Stable Fighting Strategies to Maintain Social Ranks in Captive Male Alpine

bs_bs_banner

Animal Science Journal (2012) 83, 617–622 doi: 10.1111/j.1740-0929.2011.01007.x

ORIGINAL ARTICLE Stable ﬁghting strategies to maintain social ranks in

captive male Alpine musk deer (Moschus sifanicus)asj_1007 617..622

Xiuxiang MENG, Nicholas CODY, Baocao GONG and Leilei XIANG College of Life and Environmental Sciences, Minzu University of China, Beijing, China

ABSTRACT

This study was conducted at the Xinglongshan Musk Deer Farm of China from July to September 2008. Results showed that captive male musk deer exhibit aggressive dominance behavior, by which a stable social ranking is established. Generally, there were three types of aggression in agonistic interactions among males: attacking, displacing and threatening. Threatening was more frequently observed than displacing and attacking. When in conflict with other deer, high-rank males exhibited significantly more attacking than displacing and threatening. Moreover, no attacking occurred in low-rank and middle-rank males, but these individuals initiated significantly more threatening displays than high-rank individuals. Among musk deer groups with different social ranks, there were no significant differences between threats received by middle-rank and low-rank groups, but attacks directed to high-rank males was significantly lower than displacing and threatening behaviors. On the basis of these results, it is suggested that when a captive male musk deer population is assembled, individuals should be diversified in fighting ability and level of aggression. In particular, deer with higher aggression should not be enclosed with deer with similar tendencies, but should be enclosed with individuals with lower fighting levels. This should maintain stable social structures within captive musk deer groups and improve the overall welfare of captive musk deer.

Key words: Alpine musk deer (Moschus sifanicus), dominance hierarchy, ﬁght strategy, in captivity.

INTRODUCTION Wildlife farming is generally conducted through the In social animal populations, competition may occur establishment of captive populations, and compared to among individuals for limited resources, with conse- the availability of resources in nature, the temporal quent conflicts leading to the establishment of social and spatial pattern of resource distribution in captive rank patterns within such populations (Alados & Escos populations is more even and controlled (Craig 1981). 1992). A stable dominance hierarchy can lead to more Because they mate in an enclosed space, have fixed- constant social structures and individuals with higher time feeding at a consistent location and lack the dominance rank in the hierarchy will have priority ability to escape from their relatively space-limited access to resources such as food and shelter (Clutton- enclosures, captive animals tend to more strongly Brock 1982). In both wild and captive animal popula- express fighting and dominance behavior (Fraser & tions, the creation and maintenance of social rank Broom 1990). Barroso et al. (2000) reported that the will always lead to some form of conflict. The most aggression level in a population of domestic goats common forms of conflict include aggressive interac- (Capra hircus) was far stronger when the individuals tions, antagonistic interactions and assessment of were fed at the same trough in a relatively narrow an individual’s fighting ability shown by individual space. recognition, ritual display and physical contact The five species of musk deer (Moschus spp.) are (Kaufmann 1983; Barroso et al. 2000). In conflict primitive, small, solitary forest ruminants well-known interactions, the behavior displayed is related to the dominance rank of the individual that initiates the attack and the one that receives it. For example, male Correspondence: Xiuxiang Meng, College of Life and Environmental Sciences, Minzu University of China, 27 fallow deer (Dama dama) expressed more non-contact Zhongguancun Nandajie, Beijing 100081, China. (Email: attacks toward opponent deer with a lower social [email protected]) rank, but more physical attacks between individuals of Received 31 March 2011; accepted for publication 25 similar rank (Mattiangeli et al. 1999). October 2011.

