Ambivalent signals during agonistic interactions in a captive wolf pack

Applied Animal Behaviour Science 105 (2007) 274 283 www.elsevier.com/locate/applanim Ambivalent signals during agonistic interactions in a captive wolf pack Jaume Fatjó a, *, Dorit Feddersen-Petersen b, José Luís. Ruiz de la Torre a, Marta Amat a, Monique Mets a, Barbara Braus b, avier Manteca a a Departament de Fisiologia, Facultat de Veterinària, Universitat Autònoma de Barcelona, 08193 Bellaterra, Spain b Zoologisches Institut, Christian Albrechts-Universität zu Kiel Olshausenstr. 40, D-24118 Kiel, Germany Available online 28 November 2006 Abstract A study was designed to quantify ambivalent behaviour during social aggressive interactions in wolves. Agonistic interactions in a group of six European captive wolves, consisting of three males and three females, were analyzed for bared teeth, body posture and the position of ears, tongue, lips, legs and tail. The behavioural elements in each of these categories were assumed to be neutral or to signal dominance or submission. Wolves were considered to show ambivalence when dominant and submissive signs were observed simultaneously. More than 200 aggressive interactions were videotaped and parts of them were analyzed frame by frame. Results indicated that male wolves showed high levels of ambivalence (48%) and that this behavioural trait is not linked to a particular social status. Regarding specific body signals, tail position was the most reliable indicator of status, whereas bared teeth was not linked to a particular position in a dominance relationship. The possible application of these results to understanding aggression problems in dogs is briefly discussed. # 2006 Elsevier B.V. All rights reserved. Keywords: Aggression; Ambivalence; Communication; Conflict; Hierarchy; Wolf 1. Introduction The wolf is the most gregarious species within the canidae family, the pack being its basic social unit (Mech, 1970). Only 27 packs of wolves have been studied in captivity showing many different social structures and housing conditions (Packard, 2003). * Corresponding author. Tel.: +34 935811352; fax: +34 935812006. E-mail address: jaume.fatjo@uab.es (J. Fatjó). 0168-1591/$ see front matter # 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.applanim.2006.11.009

J. Fatjó et al. / Applied Animal Behaviour Science 105 (2007) 274 283 275 Over many decades the social structure of wolf packs has been linked to the existence of a dominance hierarchy. In terms of biological adaptation, establishing a hierarchy has been understood as a way to regulate a privilege system, to assign leadership or just to avoid overt aggression in competitive situations (Schenkel, 1967; Zimen, 1982; Halfpenny, 2003). Recently the concept of a stable hierarchy for free ranging wolves has been revised and criticized. A wolf pack living in the wild is formed by the mated pair and their young offspring. Thus, a putative dominance hierarchy within the pack could be alternatively viewed as just a reflection of the age, gender and reproductive structure of the group (Mech, 1999). However, the concept of dominance hierarchy formation is still useful for evaluating the social dynamics of non-familiar packs, like those commonly found in captivity (Mech, 1999; Packard, 2003). Social rank order results from the formation of dominance relationships between all pack members. The concept of social dominance has been defined in many ways and the criteria to describe dominance relationships are also diverse. In social species of mammals dominance has been understood not only as indicator of priority access to certain resources, but in a much broader sense as an expression of the degree of social freedom each animal allows itself during an encounter (Zimen, 1982; Mason, 1993). Dominance is not an individual trait but a reflection of an asymmetrical and dynamic relationship between two individuals that could vary over time and depending on the context (Van Hooff and Wensing, 1987). The function of wolf visual communication during social interactions was formerly linked to the formation and maintenance of dominance relationships. Following Darwin s principle of antithesis two opposite motivational states, dominance and submission, were related to two opposite body postures. Each single visual signal was considered to be indicative of a dominant or a submissive attitude (Schenkel, 1947; Fox, 1970). More recent studies continue to refer to dominant and submissive postures as the best way to assess social relationships and to submissive behaviour as the most reliable indicator of the direction of a dominance relationship between two wolves (Van Hooff and Wensing, 1987). Nevertheless, communication during social interactions probably includes information about other dimensions of social relationships, like play, affiliation or sexual behaviour (Harrington and Asa, 2003). During a social interaction a wolf can be in motivational conflict and thus show ambivalent body posture. Ambivalent visual signals in canid communication have been described by many