any detailed correspondence in the structure of vocalizations, the mere presence of loud calling by one species may add significantly to the background of noise against which another must make itself heard, hence presumably the adaptive value of avoiding temporal overlap.
On Siberut Island, Indonesia, both male Mentawai langurs (Presbytis potenziani) and male Kloss's gibbons produce loud predawn spacing calls. They overlap, with gibbon calling (songs) concentrated in the hours 0300-0600 and langur calling at 0400-0600. However, calling by the langurs does not occur in prolonged choruses like the singing of the gibbons but in bouts ranging from eight seconds to eight minutes duration, during which males from different groups take turns producing calls of three to four seconds duration spaced from one to sixty seconds apart. Even during the hour before dawn, when langur calling reaches its peak, these rounds of calling (including silent intervals between calls) occupy on the average only five and a half minutes, or less than 10 percent of the hour. Hence they interfere very little with the gibbons' choruses. Since among primates on Siberut only langurs and gibbons habitually make loud spacing calls, interspecific competition among primates for the auditory environment is negligible.
Dawn bird choruses on Siberut are, by contrast, continuous and loud, involving many individuals of several different species. Tenaza noted in the field that background noise generated by the dawn bird chorus not only reduces the audible range of gibbon songs but also makes it difficult and often impossible to determine by ear the location of a singing gibbon. It therefore is inefficient for gibbons to sing during the bird choruses, and it seems likely that the temporal separation of Kloss's gibbon choruses from bird choruses (Fig. 26) is an evolutionary consequence of interspecific competition for the auditory environment.
Species-specificity is a matter of special interest in gibbon songs. In a study of the structure of female songs, we found strong resemblances in six of the eight species analyzed, with extremes in the series connected by intermediates (Fig. 23). Female songs in this group are preceded by short, monotonously repeated notes, and begin with a prolonged syllable of rising inflection. In female white-handed and darkhanded gibbons this first syllable is followed by similarly prolonged notes, which rise in pitch near or following the middle of the song, then fall in pitch at the end. In other species of this group the prolonged notes initiating female song become progressively shorter, grading into a slow trill in the Javan gibbon and into a rapid trill in the Bornean, pileated, and Kloss's gibbons. The Kloss's gibbon female trill grades back into prolonged notes that terminate the song, but in the others song ends in the trill.
Despite basic similarity among female songs in this group of gibbons, each species has a species-typical female song structure. The greatest similarities are between songs of dark-handed and white-handed gibbon females on the one hand and between those of Bornean and pileated gibbon females on the other, but these two song types are bridged rather nicely by the slowly trilled song of the female Javan gibbon. The long, typical song of female Kloss's gibbons might seem widely separated from dark- and white-handed gibbon songs, but the gap is narrowed by its trill-less variant (Fig. 16C), which resembles them rather closely in both song duration and signal structure.
Among males in these species there is less interspecific similarity in songs than we find among females. While the songs of Bornean and pileated gibbon females are nearly identical, male songs of these species differ considerably; the same is true of the dark-handed and white-handed gibbons. This perhaps is related to present and past contact between populations and selection against hybridizing individuals, leading to interpopulation divergence in male songs. Available evidence suggests that male gibbon song functions in mate attraction, as well as in territorial interaction (Aldrich-Blake and Chivers, 1973; Chivers, 1974; Tenaza, 1976). White-handed and dark-handed gibbons presently have contiguously allopatric distributions in Malaya and Sumatra (Fooden, 1969). Paul Gittins recently discovered a zone of sympatry and three hybrid pairs, two with young, near the Mudah River in Malaya (D. J. Chivers, pers. comm.). Although Bornean gibbons are restricted to Borneo and pileated gibbons to southeastern Thailand and Cambodia (west of the Mekong River), these areas have been connected by extensive land bridges during the Pleistocene and also are connected by the major faunal migration route proposed by de Terra (1943). Hence it is possible that the similarity in female songs of the two species reflects close phylogenetic relatedness, rather than convergence, and that selection against hybridization has led to the divergence in male songs. Pileated and white-handed gibbons are for the most part allopatric but have a narrow zone of sympatry in southeastern Thailand (Marshall et al., 1972). W. Y. Brockelman and J. T. Marshall have discovered interspecific pairs and hybrid individuals in this area (Brockelman, pers. comm.).
Whatever the details of their recent evolutionary divergence, the basic similarity of their songs suggests that the gibbons illustrated in Fig. 23 are more closely related to one another than any of them is to the hoolock or to the siamang. It also supports the conclusion that Kloss's gibbon should be grouped with these "typical" gibbons (Chasen and Kloss, 1927; Schultz, 1932) rather than with the siamang, where originally it had been placed (Miller, 1903). As indicated earlier, the songs of the siamang and the hoolock differ considerably from one another and from the other species considered here.
AGE CLASS AND SEX-SPECIFICITY
Any specificity in vocal morphology that relates to sex results in the encoding of additional information of potential communicative value. Some vocal categories may be completely absent from one sex or the other. This is true of the gorilla and the orang-utan, with certain adult male vocalizations absent in the female repertoire. A predominantly male role in territorial defense perhaps explains this sexual asymmetry in vocalizations, a suggestion that seems at least plausible when we consider the sharing of territorial defense by male and female gibbons with similar male and female repertoires. However, the same applies to chimpanzees, in which males are territorial and females are not.
Sexual differences in vocal behavior may also occur in the frequency of use of similar categories of vocalization. Quantitative data on chimpanzee and gorilla calling reveal many such asymmetries. Most striking is the overwhelming domination of gorilla vocalization by adult males, which applies to all but one of the fourteen vocalizations used by adult gorillas (Fossey, 1972; and Table 7). Thus, only one vocalization is used more by adult females (69 percent). Even here adult males contribute 23 percent to the total, and the overall numbers are small, this being one of the least frequently used vocalizations. There is no record of a gorilla vocalization unique to adult females.
Although there are interesting differences in the details of vocal use between male and female chimpanzees, perhaps most striking is the extent of sharing of all vocal categories between the sexes. There is no call type unique to one sex, and males and females contribute roughly equal proportions to the overall frequency of vocal behavior.
A difference in adult sexual roles in the chimpanzee and the gibbons is implied by consistent differences in male and female renditions of certain calls. Although both male and female chimpanzees engage in pant-hooting, there are differences striking enough to permit a human observer to determine the sex of the vocalizer (Marler and Hobbet, 1975). The same is true of gibbon songs (Fig. 23). The contrast has been interpreted differently in the two genera. Both male and female gibbons defend their territory against intruders of their own sex, hence the selection pressure for sexual differences in calling. Within the highly dispersed social groupings of chimpanzees, the sexual difference in pant-hooting is presumably as much an issue of communication within groups as between them.
