Social Play in the American Black Bear: Its Similarity to Canid Social Play and an Examination of its Identifying Characteristics

Transcription

1 AMER. ZOOI.., 14: (1974). Social Play in the American Black Bear: Its Similarity to Canid Social Play and an Examination of its Identifying Characteristics J. D. HENRY AND S. M. HERRERO Department of Biology and Faculty of Environmental Design, The University of Calgary, Calgary, Alberta, Canada, T2N 1N4 SYNOPSIS. Bears evolved from a canid stock at quite a recent date (early Miocene). j-'cspitc tuis recent origin, ucsrs show substantial morphological, pliysiologicai, and ecological differences when compared to modern day canids. However, the display behaviors of Canidae and Ursidae have remained remarkably similar. In this paper, the motor patterns of black bear social play are described in detail. Numerous similarities between canid and ursid social play are pointed out. Agonistic displays common to both families are also pointed out. These behavioral similarities support the principle that social behavior, particularly display behavior, will frequently be conservative in its evolution as compared to the evolution of morphology, anatomy, or ecological adaptations. Beach (1945) stressed the importance of identifying and testing the general characteristics of play. A large number of characteristics have been suggested as being diagnostic of play, but these characteristics have received very little testing. Five characteristics of social play were tested in this study, and two were found to be only partially valid for black bear social play. Extensive testing of the general characteristics of play on as wide a range of species as possible is definitely recommended for future research. INTRODUCTION Bears evolved from a canid lineage at quite a recent date. The ursid line is first identified in the fossil record during the lower Miocene, and the modern day genus Ursus did not appear until the upper Pliocene (Herrero, 1972). Colbert (1955) states that, along with Hyaenidae, the family Ursidae is the youngest and most recent family of carnivores to differentiate. Despite this recent origin, Ursidae exhibits many differences when compared with modern day canids. For example, regarding morphological differences, bears show only a rudiment of the large canid tail. Bears also exhibit a plantigrade condition and have evolved a bunodont dentition (Davis, 1964). Even the brain morphologies of Canidae and Ursidae This research was supported by N. R. C. Grant # , by the Killam Foundation, and by the Environmental Sciences Centre (Kananaskis). The co-operation of the National and Historic Parks Branch of Canada and the Calgary Zoo made this research possible. We gratefully acknowledge the financial support and co-operation received from these organizations. 371 show substantial differences (see Papez, 1929, and Davis, 1964, for reviews). For example, Ursidae manifests expansion of certain areas in the cerebral cortex which appears to be related to the enhanced manipulative abilities of the ursid forelimbs (Davis, 1949, 1964). Regarding ecological differences, many ursid species have adopted an omnivorous niche. And in cold climates, bears have evolved winter lethargy. Folk et al. (1972) and Nelson et al. (1973) review the physiological adaptations that allow bears to achieve this state of winter dormancy. Despite these morphological, ecological, and physiological differences, the social displays of Canidae and Ursidae have remained remarkably similar. This is one of the main hypotheses examined in this research. It is examined first by comparing the social play of black bears with that of canids. A subsequent article will look at agonistic behavior. If the social displays of Canidae and Ursidae are indeed similar, it illustrates the principle that social behavior can be rather conservative in its evolution when compared with the evolution of morphology,

2 372 J. D. HENRY AND S. M. HERRERO anatomy, or ecological adaptations. Hofer (1972) and Geist (1974) discuss this principle. Hofer applies it to evolution among the primates. The preceding statements assume that behavioral homologies exist between Canidae and Ursidae. Criteria indicating morphological homology were outlined by Remaine and applied to behavioral homology, first, by Baerends (1958) and, more recently, by Wickler (1967a). The conclusion that social behavioral homologies exist between Canidae and Ursidae has been made because several of these criteria can be satisfied. First, homology is indicated because of the criterion of relative position: namely, many motor patterns occupy the same relative position in the temporal sequencing of canid and ursid social behaviors. Secondly, the criterion of specific quality is satisfied: There are extensive similarities in the external appearances of motor patterns and displays in canid and ursid social behaviors. Finally, the fact that the same motor patterns and displays occur in a large number of closely related canid and ursid species indicates that these motor patterns are homologous. Numerous examples illustrating the above three criteria will be given in the descriptions of black bear social play. Beach (1945) states that a definition of play behavior must be based on its objective characteristics which, when combined, set off play behavior from non-play behavior. Bekoff (1972) has offered an operational definition of social play that is, in fact, based on several of play's suggested identifying characteristics (see also Bekoff, 1974). His definition of social play is used throughout this study. A large number of charafcteristics have been suggested as being diagnostic for solitary and social play (see Meyer- Holzapfel, 1956; Tembrock, 1958; Marler, 1966; Loizos, 1967), but these characteristics have received very little testing (Muller- Schwarze, 1968, 1971). It is imperative that objective descriptions of play exist from as wide a range of species as possible and that the suggested identifying characteristics of play be tested on this wide range of species. If play is a valid, unitary category of mammalian behavior, this testing should lead to an objective and verifiable definition of play behavior. The purpose of this paper, then, is twofold: first, to describe the social play behavior of black bears as part of an effort to describe the social behavioral repertoire of this species; and, second, to examine which of the suggested characteristics of social play are valid for black bear social play. METHODS Black bear social play was studied principally by observing black bear cubs (Ursus arnericanus) under captivity conditions. At the Calgary Zoo, three cubs (two females, one male) were raised together from approximately 3 months of age in a 12- x 12- yard arena. As a consequence of feeding and cage cleaning, the cubs received approximately li/2 hr of close contact with humans per day. All three cubs appeared healthy and showed considerable growth over the summer, but the male cub always remained somewhat smaller than the two females. The male cub did not, however, permanently occupy the lowest position in the hierarchy. Observations on social play began after the bears had been housed together for over a month. From mid-june to the end of July 1973 the cubs were observed and a total of 508 social play sequences were filmed using 8 mm cine film. For each filmed sequence, pertinent notes were also taken regarding the exact beginning and end of the sequence, the precise location of vocalizations on the film, etc. In slow motion and occasionally frame by frame, these films were analyzed for social play motor patterns, and the films were tested for the identifying characteristics of black bear social play. Additionally, for several years, freeranging black bears have been observed in several Canadian national parks. Bears were frequently observed at or around sanitary landfills where they congregated to feed on refuse, but also were observed occasionally in the backcountry, a number of miles away from these landfills. Black bears were observed during the summer and fall of 1968

