Effects of caller activity and habitat visibility on contact call rate of wild Japanese macaques (Macaca fuscata).код для вставкиСкачать
American Journal of Primatology 70:1055?1063 (2008) RESEARCH ARTICLE Effects of Caller Activity and Habitat Visibility on Contact Call Rate of Wild Japanese Macaques (Macaca fuscata) HIROKI KODA, YUKIKO SHIMOOKA, AND HIDEKI SUGIURA Primate Research Institute, Kyoto University, Inuyama, Aichi, Japan A major function of contact calls in nonhuman primates is to maintain spatial cohesion among individuals in a group. The risks of spatial/visual separation from the group are likely to affect auditory contact behavior, in particular by increasing the call rate. We tested whether the risk of separation influences coo call emission by investigating the variation in call rate among behavioral contexts in two wild populations of Japanese macaques (Macaca fuscata). We focused on caller activity and the degree of visibility within the habitat as primary potential factors mediating call rate. We first estimated the habitat visibility of the two research sites at Yakushima Island (YK) and Kinkazan Island (KZ), Japan. The habitat visibility of YK was significantly more restricted than that of KZ. We then compared the call rate of 20 adult and 12 juvenile female macaques between the two wild populations to examine the potential effects of environmental differences. Both populations had a lower call rate during grooming than during feeding and moving, which are behaviors associated to higher interindividual distances. The call rate of YK adult females was significantly greater than that of both juveniles and KZ adult females, independently of activity. The call rate increased as macaques matured in the YK population, but not in the KZ population, suggesting that different developmental processes involved in contact calling of the two populations. Our findings suggest that separation risk influences call rate, and also imply a possibility of social influence that social structure change effects on the call rates. Am. J. c 2008 Wiley-Liss, Inc. Primatol. 70:1055?1063, 2008. Key words: Japanese macaques; contact calls; habitat visibility; caller activity; call rate INTRODUCTION Contact calls are used by many social animals [Bradbury & Vehrencamp, 1998]. Primates are usually organized into a wide variety of social groups, and their behaviors within a group are coordinated. In nonhuman primate social groups, contact calls serve to maintain interindividual spatial cohesion more efficiently than by visual contact, i.e., monitoring of group members [Boinski & Garber, 2000]. Several studies have shown that adult monkeys frequently produce contact calls to prevent separation from other group members [white-faced capuchins: Boinski, 1993; Boinski & Campbell, 1995; squirrel monkeys: Boinski & Mitchell, 1992; chacma baboons: Cheney et al., 1996; Rendall et al., 2000; pygmy marmosets: Cleveland & Snowdon, 1982; Japanese macaques: Itani, 1963; rhesus macaques: Rendall et al., 1996], and that juvenile monkeys frequently emit contact calls when they are separated from their mothers [white-faced capuchins: Gros-Louis, 2002; chacma baboons: Rendall et al., 2000]. Thus, one function of these calls is to prevent separation. The degree of risk of separation primarily determines the style of auditory and visual contact communication used for spatial cohesiveness. r 2008 Wiley-Liss, Inc. The behavioral context should have a primary effect on contact calling, because the degree of spacing among individuals varies broadly according to context, such as activity levels [e.g., Stevenson, 1998]. Previous studies on the contact calls of whitefaced capuchins, i.e., trill calls, demonstrated that trills were frequently produced while traveling or during activities related to spatial separation in both adults and juveniles [Boinski, 1993; Boinski & Campbell, 1995; Gros-Louis, 2002], supporting the effect of behavioral context on contact calling. Other Contract grant sponsor: Grant-in-Aid for JSPS Fellows; Contract grant number: 15-5472; Contract grant sponsor: Grant-inAid for Young Scientists; Contract grant numbers: 19730461; and 14740419; Contract grant sponsor: Grant for the Global COE; Contract grant number: A6. Correspondence to: Hiroki Koda, Department of Behavioral and Brain Science, Primate Research Institute, Kyoto University, Kanrin 41, Inuyama, Aichi 484-8506, Japan. E-mail: email@example.com Received 6 March 2008; revised 23 June 2008; revision accepted 23 June 2008 DOI 10.1002/ajp.20597 Published online 21 July 2008 in Wiley InterScience (www. interscience.wiley.com). 