different authors as a mixture of body signals belonging to opposite emotional states (Beaver, 1999a; Harrington and Asa, 2003). Agonistic interactions have been studied extensively in wolves, but not always in great depth and detail. Another limitation of earlier research on wolf behaviour is that the majority of studies are descriptive and do not offer a model based on quantitative data (McLeod, 1996). Researchers on wolf behaviour do not usually look at specific body signals to recognize social relationships in a wolf pack, but rely on their ability to identify general patterns of communication, each of them composed of many simple body signals (Harrington and Asa, 2003). Rank order is usually established through the construction of an interaction matrix where a dominant or a submissive attitude is assigned to each wolf in every interaction. Static elements of body posture are taken into account together with more dynamic aspects of the interaction, such as which animal controls, prevents or influences the movements of the other one (Moran et al., 1981; Zimen, 1982; Harrington and Asa, 2003). Visual communication in wolves seems to be more complex than previously thought and subtle changes of facial expression and body posture could be essential for the fine-tuning of social interactions (Harrington and Asa, 2003).

276 J. Fatjó et al. / Applied Animal Behaviour Science 105 (2007) 274 283 A study was designed to quantify the frequency and general characteristics of ambivalent signals during aggressive social interactions in a captive pack of European wolves (Canis lupus lupus). It is hoped that the study of visual communication and aggression in wolves can give insights into the motivation of aggression in the domestic dog (Canis familiaris). 2. Material and methods 2.1. Subjects and housing A group of six captive European wolves (Canis lupus lupus), three males and three females were included in the study. All animals were 5-years-old hand-raised siblings, a particular structure named complex family (Packard, 2003). The hierarchy of the group was previously established using methods comparable to those described in the introduction section (Table 1). During the 12 months prior to the study no changes in the rank order were observed. The pack was kept in a 400 m 2 octagonal enclosure, in a research facility at the University of Kiel (Germany). The enclosure was divided into four areas, which were connected by openings that enabled the animals to access the separate areas freely. Different structures such as dens, tunnels and elevated platforms were available. In the middle of the enclosure was an elevated observation tower, which could be reached by a corridor and from which the animals were observed and videotaped. An indoor enclosure was located at the bottom of the observation tower. 2.2. Observational techniques The behaviour of the subjects was filmed during a period of 2 months, from early December to late January. This is the period that precedes the breeding season, when the tension in the pack rises and aggressive interactions, especially between males, are more frequent (Packard, 1989). Continuous recording was undertaken, with observation periods scheduled early in the morning and late in the afternoon, as this is the time of the day when wolves are usually more active. A behaviour sampling rule was used to record any dyadic aggressive interaction within the pack. All animals were visible all the time, except when they were inside or close to the indoor enclosure. 2.3. Analysis The data were analysed using SPSS 9.0 for Windows. Descriptive statistics were calculated to describe who interacted with whom and how often. From all the interactions collected, those recorded from the beginning to the end and where both individuals were visible all the time were selected for further analysis. In all these selected interactions the two wolves were identified and their behaviour analyzed frame by frame. For each frame and for each individual different communicative signals were evaluated and coded. A special emphasis was given to the position of the ears, tail and lips, the occurrence of teeth baring, tongue Table 1 Name and rank order of wolves included in the study Males Females Wolf name Rank order Wolf name Rank order Misha Alpha Lara Alfa Petja Beta Nina Beta Kolja Omega Tonja Omega