3 SOCIAL PLAY IN THE AMERICAN BLACK BEAR 373 at the Jasper Landfill in Jasper National Park, Alberta, and during the summer and fall of 1971 and 1972 at the Waskesiu Landfill in Prince Albert National Park, Saskatchewan. Aggressive behavior as well as social play were thoroughly studied at these times. Seventy-two social play sequences were filmed in the cub, yearling, and subadult black bears that visited these landfills, while numerous other social play sequences were described, but not filmed. These data from free-ranging black bears are used to determine whether the descriptions and conclusions made about the captive cubs appear to be valid for the behavior of freeranging black bears. THE CLASSIFICATION OF SOCIAL PLAY Free-ranging black bears, from cubs as young as 4 months old up to sub-adults estimated to be 4 years old, commonly exhibited social play. Social play usually occurred in one of three social contexts: (i) among litter mates, (ii) between a mother and her offspring, and (ill) among sibling or non-sibling sub-adults that have separated from their mothers (see also Meyer- Holzapfel, 1957). The entire play repertoire of the black bear can be classified according to the following system (after Tembrock, 1958): I. Solitary Play: Play with an inanimate object or body part, (see Leyhausen, 1949; Meyer-Holzapfel, 1957). TI. Social Play: Play with a social partner. A. Intraspecific Social Play: The social partner is a conspecific. 1. Play-fighting: Play that involves physical contact and consists primarily of agonistic-like motor patterns. 2. Locomotory Play: Play that consists mainly of chasing or climbing after another bear and involves little physical contact. Motor patterns resemble agonistic motor patterns. 3. Sexual Play: Play that involves physical contact and consists primarily of sexual-like motor patterns. B. Interspecific Social Play: The play partner is from a different species. This paper analyzes intraspecifk social play of the black bear which, by definition, implies an interaction between two or more black bears. An interaction is defined as any behavioral situation where two or more black bears react with each other in such a way as to disrupt their ongoing patterns of moving, feeding, or resting (Stonorov and Stokes, 1972). Social play was analyzed in terms of sequences. A sequence is simply an interaction that takes place over time (seconds or minutes). Social play among black bears usually occurs in discrete sequences, that is, the bears approach, interact with each other, and then usually separate bringing the interaction to an end. Sequences are made up of a series of motor patterns. A motor pattern is defined as a distinct and frequently repeated spatiotemporal pattern of muscular contractions (Hinde, 1970). The sequences of black bear social play showed an intensity gradient. This gradient was particularly conspicuous in play fighting. We tried only to divide this intensity gradient approximately in half and to classify sequences into either high or low intensity. We did so according to the following criterion: if over 50% of the motor patterns, as measured in seconds, were judged to be performed in a high intensity manner (fast running, vigorous clawing actions, intense biting actions, etc.), the play sequence was classified as a highintensity sequence. Otherwise, it was classified as a low-intensity sequence. The 508 social play sequences from the captive cubs are classified in Table 1. The TABLE 1. Quantitative classification of the 508 social play sequences from the captive cubs. In the third column, the average time duration and its standard deviation are listed for each type of social play. Play-fighting low intensity high intensity Locomotory play Sexual play Totals % Time duration n occurrence (X ± SD) % 14% 2% 100% sec 18.2 ±13.1 sec sec sec

4 374 J. D. HENRY AND S. M. HERRERO table shows that black bear intraspecific social play is predominantly play-fighting (84% of the sequences). Only 14% of the play sequences were locomotory play, and only 2% were sexual play. These findings correlate well with observations from freeranging bears. Among free-ranging cubs and sub-adults, social play was predominantly play fighting, while locomotory play and sexual play were only occasionally observed. Social play between a sow and her cub usually took the form of low intensity play fighting. One way to test the adequacy of a behavioral classification system is to measure the amount of overlap between categories of that system. 86% of the filmed social play sequences showed behavior from only one category of play (either play-fighting, locomotory play, or sexual play). Only 14% of the sequences showed behavior that combined two different types of social play. Of the four different combinations possible, only two combinations appeared: locomotory play was combined with play-fighting, and sexual play was combined with playfighting. In Table 1, these combination sequences were classified according to which type of play predominated in the sequence. We conclude that play-fighting is by far the predominant form of social play among black bears and that the other two forms of social play (locomotory play and sexual play) frequently have at least some playfighting behavior associated with them. THE EAR POSTURES, FACIAL EXPRESSIONS, AND VOCALIZATIONS OF SOCIAL PLAY Ear postures The pinnas of the black bear are large HG. 1. The smaller cub shows the relaxed openmouth face and partially flattened ears. The larger cub shows the alert face and frontal alert ears,

5 SOCIAL PLAY IN THE AMERICAN BLACK BEAR 375 FIG. 2. One cub exhibits crescent ears. The other cub is face-pawing. and mobile. The pinnas appear especially large in young black bears. Their conspicuity appears to be due to differential growth. Using analysis of variance, Sauer (1966) examined measurements from 129 black bears of various ages and found that the pinnas attain adult size when the black bear is only 1 or 2 years old. As with many canid species, the pinnas of the black bear appear to be important signals for social communication (Burghardt, 1972). Five distinct ear postures are recognized in black bear social communications. All five ear postures are observed during social play, and a description of each follows. Frontal alert: The pinnas stand erect on the head and face forward (Fig. 1). Laterad: The pinnas face laterad, halfway between erect and flat on the neck. Laterad appears to be the relaxed ear posture. Crescent posture: The pinnas face ventrad or nearly so, and stand perpendicularly out from the side of the head (Fig. 2). This ear posture can be readily identified by the crescent-like curve formed by the bear's pinnas and forehead. In young black bears, the crescent ear posture may have an exaggerated quality due to the special largeness of the pinnas. The crescent ear posture is observed during active submission, during interactions between a sow and her cub, and during social play. Flattened posture: During aggressive behavior, the black bear flattens its ears on its neck, and the pinnas disappear completely from view. Flattening the ears appears to be a strong sign stimuli in black bear social communication. For example, in social play if the physical contact becomes intense, one of the bears may flatten its ears, and this frequently leads to a quick termination of the social play sequence.

6 376 J. D. HENRY AND S. M. HERRERO FIG. 3. The darker subadult rears bipedally and shows the puckered-lip face. The other subadult ex- Partially flattened posture: The pinnas are partially flattened on the neck but are still somewhat visible (Figs. 1, 3). Specifically, the pinnas are flattened to a point so that an imaginary straight line can be drawn from the nose through the eye of the bear and along the exterior surface of the pinna (Fig. 1). If the pinnas are flattened below this imaginary line, the ear posture is classified as flat. Above this line, the ears are classified as laterad. In the black bear, the partially flat posture is seen during flight behavior and also during social play. During social play, this ear posture can be interpreted as a somewhat defensive, but play-permitting signal. Since flattening the ears tended to disrupt the social play sequence whereas partially flattened ears did not, these two ear postures are interpreted as distinct, although somewhat related, sig- \ A-J hibits partially flattened ears and the relaxed openmouth face. nals (see also Schenkel, 1947; Fox, 1970). It is important to realize that these five ear postures graded and changed into one another, and that it was difficult to determine what was communicated by the many intermediate ear postures observed. We did, however, examine the hypothesis that the ears were clearly crescent or clearly partially flat at the beginning of the social play sequence. This hypothesis was supported by 88% of the high intensity social play sequences. The bear that initiated social play usually clearly exhibited crescent ears. Its partner usually clearly exhibited partially flattened ears. Social communications by means of ear postures is further supported by the exceptions during high intensity social play sequences. In 23 sequences, the bear being approached either flattened its ears or