One way to measure the extent of sharing of a vocalization by different sex and age classes is with an index of diversity derived from information theory (Baylis, in press). Based on the formula indicated in the legend of Table 14, Hjmax is the value if all contributed equally to a given vocalization. The extent of departure of the realized values of H1 from the maximum reflects the degree of inequality of sex and age class contributions to that call type.
As can be seen from Table 13, the values for many chimpanzee vocalizations are close to the maximum and H\ is less than half the maximum for only two of fourteen sounds, laughter and lip smacking. The situation in the gorilla is very different, with Hi less than half of Hi max in twelve of fifteen sounds, an obvious reflection of the lower diversity of usage of most gorilla sounds. The usage diversity of the great majority of chimpanzee sounds exceeds that of even the most diversely used gorilla vocalization.
There is ample evidence that the social roles of primates vary with age, and data on chimpanzee and gorilla vocalization reveal striking differences in this regard. In the gorilla the domination of vocal behavior by adult males is due almost entirely to silverbacks, blackbacked males differing much less from adult females than silverbacks do in the frequency of use of different call types. Thus, there are many behavioral markers signifying the age class that silverbacked males represent. The gorilla data do not permit further characterization of intrasexual age classes, but they reveal some calls unique to infants (2) or used frequently by them (1).
Similarly in the Gombe sample of chimpanzee vocalizations, laughter is dominated by infants and juveniles of both sexes, though especially by males. Other differences in age class usage of vocal types can be seen in Figs. 27 and 28. Thus, adolescent females tend to dominate the bark vocalization, older adult males the wraaa and rough grunting, infant males the whimpering, and older adult females the pantgrunting. However, the greater use of a call by a particular age may be only statistical in nature, and thus unreliable as a marker of that particular age class to other individuals. This is true of distinctions between older and younger adult male chimpanzees, none of which is as marked as the differences between blackbacked and silverbacked gorillas. The pattern of vocal usage by older and younger adult male chimpanzees does differ, however (Fig. 28), in ways that may again correlate with more subtle changes in social role with age. Thus, in the Gombe sample, older males use the aggressive "cough" more often than younger ones, as well as the wraaa call, which has both alarm and aggressive connotations. The same applies to the rough grunting given at food, which, since older males outrank younger males, they tend to control. These preliminary data suggest that quantitative records of vocal behavior may provide a sensitive and versatile tool for investigating the details of social role in ape societies.
Diversity of usage of chimpanzee and gorilla sounds by age and sex classes.
An observer must be familiar with the vocal behavior of a particular population of a species before individuality becomes detectable unless individual differences are very prominent, as in many bird songs. Thus far individual specificity has been demonstrated in only two ape vocalizations, the pant-hooting of chimpanzees (Marler and Hobbet, 1975) and the "songs" of male Kloss's gibbon (Tenaza, 1976), though it undoubtedly occurs in vocalizations of other species of apes as well (e.g., Chivers, 1974:238). In both cases the vocalization is uttered by males and females, with individuality imposed on sexspecific variations.
In Kloss's gibbon males in adjacent territories countersing in the dark and through dense foliage over distances of 150-500 m. Females, on the other hand, meet at territorial boundaries in view of one another and engage in vigorous, mutual, visual displays while countersinging (Tenaza, 1976). Demands for individuality in female song might be less than in male song, though this has yet to be studied.
The pant-hooting of adult female chimpanzees differs consistently from that of adult and adolescent males in the absence of a "climax" section (Marier and Hobbet, 1975). Also, the average duration of female pant-hooting sequences studied was greater than that of male sequences, though the ranges overlap greatly. The pitch of the first harmonic of female loud calls—their equivalent to a climax—tends to be deeper than that of the climax calls of most males; their duration is shorter; and their shape tends to be more arched. Thus, there are ample cues available for a listening chimpanzee to establish the sex of distant pant-hooters, even if individual differences cannot be discerned. An intriguing relationship has been suggested between the duration of pant-hooting and age in males. The sequences of two older males studied were significantly shorter than those of younger males. This difference also correlates with dominance rank; higher-ranking animals had shorter pant-hooting sequences. The long sequences of the two females studied, lower than males in rank, conform to this relationship. More study is needed to establish its generality.
Individuality in the pant-hooting of seven chimpanzees at Gombe National Park was studied by calculating the significance of differences between all possible pairs (Marler and Hobbet, 1975). The four measures compared were (a) duration of pant-hooting sequences, (b) peak frequency of climax exhalation calls, (c) duration of climax exhalation calls and (d) interval between climax exhalation calls. All pairs differed significantly (p˂.05, two tailed t-test) in at least one measure, and they averaged 2.3 significantly different measures per pair analyzed, out of a possible maximum of 4.0 (Table 15). On this basis there are cues available in pant-hooting for individual recognition, although the likelihood that some properties of pant-hooting are more conspicuous than others to listening chimpanzees limits the value of a simple statistical analysis of individual differences.
A tally of those measures in which the pant-hooting of pairs of individual chimpanzees showed significant differences (p ˂.05, two-tailed t-test).
In both gibbons and chimpanzees field observers have been able to identify individuals by their calling. Whether species members do the same remains to be demonstrated, though it seems probable. The patterned sounds of panthooting and singing provide sufficient parameters to maintain individuality in addition to higher orders of distinctiveness at sexual and specific levels. Further study may reveal individuality in other simpler vocalizations.
The abundant demonstrations of group specificity manifest in bird song as local dialects (Marier and Mundinger, 1971) have yet to be paralleled in any of the apes or other primates, apart from humans. The only suggestion of something equivalent comes from a demonstration of troop differences in one class of calls of the Japanese macaque (Green, 1975b). With this exception the absence of local dialects in ape vocalizations is consistent with the failure thus far to demonstrate that individuals learn from others in vocal development.
No student of ape behavior has yet undertaken an analysis of what we might label répertoriai specificity—the degrees to which different items in a repertoire differ distinctively from one another. The problem is complex since an observer's selection of measures on which an acoustical comparison of recorded vocalizations is based might not coincide with those parameters most salient to a conspecific listener. Human judgments as to which sound features are most conspicuous, used in studies of Cercopithecus vocalizations (Marler, 1973), are probably better than a random selection, but perceptual studies with conspecific subjects are urgently needed, as well as thorough descriptive analyses.