1056 / Koda et al. primates [e.g., ring-tailed lemurs: Oda, 1996] have shown similar patterns. Forest habitat structure is also a significant factor influencing contact calls. In natural habitats, auditory signals are subject to temporal and spectral degradation, as well as frequency-dependent attenuation, when sounds travel through the environment [Wiley & Richards, 1978]. For example, birds have clearly been shown to adjust their acoustics to habitat [e.g., Brumm, 2004; Slabbekoorn & Smith, 2002]. The influence of forest habitat structure on the acoustical call structure of several nonhuman primate species has been reported [blue monkeys, gray-cheeked mangabeys, vervet monkeys, and yellow baboons: Brown et al., 1995; pygmy marmosets: de la Torre & Snowdon, 2002; Japanese macaques: Sugiura et al., 2006; Tanaka et al., 2006; blue monkeys: Waser & Brown, 1984, 1986]. Although many studies have investigated the effects of forest habitat on the acoustic features of contact calls, little information is available on the effects of habitat on call rates or patterns. An exception is a study by Boinski and Campbell , who compared the call rates of several call types in white-faced capuchins between a wet and dry forest, and found differences in vocal usage patterns between the populations. They also found that forest habitat structure, such as visibility, affected the call rate and the emission pattern. In general, behavioral context and forest habitat structure influence contact call behavior, because the risk of spatial separation from a group varies broadly according to behavior and habitat [Boinski & Garber, 2000]. We predicted that these two factors would interact to lead to variation in the contact call rate. Regarding behavioral context, activity type should greatly impact call rate. Because individuals within a group usually disperse during foraging or moving, and show spatial cohesiveness during grooming, the degree of separation risk likely differs among these activities. Habitat visibility is a major environmental factor mediating call rate. Because monkeys more often become separated from the group in a habitat with restricted visibility, they probably increase their call rate to ensure spatial cohesiveness. However, few studies have investigated the interaction between caller activity and habitat visibility. Therefore, we examined the effects of this interaction on the contact call rates of wild Japanese macaques. Japanese macaques are suitable models for this examination for many reasons. Comparative ecological and sociological studies have been performed on this species [e.g., Yamagiwa et al., 1998]; they are widely distributed from warm- to cool-temperate zones in Japan, and they inhabit forests from the seashore to alpine regions [Agetsuma & Nakagawa, 1998]. Although habitat visibility has never before been evaluated objectively, we predict different degrees of visibility in their different habitats. In Am. J. Primatol. addition, Japanese macaques emit ??coo?? calls as contact calls to maintain group cohesion. The coo call is the most frequent call type of their vocal repertoire and is uttered in a broad range of contexts [Green, 1975; Mitani, 1986; Okayasu, 1987; Sugiura, 2007b]. Several studies on wild populations have consistently suggested that Japanese macaques adjust or modify the acoustic structure of coo calls for the efficient maintenance of spatial cohesion among individuals [Koda, 2004, 2008; Sugiura, 1993, 1998, 2007a,b; Sugiura et al., 2006; Tanaka et al., 2006]. We tested whether separation risk influenced coo call emission by investigating the variation in call rate according to caller activity and habitat structure. First, we directly evaluated the habitat visibility of the two study sites using a laser range finder. We then compared the rates of coo calls of wild Japanese macaques between two populations inhabiting forests with variable visibility as well as the three major behavioral activities, i.e., feeding, moving, and grooming. Moreover, to better clarify the effects of the interaction of caller activity and habitat visibility, we also considered caller age differences. Adult females normally remain in contact with other adult females in the group, whereas juveniles normally remain in close proximity to their mothers in this species [e.g., Nakamichi, 1996], suggesting that risk of separation varies according to age. Therefore, contact call rates should be different between adults and juveniles. This basic condition of proximity should predict that the effects of the interaction between activity and habitat on call rates will differ between adult and juvenile monkeys. METHODS The research methodology complied with protocols approved by the guidelines (Guide for the Care and Use of Laboratory Primates, Second Edition) of the Primate Research Institute, Kyoto University, Japan, and adhered to the legal requirements of Japan. Study Sites We investigated two wild populations of Japanese macaque, one inhabiting Yakushima Island (YK) in southern Japan (30.41N, 130.41E) and one at Kinkazan Island (KZ) in northern Japan (38.11N, 141.31E). The total area of Yakushima Island is approximately 503 km2, and the study site is located on the northwest coast (0400 m above sea level). The mean annual temperature is approximately 211C and annual rainfall is approximately 2,600 mm. The site is covered with warm-temperate, broad-leaved forest species. The forest canopy consists mainly of Fagaceae and Lauraceae. Deciduous trees include several species of Ficus. Thus, the diversity of plant species is very high within the study area [Agetsuma & Nakagawa, 1998]. The total area of Kinkazan Island Activity and Visibility Effects on Calls / 1057 is approximately 10 km2, with a peak at 445 m above sea level. The mean annual temperature is 111C and annual rainfall is approximately 1,500 mm. The island is covered with a mixed deciduous?coniferous forest, which includes trees such as Fagus crenata, Abies firma, and Pinus thunbergii. However, the saplings of woody plant species rarely develop into mature trees owing to