© 2012 The Authors Animal Science Journal © 2012 Japanese Society of Animal Science 618 X. MENG et al. for the musk secreted by adult males, which has been dominance and fighting patterns (Rushen 1985), this used broadly in traditional Asian medicine and the research could ultimately have implications for musk perfume industry. Because of habitat degradation and deer farming, sustainable musk production and animal loss and long-term illegal hunting, musk deer have welfare. become endangered. To protect musk deer, in situ conservation measures, such as the establishment of MATERIALS AND METHODS nature reserves, have been undertaken. In addition, Experimental animals to provide ex situ protection of musk deer while maintaining sustainable musk production, musk deer This study was conducted at Xinglongshan Musk Deer Farm (XMDF) in Gansu Province, northwest China. Groups of 10 farming has been encouraged in China since the 1950s captive male alpine musk deer (Moschus sifanicus) were kept (Homes 1999; Parry-Jones & Wu 2001). As many as in two separate enclosures with five males in each enclosure. 5000 musk deer are held in more than 10 musk deer Individuals were housed in outdoor yards (100 m2), with farms and breeding centers in China (Meng et al. 2006; unrestricted access to adjoining indoor brick cells (4 m2). Sheng & Liu 2007). Neighboring enclosures were separated by wire mesh, Although behavioral studies are essential for suc- enabling olfactory and auditory communication between cessful conservation of the species, it has proven diffi- individuals, but preventing physical contact. The enclosure facilities and daily management system were the same as in cult to study the social structure and dominance Meng et al. (2003). hierarchy of wild musk deer. Due to their solitary, The musk deer were all adults but had different origins, territorial nature, along with the difficulty in accessing that is, bred in captivity, captured as fawns or adults in the their closed forested habitat (Green 1986, 1987a,b) wild, or introduced from another farm. However, the origin there are few reports on musk deer social structure in of each individual could not be determined. The detailed the wild except for some anecdotal reports and direct animal management has been reported in Meng et al. (2003). All animals were healthy and individually identified by num- descriptions of threatening and fighting behavior bered plastic ear-tags. (Zheng & Pi 1979, 1984; Ohtaishi & Sheng 1993). Through the creation of musk farming, it became pos- Behavior sampling and definition sible to study social rank and fighting patterns once Focal sampling and particular behavior recording (Altmann deer were enclosed in social groups. Some researchers 1974) were used to collect all occurrences of aggression on have observed a small number of captive musk deer a focal group during a 10-min period, with binoculars and described their general fighting patterns and (10 ¥ 42°) being used to both observe behavior and verify dominance structure (Zhang 1979, 1983), but, so far, animal identification. Recorded aggression behaviors were no quantitative research has been conducted to inves- defined as follows (i) Attacking: an obvious physical contact agonistic interaction, initiated by one individual against tigate the real characteristics of the social structure another. This category included attacks by mouth or foreleg of captive musk deer. Musk deer farming in China as well as chase attacking. (ii) Threatening: an individual encloses the normally solitary musk deer into larger standing with head up and ears erect, displaying canines to social groups to save enclosure space and reduce costs rivals with lips vibrating, and the rear part of its body wag- (Parry-Jones & Wu 2001; Meng et al. 2006). However gling. This category included threatening approaches, circling intense fighting has been witnessed in these artificial and aggressive retreats. (iii) Displacing: one individual populations, which can result in physical injury or approaches another and the latter retreats. Because conflicts among individuals typically occur during death to deer, ultimately affecting the success of musk the peak activity period of captive musk deer, behavioral production (Hu et al. 1990; Lai & Sheng 1990; Sheng & sampling was conducted from 05.00 to 08.00 hours and from Liu 2007). It was therefore hypothesized that aggres- 17.00 to 20.00 hours (Meng et al. 2002). All observations and sion levels in captive musk deer would be related to data recording were conducted by the same researcher and the fighting ability of a population’s individuals and took place 3 days a week from July to September 2008 (total that optimal social structure would lessen the aggres- of 220 observation hours). sion level, improving the success and sustainability of Statistical analysis musk deer farming (Zhang 1979). Thus far, there has been no research related to this issue, so any conclu- Musk deer dominance index (DI) was calculated as: (1/N) ¥S (W /T ) where: N = total number of opponents; sions on the topic cannot be justified. i i Wi = number of wins in agonistic interactions with opponent Therefore, it is necessary to study displays of aggres- i; Ti = total number of agonistic interactions with opponent sion and social rank patterns of musk deer in captivity. i (Bro-Jørgensen 2002). The five individuals in each enclo- In this study, it is predicted that social ranks exist in sure were categorized as high dominance-rank group (rank = stable captive musk deer populations and more than 1~2), middle dominance-rank group (rank = 3) and low one aggression type will be involved in maintaining dominance-rank group (rank = 4~5). Fight frequency was calculated as: (A/B) ¥ 100%, where: A = the number of ago- these ranks, with individuals employing fighting tech- nistic interactions initiated or received by the individual and niques according to both their own and their oppo- B = the total number of agonistic interactions initiated or nent’s social rank. As the welfare of captive animals received by all individuals in one enclosure population or and overall farm productivity is related to social group.