J. Fatjó et al. / Applied Animal Behaviour Science 105 (2007) 274 283 277 Table 2 Visual communicative signals and their ethological interpretation (neutral signals excluded) Visual signal Dominant Submissive Ears position Up Lowered Flattened Tail position Up to horizontal Down Between legs Mouth Teeth baring Tongue Licking Lips Rounded Retracted Movement Increase distance Decrease distance licking and biting. Also, the tendency of an individual to decrease or increase the distance to one another was recorded. The alternatives for each category were labelled as indicative of a dominant, a submissive or a neutral attitude, based on reference studies on wolf communication (Schenkel, 1947; Mech, 1970; Zimen, 1982; Harrington and Asa, 2003) (Table 2). According to our definition a wolf was considered to show an ambivalent posture whenever it simultaneously expressed dominant and submissive signals. Ambivalence was determined as the fraction of frames in which the animals showed an ambivalent posture divided by the total number of frames analysed. Ambivalence was calculated for the group of wolves, for each gender, and in the case of males per individual wolf and per pair of wolves. No interactions related to food were analyzed. Consequently, interactions could be related to status assertion, privileged access to certain areas or to a particular pack member. x 2 -tests were performed to compare the distribution of the frequencies of ambivalent behaviour between individuals and the association between previously established rank order and the different body signals. 3. Results 3.1. Descriptive statistics More than 200 interactions including aggression signals were collected during the observational period. The mean duration of the observed interactions was 9.7 s. The characteristics of the interactions differed between individuals as well as between different pairs of wolves. Males took part in the vast majority of observed interactions (85.8%) and the omega male was involved in most of them (76.3%). Male to male interactions were the most frequent (82%), whereas female to female and male to female accounted, respectively for 10.5% and 7.5% of all interactions. Respectively between the males, the beta and the omega wolves were the pair that interacted most (47.7%), followed by the alpha and the omega (28.7%) and the alpha and the beta (23.8%). The beta male was never observed to interact with any of the females and the beta female never did so with any of the males. All recorded female to female interactions were between the alpha and the beta female. The main reason for the lack of recorded interactions involving the omega female was that she spent most of her time inside or near the indoor enclosure.

278 J. Fatjó et al. / Applied Animal Behaviour Science 105 (2007) 274 283 Most of the time (73%) interactions took place at a very close distance but without physical contact that was only established in 19% of the recorded frames. 3.2. Ambivalence All results about ambivalence were expressed in percentage of interaction time. In the case of males, during 68.8% of the time at least one of the individuals showed some signals of ambivalence. Both wolves displayed ambivalence at the same time in 20.38% of analyzed frames. In 31.2% of the time none of the individuals acted in an ambivalent way. Overall, the mean rate of ambivalence for a single wolf was 48% of the total time of interaction. The mean rate of ambivalence for female to female interactions was 37.6% of the total time of interaction. Due to the small percentage of interactions where both females were visible and the almost complete lack of recording interactions involving the omega females, only male to male interactions were further analyzed. The percentage of ambivalence differed between individuals and between pairs of individuals. The alpha male (49.18%) and the beta male (51.36%) showed ambivalence at comparable rates (x 2 = 0.61; p = 0.435). However, the level of ambivalence displayed by the omega (36,38%) significantly differed from the alpha s (x 2 =3.76;p = 0.05) and the beta s (x 2 =5.23;p = 0.022). In some cases, the rate of ambivalence showed by each individual did differ depending on the receiver. The beta expressed much more ambivalence when interacting with the omega (61%) than with the alpha (15.4%) (x 2 = 9.45; p = 0.0021). Similarly, the level of ambivalent behaviour of the omega was higher towards the beta (44.7%) than towards the alpha (17%) (x 2 = 6.84; p = 0.0089). Only the alpha showed no significantly different levels of ambivalence when interacting with the beta (55.6%) or the omega (45.2%) (x 2 = 1.77; p = 0.184). 3.3. Specific body signals The proportion of observations in which each body signal was expressed differed between the three males. In order to link the results to a specific social status the alpha male in all interactions and the beta male when interacting with the omega were both considered dominant individuals. Conversely, the omega male in all interactions and the beta male when interacting with the alpha were considered submissive individuals. The percentage of time each wolf held its ears in a certain position is depicted in Table 3. The position of the ears was linked to the social status of the sender. In 89.5% of the time when ears were observed in a fully up position a dominant wolf was involved (x 2 = 7.94; p = 0.0048). Most of the time when ears were carried in a flattened position (78.1%) involved a submissive Table 3 Ear position Position of the ears Alfa Beta a Beta b Omega Fully up 40 0 18 2 Partially lowered 60 53 73 54 Flattened 0 47 9 43 Expressed as the percentage of time each individual held its ears in a particular position. a Data of the beta male when interacting with the alpha. b Data of the beta male when interacting with the omega.