The first question is whether vocalizations are discretely separate from one another, and, if so, what the acoustical distance is between them. The situation is complicated by the abundance of intermediate forms between the categories of many ape calls, constituting vocal systems that are graded rather than discrete (Fossey, 1972; Marler, 1965, 1969, 1976; van Hooff, 1971). Methods for interpreting such graded vocal systems remain to be developed (but see Green, 1975a). From a functional viewpoint one might expect some intraspecific vocal distinctions to be more critical than others. The difference between one call designating predators and another designating food might be highly significant, whereas the potential confusability of calls associated with varying levels of intensity within a single system such as aggressive signaling would be less critical. Presumably discrimination will be most accurate and rapid between calls that are discretely different from one another and that have considerable acoustical space separating them. At present we lack the data to test this prediction.
DISCRETE AND GRADED SIGNAL SYSTEMS
Variation from one rendition of an item from an animal's repertoire to another may take several forms. A call type that is discretely separate from other types may exhibit within-category variation. Such variation might be accidental and random, or it might be orderly and even highly organized. Among chimpanzee vocalizations laughter is perhaps the most discrete, yet it varies considerably in ways that are likely to have communicative significance.
Categories may vary so much that different types become connectable by intermediate forms. Variation of this type raises a number of questions. In the first place it renders uncertain the meaning of the original categories. Even though field observers confronted with such a graded vocal system feel reasonably confident in discerning categories, their judgments may be based on frequent usage of modal forms, intermediates often being rarer (Table 16).
Numbers of intermediate chimpanzee calls classified by the categories they fall between.
The functional significance of vocal grading is still unclear. We can distinguish at least two ways in which it may emerge during a descriptive analysis of vocal behavior. There may be adjacent graded forms or separate ones. In the former case an individual produces a string of continuously changing variants so that calling changes quickly from type A to type B in a series of small steps rather than a single jump. In such a series the differences between adjacent pairs may be slight, and one may wonder whether they would be perceptible to another animal. But if they are given in a string with brief intervals separating them, then each provides a frame of reference for the next, and directions and rates of change of vocal morphology are thus more readily detectable. It may be that this is how some gradations come to assume communicative significance. If this argument were valid, one might predict a correlation between degrees of vocal grading, and the tendency for them to be uttered in rapid strings. There are hints that this may be true of vocalizations of the chimpanzee.
Table 16 classifies 343 chimpanzee calls judged to be intermediates and classified according to the two categories they fall between. Thus, there are two entries for each intermediate call. They make up 12 percent of the 2,656 calls analyzed. The ratio of typical renditions of a call type to those judged to be intermediate in form between it and another category, shown in the third column of Table 16, provides a rough measure of the degree of variability of that call type. At one extreme is the waa bark, for which there were actually more records of intermediates than of typical forms. At the other extreme are the cough and laughter, with no intermediates at all. Intermediates were recorded for eleven of the thirteen vocal categories altogether. Thus, the great majority of chimpanzee call types grade into one another through intermediates. Two seem exempt from such grading, and two have intermediates only rarely. The four that are either discrete or close to it are rough grunting, pant-hooting, coughing, and laughter.
The first nine calls in Table 16 are especially prone to gradations with other categories. While quantitative data have yet to be analyzed, these calls are often given in rapid trains in which there are changes to and fro between several call types. The transitions are commonly bridged by a series of continuously graded intermediates, so that changes are gradual rather than abrupt. This manner of delivery might prove to be a clue to the communicative significance of graded forms in the chimpanzee vocal repertoire.
In addition to grading of "adjacent" utterances, similar variation also occurs in vocalizations rendered separately. Thus, an observer recording vocalizations given singly at different times may, upon analysis, find them to form a graded series. Given the difficulty in identifying a single signal on a continuum, one finds it hard to imagine how graded sounds given in this fashion would be used in communication, unless additional cues were provided by the signaling animal. It may be that when two animals are communicating at close range, visual signals provided simultaneously or just before or after a graded vocal signal give the additional information needed for the accurate forecast of the signaler's future behavior that a respondent would require.
VISUAL AND AUDITORY SIGNALING
A relationship has been postulated between the distance over which a communication signal normally functions and its place on a continuum from gradedness to discreteness (Marler, 1973). The longer the distance, the more likely the signal is to be discrete. One would thus expect a relationship between the loudness of a given category of ape sounds and the degree of variability, but reliable measurements of neither have yet been made. One might also be prepared for the possibility that within a graded complex, modal forms might be louder than intermediates.
We have interpreted the discreteness of longdistance signals to be a compromise between information coding and the problems of accurate identification of a signal by another, at a distance, under noisy conditions. At closer range accurate identification of intermediate forms is easier to accomplish. Perhaps even more important is the supplementation of a graded sound signal at close range by visual information, and we have seen that the apes are provided with a rich array of facial expressions, gestures, and postural and morphological signals. Appropriate coordination between auditory and visual signaling should greatly enhance the accuracy with which subtle gradations in signal form, and presumably in corresponding encoded information, might be received and used by another animal.
Such speculations cannot be tested until we can determine how an animal such as a chimpanzee processes the stimuli that conspecific vocalizations provide. In recent years a contrast has been demonstrated in our own species between the continuous processing of non-speech sounds and the categorical processing of the sounds of speech, using series of synthetic speech sounds in which all parameters are under experimental control (Studdert-Kennedy, 1975). A similar approach to analysis of the perceptual processing of vocal sounds by apes is a necessary prerequisite to understanding the communicative significance of the unusual emphasis on graded vocal sounds among higher primates.
SIGNAL REPERTOIRE SIZES
No attempt has been made to estimate the sizes of the repertoire of visual signals that apes use, but we have estimated the numbers of distinct vocalizations. In all cases initial classifications were developed in the field by ear, with some subsequent refinement as a result of sound spectrographic analysis of recordings, especially in the chimpanzee. We have noted that the estimates of repertoire size, reached by several different observers, are rather similar for the ape species studied thus far (cf. Moynihan, 1970). They range between eight and sixteen. The former is Chivers's (in press) estimate of the call repertoire of the siamang. In Kloss's gibbon only ten vocalizations have been described thus far, but we estimate that a final total of about fifteen vocal categories is likely. More studies are needed of orang-utan and siamang vocalizations before we can be sure that there are no significant differences in overall vocal repertoire size among the apes. Also, it may turn out that these estimates derived from descriptive studies alone are deceptive. There may be overemphasis on louder sounds and on modal forms along graded acoustical continua.