Table 1 Aggression initiated by captive male musk deer Rank Attacking Displacing Threatening Friedman test High-rank (n = 8) 35.71 Ϯ 7.74 30.19 Ϯ 3.87 34.11 Ϯ 6.95 c2 = 0.9; df = 2; P = 0.639 Middle-rank (n = 4) 0 26.65 Ϯ 3.84 73.35 Ϯ 3.84 c2 = 8.0; df = 2; P = 0.018 Low-rank (n = 4) 0 16.07 Ϯ 11.80 83.93 Ϯ 11.80 c2 = 6.62; df = 2; P = 0.037 Kruskal–Wallis test c2 = 12.908; df = 2; c2 = 2.014; df = 2; c2 = 10.230; df = 2; P = 0.002 P = 0.365 P = 0.006 Mann–Whitney HM: Z =-2.77, P = 0.04 HM: Z =-2.732, P = 0.04 U-test HL: Z =-2.77, P = 0.04 HL: Z =-2.39, P = 0.016 ML: Z = 0, P = 1 ML: Z =-1.176, P = 0.343 Note: data showed as percent (mean Ϯ SE). HM, high rank versus middle rank; HL, high rank versus low rank; ML, middle rank versus low rank.

The Friedman test was applied to analyze differences 100 among males of different rank in the dominance hierarchy. Attacking Displacing Threatening On finding a significant difference between groups, the 80 Wilcoxon test was used to further explore these differences. The Kruskal–Wallis test was used to explore the differences among groups in each fighting category, and the Mann– 60 Whitney U-test was utilized to determine if these differences were significant. Statistical analysis was conducted with 40 the SPSS 11.0 (SPSS Inc., Chicago, IL, USA), using two-tailed probability, with a significance level of P = 0.05. Ratio of antagonistic interaction (%) interaction antagonistic of Ratio 20