high feeding pressure by Sika deer, Cervus nippon; grasslands are widely spreading in areas of the study site [Agetsuma & Nakagawa, 1998]. Both populations were equally habituated to human observers [e.g., Yamagiwa et al., 1998]. d : Evaluation of Environmental Characteristics Among the several physical environmental characteristics that may affect macaque vocal behavior, we focused on habitat visibility, i.e., the degree of visual perspective in the forest. In other words, we examined how far we could observe certain objects (e.g., trees, rocks) from an arbitrary location in the forest. We evaluated habitat visibility using the procedures that follow. We randomly chose 50 and 80 measuring locations from the home range of the subject macaque populations of YK and KZ, respectively. The home ranges of YK and KZ were approximately 70 and 100 ha, respectively; therefore, one or two measuring locations per 2 ha were carefully chosen to avoid sampling bias. We then measured the distance between the observer and the farthest object on the ground using a laser range finder (Laser Rangefinder Nikon Laser 800S, Nikon, Tokyo, Japan; measurement range 5 10?730 m; measurement accuracy 5 0.5 m; Fig. 1) at each measuring location. The observer used the laser range finder at a height of 100 cm from the ground. In the case of the farthest objects within 10 m, we directly measured the distance using a tape measure. In our measurements, the target objects were mostly stones or trees. If there were no appropriate stones or trees in any direction, then we measured the distance to the surface of the ground. The distance was defined as the maximum visible distance (MVD), and we recorded MVDs in eight directions, i.e., 0, 45, 90, 135, 180, 225, 270, and 3151, determined using a magnetic compass at each measuring location (Fig. 1). To reduce measuring error, we measured each distance five times, and adopted the average value for the MVD in that direction. All measurements were taken on sunny days from 17 May to 3 June 2005 at YK and from 18?23 July 2005 at KZ. We defined the degree of visual perspective from each measuring location. First, the visible area from each location was calculated based on the MVDs in eight directions. We estimated the area as a polygon with eight vertices (Fig. 1). By denoting the MVD at direction y as dy, we could calculate the Fig. 1. Schematic representation of the MVDs in eight directions (0, 45, 90, 135, 180, 225, 270, and 3151) and circle approximation. The star represents the standing point of the observer for measuring MVD using a laser range finder. The eight directions were measured with a magnetic compass. The visible area of the polygon (area within bold lines) and the estimated radius r were used for the circle approximation. The area of the polygon represents the S value, which is equal in size to the gray circle with radius r. The estimated radius r was used for the subsequent analysis. MVD, maximum visible distance. area of the polygon as follows: S Ό π1=2ή sin πp=4ή P 7 iΌ0 dπp=4ήi dπp=4ήπiώ1ή , where S is the polygon dimension, i is an integer, and d0 5 d2p. We calculated 50 and 80 visible areas for YK and KZ, respectively. We then approximated the visible polygon using apcircle ????????? of radius r and calculated r as S 5 pr2 or r Ό S=p. The circle approximation allowed us to interpret the result of the visible area more easily because r represented the standardized distance based on the visible area (Fig. 1). We could treat r as the degree of visual perspective at the measuring location. The distribution pattern of r values is likely to reflect the specific features of environmental characteristics at each site. We performed a Mann?Whitney test on the r values to test which habitat had better visibility. Behavioral Observations Subject populations and animals Field observations were carried out for three groups of the YK population (Kw, Kw-A, and Kw-Z) and one group of the KZ population (A) from 2002 to 2005 during the nonmating season. Kw, which was the original group on Yakushima Island, split into two groups (Kw-A and Kw-Z) between August 2003 and April 2004. In 2002 and 2003, the Kw Group was composed of about 50 individuals (range: 48?56 individuals), including 21 adult females (Z4 years), 15 or more adult males (Z4 years), 4 juvenile females (1?3 years), 4 juvenile males (1?3 years), and some unidentified monkeys. After fissioning of the Kw Group, there were 15 individuals (6 adult females, Am. J. Primatol. 1058 / Koda et al. 2 adult males, 3 juvenile females, 2 juvenile males, and 2 infants (o0 years)) in the Kw-A Group, and up to 40 (14 adult females, Z11 adult males, 3 juvenile females, and 4 juvenile males) in the Kw-Z Group in 2004 and 2005. In 2004, the Kinkazan-A Group was composed of 39 monkeys (17 adult females, 5 adult males, 5 juvenile females, 4 juvenile males, and 8 infants). We identified all subject macaques before observations. We focused on the coo calls of females. We collected data on the frequency ratio of calls from 20 adult females and 12 juvenile females from the two study areas. Juvenile macaques were defined as being between the ages of 1 and 3 years, and adult macaques were older than 6 years. The ages of all subjects in the Kinkazan-A Group and those of all juvenile subjects in the Yakushima Kw-A and Kw-Z Groups were confirmed; however, the