7 SOCIAL PLAY IN THE AMERICAN BLACK BEAR 377 showed frontal alert ears. The other bear, instead of initiating social play, pivoted and fled in all but two of these sequences. Thus, from the ear postures alone we could successfully predict whether or not the play sequence would be initiated. The ears of both bears were most clearly crescent or partially flat at the beginning of high intensity social play sequences. However, black bears tend to exhibit these two ear postures, although in more variable forms, continuously during all social play. These two ear postures, in conjunction with other characteristics, are useful identifying characteristics of black bear social play. Facial expressions Lorenz (1953) suggests that because bears are solitary animals and have a thick skin covering their head and face that they strike without warning, that is, without first laying back their ears or giving an agonistic facial expression. Our observations, however, suggest that black bears almost always lay their ears back before attacking and that facial expressions appear to be of great importance for social communication (see Meyer-Holzapfel, 1957). There appear to be seven distinct facial expressions in the black bear. Bolwig (1964) and Fox (1970) have emphasized the similarity of facial expressions in primate and canids, and ursids also show many of these similarities. Because of these similarities, we adapted our terminology and method of study from van Hooff (1967) and Fox (1970) whenever possible. In the black bear, there are three closedmouth facial expressions which are listed in order of increasing alertness: the relaxed face, the alert face (Fig. 1), and the puckered-lip face (Fig. 3). All three are frequently seen during social play. In addition, there are four open-mouth facial expressions that are also listed in terms of increasing alertness: the relaxed openmouth face (Fig. 1), the jaw-snapping face, the jaw-gape face, and the biting face. Of these four, only the relaxed open-mouth face is commonly seen during social play. The other three are observed during aggressive behavior. Ear postures in the black bear appear to be relatively independent of facial expressions. For example, the puckered-lip face is associated with frontal alert ears during social investigation, crescent or partially flat ears during social play, and flat ears during serious aggression. Judged by the reactions of the bear receiving the signals, the same facial expression combined with different ear postures appears to have a somewhat different message content. A detailed description of the puckered-lip face and the relaxed open-mouth face follows. 1 The puckered-lip face (Fig. 3): The eyelids in this facial expression are usually wide open. The skin on the forehead is flexed and a furrow sometimes appears. The eyebrow regions are usually raised. This elevation of the eyebrow regions may be emphasized by the brown eye patches that are located above and mediad to the eyes (see Fig. 2). Cubs, yearlings, and sub-adults usually have these brown eye patches, although there are some exceptions; these eye patches frequently disappear in adult boars. The bear with a puckered-lip face alternates between a quick short stare at the opponent and a partial looking away. In this facial expression, the mouth is held tightly shut. The upper lip and the corners of the mouth are puckered out. This puckering is especially conspicuous in front (see also Meyer- Holzapfel, 1957; Burghardt, 1972). The anterior black color of the upper lip may help to emphasize this puckering (see Fig. 3). Vocalizations usually do not occur with this facial expression, although moaning is sometimes associated with it in cubs. The puckered-lip face occurs in several different social contexts. It is commonly observed during social investigation, during active submission (Schenkel, 1967), during the terminal phase of agonistic encounters, and during the initiation and termination of social play. During social play, the puckered-lip face appears in a low intensity form, that is, the forehead is usually not furrowed, the eyebrow regions are fre- 1 In a subsequent article, the other five facial expressions will be described, and a diagram illustrating all the facial expressions and ear postures of the black bear will be provided.

8 378 J. D. HENRY AND S. M. HERRERO quently not raised, and the eyelids may appear partially closed. The relaxed open-mouth face (Figs. 1, 3): This facial expression is related to the aggressive jaw-gape face, but the two are distinct because of subtle differences in the eyelids, forehead, and ears. In the relaxed open-mouth face, the eyelids are partially closed, and the forehead and eyebrow ridges are relaxed. The ears are either crescent or partially flat. The mouth is held open from 20 to 60, and the corners of the mouth are slightly elevated. The upper lip is puckered out but not as extensively as in the puckered-lip face. The lower lip is also puckered out and, along its sides, hangs down away from the teeth. The lower canines are usually clearly visible. No vocalizations except for breathing sounds are normally associated with this facial expression. The relaxed open-mouth face is usually observed during social play in the following context: Prior to biting, one bear assumes this facial expression and stares directly at the opponent. This is quickly followed by reaching to bite at the opponent, or by head-ducking as the opponent bites or paws first. Occasionally a bear will assume this facial expression and maintain it for up to 30 sec without biting. During this time the bears paw and play-wrestle with each other. The aggressive jaw-gape differs from the relaxed open-mouth face in a number of subtle ways. In both, the mouth is held open 20 to 60. In the jaw-gape face, however, the eyelids are wide open, instead of partially open; the eyebrow regions are raised instead of relaxed; and the ears are flat on the neck, instead of crescent or partially flat. The jaw-gape face is usually associated with the aggressive body posture of arching the back and strongly flexing the head and neck down. This challenge posture is rarely observed during social play. Thus, these two facial expressions can be distinguished on the basis of subtle differences in their external appearances. The relaxed open-mouth face is only observed during play behavior and can be considered a "play face" in the black bear. In the black bear this "play face" appears more closely associated with biting actions and biting intention movements than it does in many canid and primate species (see Bolwig, 1964; van Hooff, 1967; Fox, 1970). The "play face" in bears is not observed during locomotory play nor is it usually observed during the play-soliciting approach (see below). In many canid and primate species, the "play face" is observed during both of these contexts. Vocalizations Play in the black bear is distinguished by its lack of vocalizations, that is, during solitary and social play no vocalizations are given except for respiratory sounds, such as panting and breathing sounds. Serious agonistic behaviour is the antithesis of social play in that it is marked by numerous and frequent vocalizations (Darwin, 1872). Burghardt (1972) also observed that during black bear play silence was the rule and stated that "this distinguishing characteristic of true play was virtually without exception." Silence during play-fighting was observed in captive sloth bears (Mclursus ursinus), Himalayan black bears (Selenarctos thibetanus), and polar bears (Ursus maritimus) at the Calgary Zoo. Free-ranging sub-adult grizzlies (Ursus arctos) in Banff National Park were also observed to be silent during play-fighting. The rule of silence may differentiate the social play of Ursidae from that of Canidae. Bekoff (1972) reports that in the domestic dog play-soliciting gestures, such as the play bow, are often accompanied by an attention-getting bark. Although social play is silent, it may be terminated by one of the bears moaning. When physical contact in social play became intense, one bear frequently oriented towards the other bear and moaned. Moaning was followed by a quick termination of social play in 22 out of the 34 sequences in which moaning occurred. In those sequences where contact play continued, the moaning bear frequently flattened its ears and aggressively attacked its social partner. In conclusion, social play in the black bear can be identified by two constant accompanying features: (i) no vocalizations are given except for respiratory sounds, and