It seems conceivable that there are limits to the number of long-distance signals a species is likely to evolve, perhaps because of the need for a discrete organization with sufficient acoustical space between categories for unequivocal identification under difficult conditions. The number of effectively different signals that a graded vocal repertoire provides for close-range communication may prove to be much greater. Analyses of the highly graded vocal repertoire of the Japanese macaque show that very small steps along some vocal continua are associated with different circumstances of vocalization, so encoding information about different situations (Green, 1975a). Thus the complete repertoire of sound signals, each having different behavioral significance to a listening Japanese macaque, may be much larger than an estimate based only on modal forms would lead us to expect.
SYMBOLISM AND SYNTAX
What is it that apes are signaling about, and can we infer that signals serve as signs or symbols? Two obvious external referents for vocalizations are food and danger. While the circumstances in which gorilla belching is given include not only food discovery but also generally pleasurable and relaxing situations, the rough grunting of the chimpanzee seems more closely tied to food alone (Fossey, 1972; van Lawick-Goodall, 1968a). Thus, in the sense of Smith (1969) one might consider that rough grunting incorporates a message about the signaler's discovery of a favored food. Moreover, subsequent events show that rough grunting also specifies readiness of the signaler to share the food, not a necessary concomitant of all food discovery. One might think of the alarm calls of chimpanzees (see Table 1, whimper, wraaa) as encoding information about the presence of danger in varying degrees, though in none of the apes is this information as specific as in the vervet monkey, where different classes of predators seem to be specified (Struhsaker, 1967; see p.51, this volume).
To establish that such signals do in fact function as signs we must examine a complete communicative interaction to determine whether another animal's behavior is changed by receipt of the signal in an appropriate fashion. Appropriateness here may be taken to imply correspondence between signal-evoked changes in the respondent's behavior and changes directly evoked by the referent in question—food for example. The hurried approach of other chimpanzees to a rough-grunting individual, often calling similarly as they come, implies a signlike function for the vocalization.
Such behavioral exchanges can often be interpreted in other ways, however. The first detectable behavioral response to many signals is a rather generalized change in behavior. Whether it is a movement oriented to the signal source— approach or withdrawal—or scanning in that direction, or something even less specific such as a change in vigilance, this initial response will in turn change the pattern of stimulation that the respondent experiences. It is often very difficult in practice to separate the role of such stimulation changes from that of initial signal reception in determining the recipient's eventual behavioral change.
Utterance of a similar call as an early response to rough grunting might seem to provide a basis for inferring that it is recognized as a food sign, since chimpanzees do not seem to give this call in any other circumstances. However, further reflection suggests that the capacity to perceive a signal and respond in kind demonstrates only an ability to identify the signal type. More information is needed to infer that it functions as a sign, just as a child must be required to do more than simply voice a written word to reveal that its symbolic meaning is understood.
Laboratory experimentation with chimpanzees taught to use languagelike systems of communication in human interchanges can overcome the uncertainties of field observation (Gardner and Gardner, 1971; Premack, 1971; Rumbaugh et al., 1974; Gill and Rumbaugh, in press). The results reveal a capacity to use artificial signals—whether hand signs, plastic wordtokens or computer keyboard symbols—as both signs and symbols. The experiments show not just a triggering of an appropriate behavioral event but also some demonstration that the underlying conception is understood (Langer, 1942; von Glasersfeld, in press; Premack, 1971). However, as Premack (1975) has indicated, "the whole topic of representational process or symbolization is in a highly unsatisfactory state and seems likely to remain so until some relevant operational criteria are provided." Research on the capacities of chimpanzees to use languagelike systems promises not only insights into the use of representational processes by animals, but also a more objective view than we have had yet of the nature and significance of such processes in our own species.
Preoccupation with animal symbolism should not lead us to neglect the social value of nonsymbolic communication to animals. In a discussion of communication by such means, Premack gives the following human analogy. He asks us to consider two ways in which we might benefit from his knowledge of present conditions at some place he has just visited and we have not.
I could return and tell you, "The apples next door are ripe." Alternatively, I could come back from next door chipper and smiling. On still another occasion I could return and tell you, "A tiger is next door." Alternatively, I could return mute with fright, disclosing an ashen face and quaking limbs. The same dichotomy could be arranged on numerous occasions. I could say, "The peaches next door are ripe," or say nothing and manifest an intermediate amount of positive affect since I am only moderately fond of peaches. Likewise, I might report, "A snake is next door," or show an intermediate amount of negative affect since I am less shaken by snakes than by tigers. [Premack, 1974:1]
With foreknowledge of his fears and appetites we can learn much of value without any symbolic communication if we know where and when the situation mirrored in his affective signaling behavior was experienced. While the distinction between affective and symbolic communication may not be a sharp one (Marler, this volume, p.54), similar arguments help us understand the communicative value of signals with nonspecific referents (a possible paraphrase of "affective signals") as contrasted with signals that symbolize a specific class of reference.
Menzel (1971, 1973, 1974) has demonstrated the efficiency with which captive chimpanzees can choose which companion to follow to hidden food, the companion having been shown the site by the experimenter beforehand. Respondents even developed skill in overtaking the leader at the last minute and reaching the food first. However, it is not always clear that the term "communication" is appropriate for such interchanges. To judge from Menzel's descriptions, respondents may derive their cues not only from observing the leader's signaling behavior but also by directly perceiving its eager efforts to regain the hidden food itself.
The relationship between the two animals Rock and Bell is illuminating in this regard (Menzel, 1974:134-35). Rock was dominant over Bell and if she led him to food he would attack her and take it as soon as she uncovered it. As tests continued, Bell became more and more devious about approaching the food when she was the experimental "leader" and Rock was present. As Menzel describes later trials,
Bell accordingly stopped uncovering the food if Rock was close. She sat on it until Rock left. Rock, however, soon learned this, and when she sat in one place for more than a few seconds, he came over shoved her aside, searched her sitting place, and got the food. Bell next stopped going all the way. Rock, however, countered by steadily expanding the area of his search through the grass near where Bell had sat. Eventually Bell sat farther and farther away, waiting until Rock looked in the opposite direction before she moved towards the food at all—and Rock in turn seemed to look away until Bell started to move somewhere. On some occasions Rock started to wander off, only to wheel around suddenly precisely as Bell was about to uncover the food. Often Rock found even carefully hidden food that was thirty feet or more from Bell, and he oriented repeatedly at Bell and adjusted his place of search appropriately if she showed any signs of moving or orienting in a given direction. If Rock got very close to the food, Bell invariably gave the game away by a "nervous" increase in movement. However on a few trials she actually started off a trial by leading the group in the opposite direction from food, and then, while Rock was engaged in his search, she doubled back rapidly and got some food. In other trials, when we hid an extra piece of food about 10 feet away from the large pile, Bell led Rock to the single piece, and while he took it she raced for the pile. When Rock started to ignore the single piece of food to keep his watch on Bell, Bell had temper "tantrums."