ages of all adult female subjects in the Yakushima Groups were estimated from their appearance, a technique that is widely accepted [e.g., Yamagiwa et al., 1998]. We chose six female macaques from Kw, three from KwA, four from Kw-Z, and seven from A as subjects. Moreover, we chose six juveniles from both the Kw-Z and A Groups. As the individual differences of call rates are likely large, therefore we made a great effort to collect data from a large number of female macaques. The age and observational period of all subjects are listed in Table I. Data collection We observed subject animals using the focal animal sampling method, with a given observational session lasting more than 30 min. Observations were carried out by the three authors, with an agreement of behavioral sampling definition before observation. We performed observations so that the total observation time was more than 4 hr for each subject (for total time per subject, see Table I). Observations were during both morning and afternoon to avoid the sampling bias for each subject. Vocalizations were TABLE I. The Population, Group, Sex, Age, Age Class, and Total Observation Time of the 32 Subject Animals Subject Population Age (years)a Age class Group Total observation time (min) Kinako Mao Rabi Shima Nene Otoha Ibis Haro Kiki Kurara Shifu Atena Mariko Bera Chiro Mari Chiara Neri Julia Jun Anie Blanche Chocolat Doris Hanna Miyo Nobuko Raffie Sara Anne Fumiko Zina Kinkazan Kinkazan Kinkazan Kinkazan Kinkazan Kinkazan Kinkazan Kinkazan Kinkazan Kinkazan Kinkazan Kinkazan Kinkazan Yakushima Yakushima Yakushima Yakushima Yakushima Yakushima Yakushima Yakushima Yakushima Yakushima Yakushima Yakushima Yakushima Yakushima Yakushima Yakushima Yakushima Yakushima Yakushima 1 1 1 1 3 3 11 14 15 17 18 19 20 1 1 1 2 2 3 Middle Middle Middle Middle Middle Middle Middle Middle Middle Middle Old Old Old Juvenile Juvenile Juvenile Juvenile Juvenile Juvenile Adult Adult Adult Adult Adult Adult Adult Juvenile Juvenile Juvenile Juvenile Juvenile Juvenile Adult Adult Adult Adult Adult Adult Adult Adult Adult Adult Adult Adult Adult A A A A A A A A A A A A A Kw-Z Kw-Z Kw-Z Kw-Z Kw-Z Kw-Z Kw-Zb Kw-A Kw-Zb Kw-Z Kw-Zb Kw-Zb Kw-Z Kw-Z Kw-Z Kw-Z Kw-Ab Kw-A Kw-Zb 437 427 423 449 433 425 1,316 1,003 1,620 1,277 1,457 1,722 824 278 259 242 501 243 241 732 241 813 741 843 758 778 311 240 438 365 242 241 a Observation period 2004 2005 2005 2004 2004 2004 2004 2004 2004 2004 2004 2004 2004 2005 2005 2005 2004 2004 2004 2003 2004 2003 2004 2003 2003 2004 2004 2004 2004 2002 2004 2003 July July July July July July July July July July July July July May May May April?May April?May April?May July, 2005 May August July, 2005 May August, 2005 May July, 2005 May July, 2005 May April?May, 2004 August, 2005 May April?May August August, 2005 May July August July The ages of all subjects in the Kinkazan-A Group and all juvenile subjects in Yakushima Kw-A and Kw-Z Groups were confirmed. However, the ages of all adult female subjects in the Yakushima population were estimated. This was the Kw Group during observations. After group fission between August 2003 and April 2004, the subjects belonged to Kw-A or Kw-Z as indicated. b Am. J. Primatol. Activity and Visibility Effects on Calls / 1059 Data analysis For each subject, the call rate during each activity was calculated as a frequency ratio, which was the number of calls (times) during each activity per observation time of each activity (min). To examine the habitat, age class, and activity effects on the call rate, we performed a four-way mixedmodel analysis of variance (ANOVA) using the general linear mixed-model (GLMM) procedure with population (YK or KZ), age class (adult or juvenile), and activity (feeding, moving, or grooming) as fixed effects, and individuals as a random effect. Least significant differences (LSDs) were calculated for post hoc comparisons among the significant main effects if interactions were not significant. Moreover, post hoc comparisons were performed using the simple main effect test if some interactions were significant. Because the grooming activity of two subjects (Kinako, a KZ juvenile, and Julia, a YK juvenile) could not be observed, we treated these data as missing values in the GLMM analysis. All statistical procedures were performed using SPSS 13.01. In all ANOVAs we used the type III sums of squares for all main effects [Grafen & Hails, 2002]. The significance level was set at Po0.05. RESULTS Habitat Visibility The medians of the estimated radius r in the YK and KZ habitats were 20 and 42.5 m, respectively (Fig. 2). A Mann?Whitney U-test showed significant differences between the two sites, indicating better visibility at KZ than at YK (U 5 40, N1 5 50, N2 5 80, 100 r (m) 150 Habitat visibilities 50 counted using continuous recording methods. We recorded the types of activities in three main categories, i.e., feeding, moving, and grooming, using instantaneous sampling methods with 1-min intervals. Feeding included searching for, handling, and processing food, and moving within a food patch. Grooming comprised social grooming, self-grooming, and other grooming activities before grooming such as approaching a grooming partner or soliciting grooming. Functionally, self-grooming is not allo-grooming. However, self-grooming also frequently occurred when monkeys could not find a grooming