Effects of group size on birth rate, infant mortality and social interactions in Formosan macaques at Mt Longevity, Taiwan M.J. Hsu 1, J.F. Lin 1,2 and G. Agoramoorthy 3,4 1 Department of Biological Sciences, National Sun Yat-sen University, Kaohsiung 804, Taiwan 2 Shi-Pu Junior High School, Kaohsiung 840, Taiwan 3 Department of Pharmacy, Tajen University, Yanpu, Pingtung 907, Taiwan Received 5 October 2004, accepted 16 November 2005 In social animals, the influence of life-history traits on group structure and size is potentially important in the context of the evolution of social systems. Birth rate and infant survival have been differently predicted as a function of group size by models of the evolution of group living in primates. A wild population of the Formosan macaque, Macaca cyclopis that inhabits Mt Longevity, Taiwan has been monitored since July 1993 to collect demographic, reproductive and social behaviour data. Little is known about the effects of troop size on population growth in Formosan macaques. In this paper, we have presented for the first time 5-years of data (1997-2001) on the troop dynamics of 14 social troops. We also examined the effects of troop size on birth rate, infant mortality and inter-troop social interactions. The annual average troop size increment fluctuated from 1.94% to 10.97%. The annual average troop size of 14 groups was 43.57 ± 15.88 individuals in 1997, and increased to 55.97 ± 16.86 individuals in 2001. The per-capita rates of troop size increase was negatively correlated with troop size without infants (F1,12 = 5.57, P < 0.05). The average birth rate (birth per adult female) was 0.78 (± 0.07, n = 14) and it was not linearly related to troop size, nor the number of adult females (P > 0.33). The average infant mortality was 0.16 (± 0.09, n = 14), which was relatively constant regardless of the troop size category (small, medium or large). Intertroop dominance was closely dependent on the number of adult males and females rather than group size. Thus larger group size imposed an advantage on habitat utilization without the appearance of low birth rate or high infant mortality. On the other hand, infant mortality increased and infant survival per adult female decreased when the percentage of adult females increased in troops. Therefore, competition among adult females plays an important role partially supporting the predation — intra-group feeding competition hypothesis. key words: Formosan macaque, Macaca cyclopis, Taiwan, troop size, infant mortality, birth rate, social interaction. Ethology Ecology & Evolution 18: 3-17, 2006 4 Corresponding author (E-mail: agoram@mail.nsysu.edu.tw). 4 M.J. Hsu, J.F. Lin and G. Agoramoorthy Introduction . . . . . . . . . . . . . . . . . 4 Materials and methods . . . . . . . . . . . . . . 5 Results . . . . . . . . . . . . . . . . . . 7 Troop size and growth rate . . . . . . . . . . . . . 7 Natality . . . . . . . . . . . . . . . . . . 9 Infant mortality . . . . . . . . . . . . . . . 10 Inter-troop interactions . . . . . . . . . . . . . . 11 Discussion . . . . . . . . . . . . . . . . . . 13 Troop size and effect on demography . . . . . . . . . . 13 Social interactions . . . . . . . . . . . . . . . 15 Acknowledgments . . . . . . . . . . . . . . . . 15 References . . . . . . . . . . . . . . . . . 16 INTRODUCTION Most diurnal primates exhibit a wide range of social grouping patterns. The groups may vary widely in size, age and sex classes and cohesiveness. The benefits that result from group living and the extent to which animals of different ages and sexes share these benefits, remain an area of contention (Chapman et al. 1995, Lefebvre et al. 2003). Macaques are an excellent taxonomic group to address such issues since most species form cohesive groups and exhibit variation in group size and structure. Birth rate and immature survival rate have been differently predicted as a function of group size by models of the evolution of group living in primates. According to Wrangham’s (1980) inter-group feeding competition hypothesis, large groups have an advantage in accessing resources. This hypothesis predicts that birth rate may exhibit a humped curve against group size of female-bonded primates, and it indicates that the infant per female rate is expected to be higher in medium-sized groups and lower in large- and small-sized groups (van Schaik 1983). In contrast, the predation — intra-group feeding competition hypothesis proposed by van Schaik (1983) suggests that the birth rate may decrease with group size when intra-group competition becomes very intense (Sterck et al. 1997). This