J. Fatjó et al. / Applied Animal Behaviour Science 105 (2007) 274 283 279 Table 4 Tail position Position of the tail Alfa Beta a Beta b Omega Fully up 25 0 9 0 Above the horizontal 57 0 27 0 Slightly lowered 18 6 60 7 Lowered 0 94 3 93 Expressed as the percentage of time each individual held its tail in a particular position. a Data of the beta male when interacting with the alpha. b Data of the beta male when interacting with the omega. individual (x 2 = 5.53; p = 0.0187). However, ears in a partially lowered position were equally observed in dominant (47%) and submissive (53%) animals (x 2 = 1.78; p = 0.182). Regarding the tail, the percentage of time each wolf held its tail in a certain position is depicted in Table 4. Compared to the ears, a clearer relationship between tail position and rank order could be established. Tail held in a position above the horizontal was linked to a dominant individual 99.7% of the time this signal was observed (x 2 = 12.2; p = 0.0005). A slightly lowered tail, less than 258 below the horizontal, was also typically linked to dominant individuals in 90.14% of the time (x 2 = 9.14; p = 0.024). Most of the time (90.3%) a fully lowered position of the tail was identified a submissive wolf was involved (x 2 = 11.67; p = 0006). Although around the limit of statistical significance, baring teeth was observed at comparable rates in all individuals, submissive (66%) and dominant (34%) (x 2 = 3.77; p = 0.052). On the contrary, the position of the lips during teeth exposure was linked to social status. Eighty seven percent of the time lips were observed in a rounded shape a dominant wolf was identified (x 2 = 4.2; p = 0.014), whereas retracted lips were linked 94% of the time to a submissive animal (x 2 = 5.5; p = 0.009). Tongue licking was intermittently observed in all individuals at comparable rates. 4. Discussion In terms of social behaviour and aggression the dynamics of the pack was consistent with previous descriptions of the behaviour of captive wolves, particularly one that comes from a pack with the same structure based on hand-raised siblings (Zimen, 1982). Care should be taken when extrapolating our results to other groups of wolves, since the social behaviour of social species could be markedly affected by environmental factors (Wrangham and Rubenstein, 1986). In fact, one of the most controversial issues in the study of wolf behaviour is the suitability of captive wolf studies to reflect the natural behaviour of the species (Mech, 1999). Aggression seems to occur at a much lower level in natural wolf packs as severe conflicts are resolved by dispersion (Mech and Boitani, 2003). However, the study of certain aspects of wolf communication and social interactions can only be done currently at a practical level in captivity and its integration with results from field studies seems to be the most rapid and thorough way to understand the behaviour of this species (Packard, 2003). Most agonistic interactions were between animals of the same gender and particularly between males. This probably reflects a separate ranking order, a different social strategy for males and females or just the existence of different patterns of social interaction based on gender and pack composition. An alternative explanation is that most aggressive interactions before the breeding season occurred between males (Zimen, 1982; Derix, 1994; Packard, 2003). Also, only

280 J. Fatjó et al. / Applied Animal Behaviour Science 105 (2007) 274 283 dyadic interactions were analyzed. However, dominance relationships between two individuals depend on their relationships with other pack members (Zimen, 1982). According to our results ambivalence seems to be a very common behavioural expression in captive wolves and not exclusive to a particular social status. In fact, higher levels of ambivalence were observed in high ranking individuals, the alpha when interacting with any other male and the beta when interacting with the omega. Ambivalence in a dominant individual can be understood in different ways. Schenkel described a dominant wolf showing signs of ambivalence as an individual that is tolerant but fails to display its superiority to the other wolf (Schenkel, 1967). The expression of ambivalence by a dominant individual could also reflect