partner. They remained alone and allo-grooming during the time that other members allo-groomed at a particular grooming site. Therefore, self-grooming was treated as grooming in the same way as allo-grooming was in our current analysis. Moving indicated movement between food patches and resting sites. Activities that were not classified as above were treated as undefined activities in the analysis. We excluded all vocalization associated with agonistic interactions or aggressive events, such as screaming, as well as controllable sounds such as sneezing. Yakushima Kinkazan Fig. 2. Estimated visible distance radius r of Japanese macaque habitat in Yakushima and Kinkazan. Box plots represent the medians (horizontal bold lines), 25th and 75th percentiles (bottom and top of box), the 1.5 interquartile range (whiskers), and outliers (circles). Po0.0001; Fig. 2). The statistical results and the distribution patterns of the radii of visibility (r values) clearly showed two characteristic properties of habitat visibility in the YK and KZ forests. First, the visual perspectives were extremely limited throughout the YK forest. Visibility in YK had a short perspective distance (median: 20 m) and small range (25th?75th percentile: 16?22 m; Fig. 2). The small range suggests poor visibility. These environmental properties likely restrict visual contact among macaques to about 20 m throughout the YK forest. In contrast, visibility in KZ had a longer perspective distance (median: 42 m) and greater range (25th?75th percentile: 35?56 m; Fig. 2). However, there were many positive outliers for habitat visibility in KZ, suggesting extremely high visibility at this location (Fig. 2). Therefore, both physical characteristics (i.e., distance and range) implied poorer habitat visibility in the YK forest than in the KZ forest. Call Rate The mean (7SE) calling rates of all observation times for the four subject groups (YK adult, YK juvenile, KZ adult, and KZ juvenile) were 0.95 (0.11), 0.51 (0.069), 0.48 (0.059), and 0.37 (0.055) times per minute, respectively. There were significant main effects of population, age class, and activity on the call rates of monkeys (Fig. 3, Table IIa). However, there was a significant two-way interaction between population and age class (Table IIa). No other interactions were significant. The lack of significant interactions involving activity suggested that population levels and age Am. J. Primatol. 1060 / Koda et al. Fig. 3. Call rates of Japanese macaques for each activity. Mean values with standard errors represent averaged values among each subject?s mean call rate. Closed circles represent the Yakushima population and open triangles represent the Kinkazan population. TABLE II. (a) Summary of Mixed-Model ANOVA Results of the Effects of Population, Age Class, and Activity on the Call Rate of Japanese Macaques Using the GLMM Procedure, (b) Summary of Post Hoc Least Significant Differences for Pair-Wise Comparisons of the Effect of Activity on the Call Rate of Japanese Macaques Using the GLMM Procedure, and (c) Summary of Tests of the Simple Main Effects of Population and Age Class on the Call Rate of Japanese Macaques Using the GLMM Procedure F P 10.0 11.8 21.2 6.94 2.63 0.36 0.72 0.004 0.002 o0.0001 0.013 0.081 0.70 0.49 Mean difference 95% Confidence interval df P Feeding vs. moving Feeding vs. grooming Moving vs. grooming 0.058 0.397 0.338 ?0.067 to 0.184 0.266?0.527 0.208?0.468 54.5 55.7 55.7 0.356 o0.0001 o0.0001 (c) Simple main effects (specific level of the other factor) df F P Population (adult females) Population (juveniles) Age class (Yakushima population) Age class (Kinkazan population) 1, 1, 1, 1, 21.6 0.12 20.9 0.29 o0.0001 0.74 o0.0001 0.59 Source df (a) Population Age class Activity Population Age class Population Activity Age class Activity Population Age class Activity 1, 1, 2, 1, 2, 2, 2, (b) Pair-wise activities 28.8 28.8 55.3 28.8 55.3 55.3 55.3 27.9 29.4 28.9 28.7 ANOVA, analysis of variance; GLMM, general linear mixed model. class did not influence the activity types differently. We performed post hoc pair-wise comparisons to calculate the LSDs among the three levels of activity. The call rate was significantly lower during grooming than during feeding or moving (Fig. 3, Table IIb). Because there were no significant interactions with activity, the effects of population and age class did not differ among the three levels of activity. In Am. J. Primatol. contrast, the significant interaction of population and age class suggested that the effect of population differed between the two levels of age class and vice versa. Therefore, we performed subsequent simple main effect tests for each level of population and age class. There was a significant effect of population for adult females and a significant effect of age class for the YK population (Table IIc). The results of both Activity and Visibility Effects on Calls / 1061 ANOVA and simple main effect tests showed a higher call rate only by the YK adult female subjects. DISCUSSION Effects of Caller Activity on Call Rate Caller activity is generally one of the primary factors regulating the call rate in a wild social group [Boinski, 1993; Boinski & Campbell, 1995; Gros-Louis, 2002]. Call rate during grooming was lower than that during feeding or moving. Because macaques disperse more and range farther during feeding and