hypothesis predicts that infant per female rate should decrease when group size increases. In addition, when natural predators occur, mortality among immature individuals should be higher in smaller-sized groups of social primates (van Schaik 1983). Nevertheless, the controversy continues concerning the interpretation of the correlation between reproductive parameters and troop size in female-bonded primates (Janson & Goldsmith 1995, Kumar 1995, Takahata et al. 1998a). Among the 21 extant species in the widespread genus Macaca, the Formosan macaque, Macaca cyclopis, is one of the least known. It is endemic to the island of Taiwan (area 36000 km2) and only a few field studies have attempted to understand its population dynamics, reproduction and ecology (reviewed in Hsu & Lin 2001). Despite its wide distribution at various altitudes ranging from 100 to 3300 m, hunting pressure was substantial with mass captures of about 1000 macaques annually for the purpose of medical research, the preparation of skeletons, Chinese medicine (skeletal jelly), and the pet trade in the past (Peng et al. 1973, Masui et al. 1986). Illegal poaching still continues despite the fact that this species is listed in the IUCN (2003) Red List of Threatened Animals and protected by Taiwan’s Wildlife Conservation Law (1989). Information on troop dynamics, life histories, social behavior and reproduction are therefore important for the conservation of this endemic species in Taiwan. Effects of group size in wild Formosan macaques 5 We have been monitoring a population of M. cyclopis that inhabits the lowland rainforest at Mt. Longevity, southern Taiwan since July 1993 (Hsu et al. 2000, 2001, 2002; Hsu & Lin 2001). The majority of the troops have a multi-male/multi-female structure, which makes the population interesting for socio-ecological studies. Data on birth seasonality and inter-birth intervals of this population were published (Hsu et al. 2001), but little is known about the effects of troop size on population growth. The aim of this paper is to present data on the demographic parameters of multi-male/multi-female social troops of wild Formosan macaques that inhabit Mt. Longevity, Taiwan. We examined the effects of troop size on birth rate, infant mortality and inter-troop social interactions based on the predictions of the existing hypotheses concerning the evolution of group living in primates (Wrangham 1980, van Schaik 1983). We attempted to test whether intra-group female competition reduced birth rate or increased infant mortality in social troops of wild Formosan macaques. Materials and methods Study site Mt Longevity (22°39’N, 120°15’E) is located in Kaohsiung City adjacent to the Taiwan Strait and it has been isolated from other mountains over the last 50 years by to urban development. Its natural forest has been preserved since the beginning of the last century mainly because it was an active military base during the Japanese occupation of Taiwan (Hsu et al. 2001). The public has had access to this mountain only since 1989 and the northern part is still an active military base to which civilians including researchers have no access. The protection of the natural forest by the Taiwan military has left the habitat less disturbed. The flora includes 215 species in 73 families and 176 genera. The total surface area of Mt Longevity has been estimated as 35 km2. We used a topographic map to calculate the total area used by all macaque troops that includes several raised coral reefs, undulating hillocks and valleys. The study site, excluding the restricted area occupied by the military, covers about 25 km2. According to the records of the Central Weather Bureau in Kaohsiung, the average annual precipitation between 1997 and 2001 was 2288 mm (SD ± 465), concentrated from June to August. The wet season, with monthly average above 100 mm, was from May to October (Fig. 1). The dry season started in November and ended in March with a monthly average rainfall below 50 mm. This pattern fits the characteristics of the seasonal variations in rainfall demonstrated by Walter & Leith (1967). The average monthly temperature was coldest in January (19.8 oC) and highest in July (28.8 oC). The average monthly humidity was above 75% throughout the year. Data collection A thorough search of the forest, using trails, footpaths and line transects, enable us to identify and follow 14 troops from 1997-2001 (Table 1). Some of these troops had also been monitored since 1994 (Hsu & Lin 2001). We classified individuals into broad age and sex classes based on direct measurement of chronological age from