a state of social stress and instability around its rank position. In fact, aggression between conspecifics is one of the main sources of stress in mammals and is not necessarily linked to a particular rank position (Blanchard and Blanchard, 2006). A study to measure stress in a wild colony of wolves points out that glucocorticoids were comparatively more elevated in alpha individuals (Halfpenny, 2003). Tongue licking is considered a clear signal of conflict in wolves and dogs and was observed in all members of the pack at comparable rates (Beaver, 1999a; Harrington and Asa, 2003). An alternative explanation is that ambivalent signals could be a strategy to avoid the escalation of aggression, which would agree with current views of social behaviour in wolves that emphasise the notion of social aggression as a balance between cohesion and conflict (Packard, 2003). Aggressiveness and dominance are not equivalent terms and no linear relationship seems to exist between aggression and social status (Reisner, 2002). The results of our study point in the same direction, as a higher tendency to expose the teeth, probably the most clear single indication of an aggressive attitude, was identified in the submissive individuals. Further, at least in some circumstances, when a submissive individual became more offensive during an interaction, the dominant responded by decreasing its level of assertiveness, for instance, by lowering its ears or by showing a diverted eye look. Although, this finding comes from non quantitative data it deserves further investigation. In any case, the ability of the dominant wolf to reduce overall aggression in the face of a threat from a submissive individual could evolve as a mechanism to keep aggression in a highly ritualized way. In fact, overt aggression was rarely observed in the pack included in this study and never led to open wounds in any of the wolves involved in agonistic encounters. The interpretation of ambivalent behaviour and particularly aggression in a submissive individual could be linked to the concept of submissiveness itself. Following the early definition by Lorenz, submission may be defined as any behaviour that appeases and blocks the aggression of the opponent (Blanchard and Blanchard, 2006). In the case of a truly submissive behaviour, all intentional movements of aggression should be absent. According to this definition, growling and teeth bared paired with an overall submissive posture should be understood more in terms of defensive aggression rather than a truly submissive display (Schenkel, 1967; Harrington and Asa, 2003). Furthermore, the meaning of aggression could vary depending on whether it is linked to active or passive submission. Aggression showed by a submissive individual approaching a dominant could be considered an obtrusive attitude by the former, whereas aggressive signals displayed during passive submission could express helplessness (Schenkel, 1967). For reasons already outlined in relation to ambivalent behaviour of dominant wolves, an alternative explanation is that dominance and submission belong to an axis different from aggression. The former helps to provide information about status, whereas the latter could be understood as a way to avoid contact or to defend a particular resource (Reisner, 2002). Thus the

J. Fatjó et al. / Applied Animal Behaviour Science 105 (2007) 274 283 281 predisposition to reduce or to escalate aggression could be an individual temperament trait not related to the tendency to show dominant or submissive behaviour (Packard, 2003). Regarding the meaning of individual body signals, the position of the tail seemed to be the most reliable indicator of social status, as already reflected by other authors (Harrington and Asa, 2003). Taken alone, a tail held in a horizontal or above the horizontal position was a good predictor of dominance, whereas a marked lowered tail consistently reflected submission. Ears held in a partially lowered position appeared not to be linked to any particular position within a dominance relationship. A possible explanation for this finding is that lowering the ears could have another biological function besides expressing submission, e.g. protecting the animal from injury when aggression escalates (Beaver, 1999b). As already