9 SOCIAL PLAY IN THE AMERICAN BLACK BEAR 379 (ii) the ears are continuously either in the crescent or partially flattened ear postures. No other type of black bear social behavior that we have observed shows these two features continuously during the behavioral sequence. THE MOTOR PATTERNS OF SOCIAL PLAY Table 2 enumerates 30 different motor patterns that are frequently observed during black bear social play. Twenty-two of these motor patterns are similar to motor patterns observed in various canids (Tembrock, 1958; Fox, 1969, 1970, 1971; Bekoff, 1972). This fact illustrates the extensive similarity between canid and ursid social behaviors. This section describes the motor patterns characteristic of locomotory play, play fighting, and sexual play in that order. Motor patterns of locomotory play In the black bear, there are two types of locomotory play. The first kind involves chasing after a conspecific. The second type involves seizing an object. Chasing usually begins by one animal walking up to another animal from behind and play nipping that animal. The play nip is a quick, painful bite usually delivered to the flank, rump, or hindlimb of the opponent (see Table 3). The attacking bear immediately flees. The bear that was bitten usually pivots around, obviously aroused by the bite and makes high intensity biting or clawing actions at the fleeing bear. About 25% of the time, it chases after the attacker. Alternately, chasing was initiated by one bear running at a second bear who fled before physical contact was made. The first bear then simply continued to chase after the fleeing bear. During chasing behavior, both bears exhibit the alert or puckered-lip face. The fleeing bear usually has partially flat ears while the chasing bear usually has crescent ears. Chasing sequences among free-ranging sub-adults have been observed to last for over 2 min. Infrequently, a climbing chase was observed while both bears were in a tree. TABLE 2. The motor patterns of black bear social play. Motor patterns of play-fighting: Playful approach Rearing Head butting * Pawing * Face-pawing * Lunging * Clawing Inhibited clawing Hind-leg-clawing Rolling over Play bites Biting intention movement Inhibited bite Neck-bite-hold Head shake Muzzle seizure Standing over Looking away Licking-cut-off Head and neck extension Motor patterns of locomotory play: Play nip Running chase Climbing chase Seizing the object Motor patterns of sexual play: Vulva stimulation Shoulder-scruff-bite Clasping with forepaws Pelvic thrusts Mounting Inguinal presentation * Indicates that motor pattern is also observed in canid species (see text). The second type of locomotory play involves seizing an object. This type of locomotory play has been described in several other carnivore species. Chrisler (1958) described it among young semi-captive wolves, and Schaller (1972) described it between a lioness and two cubs. In the captive bear cubs, the object seized was usually a new object in the arena. It was usually chewable, but not edible, for example, a bone, a branch, or a feather. Specifically, one cub approached another cub that was chewing on a new object. The cub approached at a walk and then gently pawed or smelled at the object in the other bear's mouth. The object was seized when it dropped to the ground, or it was seized directly from the mouth of the other bear. The bear then fled with the object and began to chew on it. The second bear usually did not immediately chase after the fleeing bear, but within a few minutes it approached and

10 380 J. D. HENRY AND S. M. HERRERO seized the object back in a similar manner. During one occurrence involving a feather, reciprocal seizing of the object continued off and on between the three captive cubs for over 20 min. Motor patterns of play-fighting Play-fighting sequences started with one bear approaching another bear, usually in a walk (80% of the sequences), infrequently in a run (20%). This bear exhibited a puckered-lip face with ears usually crescent (63%), infrequently partially flat (37%). The bear approached and, when close to the other bear, pawed, bit, head butted, or reared bipedally. The other bear, almost simultaneously, reared, bit, or pawed back at the initiator. The playful approach followed by pawing, biting, head butting, or rearing appears to be the play soliciting gesture of the black bear. The black bear does not show the play bow, exaggerated approach, or play dance that are play-soliciting behaviors in canids (Fox, 1970; Bekoff, 1972). Play-fighting sequences ended with both bears shutting their mouths, assuming a low intensity puckered-lip face, and walking or running away from each other. In 50% of the play-fighting sequences, this moving apart was preceded or accompanied by a cut-off or submissive display, for example, looking away, licking cut-off, head and, neck extension, or shrugging away (see Fox, 1970, for a description of these behaviors in canids). The 20 different motor patterns commonly observed during play-fighting are described under two headings. These headings relate to the double weapon system of the black bear. Canid species use only the teeth to inflict injuries during social combat (Fox, 1969, 1970). Ursid species, on the other hand, are equipped with two weapon systems for social combat, the teeth and the claws. This dual weapon system of the bears has influenced the appearance of both serious agonistic behavior and play-fighting. The dual weapon system has apparently led to the development of new motor patterns while modifying the external forms of others. Motor patterns associated with biting actions. During play-fighting, different kinds of biting actions were observed. The neckbite-hold was the most intense type of biting observed and differs from the ordinary playbite in a number of features. During the neck-bite-hold, the opponent's neck was bitten and held for up to 10 sec, instead of being immediately released as in the playbite. Pawing actions, attempting to pin the opponent to the ground, were observed with neck-bite-holds, but not with ordinary playbites. Vigorous head shaking was often observed just prior to the release of the neckbite-hold. Head shaking was only infrequently associated with ordinary play-bites. Also the orientation of the neck-bite-hold was quite regular: 74% weie oriented towards the lateral neck surface of the opponent (Table 3). The orientation of the play-bite was more variable: 14 different body parts were bitten in 146 different playbites (Table 3). Biting attempts are complete biting actions where the animal missed or had its head blocked by the opponent. Biting attempts are to be distinguished from biting intention movements. Biting intention movements are incomplete biting actions, performed once or repeatedly, but always in an incomplete manner. They usually consisted of one bear opening its mouth and orienting it towards, without reaching to bite, the other bear. At their highest degree of completion, biting intention movements take the form of the inhibited bite. In the inhibited bite, the jaws are placed around the opponent's neck or limb but are not shut or are shut very gently. The jaws are held there for 1 to 2 sec, and then withdrawn. Inhibited biting is also observed during serious agonistic encounters in the black bear. Bekoff (1972) suggests that in canids inhibited biting appears to be a partially learned response as a result of a conflict situation: the animal wishes to bite but fears a counter-attack (see also Poole, 1966; Fox, 1971). During play-fighting, young black bears also show muzzle seizure behavior (Fig. 4). During this behavior, two bears come to-

11 SOCIAL PLAY IN THE AMERICAN BLACK BEAR 581 FIG. 4. Two of the captive cubs engage in muzzle seizure behavior. gether in a face-to-face orientation, interlock their jaws, and hold them interlocked for up to 15 sec. Gentle squeezing actions are frequently observed. Muzzle seizure is observed in the wolf and coyote (Fox, 1970, 1971). In the wolf, it serves a number of different functions, for example, a greeting behavior, a dominance display, a play activity among young wolves, and a form of "ritualized aggression" (Fox, 1971). In the black bear, muzzle seizure has never been observed to form a type of "ritualized aggression." It is usually observed only during low intensity play-fighting amng cubs and sub-adults, or infrequently as a greeting behavior between a sow and her cub. Motor patterns associated with pawing and clawing actions. Davis (1949, 1964) states that bears, as compared to canids, show an increased manipulative ability of the forequarter and forepaws. He supports this statement by describing in detail the morphological adaptations exhibited by bears for this increased manipulation of the forepaws. During play-fighting, increased manipulative ability of the forepaws is evident in both pawing and clawing actions. Pawing actions are variable, low intensity arm movements that are often aimed at knocking the opponent's head or arm out of the way and thus blocking an attack. Figure 2 shows face-pawing, a type of pawing frequently observed during black bear social play, and often used to initiate the social play sequence. Pawing actions also occur during lunging when one bear extends his forelimbs and lunges at the opponent's shoulder region in an attempt to knock the social partner over. Knocking the social partner over is also