known birth year or based on physical characteristics and body size (National Research Council 1981, Hsu & Lin 2001). Most females give birth to their first infants around 5 years of age and we considered adult females were equal to, or older than, 5 years. Subadult males were between 5-6 years, and their secondary sexual characteristics not as fully developed as they are in adult males. Adult males were equal to, or older than, 7 years. Individuals aged between 1 and 4 years were 6 M.J. Hsu, J.F. Lin and G. Agoramoorthy considered juveniles. Infants of both sexes were typically nursing, cared for by their mothers, and weaned at around six months of age (Hsu et al. 2000). Artificial feeding of monkeys at Mt Fig. 1. — Monthly variations in rainfall and temperature at Mt Longevity Taiwan between 1997 and 2001. Table 1. Demographic parameters including troop size, birth rate and infant mortality in Formosan macaques at Mt Longevity (1997-2001). Data were indicated as mean (± SD). Troop name No. adult males No. adult females Troop size (without infants) Troop size Birth rate (%) Infant mortality (%) E 4.2 ±1.0 10.8 ±3.5 28.1 ±6.5 32.7 ±8.5 79.2 ±16.0 28.7 ±29.9 Ia 4.4 ±0.7 8.4 ±1.4 29.4 ±2.1 33.1 ±2.5 77.1 ±16.7 17.7 ±26.7 N 4.7 ±1.0 9.3 ±1.8 31.0 ±1.6 34.9 ±1.6 64.1 ±18.4 5.0 ±11.2 M 3.9 ±1.3 9.5 ±1.0 31.8 ±3.6 36.4 ±3.4 73.7 ±14.9 3.3 ±7.5 A 5.2 ±1.4 9.6 ±0.4 34.1 ±1.2 38.3 ±1.5 83.3 ±8.4 22.1 ±18.4 D 5.0 ±0.5 10.3 ±2.2 35.4 ±8.1 40.6 ±8.5 85.5 ±13.5 6.0 ±13.4 O 6.9±0.9 12.6±1.5 42.0±6.5 47.8±7.1 73.3±18.1 4.4±9.9 J 5.4 ±1.0 13.5 ±2.7 43.1 ±6.1 49.5 ±7.2 90.8 ±8.6 21.8 ±11.8 C 7.3 ±1.2 16.1 ±2.2 46.5 ±3.2 51.6 ±3.5 68.7 ±9.4 27.3 ±14.9 B 8.7 ±1.1 16.2 ±3.4 48.5 ±6.7 54.8 ±8.6 79.2 ±10.2 26.0 ±15.5 K 9.3 ±1.8 16.8 ±3.1 52.8 ±5.1 61.0 ±5.8 80.5 ±18.8 20.3 ±11.9 F 8.1 ±1.7 18.7 ±4.1 59.2 ±4.9 67.2 ±6.8 77.1 ±5.2 25.6 ±12.7 G 7.2 ±1.7 18.9 ±2.6 60.7 ±4.4 69.8 ±5.9 79.1 ±9.2 10.2 ±6.6 I 7.8 ±1.5 20.0 ±4.4 66.7 ±9.6 77.1 ±10.7 85.0 ±10.2 12.5 ±12.0 Effects of group size in wild Formosan macaques 7 Longevity is legally prohibited but some troops have home ranges that overlap with tourist trails and receive occasional food items from visitors. We conducted a systematic census to record troop composition changes twice a week between January 1997 and December 2001. We included a temporary all-male group in the troop with which it was most closely associated. We used focal animal sampling and scan sampling (Lehner 1996) to record data on social interactions. We recorded inter-troop interactions that included fight, chase, and displacement during close encounters between 1998 and 2001 to determine winner-loser or uncertainty and to establish an inter-troop dominance hierarchy (Lehner 1996). Calculation and analysis Troop size at the end of each year was compared with that at the beginning by calculating the annual per-capita rate of troop size increase. The mean annual per-capita rate of troop size increase from 1997 to 2001 was calculated as the average annual growth rate. The mean size of a troop for 5 years was the average of the annual average troop sizes from 1997-2001. The time of conception was backdated from birth, assuming a gestation length of 163 days (Hsu et al. 2001). All means are presented as the values ± 1 standard deviation. Birth rate was calculated as the total number of births divided by the availability of sexually mature females (≥ 5 years) in a troop, which is the crude birth rate. Infant weaning rate was the number of infants weaned divided by the available number of sexually mature females (≥ 5 years) in a troop. We calculated the percentage of winning events during aggressive conflicts as the total number of winning events plus half of the counter attack events divided by the total number of aggressive interactions of a troop. We conducted all statistical analyses through Statistical Analysis System software (SAS Institute Inc. 2000). We calculated Spearman rank correlation coefficients among demographic parameters (Table 1). The Spearman correlation coefficient indicated that the mean number of adults (males and females), troop sizes with or without infants, were significantly correlated to each other (P < 0.001). Therefore, we used only one independent variable (base on the higher R2) in various regression analyses to test the relationships with dependent variables. Dependent variables included the number of births, the number of surviving infants, the birth rate, the infant mortality, the infant survival rate, the number of infants weaned per adult female, and the percentage of winning shown by the troop. We used analysis of variance (ANOVA F test) within various General linear models to test the effects of the independent variables (troop size with or without infants, number of adult females or proportion of adult females) and categories (troop, year) on the dependent variables. Results Troop size and growth rate The annual average troop size for 14 troops fluctuated from an increment of 10.97% to as little as 1.94%. In 1997, the troop size was 43.57 ± 15.88 individuals (range 20.4-78.3) but it increased to 48.35 ± 16.96 by 1998. Subsequently, a slight increased was recorded between 1998 and 2000 (2.61 % to 1.94%), but it stayed as 49.61 ± 12.01 to 50.57 ± 12.65 in 2000. Nonetheless, the annual average troop size was 55.97 ± 16.86 individuals (range 34.8-88.4) during 2001, which was an increment of 10.68% compared to the previous year. Most troop sizes were between 30 and 70, which included 85.7% of troop sizes (Fig. 2). The largest troop (I) recorded in our study had 100 individuals in July 1998. 8 M.J. Hsu, J.F. Lin and G. Agoramoorthy All the 14 troops had a multi-male/multi-female structure with an average adult sex ratio (AM/AF) of 0.48 (± 0.07, range 0.38-0.57). When subadult males, juveniles and infants were included, the mean ratio of total males to females per troop was 1.02:1. The mean proportion of adult females in 14 troops was 27.4% (± 2.2, range 25.2-32.6%). Even though it was the highest in the smallest troop (E), the mean proportion of adult females was not linearly correlated (P > 0.51) with mean troop size (range 32.7 to 77.1, Table 1). The per-capita rate of troop size increase fluctuated widely, from 1.61 to 0.80 (Fig. 3). It was negatively correlated with troop size (with and without infants) in 1997-1999 (P < 0.05) and the relationship did not change significantly within years (P > 0.05). Troop size without infants (F1, 55 = 36.79, P < 0.001) and troop (F13, 55 = 2.55, P < 0.01) accounted for 52.0% of the variations of per-capita rate increase. Therefore, we further used mean troop size without infants as an independent variable and found the significant effect of that on mean per-capita rate increase in troops (F1, 12 = 5.57, P < 0.05). The regression model to estimate the per-capita rates of troop size increase was 1.22867-0.00316 multiple (troop size without infants). The annual mean per-capita rate of troop size increased in 1997/1998 (1.12 ± 0.11) but slowed down during 1998/1999 (1.04 ± 0.11) and 1999/2000 (1.03 ± 0.08). Nevertheless, it later increased to 1.11 ± 0.08 in 2000/2001. Overall, the average percapita rate of troop size increase was 1.07 (± 0.05, n = 14) during 1997-2001. The relation of per-capita rate of troop size increase was not consistently related to the number of adult females; the relationship correlated negatively in 1997 (rs = – 0.73, P < 0.01) but positively in 2000 (rs = 0.61, P < 0.05). The per-capita rate of troop size increase was not significantly correlated with the troops’ birth rate from 1997- 2001 (P > 0.18, n = 14 for each year). Fig. 2. — The frequency distributions of troop sizes in the Mt Longevity Formosan macaque as of December each year between 1997 and 2001. Effects of group size in wild Formosan macaques 9 Natality The birth peak occurred in May (47.7% of 705 births), and the majority of births (81.1%; n = 519) occurred within a period of 2 months (April 15-June 15). The overall mean birth date was May 11 ± 3 days and the average median date was May 8 ± 5 days. The mean spread for five birth seasons was 142 ± 28 days. Only two births occurred in February and three in August. The earliest birth date was recorded on 17 February (1999) and the latest was on 15 August (2001). The conception period ranged mainly from November to December (80.3%). The mean number of births per troop was significantly correlated with the mean number of adult females (F1, 12 = 208.3, R2 = 0.95, P < 0.001) and mean troop size without infants (F1, 12 = 112.4, R2 = 0.90, P < 0.001). Since the intercept did not significantly differ from zero, the best regression line to estimate the number of births was: = 0.7905 • (number of adult females). The mean birth rate was neither significantly related to average troop size with or without infants (both rs = 0.24, P > 0.40), nor with the number of adult females (rs = 0.28, P > 0.33). In addition, the mean birth rate did not show a humped curve against group size and the variations in birth rates were larger in small and medium-sized troops (Fig. 4). The average birth rate (birth per adult female) was 0.78 (± 0.07, Table 1). Furthermore, the mean proportion of infants within a troop did not correlate with neither the number of adult females (P > 0.51) or the troop size without infants (P > 0.88). The average proportion of infants in 14 troops was 16.7% (± 1.8, range 13.3-19.5%). The mean proportion of infants was not significantly correlated with the mean birth rate (P > 0.07). Fig. 3. — The relationship between average troop size and the per-capita rates of size increase in 14 troops of Formosan macaques at Mt Longevity during the period 1997 to 2000. 