RESULTS 0 General conflict pattern in captive male High-rank Middle-rank Low-rank musk deer Figure 1 Aggression received by captive male musk deer (Note: the bars show the mean and SE). During the non-mating season, captive adult male musk deer expressed three aggressive interactions – attacking, displacing and threatening. Threatening (Kruskal–Wallis test, c2 = 12.908; d.f. = 2; P = 0.002), was more common (56.38 Ϯ 7.28%) than displacing with high-rank males attacking (35.71 Ϯ 7.74%) (25.78 Ϯ 3.66%) and attacking (17.86 Ϯ 5.94%). significantly more than middle-rank males (0, Mann– This difference was highly significant (Friedman test, Whitney U-test, Z =-2.77, P = 0.04) and low-rank c2 = 10.38; df = 2; P = 0.006). The post hoc test showed males (0, Mann–Whitney U-test, Z =-2.77, P = 0.04). that the differences between threatening and attack- The frequency of displacing was not significantly dif- ing (Wilcoxon, Z =-2.484, P = 0.013) and between ferent among males of different ranks (Kruskal–Wallis threatening and displacing frequencies (Wilcoxon, test, c2 = 2.014; df = 2; P = 0.365). However, there Z =-2.558, P = 0.011) were significant, but that was a significant difference in threatening initiated attacking was not significantly different from displac- by males of different ranks (Kruskal–Wallis test, ing frequency (Wilcoxon, Z =-0.890, P = 0.373). c2 = 10.230; df = 2; P = 0.006), with the frequency of threatening in high-rank males (34.11 Ϯ 6.95%) Relationship between fighting types and being significantly lower than that in middle-rank the dominance hierarchy rank of the (73.35 Ϯ 3.84%, Mann–Whitney U-test, high rank vs. fight initiator middle rank (HM), Z =-2.732, P = 0.04) and low-rank Ϯ As showed in Table 1, the fight types initiated by males males (83.93 11.80%, Mann–Whitney U-test, HM, of high rank were distributed evenly, and there was Z =-2.39, P = 0.016). The difference of threatening no significant difference in frequency among attack- frequency between middle-rank and low-rank ing (35.71 Ϯ 7.74%), displacing (30.19 Ϯ 3.87%) males was not significant (Mann–Whitney U-test, Z = and threatening (34.11 Ϯ 6.95%) (Friedman test, P = -1.176, P = 0.343). 0.64). However, males with middle and low ranks Relationship between fighting types expressed more threatening displays (middle-rank, 73.35 Ϯ 3.84%; low-rank, 83.93 Ϯ 11.80%) without and dominance hierarchy rank of the attacking, all differences being statistically significant fight receiver (Friedman test, middle-rank, P = 0.018; Friedman test, Aggression directed toward opposing ranks of low-rank, P = 0.037). captive male musk deer are shown in Figure 1. In Significant differences were also found in frequency aggression directed toward middle-rank males, attack- of attacking initiated by males of different ranks ing (16.40 Ϯ 8.52%) and displacing (24.23 Ϯ 2.92%)