12 382 J. D. HENRY AND S. M. HERRERO accomplished by head butting. In this motor pattern, one bear walks into the other bear with its forehead and gives a quick head lift movement trying to knock the other bear off balance. Head butting is frequently observed at the initiation of the social play sequence. Clawing actions as compared to pawing actions are vigorous swiping movements aimed at various parts of the opponent's body (Table 3). Observations support that clawing actions can inflict serious injuries. During several aggressive encounters, we observed an adult bear inflict 4- to 6-inchlong cuts in the side or rump of a rival bear. During play fighting, the black bear appears to use a form of inhibited clawing. In this clawing, the claws are not flexed down, and the opponent (or the observer's hand) is consistently hit with the foot pad instead of with the claws. We have described the neck-bite-hold observed during play-fighting. In order to break the neck-bite-hold the bear frequently rolls over. Hip slamming which many canids use to break neck-bite-holds (Fox, 1970, 1971) has not been observed in the black bear. If rolling over does not break the hold, hind-leg-clawing is frequently observed. Specifically, the bear that has rolled over lifts its hind feet and alternately claws at the forehead of the opponent. Hind-legclawing is similar to a motor pattern seen in the domestic cat (Leyhausen, 1973), but this claw-related motor pattern has never been described in canid species. Burghardt and Burghardt (1972) describe rolling over as a play-invitation behavior in black bears. We observed rolling over used in the context of breaking neck-biteholds. We also infrequently observed it when a larger bear playing with a smaller bear rolled over apparently to assume a subordinate role. In both of these contexts, rolling over did not occur until well after the play sequence had begun. In less than 1% of the sequences (3 out of 508) was rolling over observed at the initiation of the social play sequence. In many play-fighting sequences, reamig initiates the sequence but it may also appear repeatedly throughout the sequence. Rearing occurs when the bear lifts its forepaws off the ground, and while standing on its hind legs, or sitting, holds its forepaws free of the ground for up to 20 sec (see Fig. 3). Rearing frees the forepaws for manipulative actions, and rearing and pawing are closely associated in many playfighting sequences. For example, repeatedly during a sequence, both bears may rear and paw at each other trying to knock each other over (Meyer-Holzapfel, 1957). These extended rearing-pawing matches are one of the most common activities during play-fighting, but they have never been observed during serious agonistic encounters. Rearing also frequently precedes standing over. Standing over occurs when a bear approaches a conspecific from the side, rears, and places its paws on the conspecific's shoulder or back. Fox (1971) describes standing over in various canids. Among the captive black bear cubs, standing over usually preceded play fighting, but in four sequences standing over led to sexual play. By way of summary, the typical playfighting sequence usually exhibits four phases: initiation, pushing-pulling, bitingclawing, and finally, termination. One bear approaches another bear and initiates play by rearing, pawing, biting, head butting, or, infrequently, lunging at that animal. Frequently, both animals rear and an extended pawing match follows. The purpose of pawing appears to be to knock the social partner off balance so that a biting or clawing action can be delivered. Biting intention movements and face-pawing actions are repeatedly observed during these extended rearing-pawing matches. Once the social partner is pushed or pulled into a vulnerable position, a play-bite or neckbite-hold is frequently delivered. The bitten animal then rolls over and may hind-legclaw in order to break the neck-bite hold. Once freed, both bears frequently return to the rearing-pushing-pulling phase until one of the bears once again becomes vulnerable to a bite or clawing action. This alternation between the pushing-pulling and biting-clawing phases continues on the average for about 20 sec (see Table 1) until

13 SOCIAL PLAY IN THE AMERICAN BLACK BEAR 383 TABLE 3. The orientation of 200 biting actions and 200 clawing actions during black bear play fighting. Number in parentheses is percentage of actions oriented towards that specific body region. Play bites Bite-and-holds Play nips Clawings Lateral face Forehead Lateral neck Shoulder Ear Ventral neck Chest Foieliiu'u Abdomen Lateral body Dorsal body Rump Hindlimb Foot Total (31%) (6) (13) (3) (3) ( 1) ( 1) ( 3) 18 (2) (6) (8) (16) (100%) one of the bears closes its mouth, frequently gives a cut-off display, pivots, and flees. Flight behavior by one of the bears usually terminates the play-fighting sequence. The initiation and termination phases of the play-fighting sequences are comparatively regular. The pushing-pulling phase alternating with the biting-clawing phase shows a great amount of variability in the temporal sequencing of motor patterns. Many of the canid motor patterns that black bears do not show are highly cursorial ones, for example, hip slamming, the play dance, the play bow, and the exaggerated approach (Fox, 1970; Bekoff, 1972). These may represent motor patterns of more recent origins that evolved in Canidae after the ursid line had differentiated (see Lorenz, 1941, for an evolutionary study of motor patterns in Anatinae). Table 3 examines the orientation of 200 TABLE 4. The orientation of 129 biting actions and 42 clawing actions during adult black bear aggressive behavior. Number in parentheses is percentage of actions oriented towards that body region. Lateral face Lateral neck Shoulder Forelimb Lateral body Rump Hindlimb Foot Total Biting actions (70%) (21) (2) (2) (2) (3) (100%) Clawing actions (10%) ( 7) (52) (10) (19) (2) (100%) (12%) (74) (8) (3) (3) (100%) (7%) (7) (13) (33) (27) (13) (100%) (28%) (26) (14) (25) / 9\ \ *V (3) < 1) (100%) biting actions and 200 clawing actions observed during play-fighting in the captive cubs. Table 4 examines the orientation of 129 biting actions and 42 clawing actions observed during aggressive behavior in adult black bears. Adult aggressive behavior was identified by the complete flattening of the ears and numerous vocalizations that accompanied it. Complete biting actions are relatively rare in adult black bear aggressive behavior. Thus, the biting actions contained in Table 4, are composed of 100 high intensity biting intention movements, 17 inhibited bites, and 12 complete, uninhibited bites. In the black bear, there appears to be a strong tendency to bite at the side of the face or the lateral neck surface of the opponent. In play-fighting 49% of the biting actions (98 out of 200) were directed towards these two regions. The remaining biting actions in play-fighting were distributed among 12 other body regions. Within the limitations of the comparison, Table 4 suggests that this tendency to bite at the side of the face or lateral neck surface appears even stronger in adult aggressive behavior; 91% of the biting actions in aggression were oriented towards these two body regions. Biting actions during adult aggressive behavior, therefore, appear to be less variable in their orientation than biting actions during social play. In play-fighting, there appears to be a strong tendency to claw at the side of the