The actions of Bell to which Rock is responding seem in no way specialized for a communicative function, failing to satisfy what many regard as a critical criterion (Hockett, 1961). It is as appropriate to question whether Bell is "communicating" with Rock as to ask whether a mouse racing for its burrow "communicates" to an owl the appropriate trajectory for a successful attack upon it. Nevertheless, such interactions are fascinating for the perspicacity they reveal in one animal's ability to predict what another is going to do, surely a key element in the evolution of communication.
SIGNALS AND SPACING
Communicative interchange often has repercussions on the relative spacing of the participants. They may come together for a period of time, engage in some activity that requires proximity, then separate again afterward. Alternatively, the consequence of communication may be repulsion, or the maintenance of spacing that might otherwise vary. By its impact on spacing, social communication has immediate and farreaching ecological consequences, becoming in turn the focus of strong selective influences. Yet we are still ignorant of the ways in which signaling behavior controls the spatial organization of ape populations. Approaches that show promise, such as Rummer's analyses of spacing mechanisms in baboons and Waser's demonstration by field playback techniques of the role of the greycheeked mangabey whoop-gobble in intertroop spacing, have yet to be systematically applied to studies of ape behavior (Kummer, 1971; Waser, 1975). However, gibbons have been induced to approach and withdraw by tape recorded calls played back in the field (Brockelman and Kobayashi, 1971; Mackinnon and Chivers, in prep.). The inferences to be drawn from the descriptive studies reviewed here are speculative and limited in scope.
Although the apes display a broad spectrum of patterns of social organization, the most striking conclusion to emerge from this survey of signaling behavior is the degree of correspondence between species. Their vocal repertoire sizes are similar. We have seen that the chimpanzee and the gorilla, whose social systems differ strikingly, share many sound signals so intimately that one can discern homologies between a significant portion of their repertoires. As other ethologists have concluded in comparative studies of communicative behavior, so in the apes one must look to changes in the pattern of usage of signals to explain the divergent communicative organization of species that share common origins.
If we compare the patterns of social organization in chimpanzee and gorilla in terms that might relate to their systems of vocal communication, they seem to have more in common in patterns of between-group than within-group relationships. There is territoriality in both species, and adult males take a prime role in territorial defense in both the chimpanzee and the gorilla. They differ in that the chimpanzee territory is maintained by a group of males while in the gorilla the onus tends to fall on one silverbacked male. The two species also differ in the closeness of identification of adult females with a particular group and its home range and territorial commitments. However, in neither species is there evidence as yet of a female role in territorial maintenance. Thus while intergroup relationships in chimpanzee and gorilla do differ in a number of significant respects they also have major features in common.
With regard to within-group relationships, the contrast between the two species is more striking. Groups of gorillas are relatively compact and coherent. Individuals may get out of visual contact with companions feeding in dense vegetation, and vocalizations, especially belching, are used to maintain contact in this circumstance (Harcourt, pers. comm.). The belching vocalization is a dominant component in the vocal behavior of habituated gorillas. The distances involved are small, and mechanisms for reestablishing contact seem relatively direct and uncomplicated.
Chimpanzee group members are often much more dispersed. Some adult males and females spend as much as 80 to 95 percent of their time alone (Wrangham, pers. comm.). To communicate with other group members, they must often signal vocally over great distances, with companions out of sight to them. During long periods of isolation any adult may encounter predators. In rejoining group members all must be ready to engage in the signal exchanges required for reestablishment of relationships with fellow group members after long periods of visual if not vocal separation. These are occasions of high arousal and involve a variety of extended vocal and visual signaling.
Thus patterns of chimpanzee and gorilla social organization differ most strikingly in the organization of within-group relationships. The chimpanzee has a large, dispersed social group, containing several adult males, with members recombining from day to day in different sub-groupings of adult males, females, and young. They also spend much time alone. Group members are often separated by long distances. The gorilla has smaller, more coherent social groups, usually with only one fully adult male. Withingroup vocal signaling is over much shorter ranges. The rate of within-group reunions, and the level of social arousal associated with them, is much lower than in the chimpanzee. The compactness of the group is such that they confront such exigencies as predator detection, dissemination of alarm, and defense as a group rather than on an individual basis.
Given these differences in social organization in the two species, can they be correlated with any of the differences in vocal behavior? If so, can a case be made for the relationship being a causal one? Given the compact gorilla social group, with relatively stable social composition, it is possible for one sex- and age-class to assume responsibility for many of the communicative decisions required for the maintenance of withingroup social organization, defense against predators, as well as the maintenance of relations with other social groups and solitary individuals. An argument such as this makes the domination of gorilla social behavior by silverbacked males to some extent intelligible. However, although gorilla group composition and coherence permits the silverbacked male to assume the main prerogative of vocal signaling, it is by no means clear that this is a required consequence.
A significant and perhaps major proportion of chimpanzee vocal behavior is occasioned by renewed contacts between within-group social sub-units. The highly dispersed, fluid nature of chimpanzee society must favor competence to cope with major environmental and social contingencies on an individual basis in all sex and age groups other than dependent infants. The obvious exception here is territorial defense, which is a solely male responsibility, as in the gorilla. The major vocalization involved in territorial exchanges, pant-hooting, also serves other functions incorporating long-distance vocal signaling, hence its presence in the female repertoire as well. Moreover we have noted that the female form of this vocalization is discriminable from that of adult males. The other acoustical signal involved in between-community interactions, drumming, is restricted to adult males.
There is ample evidence that chimpanzee subgroupings signal vocally to one another over distances of hundreds of meters, something that probably never occurs within a gorilla group (Harcourt, pers. comm.). An obvious case is meat-eating, where loud sustained vocalizations attract distant group members to the event (van Lawick-Goodall, 1968a, 1971; Teleki, 1973b). Coordination of movements and other behaviors within the community is presumably modulated by these exchanges, in which all animals other than infants may participate on occasion, presumably to the benefit of the group as well as to themselves. The remarkable spread of the use of all vocalizations throughout the chimpanzee community membership is perhaps interpretable in these terms, in striking contrast to the domination of gorilla vocal behavior by one sex- and age-class, the silverbacked males.