moving than during other activities, coo calls probably become more important. In contrast, macaques usually groom within close proximity to each other at grooming sites [e.g., Kawai, 1969]. Because of lower separation risk, auditory contact during grooming is less necessary than during feeding and moving. These results are supported by previous studies that also reported that wild Japanese macaques emit coo calls more frequently when they are moving and feeding than when they are grooming [Itani, 1963; Okayasu, 1987; Sugiura, 2007b]. Of further importance is the nonsignificant difference in call rate between feeding and moving. This suggests that caller activity certainly influences call rate, but also that the degree of spatial dispersion in the group is the primary factor influencing call rate, in agreement with previous studies [Boinski, 1987, 1991, 1993; Boinski & Campbell, 1995, 1996; Boinski & Garber, 2000; Boinski & Mitchell, 1992; Boinski et al., 1994]. The acoustic variation in the calls of wild Japanese macaques also supports this idea; the acoustic features of the coo call increase detectability and locatability in accordance with proximity rather than caller activity [Sugiura, 2007b]. The variation in call rate by activity in this study likely reflects the potential variation in spatial dispersion. Habitat Visibility and Age Effects The differences in call rate between the populations are primarily explained by the variation in habitat visibility. Because visibility is restricted in the YK forest compared with KZ, macaques are probably exposed to higher risk of separation from the group. To maintain group cohesion, they likely emit coo calls more frequently than KZ. In the KZ habitat, it is possible for macaques to remain in contact both visually and by calling. Consequently, the adult call rate was greater at YK than at KZ. The difference in adult female call rates between the populations is probably provoked by the difference in habitat visibility. As visual contact is impeded among macaques at YK, spatial cohesion is probably maintained to a large extent by auditory contact achieved by exchanging coo calls among individuals. Although call emissions by adult female macaques were affected by the environment, this phe- nomenon was not observed in juvenile macaques, likely because juveniles normally maintain close proximity to their mothers [Nakamichi, 1996], whereas adult female macaques maintain contact with other adult females within the group [Mitani, 1986]. Because the distances between adult female members are relatively greater than those between juveniles and their mothers, adult female contact with other members is more strongly influenced by habitat visibility. Contrary to adult females, juveniles are usually near their mothers, and their contact is not affected by habitat visibility. Thus, their call rates did not significantly differ between the two populations. These findings suggest that call rate is generally affected by habitat visibility in wild Japanese macaques. Separation Risk as a Fundamental Cause of the Underlying Variation in Call Rates Our results consistently showed that wild Japanese macaques increase their call rate when they risk losing the group, suggesting their sensitivity to spatial separation. For spatial cohesion in primate social groups, we propose a fundamental cause, namely separation risk. Indeed, separation risk could explain the variation. For example, visual separation during group travel leads to an increase in the call rate, and in several primate species, the call rate increases as distance among group members increases [Boinski, 1991, 1993; Boinski & Campbell, 1995; Boinski & Mitchell, 1992; Cheney et al., 1996; Rendall et al., 2000; Snowdon & Hodun, 1981]. These findings, along with our current results, suggest that auditory contact effectively serves to maintain spatial cohesion, especially to compensate for visual contact under conditions of low visibility. Our study indicates that Japanese macaques are sensitive to spatial separation, which is reflected in their calls, in particular in acoustic modifications of their contact calls according to the degree of spatial separation. In both naturalistic observations and experimental settings, increasing the distance between caller and recipient(s) led to more detectable, longer-lasting, and/or higher-pitched calls [Masataka & Symmes, 1986; Oda, 1996; Snowdon & Hodun, 1981; Sugiura, 2007b]. These findings both indicate sensitivity to separation risk. Boinski and Garber  proposed a similar idea that vocal communication by contact calling is primarily determined by the ??motivation to coordinate troop travel.?? They suggested that increased motivation to coordinate troop travel leads to higher call rates, louder calls, longer travel signals, more fluctuation and modulation of acoustic signals, and more exaggerated movements and monitoring of recipient responses. Motivation to coordinate troop travel could be interpreted as sensitivity to separation risk. Such internal motivation or sensitivity likely determines Am. J. Primatol. 