14 384 J. D. HENRY AND S. M. HERRERO face, forehead, or shoulder region of the social partner. These three body regions account for 79% of the clawing actions observed during play-fighting. Four other body regions received the remainder of the clawing actions. In adult aggressive behavior, one body region of the opponent, the shoulder region, received 52% of the aggressive clawing actions. Thus, as with biting actions, clawing during adult aggressive behavior appears less variable in its orientation than it does during social play. Fox (1969) suggests that canid aggressive behavior is ritualized because the attacker tends to bite at certain, restricted body regions of the opponent. The above data suggest that this may also be true for black bear aggressive behavior. Also the data suggest that the side of the face, forehead, lateral neck, and shoulder regions of the black bear should be examined for defense mechanisms against aggressive clawing and biting. Defense mechanisms may take the form of thick dermal shields (see Geist, 1964), or long, thick protective coats of hair (see Schallar, 1972). The long thick fur on the dorsal and lateral surfaces of the bear's neck may function as such a defense mechanism. Motor patterns of sexual play In the black bear, as in many species, sexually immature animals on occasion exhibit sexual behavior, and this type of behavior has been termed sexual play. We may validly ask if sexual play is really play behavior; and, if so, by what criteria and characteristics is it identified as such. We will examine this question by comparing the sexual behavior shown by black bear cubs with the sexual behavior as shown by adult black bears. Sexual play was infrequently observed in the social play of black bear cubs and subadults (see, for example, Table 1). During the sexual play of captive and free-ranging cubs, a male cub usually tried to mount a male or female littermate. The male cub approached, ears crescent, giving no vocalization. The male cub then reared and frequently showed standing over by placing his forepaws on the littermate's back. Next, the male cub moved into mounting position, bit the littermate's shoulder scruffskin, and, finally, clasped the littermate bilaterally with his forepaws. At this point, the littermate usually rolled over on its back and initiated play-fighting by pawing and biting at the male cub's face. Infrequently, pelvic thrusts were seen during sexual play. Since the baculum of cubs is short and weighs appi'oximately half a gram (Rausch, 1961), it is doubtful that intromission ever takes place. Vulva stimulation was also observed during sexual play. The male cub usually approached a female littermate from behind and smelled the base of her tail and rump without touching them. Then the male cub moved his nose ventrad and smelled and gently bit and licked the vulva region of the female. Frequently, the female cub was lying on her side and, when the male cub touched her vulva, the female cub inguinal presented, that is, she rolled one leg laterad exposing the inguinal region (see Fox, 1971, for a description of inguinal presentation in canids). In adult sexual behavior, vulva stimulation has always been observed to precede the first mounting attempt. In sexual play, mounting was often attempted without vulva stimulation preceding it. Littermates usually responded to mounting and vulva stimulation by initiating play-fighting. They frequently sat or rolled over on their backs in order to break the mounting attempt. Then they pawed and bit at the male cub's face in a low intensity manner. Rolling over and pawing at the male's face are also observed in non-receptive adult females in response to the serious mounting attempts of adult boars. The non-receptive adult female during this reaction frequently shows partially flat ears, a lack of vocalizations, and the relaxed open-mouth face. It is one of the few consistent occurrences of play-like behavior in the behavioral repertoire of adult black bears. From the above descriptions, it is apparent that the sexual behavior shown by

15 SOCIAL PLAY IN THE AMERICAN BLACK BEAR 385 cubs exhibits a number of features only observed in other types of social play. For example, the male cub in approaching a littermate gives no vocalization and his ears are usually clearly crescent. During adult sexual behavior, the boar approaching a female frequently makes a throat-clunking vocalization and his ears, although variable, are frequently frontal or laterad. Furthermore, during sexual play the male cub usually performs only part of the normal adult sexual sequence; for example, mounting is frequently attempted without vulva stimulation preceding it. Marler (1966) states that partially performed sequences are frequently a characteristic of social play. These characteristics support the conclusion that the sexual behavior shown by cubs is indeed a form of social play. HYPOTHESIS TESTING AND DISCUSSION Beach (1945) stresses that play must be defined according to its general, objective characteristics. A number of researchers have speculated about what the general characteristics of play might be (see Meyer- Holzapfel, 1956; Tembrock, 1958; Marler, 1966; Loizos, 1967). Muller-Schwarze (1968) showed the value of testing this speculation. He found that several of the characteristics as put forth by Meyer-Holzapfel (1956) were not valid for the social play of blacktailed deer (Odocoileus hemionus columbianns). Extensive testing of this speculation over a wide range of species is necessary if the valid diagnostic characteristics of play are to be identified. In this section, five suggested characteristics are tested for their validity in black bear social play. Hypothesis about the characteristics of social play can be grouped under two headings: (i) characteristics about the motor patterns of social play, and (ii) characteristics about the sequencing of social play. The following five hypotheses are examined under these two headings. Characteristics of the motor patterns of social play Hypotheses about the motor patterns of social play compare the appearance of the motor patterns during social play to the same motor patterns as they appear in nonplayful behavior. We will discuss three of these hypotheses. Hypothesis 1: Motor patterns during play often have an incomplete appearance when compared to their appearance during non-playful behavior. For example, during play-fighting, in dogs, snarling may be dissociated from the piloerection which inevitably accompanies it during true fighting (Marler, 1966). This hypothesis appears valid for much of black bear social play. The following examples illustrate this point. First, the above example about piloerection appears valid for black bear social play. In black bears, piloerection of the dorsal and lateral neck fur is occasionally observed during real fighting but has never been observed during play-fighting. Second, during black bear social play many behaviors lack the vocalizations that normally accompany these behaviors during non-playful behavior (see p. 378). Third, many threat displays are omitted from play-fighting. For example, in an adult agonistic encounter, an approaching bear frequently stops abruptly and paw swats the ground (see also Jonkel and Cowan, 1971). Later in the encounter, both bears usually display the challenge postiure (see p. 378). At the end of the encounter, a bear moving away frequently jaw snaps repeatedly at the opponent. These three threat displays are rarely observed during play-fighting. Poole (1966) also observed that in the play-fighting of polecats (Putorius putorins) most agonistic behaviors were present except four threat behaviors which were omitted. Thus, in at least two species of Carnivora, certain threat displays appear to be selectively omitted from play-fighting behavior. The above examples serve to illustrate that the first hypothesis appears valid for much of black bear social play: many motor patterns are given in an incomplete manner when compared to their appearance during non-playful behavior. Hypothesis 2: Social play may be identified by certain motor patterns which are observed only during social play (Marler,