10 M.J. Hsu, J.F. Lin and G. Agoramoorthy Infant mortality A total of 88 infants died within 6.5 months of age with 39.8% of deaths occurring on the day of parturition and a further 21.6% occurred within 2 weeks of age. In total, 69 infants died within 2 months of age and 1/3 of them (n = 23) could be associated with an alpha male change during the previous mating season. However, the alpha male change in troops did not itself increase infant mortality. The annual infant mortality between 1997 and 2001 were not correlated with troop size. Even after we separated troop sizes into three categories (small, medium and large), the relationships between troop size and infant mortality were not consistent. Infant mortality was highest in medium-sized troops in 1997, but it became the lowest in 2000 (Fig. 5). However, no significant difference was found in mean infant mortality among these three categories of troop sizes (Wilcoxon test, P > 0.10). The average infant mortality for 14 troops was 0.16 ± 0.09 (Table 1). Mean infant mortality from the 14 groups was significantly correlated with the percentage of adult females per troop (rs = 0.635, P < 0.05; Fig. 6), but not correlated with troop size (P > 0.77), number of adult females (P > 0.25) nor number of adult males (P > 0.32). The regression line to estimate infant mortality equals – 56.28 + 2.651 multiple (percentage of adult females in a troop) and infant mortality was not linearly correlated with mean birth rate (P > 0.48, n = 14). The average number of infants weaned per adult female (infant weaning rate) was 0.65 ± 0.08 (n = 14). It did not linearly correlate with mean troop size (with or without infants) in 14 troops (P > 0.51). The mean infant weaning rate was negaFig. 4. — The relationship between mean birth rate and the average number of individuals in the troop expressed in terms of the number of adult females, troop size without infants, and total troop size for 14 troops of Formosan macaques at Mt Longevity between 1997 and 2001. Effects of group size in wild Formosan macaques 11 tively correlated with the mean percentage of adult females (rs = – 0.71, P < 0.01) and adult males (rs = – 0.55, P < 0.05, Fig. 7). The best regression model to estimate infant weaning rate was 1.3454 - 0.02522 • (percentage of adult females per troop) (F1, 12 = 12.34, R2 = 0.51, P < 0.005). The proportion of adult males per troop also had negative effect (Fig. 7) on the mean infant weaning rate (F1, 12 = 4.79, R2 = 0.28, P < 0.01), but the effect was not significant when the percentage of adult females was included in the regression model (P > 0.08). Inter-troop interactions Between 1998 and 2001, we recorded 3050 troop interactions among 14 social troops. The average number of troops interacted with was 6.6 ± 2.5 troops (n = 14). Direct aggression accounted for 78.5% to 88.5 % of all interactions. Counter attacks occurred on only a few occasions (1.55-8.3%), and no winning troop was determined. Occasionally, two troops were seen close to each other peacefully without apparent aggression (11.5-21.5%). At Mt Longevity, the dominance rank in troops changed periodically depending on the year, location and number of individuals involved. In 1998, troop I was the most dominant troop in the southern part of Mt Longevity and it displaced other troops in 89.8% of the aggressive inter-troop interactions. The following year, the number of adults in troop K was larger than troop I, and K became the dominant troop in the southern part displacing other troops in 83.3% of the aggressive interactions. However, in 2001 the situation in the southern part reversed when troop I, with more adults than troop K regained the dominant status. On the other hand, in the northern part of Mt Longevity, troop B was dominant in 1998 while troop G held the dominant status during 1999 and 2001 respectively. Small troop size, i.e. troops that had few adult males and females, imposed some disadvantages in inter-troop interactions (Fig. 8). The percentage of winning Fig. 5. — The average infant mortality of small, medium and large-sized troops of Formosan macaques at Mt Longevity from 1997 to 2001. 