1062 / Koda et al. the type of vocal communication used in wild primate social groups. Further Directions Another important contribution of this study is the first systematic evaluation of habitat visibility. Our procedure successfully described the qualitative differences in habitat visibility between the two research sites, considering previous data on vegetation sampling [Maruhashi et al., 1998]. They carried out vegetation sampling at YK and KZ Islands, and found that the density of trees at YK was 13.4 times that at KZ, and that the density of small trees was particularly greater at YK, suggesting lower visibility there. Furthermore, grasslands were spreading at the KZ study site. The high density of trees corresponds to our results for YK habitat visibility, whereas the low density of trees and grasslands corresponds to our results for KZ. The importance of forest structure on contact calls has been strongly suggested [Boinski & Garber, 2000], but habitat visibility has never before been considered. Our procedure could be applied together with geographic information system techniques in future studies on audiovisual contact communication in wild populations. Finally, another factor that should not be overlooked is social influence or effect of individual life histories on matrilineal society. The social framework influences intragroup vocal communication [e.g., Snowdon & Hausberger, 1997]. For example, call usage is developmentally acquired via social learning from adults [e.g., Seyfarth & Cheney, 1986]. Social status also likely determines the pattern of vocal communication. Adult male primates, including Japanese macaques, usually produce fewer contact calls than adult females [Boinski, 1991; Boinski & Garber, 2000; Mitani, 1986]. Such sex differences in contact calls are no doubt provoked by sex differences in social status or role for group cohesiveness. We should extend our current findings in future research by considering social status, social life history, and social dynamic change. For example, our data might have been affected by variable social structures of the four subject groups, as we can expect that the call rate can be influenced by the numbers of potential partners among a group. Increasing the call rate in YK adult females might have been provoked by the social structure of large Kw-Z Group. Although the juvenile call rate is generally higher than that of the adult females in other primate species [e.g., Boinski & Campbell, 1995, 1996; Gros-Louis, 2002], our contrary result for YK might be explained by such a social structure variation. To examine such social influences on call rates, we have to investigate more groups of a greater variety of compositions and sizes. The comparison of call rates of troops at prefission Am. J. Primatol. and postfission is also appropriate. Although we had insufficient data to address this factor, members of the split group (Kw-A), which was smaller than the original group (Kw), might have gradually decreased their call rates because this split small group also seems to have become more spatially cohesive. Future research should consider other factors such as social influences on vocal communication of wild populations. ACKNOWLEDGMENTS We appreciate helpful comments from Professor Nobuo Masataka and Dr. Yasuyuki Muroyama throughout the study, and we thank Professor Kosei Izawa, Dr. Masaki Shimada, Dr. Yamato Tsuji, and the researchers of Kinkazan Island for information on the Kinkazan-A Group. We also thank the researchers and residents of Yakushima Island for their help, the Yakushima Environment Conservation Center for permission to perform the research, and the field research center of KUPRI for permission to stay at the field station in Yakushima Island during the research. The manuscript was considerably improved thanks to Dr. M. A. Huffman and two anonymous referees. The research methodology complied with protocols approved by the guidelines (Guide for the Care and Use of Laboratory Primates, Second Edition) of the Primate Research Institute, Kyoto University, Japan, and adhered to the legal requirements of Japan. This work was supported by a Grant-in-Aid for JSPS Fellows to H. K. (15-5472), a Grant-in-Aid for Young Scientists (B) to H. K. (19730461) and H. S. (14740419), and a Grant for the Global COE (A6) to Kyoto University, MEXT, Japan. REFERENCES Agetsuma N, Nakagawa N. 1998. Effects of habitat differences on feeding behaviors of Japanese monkeys: comparison between Yakushima and Kinkazan. Primates 39:275?289. Boinski S. 1987. Birth synchrony in squirrel monkeys (Saimiri oerstedi)?a strategy to reduce neonatal predation. Behav Ecol Sociobiol 21:393?400. Boinski S. 1991. The coordination of spatial position?a field study of the vocal behavior of adult female squirrel monkeys. Anim Behav 41:89?102. Boinski S. 1993. Vocal coordination of troop movement among white-faced capuchin monkeys, Cebus capucinus. Am J Primatol 30:85?100. Boinski S, Campbell AF. 1995. Use of trill vocalizations to coordinate troop movement among white-faced capuchins? a 2nd field test. Behaviour 11?12:875?901. Boinski S, Campbell AF. 1996. The huh vocalization of whitefaced capuchins: a spacing call disguised as a food call? Ethology 102:826?840. Boinski S, Garber P, editors. 