mentioned, baring teeth was a response not related clearly to social rank although a higher tendency to show this behaviour was observed in submissive individuals. This finding is consistent with other studies suggesting that aggression displays are not a good way to determine the social rank of wolves (Van Hooff and Wensing, 1987). In fact, studies conducted in primates have described top rank individuals as those showing less signals of aggression and higher levels of tolerance to infant play (McGuire and Troisi, 1998). A more specific indication about an individual s rank could be obtained by combining teeth exposure with the position of the lips. Lips in a rounded shape appeared to be associated with dominant animals whereas a retracted position is more typical of submissive individuals, as also shown in our results (Harrington and Asa, 2003). The wolf has been commonly taken as a model to understand the biology and behaviour of the domestic dog (Scott, 1967). However, recent data suggest that marked differences in behaviour could have been introduced through the process of domestication, especially regarding cognitive abilities and interspecific behaviour towards human beings (Topál et al., 2005). Also, some communication signals like tail wagging could be differentially presented by dogs and adult wolves (Gacsi et al., 2005). Although some differences exist due to the process of domestication, it is widely believed that the basic motivation underlying aggressive behaviour and its associated body postures are shared by the dog and its nearest phylogenetic relative, the wolf (Vilà et al., 1999; Reisner, 2002). In dogs, the classification of aggression seems to be based on two main diagnostic criteria: the context in which aggression occurs and the dog s body language (Overall, 1997; Lindsay, 2001; Mertens, 2002). For many years, most cases of dog aggression towards family members have been linked to an underlying hierarchical conflict between the dog and one or more members of the human family (O Farrell, 1992; Borchelt and Voith, 1996; Beaver, 1999b). The diagnosis of dominance-related aggression is primarily based on the situations in which aggression occurs as well as in the observation of a dominant body posture (O Farrell, 1992; Askew, 1996; Borchelt and Voith, 1996). The aforementioned paradigm for dominance-related aggression has been challenged during the past few years by an increasing number of authors. The main reason is that a significant proportion of dogs suspected to be dominant show ambivalent signals during aggressive episodes that can be better understood in terms of social stress or just as an active avoidance reaction (Overall, 1997; Guy et al., 2001; Reisner, 2002). Our results about ambivalent signals during aggressive interactions in wolves can potentially contribute to the discussion about the motivational basis of some forms of aggression problems in dogs, but ethological studies of aggressive interactions involving dogs are required. Further research is also needed to confirm our results in other groups of wolves and taking into consideration acoustic signals.

282 J. Fatjó et al. / Applied Animal Behaviour Science 105 (2007) 274 283 Acknowledgment The authors wish to thank the Affinity Foundation for the support given to the making of this study. References Askew, H.R., 1996. Treatment of Behaviour Problems in Dogs and Cats. Blackwell Science, Berlin, pp. 109 124. Beaver, B., 1999a. Canine Behaviour: a Guide For Veterinarians. W.B. Saunders, Philadelphia, pp 106 136. Beaver, B., 1999b. Canine Behaviour: a Guide For Veterinarians. W.B. Saunders, Philadelphia, pp. 137 199. Blanchard, C.D., Blanchard, R.J., 2006. Stress and aggressive behaviors. In: Nelson, R.J. (Ed.), Biology of Aggression. Oxford University Press, New York, pp. 275 291. Borchelt, P.L., Voith, V.L., 1996. Dominance aggression in dogs. In: Voith, V.L., Borchelt, P.L. (Eds.), Readings in Companion Animal Behaviour. Veterinary Learning System, Trenton, pp. 230 239. Derix, R.R.W.M, 1994. The Social Organization of Wolves and African Wild Dogs. (Die soziale Organisation von Wölfen und afrikanischen Wildhunden). PhD Thesis, Universität von Utrecht, Niederlande. Fox, M.W., 1970. A comparative study of the development of facial expressions in canids: wolf, coyote and foxes. Behaviour 36, 49 73. Gacsi, M., Gyori, B., Miklosi, A., Viranyi, Z., Kubinyi, E., Topal, J., Csanyi, V., 2005, Sep. Species-specific differences and similarities in the behavior of hand-raised dogs and wolf pups in social situations with humans. Dev. Psychobiol. 