12 M.J. Hsu, J.F. Lin and G. Agoramoorthy and number of adults (troop size) were significantly correlated with each other (P < 0.001). Therefore, we used one independent variable each time in a regression to estimate the percentage of winning. The combination of adults (adult males and females) within a troop best explain variations in the percentage of winning (R2 = 0.54-0.70) in 3 out of 4 years. However in 2000, the number of adult males best Fig. 6. — The relationship between average infant mortality and the mean proportions of adult females in 14 troops of Formosan macaques at Mt Longevity from 1997 to 2001. Fig. 7. — The relationship between the mean number of infants surviving per adult female and the mean percentage of adult females or adult males in 14 social groups of Formosan macaques at Mt Longevity between 1997 and 2001. Effects of group size in wild Formosan macaques 13 explain the variations in the percentage of winning (R2 = 0.72), but, it was not a good indicator in 1998-1999. Therefore we chose the mean number of adults as the best indicator to estimate the percentage of winning during inter-troop interactions (Fig. 8, F1, 13 = 31.6, R2 = 0.72, P < 0.001). Since the intercept of this regression (Y = – 27.37 + 3.53 * X, Fig. 8) did not significantly differ from zero (P > 0.07), a regression model without intercept was calculated as the prediction for the percentage of winning = 2.3719 • (number of adults of a troop). Discussion Troop size and effect on demography The general theoretical context used to explain primate group size variations is associated with individuals maintaining group size to maximize inclusive fitness (Wrangham 1980). However, short-term seasonal stress may intermittently alter the cost associated with optimal group size (Pride 2005). Even though birth rate in mammals plays a key factor in determining reproductive output, which is influenced by both ecological and social components (Dunbar 1988, Hill et al. 2000), troop size may also influence reproductive parameters (Takahata et al. 2005). The per-capita rate of troop size in M. cyclopis at Mt Longevity fluctuated widely and, in general, it was negatively correlated with troop size without infants, indicating that large troop size was related to small or negative increments. Although troop size ranged widely, it did not increase above 100 individuals per troop. The emigration of both males and females and a decrease in infant weaning rate contributed Fig. 8. — The relationship between the average percentage of winning during interactions and the average number of adult females and males in 14 social groups of Formosan macaques at Mt. Longevity from 1998 to 2001. Dotted line: Y = – 27.37 + 3.53 * X. 14 M.J. Hsu, J.F. Lin and G. Agoramoorthy to limiting troop size increase between 1997 and 2001. However, the widely variable range of troop size recorded in our study is consistent with other species such as Macaca nemestrina (Caldecott 1986), M. sylvanus (Menard & Vallet 1996), and M. mulatta (Southwick & Siddiqi 1988). Troop size exceeding 70-80 individuals is not common among Macaca cyclopis. A large wild troop of M. cyclopis with 70 individuals was reported once in another site, Chichi Dashan in Taiwan (Liu 1997). At Mt Longevity, the troop size seldom exceeded 80 individuals and is not as large as the provisioned troops of M. mulatta at Cayo Santiago (Berman et al. 1997), at Kowloon, Hong Kong (Burton & Chan 1996), and of M. fuscata at Mt Ryozen, Japan (Sugiyama & Ohsawa 1982). Recent studies of M. fuscata and M. silenus found no significant relationship between infant mortality and either troop size or the number of adult females (Takahata et al. 1998a, Kumar 1995), which is consistent with our study. Neither troop size without infants nor number of adult females in M. cyclopis at Mt Longevity had significant effects on the ratios of infants to adult females (crude birth rate) and infant mortality. However, the proportion of adult females had a significant effect on both infant mortality and infant weaning rate. In contrast, the crude birth rate of wild M. fuscata was positively correlated with troop size at Yakushima and at Kinkazan (Japan) and negatively correlated with the number of adult females (Takahata et al. 1998a). While smaller troops had a disadvantage in inter-troop competition, the rate of troop size increase was high, possibly due to the smaller denominator. However, the proportion of adult females in a troop has an important influence on infant mortality and infant weaning, indicating that adult female competition has negative effects on the rate of