Animal Science Journal (2012) 83, 617–622 © 2012 The Authors Animal Science Journal © 2012 Japanese Society of Animal Science 620 X. MENG et al. occurred less frequently than threatening displays 1990; Kristensen et al. 2001). For example, red deer (59.37 Ϯ 20.43%), while aggression directed toward (Cervus elaphus) possess the ability to employ diffe- low-rank males was fairly evenly distributed among rent conflict strategies according to the context and attacking (34.45 Ϯ 12.31%), displacing (27.92 Ϯ environment based on social recognition (Bebié & 3.84%) and threatening (37.63 Ϯ 11.68%). None of McElligott 2006). the above-mentioned differences were significant Although wild musk deer are solitary and territorial, (Friedman test: middle-rank, P = 0.202 > 0.01; low- in artificial populations established by farming, indi- rank, P = 0.732 > 0.01). There were significant differ- vidual musk deer express aggression, through which ences among aggression types directed against high- the stable social ranking and individual dominance rank males (Friedman test, P = 0.002 < 0.01) post hoc hierarchy is built, as with other social ungulates tests showing attacking (4.26 Ϯ 3.52%) was signifi- (Zhang 1979, 1983; Sheng & Liu 2007). The results cantly less frequent than displacing (24.60 Ϯ 6.42%) of this study showed that three types of aggression (Wilcoxon, Z =-2.032, P = 0.042 < 0.05) and threat- were expressed in conflict among individuals, and ening (71.16 Ϯ 7.63%) (Wilcoxon, Z =-2.371, P = that lower-intensity aggression, including threatening 0.018 < 0.05). The difference between displacing and displacing, was expressed more frequently than and threatening was also significant (Wilcoxon, higher-intensity attacking. The ecological functions of Z =-2.207, P = 0.027 < 0.05). conflict in animal populations are to give more domi- Differentiating the aggression types, the differences nant individuals priority access to resources such as of aggression between males with opposing ranks mates, food and shelter (Craig 1981). We found that were all insignificant (Kruskal–Wallis test: attacking, male musk deer were kept in single-sex enclosures, P = 0.151 > 0.05; displacing, P = 0.837; threatening, and provided with enough food in a timely manner. P = 0.121 > 0.05). Therefore, for deer that won conflicts, the rewards were resources which are both infinite and non- exclusive at XMDF. For individuals using high- DISCUSSION intensity types of aggression, such as the so-called Dominance behavior is common in populations of contact interaction (Clutton-Brock 1982), the reward social animals, in which one individual dominates of winning might not be greater than its cost, that is, another according to a dominance hierarchy greater risk of injury and greater time-energy budget- (Kaufmann 1983). If individuals of solitary wildlife ing. Accordingly, the captive male musk deer in this species with strong territorial behavior are enclosed study tended to express less intense aggression, that is, together to form an artificial social group, common- non-contact interactions such as displacing and threat- place in wildlife farming and domestication, a similar ening. In contrast, Zhang (1979) reported that hard social rank pattern develops. Through the establish- conflicts occurred in a newly established musk deer ment of a stable social hierarchy, dominant individuals group that caused serious injury and even death gain priority access to resources, effectively reduc- having an adverse affect on musk production. In this ing the overall aggression level in the population. study, the musk deer group was well established and However, conflict is not avoided altogether as continu- remained stable, so conflict patterns differed from ous expressions of antagonistic behavior are required those in the newly established population (Zhang to maintain stable social ranking. The type of antago- 1979). nistic behavior will be influenced by many factors, Sheng and Liu (2007) described that in conflict, such as the dominance ranks of the individual initia- submission could be expressed by the defeated musk ting the aggression and the individual receiving it deer to reduce the risk of being continuously attacked. (Alados & Escos 1992; Mattiangeli et al. 1999). A similar pattern was witnessed in this study. In con- Conflict in social populations has the potential flicts among alpine musk deer at XMDF, the losing to cause physical injury to the individuals involved, individual was always displaced by or escaped from including both the fight initiator and receiver the winner, and the conflict duration was only short. (Clutton-Brock 1982). To reduce the overall risk of This behavioral response pattern is consistent with injury and time and energy invested in conflict, studies on the domestic pig (Sus scrofa domesticus) and individuals tend to express a more optimal, lower- captive red deer (Cervus elaphus) (Rushen 1985; Bebié intensity aggression strategy, involving threatening, & McElligott 2006). displacing and ritual displaying, far more frequently Observations of captive Cuvier’s gazelles (Gazella than high-intensity physical aggression (Kaufmann cuvieri) and Dama gazelles (Nanger dama; formerly 1983; Mattiangeli et al. 1999). In captive populations, Gazella dama) indicated that maintenance of social social recognition is very important to maintain popu- ranks was based on the continuous expression of lation stability, and the inability to recognize and aggressive behavior, and the type of aggression discern the dominance hierarchy and fighting ability expressed was related to individual characteristics such of other individuals can cause conflict (Fraser & Broom as rank in the dominance hierarchy (Alados & Escos