We can also speculate about the adaptive significance of similarities that exist in the vocalizations of the chimpanzee and the gorilla. In both, the onus of territorial maintenance falls primarily on mature adult males. In both species, the vocalizations most commonly enjoyed in intergroup encounters, pant-series and pant-hooting, are given most frequently by the age-class responsible for territorial maintenance, mature males, even though the asymmetry in usage is much less striking in the chimpanzee than in the gorilla. The similarity of the non-vocal sounds used in such encounters, chest-beating and drumming, can only result from convergence. It remains to be determined whether such mechanical sounds have a unique advantage in this context, such as better carrying-power than vocalizations, or whether interactions between the two species at some stage of their history might have favored interspecific territoriality, and thus convergence on similar signals for intergroup spacing. If details of vocal behavior are shaped by adaptive social function with any degree of precision, and we should remember that this is by no means a proven assumption, then we might infer from the remarkable degree of correspondence in the morphology of vocalizations of the two species that the major environmental and social contingencies met in within-group behavior of the two species are similar in nature, although differing in the distribution of responsibilities through the group. Some of the quantitative differences in ranking of use of equivalent vocalizations may also be adaptive. The need to maintain greater group coherence in the gorilla perhaps explains the more frequent use and broader array of contexts of gorilla belching, as compared with the less frequently used and more restricted chimpanzee rough grunting. Belching often seems to serve foraging gorillas as a "keeping in contact" signal (Harcourt, pers. comm.). Chimpanzees have no call that functions as a general, close-range, contact signal. In spite of the differences in their social organization it is interesting to note that several equivalent vocal pairs in the two species rank rather similarly in frequency of usage.
These then are the kinds of correlations between social organization and vocal signaling that throw some light on the behavioral similarities and differences in the apes. While the data are still imperfect, we can begin to see more clearly what additional observations are desirable. Data on the intensity and range of reception of vocalizations with different functions are needed. Above all, new approaches should be sought to characterize the functions of different vocalizations, so that more subtle interspecies comparisons of the proportions of a signal repertoire devoted to different kinds of adaptive tasks may be possible. For instance we have noted that apart from the gorilla roar, the only other gorilla call that definitely seems to lack a counterpart in the chimpanzee repertoire is the growl. If an aggressive function can indeed be assigned to this call, can we infer that there is a need for a closerange within-group aggressive vocalization in the gorilla which is absent in the chimpanzee, or met in a different way? Within-group aggression is often a noisier and more highly aroused interaction in the chimpanzee, including elaborate and highly ritualized aggressive displays. Playback of recorded vocalizations applied recently to primates in the field by Waser (1974, 1975) in studies of vocal communication in mangabeys is one approach to such functional questions.
Another issue on which information is urgently needed is the distribution of kin relationships in ape societies. The extent of personal acquaintanceships is another issue. One could imagine that the chimpanzee's social system, more so than the gorilla's, requires familiar acquaintanceship to be spread farther through the population. Chimpanzee vocal behavior could be one means of establishing and maintaining this larger number of individual acquaintances. At least one call, which also happens to be the most frequently used, does bear information about individual identity. By contrast one might speculate that in gibbons, the sphere of acquaintanceship would be more limited than in the chimpanzee. However, observations of young Kloss's gibbons settling adjacent to their natal territories suggest that frequent use of loud vocalizations might serve to maintain kinship ties as well as mediating territorial competition. Thus relationships between neighboring gibbon groups may be more complex than those of the gorilla, with the social network in which an animal is involved with others on an individual basis extending beyond the family group into neighboring ones. We have noted in Kloss's gibbon the use of alarm calls so loud that they seem designed to reach farther than the limits of a typical home range. They effectively link otherwise competitive social units in a cooperative association for avoidance of predators. Knowledge of how kin are distributed in other ape populations will have a significant bearing on our understanding of how this and other signaling behavior affects social organization.
The list of important but unanswered questions about social communication in apes is a long one. Although enormous strides have been made in our knowledge about the behavior of apes, we are still far from achieving any kind of general theory about how variations in communicative behavior mediate variations in the social organization. Even our descriptions of signaling are incomplete, and we are largely ignorant of its effects on the behavior of others. The further analysis of ape communication presents challenges for which the next generation of field observers and experimenters will have to develop new methods of study and analysis if the many difficulties are to be overcome.
The difficulties in understanding communication in free-living primates are not only practical but logical. It is hard to understand a communicative system without participating fully in it oneself. In the process of disentangling the relationships between social organization and signaling behavior there is a prospect of learning more about the general principles underlying animal communication, undoubtedly more complex and subtle than is often supposed.
Aldrich-Blake, F. P. G., and Chivers, D. J., 1973. On the genesis of a group of siamang. Am. J. Phys. An-throp., 38:631-36.
Andrew, R. J., 1963. The origin and evolution of the calls and facial expressions of the primates. Behaviour, 20:1-109.
Baldwin, L. A., and Teleki, G., 1976. Patterns of gibbon behavior on Hall's Island, Bermuda. Gibbon and Siamang, 4:21-105.
Baylis, J., in press. A quantitative study of long-term courtship: II. The dynamics of courtship in two new world cichlids. Behaviour.
Berkson, G.; Ross, B. A.; and Jatinandana,S.; 1971. The social behavior of gibbons in relation to a conservation program. Primate Behavior, 2:225-55.
Boutan, L., 1913. Le pseudo-language, observations effectuées sur un anthropoide: le Gibbon (Hylobâtes leucogenys Ogilby). Act. Soc. Linn. Bordeaux, 47:5-81.
Brockelman, W. Y., and Kobayashi, N. K., 1971. Live capture of free-ranging primates with a blowgun./. Wildl. Mgmt., 35:852-55.
Brockelman, W. Y.; Ross, B. A.; and Pantuwatana, S.; 1974. Social interactions of adult gibbons (Hylobates lar) in an experimental colony. Gibbon and Siamang, 3:137-56.
Carpenter, C. R., 1964. Naturalistic Behavior of Nonhuman Primates. University Park, Pa.: Pennsylvania State University Press.
Chasen, F. N., and Kloss, C. B., 1927. Spolia Mentawiensia, mammals. Proc. Zool. Soc. Lond., pp.797-840.
Chevalier-Skolnikoff, S., 1974. Facial expression of emotion in nonhuman primates. In: Darwin and Facial Expression, P. Ekman, ed. New York: Academic Press.