16 386 J. D. HENRY AND S. M. HERRERO 1966). In the black bear there are several motor patterns that are observed almost exclusively during social play. These are head butting, muzzle seizure, the play nip, and hind-leg-clawing. These four motor patterns which are frequently observed during black bear play-fighting are extremely rare in the agonistic behavior of cubs, subadults, or adult black bears. Marler (1966) gives several examples of motor patterns unique to social play, and all of the examples are play-soliciting behaviors. In the black bear, only head butting and the play nip appear to serve as play-soliciting behaviors (see p. 380). The other two motor patterns unique to black bear social play usually do not occur until well after the social play sequence has begun. Hypothesis 3: Social play is characterized by the exaggerated and uneconomical quality of the motor patterns involved (Loizos, 1967). Loizos states that this exaggerated and uneconomical quality of the motor patterns is "what all playful activity has in common." There is an important need to state more objectively what "exaggerated" and "uneconomical" mean in terms of observable behavior. These terms are vague and somewhat subjective, and, therefore, difficult to test (Beach, 1945). We interpreted "uneconomical" to mean that motor patterns during social play are more variable in their orientation than they are during non-playful behavior. We interpreted "exaggerated" to mean that social play motor patterns are performed in an enlarged or prolonged manner as compared to non-playful motor patterns. From this enlarged or prolonged quality, we would expect social play motor patterns to exhibit an increased time duration as compared to non-playful motor patterns. Tables 3 and 4 offer evidence that during social play biting and clawing actions are more variable in their orientation than are adult aggressive bitings and clawings (see p. 383). Mounting also appears more variable in its orientation during sexual play than during adult sexual behavior. There is evidence, therefore, to conclude that certain motor patterns during social play are less economical in their orientation than they are during adult non-playful behavior. Regarding the enlarged or prolonged quality of motor patterns, our observations suggested that the majority of social play motor patterns were not performed in an enlarged manner when compared to the same motor pattern in adult non-playful behavior. We tested this point by timing the filmed motor patterns. We selected and timed six motor patterns that appear in both high intensity adult aggressive behavior and in high intensity play-fighting (Table 5). Filmed sequences were selected by random numbers, and 30 timings for each motor pattern were obtained. The means and standard deviations for each motor pattern are reported in Table 5. Differences between means were tested for using the Student's t test (2 tail test, a =.05). From these tests (see Table 5.), we conclude that motor patterns in play fighting do not show a consistent increased duration when compared to their adult aggressive counterparts. Loizos' hypothesis, therefore, appears only partially valid for black bear social play. Certain motor patterns appear more variable in their orientation during social play, but our tests and observations suggest that motor patterns are not performed in an enlarged or prolonged manner in social play when compared to the same motor patterns in adult non-playful behavior. TABLE 5. The duration of motor patterns in black bear play fighting and aggressive behavior (in sec). Biting Clawing Rearing Lookingaway Lickingcut-off Bitingintentionmovement Play fighting t ± ± ± ± 0.26 Aggression i ± ± ± ±1.29 Conclusion a = 0.05 P.F. = Agg. P.F. = Agg. P.F. > Agg. P.F. = Agg. P.F. < Agg. P.F. < Agg.

17 SOCIAL PLAY IN THE AMERICAN BLACK BEAR 387 Characteristics of social play sequences We have examined three hypotheses concerning the motor patterns of social play. A number of other hypotheses concerning the sequences of social play have also been advanced. We will examine two of these hypotheses. Hypothesis 4: A characteristic of play is that certain motor patterns are repeated more frequently in the play sequences than during the non-play sequences (Meyer-Holzapfel, 1956; Loizos, 1967). We tested this hypothesis in the following way: Motor patterns that occur more than once in either a play-fighting sequence or in an adult aggressive sequence are listed in Table 6. Biting and clawing actions in this table include both inhibited and uninhibited types of biting and clawing. Filmed sequences were chosen at random and used once. This included both play sequences from cubs and aggressive sequences from adult black bears. For each sequence, one of the motor patterns from Table 6 was chosen and the frequency of it in the sequence was counted. For each motor pattern 30 non-zero frequencies were obtained (with 3 exceptions for adult aggressive behavior; see Table 6). The means and standard deviations of these frequencies are presented in Table 6. Using a Student's t test (Li, 1961), we tested for differences between means (2 tail test, a =.05). We conclude that four of the six motor patterns examined are repeated significantly more frequently during play-fighting than during adult aggression. Greater frequency of a motor pattern per sequence may result simply because playfighting sequences last longer than aggressive sequences. To test this point, we recorded the time duration of each play and aggressive sequence used. Using a t test (2 tail test, a =.05), we found that in five out of the six tests play-fighting sequences did not last significantly longer than sequences of adult aggression. The one exception involved the looking away behavior: Play-fighting sequences that showed looking away last significantly longer than adult aggressive sequences that showed this TABLE 6. The frequency of motor patterns in black bear play fighting and aggressive behavior. The number in the parentheses indicates that the sample size was different from 30. Biting Clawing Rearing Lookingaway Lickingcut-off Bitingintentionmoveraent Play Conclusion fighting Aggression a = ± ± 0.93 P.F. > Agg. (n - 19) 3.77 ± ±1.61 P.F. > Agg. (n = 23) 2.53 ± ± 1.40 F.F. = Agg. (n = 23) 2.30 ± ±1.11 P.F. = Agg ± ±0.76 P.F. > Agg ± ±2.11 P.F. > Agg. display. It appears that greater repetition of the motor patterns was not due to the play-fighting sequences lasting longer than the adult aggressive sequences. We conclude, therefore, that, within the limitations of the test, greater repetition of certain motor patterns appears to be a valid characteristic of black bear social play. Hypothesis 5: In play, normal temporal groupings of functionally related motor patterns break down, so that different kinds of behaviors intermingle in the same sequence; for example, prey capture behavior and sexual behavior may be combined in the same play sequence (Meyer-Holzapfel, 1956). Predatory behavior is rarely observed in adult black bears, and it also appears to be rare in black bear social play. Thus, there is only one possible way that black bears could manifest this characteristic, namely, by combining sexual motor patterns with agonistic motor patterns. We examined the filmed sequences from the captive cubs and from adult black bears for this combination. Regarding sexual behavior, the sample size is small, but a general trend appears to be indicated. Three-fourths of the sexual play sequence (9 out of 12) combined sexual motor patterns with agonistic motor patterns. However, about three-fourths of the adult sexual sequences (10 out of 12) also showed this combination. As mentioned earlier, this is because both cubs and nonreceptive adult females respond to courting