2000. On the move: how and why animals travel in groups. Chicago: The University of Chicago Press. Boinski S, Mitchell CL. 1992. Ecological and social factors affecting the vocal behavior of adult female squirrel monkeys. Ethology 92:316?330. Boinski S, Moraes E, Kleiman DG, Dietz JM, Baker AJ. 1994. Intragroup vocal behavior in wild golden lion tamarins, Activity and Visibility Effects on Calls / 1063 Leontopithecus rosalia?honest communication of individual activity. Behaviour 1?2:53?75. Bradbury JW, Vehrencamp SL. 1998. Principles of animal communication. Sunderland, MA: Sinauer. Brown CH, Gomez R, Waser PM. 1995. Old World monkey vocalizations?adaptation to the local habitat. Anim Behav 50:945?961. Brumm H. 2004. The impact of environmental noise on song amplitude in a territorial bird. J Anim Ecol 73:434?440. Cheney DL, Seyfarth RM, Palombit R. 1996. The function and mechanisms underlying baboon ?contact? barks. Anim Behav 52:507?518. Cleveland J, Snowdon CT. 1982. The complex vocal repertoire of the adult cotton-top tamarin (Saguinus oedipus oedipus). Z Tierpsychol 58:231?270. de la Torre S, Snowdon CT. 2002. Environmental correlates of vocal communication of wild pygmy marmosets, Cebuella pygmaea. Anim Behav 63:847?856. Grafen A, Hails R. 2002. Modern statistics for the life sciences. Oxford: Oxford University Press. Green S. 1975. Variation of vocal pattern with social situation in the Japanese monkey (Macaca fuscata): a field study. In: Rosenblum LA, editor. Primate behavior. New York: Academic Press. p 1?102. Gros-Louis J. 2002. Contexts and behavioral correlates of trill vocalizations in wild white-faced Capuchin monkeys (Cebus capucinus). Am J Primatol 57:189?202. Itani J. 1963. Vocal communication of the wild Japanese monkey. Primates 4:11?66. Kawai M. 1969. Ecology of Japanese monkeys (Nihonzaru no seitai). Tokyo: Kawade Shobo (in Japanese). Koda H. 2004. Flexibility and context-sensitivity during the vocal exchange of coo calls in wild Japanese macaques (Macaca fuscata yakui). Behaviour 141:1279?1296. Koda H. 2008. Short-term acoustic modifications during dynamic vocal interactions in nonhuman primates?implications for origins of motherese. In: Masataka N, editor. The origins of language. Tokyo: Springer. p 59?73. Maruhashi T, Saito C, Agetsuma N. 1998. Home range structure and inter-group competition for land of Japanese macaques in evergreen and deciduous forests. Primates 39:291?301. Masataka N, Symmes D. 1986. Effect of separation distance on isolation call structure in squirrel monkeys (Saimiri sciureus). Am J Primatol 10:271?278. Mitani M. 1986. Voiceprint identification and its application to sociological studies of wild Japanese monkeys (Macaca fuscata yakui). Primates 27:397?412. Nakamichi M. 1996. Proximity relationships within a birth cohort of immature Japanese monkeys (Macaca fuscata) in a free-ranging group during the first four years of life. Am J Primatol 40:315?325. Oda R. 1996. Effects of contextual and social variables on contact call production in free-ranging ringtailed lemurs (Lemur catta). Int J Primatol 17:191?205. Okayasu N. 1987. Coo sound communication. Quaternary Anthropol 19:12?30 (in Japanese). Rendall D, Rodman PS, Emond RE. 1996. Vocal recognition of individuals and kin in free-ranging rhesus monkeys. Anim Behav 51:1007?1015. Rendall D, Cheney DL, Seyfarth RM. 2000. Proximate factors mediating ??contact?? calls in adult female baboons (Papio cynocephalus ursinus) and their infants. J Comp Psychol 114:36?46. Seyfarth RM, Cheney DL. 1986. Vocal development in vervet monkeys. Anim Behav 34:1640?1658. Slabbekoorn H, Smith TB. 2002. Habitat-dependent song divergence in the little greenbul: an analysis of environmental selection pressures on acoustic signals. Evolution 56:1849?1858. Snowdon C, Hausberger M, editors. 1997. Social influences on vocal development. New York: Cambridge University Press. Snowdon CT, Hodun A. 1981. Acoustic adaptations in pygmy marmoset contact calls?locational cues vary with distances between conspecifics. Behav Ecol Sociobiol 9: 295?300. Stevenson PR. 1998. Proximal spacing between individuals in a group of woolly monkeys (Lagothrix lagotricha) in Tinigua National Park, Colombia. Int J Primatol 19:299?311. Sugiura H. 1993. Temporal and acoustic correlates in vocal exchange of coo calls in Japanese macaques. Behaviour 124:207?225. Sugiura H. 1998. Matching of acoustic features during the vocal exchange of coo calls by Japanese macaques. Anim Behav 55:673?687. Sugiura H. 2007a. Adjustment of temporal call usage during vocal exchange of coo calls in Japanese macaques. Ethology 113:528?533. Sugiura H. 2007b. Effects of proximity and behavioral context on acoustic variation in the coo calls of Japanese macaques. Am J Primatol 69:1412?1424. Sugiura H, Tanaka T, Masataka N. 2006. Sound transmission in the habitats of Japanese macaques and its possible effect on population differences in coo calls. Behaviour 143: 993?1012. Tanaka T, Sugiura H, Masataka N. 2006. Cross-sectional and longitudinal studies of the development of group differences in acoustic features of coo calls in two groups of Japanese macaques. Ethology 112:7?21. Waser PM, Brown CH. 1984. Is there a sound window for primate communication? Behav Ecol Sociobiol 15:73?76. Waser PM, Brown CH. 1986. Habitat acoustics and primate communication. Am J Primatol 10:135?154. Wiley RH, Richards DG. 1978. Physical constraints on acoustic communication in atmosphere?implications for evolution of animal vocalizations. Behav Ecol Sociobiol 3:69?94. Yamagiwa J, Izawa K, Maruhashi T. 1998. Long-term studies on wild Japanese macaques in natural habitats at Kinkazan and Yakushima: preface. Primates 39: 255?256. Am. J. Primatol.