1992; Alvarez 1993). In this study, the types of aggres- remained substantially unchanged. The establishment sion used differed among individual groups according of captive musk deer groups has been based solely on to social ranks. High-ranking individuals expressed rudimentary knowledge and observation of wild musk both low-intensity aggression and high-intensity deer. Until now, it was believed that musk deer with aggression, such as attacking. However, low-ranking low rank in the dominance hierarchy (and therefore, deer expressed lower-intensity aggression significantly poor fighting ability) would be excessively attacked. more, but with almost no high-intensity aggression. Consequently, farm managers enclosed individuals This is a pattern mirrored in captive domesticated with similar fighting ability and dominance together sheep (Ovis aries) (Barroso et al. 2000). in order to achieve mutual fight-ability, which they Moreover, because the recipient of the aggressive believed would counteract fighting. However, these display is often of a different rank in the dominance actions cause the opposite to happen, with the hierarchy, the individual initiating the fight uses types increased conflict rate causing physical injuries and of aggression of contrasting intensities. Mattiangeli even death in the captive populations. Based on the et al. (1999) reported that captive fallow deer with results of this study, when captive groups of musk deer higher rank in the dominance hierarchy tended to are established on farms they should include individu- exhibit less intense aggression, and individuals with als of different fighting abilities and different ranks similar rank and fighting ability tended to engage in within the dominance hierarchy, especially among stronger attacks. This was not observed in our study. males, which are more aggressive than females (Zhang Instead, we found that XMDF male musk deer tended 1983; Green 1987a,b). Individuals should be reformed to exhibit fewer high-intensity aggressive acts and after the captive group has been established for a more low-intensity aggression (such as threatening period of time, with a balanced dominance hierarchy and displacing) toward higher-rank individuals, but of deer of different fighting abilities and dominance exhibited an even distribution of aggression types ranks. By doing so the level of aggression among indi- toward lower-rank individuals. viduals would be lower and a stable social rank pattern In this study, calculation of the dominance index could be established and maintained in the musk and the ranking of dominance hierarchy utilized only deer population. This would benefit the animals’ the frequency of aggression displays, not the duration welfare, improve farming operations, and increase of each conflict. This may be the reason behind differ- overall musk production. ences between our findings and the fallow deer work of Mattiangeli et al. (1999). When individuals exhibit the same aggression level, the duration of conflict ACKNOWLEDGMENTS might vary widely, affecting the time and energy This research was supported by Nature Science Foun- cost of said conflict. We observed that when conflict dation of China (31170364, 30970374), Technology occurred between two male deer with a similar domi- Foundation for Selected Overseas Chinese Scholar nance rank, they displayed mutual circling and threat- (Ministry of Human Resources and Social Security ening behavior that lasted for an extended period, of the People’s Republic of China), e ‘985 Project’ potentially increasing the time and energy budget of of Minzu University of China (MUC98504-14, the conflict. However, it was far more common for MUC98507-08) and Program for New Century Excel- male musk deer to initiate attacking toward individu- lent Talents in University (NCET-08-0596). We are als of lower rank, causing the receiver of the attack to especially grateful to Dr. Kurt Johnson who helped to escape quickly, ceasing the aggression display. Thus, correct the English and improve the manuscript. the related time-energy cost and the risk of injury would be lower than similar-rank conflicts. We believe REFERENCES the findings of this study are not contradictory to the cost-benefit theory that suggests intense aggres- Alados CL, Escos J. 1992. The determinants of social status sion should occur only with high victory probabi- and the effect of female rank on reproductive success in lity and the associated quality of rewards (Alvarez Dama and Cuvier’s gazelles. Ethology, Ecology and Evolution 4, 151–164. 1993). These observations, while similar to the results Altmann J. 1974. Observational study of behavior: sampling reported for fallow deer (Clutton-Brock 1982), need methods. Behavior 49, 227–267. to be studied further to determine the details and the Alvarez F. 1993. Risks of fighting in relation to age and ecological implications to energy and time budgets. territory holding in fallow deer. Canadian Journal of Accordingly, in future studies of social rank and domi- Zoology 71, 376–383. nance, in addition to calculating direct dominance by Barroso FG, Alados CL, Boza J. 2000. Social hierarchy in the domestic goat: effect on food habits and production. fight frequency, the duration of aggression displays Applied Animal Behavior Science 69, 35–53. should also be considered. Bebié N, McElligott AG. 2006. Female aggression in red deer: Since the creation of China’s musk deer farming does it indicate competition for mates? Mammalian Biology industry, the animal husbandry systems adopted have 71, 347–355.