47 (2), 11 22. Guy, N.C., Luescher, U.A., Dohoo, S.E., Spangler, E., Miller, J.B., Dohoo FI.R., Bate, L.A., 2001. A case series of biting dogs: characteristics of the dogs, their behaviour, and their victims. Appl. Anim. Behav. Sci. 74, 43 57. Halfpenny, J.C., 2003. Yellowstone Wolves in the Wild. Riverbend Publishing, Helena, pp. 51 60. Harrington, F.H., Asa, C.S., 2003. Wolf Communication. In: Mech, D., Boitani, L. (Eds.), Wolves: Behavior, Ecology and Conservation. University of Chicago Press, Chicago, pp. 66 103. Lindsay, S., 2001. Applied Dog Behaviour and Training Vol. 2: Etiology and Assessment of Behaviour Problems. Iowa State University Press, Ames, pp. 229 272. Mason, W.A., 1993. The nature of social conflict: the psycho-ethological perspective. In: Mason, W.A., Mendoza, S.P. (Eds.), Primate Social Conflict. State University of New York Press, Albany, pp. 13 47. McGuire, M., Troisi, A., 1998. Darwinian Psychiatry. Oxford University Press, New York, pp. 89 117. McLeod, P.J., 1996. Developmental changes in associations among timber wolf (Canis lupus) postures. Behav. Proc. 38, 105 118. Mech, L.D., 1970. The Wolf: The Ecology and Behavior of an Endangered Species. The Natural History Press, New York. Mech, L.D., 1999. Alpha status, dominance, and division of labor in wolf packs. Can. J. Zool. 77, 1196 1203. Mech, D., Boitani, L., 2003. Wolf social ecology. In: Mech, D., Boitani, L. (Eds.), Wolves: Behavior, Ecology and Conservation. University of Chicago Press, Chicago, pp. 1 34. Mertens, P., 2002. Canine aggression. In: Horwitz, D., Mills, D., Heath, S. (Eds.), BSAVA Manual of Canine and Feline Behavioural Medicine. BSAVA, Gloucestershire, pp. 195 215. Moran, G., Fentress, J.C., Golani, I., 1981. A description of relational patterns during ritualized fighting in wolves. Anim. Behav. 29, 1146 1165. O Farrell, V., 1992. Manual of Canine Behaviour. British Small Animal Veterinary Association, Glouchestershire. Overall, K.L., 1997. Clinical Behavioral Medicine for Small Animals. Mosby, St. Louis, pp 88 137. Packard, J.M., 1989. Olfaction, ovulation, and sexual competition in monogamous mammals. In: Lakoski, J.M., Perez- Polo, J.R., Rassin, D.K. (Eds.), Neural Control of Reproductive Function. Alan R. Liss, Inc., New York, pp. 525 543. Packard, J., 2003. Wolf behavior: reproductive, social, and intelligent. In: Mech, D., Boitani, L. (Eds.), Wolves: Behavior, Ecology and Conservation. University of Chicago Press, Chicago, pp. 35 65. Reisner, I., 2002. An overview of aggression. In: Horwitz, D., Mills, D., Heath, S. (Eds.), BSAVA Manual of Canine and Feline Behavioural Medicine. BSAVA, Gloucestershire, pp. 181 194. Schenkel, R., 1947. Ausdrucks-studien an Wölfen. Behavior 1, 81 129. Schenkel, R., 1967. Submission: its features and function in the wolf and dog. Am. Zoologist 7, 319 329. Scott, J.P., 1967. The evolution of social behavior in dogs and wolves. Am. Zool. 7, 373 381. Topál, J., Gácsi, M., Miklósi, A., Virányi, Z., Kubinyi, E., Csányi, V., 2005. Attachment to humans: a comparative study on hand-reared wolves and differently socialized dog puppies. Anim. Behav. 70, 1367 1375.

J. Fatjó et al. / Applied Animal Behaviour Science 105 (2007) 274 283 283 Van Hooff, J.A.R.A.M., Wensing, J., 1987. Dominance and its behavioral measures in a captive wolf pack. In: Frank, H. (Ed.), Man and Wolf: Advances, Issues and Problems in Captive Wolf Research. Junk, Dordrecht, pp. 219 252. Vilà, C., Maldonado, J.E., Wayne, R.K., 1999. Phylogenetic relationships, evolution and genetic diversity of the domestic dog. J. Hered. 90, 71 77. Wrangham, R.W., Rubenstein, D.I., 1986. Social evolution in birds and mammals. In: Rubenstein, D.I., Wrangham, R.W. (Eds.), Ecological Aspects of Social Evolution: Birds and Mammals. Princeton University Press, Princeton, pp. 452 470. Zimen, E., 1982. A wolf pack sociogram. In: Harrington, F.H., Paquet, P.C. (Eds.), Wolves of the World: Perspectives of Behaviour, Ecology and Conservation. Noyes Publications, Park Ridge, NJ, pp. 282 322.