18 388 J. D. HENRY AND S. M. HERRERO males by initiating play-fighting. Thus, the combination of sexual with agonistic motor patterns cannot be used to differentiate sexual play in cubs from sexual behavior in adult black bears. Regarding agonistic behavior, only 0.4% of adult black bear agonistic sequences (1 out of 247) showed sexual motor patterns combined with agonistic motor patterns. The one sequence that showed this combination involved an agonistic encounter between an adult male and an adult female black bear. The male emerged from the encounter dominant, and immediately after the encounter, tried to mount the female. His attempt was unsuccessful because the female fled. Since only 1 out of 247 sequences showed sexual behavior combined with agonistic behavior, we may conclude that this combination is almost non-existent in adult black bear agonistic behavior. It also appears that black bears do not use a ritualized mounting behavior as a dominance display. This type of dominance display is frequently observed in primate species (see Wickler, 19676). In black bear cubs, 3.9% of the playfighting sequences (17 out of 428) showed agonistic behavior combined with sexual behavior. We, therefore, conclude as follows: Agonistic motor patterns combined with sexual motor patterns is almost nonexistent in adult black bear agonistic behavior. In cub play-fighting, the vast majority of the sequences (411 out of 428) show only agonistic motor patterns; however, a small percentage (3.9%) show agonistic behavior combined with sexual behavior. In only this very restricted sense is Meyer- Holzapfel's hypothesis a valid identifying characteristic of black bear social play. CONCLUSIONS The preceding tests show that several of the suggested characteristics of social play are only partially valid for black bear social play. Extensive testing of the suggested characteristics of play on as wide a range of species as possible is definitely indicated for future research. These tests are crucial if the valid diagnostic characteristics of play and social play are to be known. Beach (1945) and Marler (1966) have correctly stressed the need for this type of research. Repeatedly during this paper we have pointed out similarities between canid and ursid social behavior. Geist (1974), among others, has stated that social behavior, and particularly display behavior, can be expected to be rather conservative in its evolution. The conservatism of black bear display behavior is illustrated by the following example. Black bear and canids share the following active submissive displays: face-pawing, licking intention movements, and nose-stab movements oriented towards the dominant animal's mouth (Fox, 1971). These displays in Canidae have been derived from soliciting behaviors, that is, from when young canids beg food from their parents (Fox, 1971). Bears have retained these active submissive displays even though the regurgitation response of Canidae has been lost from the bear's behavioral repertoire. The extensive similarities between canid and ursid displays tend to support the principle of the evolutionary conservatism of social behavior. This principle has important implications for behavioral research and ultimately for our understanding of man's future societies. The principle deserves to be thoroughly tested. REFERENCES Baercnds, G. P Comparative methods and the concept of homology in the study of behavior. Arch. Neer. Zool. 13: Beach, F. A Current concepts of play. Amer. Natur. 79: Bekoff, M The development of social interaction, play, and metacommunication in mammals: an ethological perspective. Quart. Rev. Biol. 47: Bekoff, M Social play and play-soliciting by infant canids. Amer. Zool. 14: Bolwig, N Facial expressions with remarks on parallel development in certain carnivores. Behaviour 22: Burghardt, G. M., and L. S. Burghardt Notes on the behavioral development of two black bear cubs: the first eight months, p In S. M. Herrero [ed.], Bears their biology and management. IUCN Publ. No. 23. Morges, Switzerland. Colbert, E. H Evolution of the vertebrates. John Wiley and Sons, Inc., New York.

19 SOCIAL PLAY IN THE AMERICAN BLACK BEAR 389 Crisler, L Artie wild. Balantine Books, New York. Darwin, C The expression of emotions in man and the animals reprint. Univ. Chicago Press, Chicago. Davis, D. D The shoulder architecture of bears and other carnivores. Fieldiana Zool. 31: Davis, D. D The giant panda: a morphological study of evolutionary mechanisms. Fieldiana Zool. Mem. Vol. 3. Chicago Natural History Museum. Folk, G. E., M. A. Folk, and J. J. Minor Physiological conditions of three species of bears in winter dens, p In S. M. Herrero [ed.], Bears their biology and management. IUCN Publ. No. 23. Morges, Switzerland. Fox, M. W The anatomy of aggression and its ritualization in canids. Behaviour 35: Fox, M. W A comparative study of the development of facial expressions in canids: wolf, coyote and foxes. Behaviour 36: Fox, M. W Behaviour of wolves, dogs, and related canids. Jonathan Cape, Ltd., London. Geist, V On the rutting behavior of the mountain goat. J. Mammalogy 45: Geist, V On the relationship of social evolution and ecology in ungulates. Amer. Zool. 14: Herrero, S. M Aspects of evolution and adaptation in American black bears (Ursus americanus Pallas) and brown and grizzly bears ( /. arctos Linne) of North America, p In S. M. Herrero [ed.], Bears their biology and management. IUCN Publ. No. 23. Morges, Switzerland. Hinde, R. A Animal Behaviour. 2nd ed. McGraw-Hill Book Co., San Francisco. Hofer, H Prolegomena primatologiae, p In H. Hofer and G. Altner [ed.], Die Soderstellung des Menschen. G. Fischer Verlag. Stuttgart. Jonkel, C. J., and I. McT. Cowan The black bear in the spruce-fir forest. Wildlife Monogr. 27: Leyhausen, P Beobachtungen aneinem jungen Schwarzbaren (Ursus americanus Pall.). Z. Tierpsychol. 6: Leyhausen, P Vcrhaltensstudien an Katzen. Beiheft 2 der Z. Tierpsychologie. 3rd ed. Li, J. C. R Introduction to statistical inference. J. W. Edwards Inc., Ann Arbor, Michigan. Loizos, C Play behaviour in higher primates: a review, p In D. Morris [ed.], Primate ethology. Aldine, Inc., Chicago. Lorenz, K Vergleichende Bewegungsstudien an Anatinen. J. Ornithol. 89: Lorenz, K Man meets dog. Methuen, London. Marler, P The characteristics of play behavior, p In P. Marler and W. J. Hamilton, Mechanisms of animal behavior. John Wiley and Sons, Inc., New York. Meyer-Holzapfel, M Uber die bereitschaft zu spiel- und Instinkthandlungen. Z. Tierpsychol. 13: Meyer-Holzapfel, M Das Verhalten der Baren (Ursidae). Handb. Zool. VIII, Part 10 (17): Muller-Schwarze, D Play deprivation in deer. Behaviour 31: Muller-Schwarze, D Ludic behavior in young mammals, p In M. B. Sterman, D. J. McGinty, and A. M. Adinolfi [ed.], Brain development and behavior. Academic Press, New York. Nelson, R. A., H. W. Wahner, and J. D. Jones Metabolism of bears before, during, and after winter sleep. Amer. J. Physiol. 224: Papez, J. W Comparative neurology. Crowell, Inc., New York. Poole, T. B Aggressive play in polecats. Symp. Zool. Soc. London 18: Rausch, R. L Notes on the black bear, Ursus americanus Pallas, in Alaska, with particular reference to dentition and growth. Z. Saugetierk. 26: Sauer, P. R Growth of black bears from the Adirondacks. Proc. N.E. Sect. Wildlife Soc. 33 p. Mimeo. Schaller, G. B The Serengeti lion. Univ. Chicago Press. Chicago. Schenkel, R Ausdrucksstudien an Wolfen. Behaviour 1: Schenkel, R Submission: its features and functions in the wolf and dog. Amer. Zool. 7: Stonorov, D., and A. W. Stokes Social behavior of the Alaska brown bear, p In S. M. Herrero [ed.], Bears their biology and management. IUCN Publ. No. 23. Morges, Switzerland. Tembrock, G Spielverhalten beim Rotfuchs. Zool. Beitr. Berlin 3: van Hooff, J. A. R. A. M The facial displays of the Catarrhine monkeys and apes, p In D. Morris [ed.], Primate ethology. Aldine, Inc., Chicago. Wickler, W. 1967a. Vergleichende Verhaltensforschung und Phylogenetik, p In G. Heberer [ed.]. Die Evolution der Organismen I. 3rd ed. G. Fischer. Stuttgart. Wickler, W Socio-sexual signals and their intraspecific imitation among primates, p In D. Morris [ed.], Primate ethology. Aldine, Inc., Chicago.