Bro-Jørgensen J. 2002. Overt female mate competition and Mattiangeli V, Mattiello S, Verga M. 1999. The fighting tech- preference for central males in a lekking antelope. Pro- nique of male fallow deer (Dama dama): an analysis of ceedings of the National Academy of Sciences of the United States agonistic interactions during the rut. Journal of Zoology of America 99, 9290–9293. 249, 339–346. Clutton-Brock TH. 1982. The function of antlers. Behavior 78, Meng X, Yang Q, Feng Z, Xia L, Wang P, Jiang Y, Bai Z, Li G. 108–125. 2002. Seasonal active patterns of active musk deer. Acta Craig JV. 1981. Domestic Animal Behavior. Prentice-Hall, Inc., Theriologica Sinica 22, 87–97. (In Chinese with English Englewood Cliffs, NJ. abstract) Fraser AF, Broom DM. 1990. Farm Animal Behavior and Meng X, Yang Q, Xia L, Feng Z, Jiang Y, Wang P. 2003. The Welfare. CAB International, Wallingford, UK. temporal estrous patterns of female alpine musk deer in Green MJB. 1986. The distribution, status and conservation captivity. Applied Animal Behavior Science 82, 75–85. of the Himalayan musk deer (Moschus chrysogaster). Biologi- Meng X, Zhou C, Hu J, Li C, Meng Z, Feng J, Zhou Y, Zhu Y. cal Conservation 35, 347–375. 2006. The musk deer farming in China. Animal Science 82, Green MJB. 1987a. Scent-marking in the Himalayan musk 1–6. deer. Journal of Zoology 1, 721–737. Ohtaishi N, Sheng H. 1993. Deer of China: Biology and Manage- Green MJB. 1987b. Diet composition and quality in Hima- ment. Elsevier, Amsterdam. layan musk deer based on faecal analysis. Journal of Wild- Parry-Jones R, Wu JY. 2001. Musk Deer Farming As A Conser- life Management 51, 880–892. vation Tool in China. TRAFFIC, East Asia, Hong Kong. Homes V. 1999. On the Scent: Conservation Musk Deer, the Uses Rushen JP. 1985. Stereotypies, aggression and the feeding of Musk and Europe’s Role in Its Trade. TRAFFIC Europe, schedules of tethered sows. Applied Animal Behavior Science Brussels. !4, 137–147. Hu JH, Zhang E, Sheng H. 1990. Behavioral patterns during Sheng H, Liu Z. 2007. The Musk Deer in China. Shanghai mating season in captive forest musk deer. Journal of East Science & Technology Press, Shanghai, China. (In China Normal University (Natural Science Edition) 1, 21–25. Chinese) (In Chinese with English abstract) Zhang B. 1979. The Farming of Musk Deer in China. Agriculture Kaufmann JH. 1983. On the definition and functions of Press, Beijing. (In Chinese) dominance and territoriality. Biology Review 58, 1–20. Zhang B. 1983. Musk deer: their capture, domestication Kristensen HH, Jones RB, Schofield CP, White RP, Wathes and care according to Chinese experience and methods. CM. 2001. The use of olfactory and other cues for a social Unasylva 35, 16–24. recognition by juvenile pigs. Applied Animal Behavior Zheng SW, Pi NL. 1979. Study on ecology of the musk deer Science 72, 321–333. (Moschus sifanicus). Acta Zoologica Sinica 25, 176–186. (In Lai J, Sheng H. 1990. Studies on the scent-mark in male Chinese with English abstract) forest musk deer during mating season. Journal of East Zheng SW, Pi NL. 1984. The musk sac, musk, quantity and China Normal University (Natural Science Edition) 1, 8–13. hunting of alpine musk deer (Moschus sifanicus). Acta The- (In Chinese with English abstract) riologica Sinica 4, 35–42. (In Chinese with English abstract)