Chivers, D. J., 1971. Spatial relations within the siamang group. Proc. 3rd Int. Cong. Primat., Zurich, 3:14-21.
Chivers, D. J., 1972. The siamang and the gibbon in the Malay Peninsula. Gibbon and Siamang, 1:103-35.
Chivers, D. J., 1974. The Siamang in Malaya: A Field Study of a Primate in Tropical Rain Forest, H. Hofer and A. H. Schultz, eds. Contributions to Primatology, vol. 4. Basel: Karger.
Chivers, D. J., in press. Communication within and between siamang family groups. Behaviour.
Eisenberg, J. F., 1966. The social organization of mammals. Handbuch der Zoologie, Band 8, Leiferung 39, 10(7): 1-92.
Ellefson, J. O., 1967. A natural history of gibbons in the Malay Peninsula. Ph.D. diss., University of California, Berkeley.
Ellefson, J. O., 1974. A natural history of whitehanded gibbons in the Malayan Peninsula. Gibbon and Siamang, 3:1-136.
Fisler, G., 1969. Mammalian organizational systems. Contrib. Sei. Los Angeles Co. Mus. Nat. Hist., no. 167.
Fooden, J., 1969. Color-phase in gibbons. Evolution, 23:627-44.
Fossey, D., 1972. Vocalizations of the mountain gorilla (Gorilla gorilla beringei). Anim. Behav., 20:36-53.
Fossey, D., 1974. Observations on the home range of one group of mountain gorillas (Gorilla gorilla beringei). Anim. Behav., 22:568-81.
Galdikas-Brindamour, B., 1975. People of the forest. Nat. Geogr. Mag., 148:444-73.
Gardner, B. T., and Gardner, R. A., 1971. Two-way communication with an infant chimpanzee. In: Behavior of Nonhuman Primates, A. M. Schrier and F. Stollnitz, eds. New York: Academic Press.
Gautier, J. P., 1971. Etude morphologique et fonctionnelle des annexes extra-laryngées des Cercopithecinae; liaison avec les cris d'éspacement. Biol. gabon., 7:229-67.
Gill, T. V., and Rumbaugh, D. M., in press. Mastery of naming skills by a chimpanzee. J. of Human Evol.
Glasersfeld, E. von, in press. Signs, communication and language. J. of Human Evol.
Green, S., 1975a. Variation of vocal pattern with social situation in the Japanese monkey (Macaca fuscata J: a field study. In: Primate Behavior: Developments in Field and Laboratory Research, vol. 4, L. A. Rosenblum, ed. New York: Academic Press.
Green, S., 1975b. Dialects in Japanese monkeys: vocal learning and cultural transmission of locale-specific vocal behavior. Z. Tierpsychol., 38:304-14.
Harrisson, B., 1960. A study of orang-utan behaviour in semi-wild state, 1956-1960. Sarawak Mus. J., 9:422-47.
Hess, J. P., 1973. Some observations on the sexual behaviour of captive lowland gorillas, Gorilla g. gorilla (Savage and Wyman). In: Comparative Ecology and Behaviour of Primates, R. P. Michael and J. H. Crook, eds. New York: Academic Press.
Hockett, C., 1961. Logical considerations in the study of animal communication. In: Animal Sounds and Communication, W. E. Lanyon and W. N. Tavolga, eds. Washington, D.C.: Am. Inst. Biol. Sei. Publ. no. 7.
Horr, D. A., 1972. The Borneo orang-utan. Population structure and dynamics in relationship to ecology and reproductive strategy. Unpubl. paper read to Am. Assoc. Phys. Anthrop.
Huber, E., 1931. Evolution of Facial Musculature and Facial Expression. Baltimore: Johns Hopkins Press.
Kummer, H., 1971. Spacing mechanisms in social behavior. In: Man and Beast: Comparative Social Behavior, J. F. Eisenberg and W. S. Dillon, eds. Washington, D.C.: Smithsonian Institution Press.
Lamprecht, J., 1970. Duettgesang beim Siamang, Symphalangia syndactylus (Hominoidea, Hylobatinae). Z. Tierpsychol., 27:186-204.
Langer, S., 1942. Philosophy in a New Key. New York: Mentor Books.
Lieberman, P., 1968. Primate vocalizations and human linguistic ability. Acoust. Soc. Amer., 44:1574-84.
Lieberman, P.; Crelin, E. S.; and Klatt, D. H.; 1972. Phonetic ability and related anatomy of the newborn and adult human, Neandertal man, and the chimpanzee. Amer. Anthropologist, 74:287-307.
Mackinnon, J., 1974. The behaviour and ecology of wild orang-utans (Pongo pygmaeus). Anim. Behav., 22:3-74.
Mackinnon, J., 1975. In Search ofthe Red Ape. New York: Ballantine Books.
Marier, P., 1955. Characteristics of some animal calls. Nature (London), 176:6-7.
Marler, P., 1965. Communication in monkeys and apes. In: Primate Behavior, I. DeVore, ed. New York: Holt, Rinehart Sc Winston.
Marler, P., 1968. Aggregation and dispersal: two functions in primate communication. In: Primates. Studies in Adaptation and Variability, P. C. Jay, ed. New York: Holt, Rinehart & Winston.
Marler, P., 1969. Vocalizations of wild chimpanzees, an introduction. In: Proc. 2nd Int Cong. Primat., Atlanta, 1968, 1:94-100. Basel: Karger.
Marler, P., 1973. A comparison of vocalizations of red-tailed monkeys and blue monkeys: Cercopithecus ascanius and C. mitis in Uganda. Z. Tierpsychol 33:223-47.
Marler, P., 1976. Social organization, communication and graded signals: vocal behavior of the chimpanzee and the gorilla. In: Growing Points in Ethology, P. Bateson and R. A. Hinde, eds. Cambridge: Cambridge University Press.
Marler, P., and Hobbet, L., 1975. Individuality in a long-range vocalization of wild chimpanzees. Z. Tierpsychol., 38:97-109.
Marler, P., and van Lawick-Goodall, J., 1971. Vocalizations of Wild Chimpanzees (sound film). New York: Rockefeller University Film Service.
Marler, P., and Mundinger, P., 1971. Vocal learning in birds. In: Ontogeny of Vertebrate Behavior, H. Moltz, ed. New York: Academic Press.
Marshall, J. T., Jr.; Ross, B. A.; Chantharojvong, S.; 1972. The species of gibbons in Thailand. / Mammal., 53:479-86.
Menzel, E. W., 1971. Communication about the environment in a group of young chimpanzees. Folia Primat., 15:220-32.
Menzel, E. W., 1973. Leadership and communication in young chimpanzees. In: Precultural Primate Behavior, E. W. Menzel, ed. Symp. 4th Int. Cong. Primat. 1:192-225. Basel: Karger.
Menzel, E. W., 1974. A group of young chimpanzees in a one-acre field. In: Behavior of Nonhuman Primates, A. M. Schrier and F. Stollnitz, eds. New York: Academic Press.
Michael, R. P., 1969. The role of pheromones in the communication of primate behaviour. In: Proc. 2nd Int. Cong. Primat., Atlanta, 1968. 1:101-107. Basel: Karger.
Miller, G. S., 1903. Seventy new Malayan mammals. Smithsonian Misc. Colins., 45:1-73.
Moynihan, M., 1970. Control, suppression, decay, disappearance and displacement of displays./ Theoret. Biol., 29:85-112.
Negus, V. E., 1949. The Comparative Anatomy and Physiology of the Larynx. New York: Grune and Stratton.
Nishida, T., and Kawanaka, K., 1972. Inter-unit-group relationships among wild chimpanzees of the Mahali mountains. Kyoto Univ. Afr. Stud., 8:131-69.
Premack, D., 1971. On the assessment of language competence in the chimpanzee. In: Behavior of Nonhuman Primates, A. M. Schrier and F. Stollnitz, eds. New York: Academic Press.
Premack, D., 1974. Concordant preferences as a precondition for affective but not for symbolic communication (or how to do experimental anthropology). In: Experimental Behaviour: A Basis for the Study of Mental Disturbance, J. H. Cullen, ed. New York: John Wiley and Sons.
Premack, D., 1975. Symbols inside and outside of language. In: The Role of Speech in Language, J. F. Kavanagh and J. E. Cutting, eds. Cambridge: MIT Press.
Reynolds, V., and Reynolds, F., 1965. Chimpanzees in the Budongo Forest. In: Primate Behavior: Field Studies of Monkeys and Apes, I. DeVore, ed. New York: Holt, Rinehart & Winston.
Rodman, P. S., 1973. Population composition and adaptive organisation among orang-utans of the Kutai Reserve. In: Comparative Ecology and Behaviour of Primates, R. P. Michael andj. H. Crook, eds. New York: Academic Press.
Rumbaugh, D. M.; von Glasersfeld, E.; Warner, H.; Pisani, P.; and Gill, T. V.; 1974. Lana (chimpanzee) learning language: a progress report. Brain and Language, 1:205-212.
Schaller, G. B., 1963. The Mountain Gorilla. Chicago: University of Chicago Press.
Schultz, A. H., 1932. The generic position of Symphalangia klossii. J. Mammal. 13:368-69.
Simpson, M.J. A., 1973. The social grooming of male chimpanzees. In: Comparative Ecology and Behaviour of Primates, R. P. Michael and J. H. Crook, eds. New York: Academic Press.
Smith, J. M., 1965. The evolution of alarm calls. Amer. Natural., 99:59-63.
Smith, W. J., 1969. Messages of vertebrate communication. Science, 165:145-50.
Stebbins, W. C., 1971. Hearing. In: Behavior of Nonhuman Primates, A. M. Schrier and F. Stollnitz, eds. New York: Academic Press.
Struhsaker, T., 1967. Auditory communication among vervet monkeys (Cercopithecus aethiops).In: Social Communication Among Primates,S. A. Altmann, ed. Chicago: University of Chicago Press.
Studdert-Kennedy, M., 1975. Speech perception. In:Contemporary Issues in Experimental Phonetics, N. J. Lass.ed. Springfield, Illinois: Thomas.
Teleki, G., 1973a. Group response to the accidental death of a chimpanzee in Gombe National Park, Tanzania. Folia Primat., 20:81-94.
Teleki, G., 1973b. The predatory behavior of wild chimpanzees. Lewisburg, Pa., Bucknell University Press.
Tenaza, R. R., 1975. Territory and monogamy among Kloss' gibbons (Hylobates klossii) in Siberut Island, Indonesia. Folia Primat., 24:60-80.
Tenaza, R. R., 1976. Songs and related behavior of Kloss' gibbons (Hylobates klossii) in Siberut Island, Indonesia. Z. Tierpsychol.40:37-52.
Terra, H. de, 1943. Pleistocene geology and early man in Java. Trans. Am. Phil. Soc., n.s., 32:437-64.
van Hooff, J. A. R. A. M., 1962. Facial expressions in higher primates. Symp. Zool. Soc. Lond., 8:97-125.
van Hooff, J. A. R. A. M., 1967. The facial displays of the Catarrhine monkeys and apes. In: Primate Ethology, D. Morris, ed. London: Weidenfeld and Nicolson.
van Hooff, J. A. R. A. M., 1970. A component analysis of the structure of the social behaviour of asemicaptive chimpanzee group. Experientia, 26:549-50.
van Hooff, J. A. R. A. M., 1971. A structural analysis of the social behaviour of a semi-captive group of chimpanzees. In: Aspecten van het sociale Gedrag en de Communicatie bij Humane en Hogere Niethumane Primaten. Diss., University of Utrecht, pp.11-127.
van Lawick-Goodall, J., 1968a. A preliminary report on expressive movements and communication in the Gombe Stream chimpanzees. In: Primates. Studies in Adaptation and Variability, P. C. Jay, ed. New York: Holt, Rinehart8c Winston.
van Lawick-Goodall, J., 1968b. The behaviour of freeliving chimpanzees in the Gombe Stream Reserve. Anim. Behav. Monog., 1:161-311.
van Lawick-Goodall, J., 1971. In the Shadow of Man.London: Collins.
von Glasersfeld, E., in press. Signs, communication and language. J. of Human Evol.
Waser, P., 1974. Intergroup interactions in a forest monkey: the mangabey Cercocebus albigena. Ph.D. diss., The Rockefeller University.
Waser, P. M., 1975. Experimental playbacks show vocal mediation of intergroup avoidance in a forest monkey. Nature, 255:56-58.
Wickler, W., 1967. Socio-sexual signals and their intraspecific imitation among primates. In: PrimateEthology, D. Morris, ed. London: Weidenfeld and Nicholson.
Wrangham, R. W. (1974). Artificial feeding of chimpanzees and baboons in their natural habitat. Animal Behaviour, 22:83-93.
Wrangham, R. W. (1975). Behavioral ecology of chimpanzees in Gombe National Park, Tanzania. Ph.D. thesis. University of Cambridge.