infant weaning. Our data further indicate the negative effect of a large proportion of adult females on infant weaning rate, partially supporting the predation — intra-group feeding competition hypothesis proposed by van Schaik (1983). Our data suggest that the population density of Formosan macaques is not extremely high at Mt Longevity because the birth rate is not negatively correlated with group size. Predator avoidance behaviour has been known to function as a selective pressure to favor group living in primates and influencing group size (Treves & Chapman 1996). According to van Schaik (1983), the survival ratio of immature individuals may increase with larger group size because predation pressure is predicted to be low in large groups. During our study, we did not observe natural predation at Mt Longevity since the only probable natural predator of monkeys in Taiwan, the clouded leopard, has not been sighted for 15 years. This large cat is believed to be extinct in the wild (Hsu & Agoramoorthy 1997). However, poaching and illegal trapping remain a threat for monkeys at Mt Longevity (Agoramoorthy 2002). In addition, feral dogs have been observed to chase and attack monkeys (M.J. Hsu unpublished data). Food provisioning has been observed to influence troop size and birth rate in M. fuscata in Japan. The birth rate of M. fuscata was reported to be high in provisioned troops (0.54-0.59) compared to wild troops (0.27-0.35, Sugiyama & Ohsawa 1982, Wolfe 1986, Koyama et al. 1992, Takahata et al. 1998b). However, the birth rate (0.78 infants per adult female per year) and overall annual growth rate (7%) in our study of M. cyclopis were actually less than those of non-provisioned troops reported previously in a similar eco-region of Taiwan (0.8 and 25.2%, respectively, Wu & Lin 1992). As a matter of fact, the birth rate of M. cyclopis is similar to that of the closely related M. mulatta (0.77-0.90, Southwick et al. 1980; 0.78, Jiang 1988). Food provisioning has been legally prohibited at Mt Longevity, but visitors sometimes provide small amount of food to monkeys during weekends and holidays Effects of group size in wild Formosan macaques 15 along tourist trails. In fact, the forest at Mt Longevity harbours plant species such as Broussonetia papyrifera, Ficus microcarpa, F. septica, Malaisia scandens (Moraceae), and Passiflora suberosa (Passifloraceae), which constitute the main natural food source for monkeys throughout the year (M.J. Hsu unpublished data). Thus food provisioning has no significant effects on troop size and birth rate regardless of whether tourists have less or more frequent contact with monkeys. Social interactions Inter- and intra-troop competitions appears to play a major role in limiting troop size increase in M. cyclopis. In large troops, competition among females (different matrilineal) for resources, and competition among males to improve mating access usually increase aggression. Our study found that troops containing over 20 adults usually had a higher percentage of winning in inter-troop interactions. In small-sized troops, the intra-troop competition among females may be reduced but their infants might be extremely vulnerable to attacks during alpha male changes and inter-troop interactions. Thus our data suggest that inter-troop dominance tends to closely depend on the number of adults rather than group size as reported in M. mulatta (Vessey 1968) and M. fuscata (Kawanaka 1973, Sugiura et al. 2000). Infanticide in macaques has been rarely observed, but aggression from newly established alpha males or dominant males towards infants leading to death has been reported in the closely related M. mulatta (Ciani 1984), and M. fuscata (Soltis et al. 2000) in the wild. Although infanticide occasionally occur in this population (G. Agoramoorthy unpublished data), we did not record infanticide during this study of 14 social troops between 1997 and 2001, and infant mortality did not increase in troops as a consequence of alpha male changes. While we did not find a relationship between birth rate and troop size changes, a higher variation in birth rates occurred in small and medium-sized troops. This might have been due to a reduced availability of food resources for low-ranking females as well as inter-troop competition since home range is usually related to troop size in non-human primates including macaques (Takasaki 1984, Maruhashi et al. 1998). Inter-troop resource competition appears to favor group living and influence troop composition among non-human primates (Treves & Chapman 1996). In general, the per-capita growth rate of troops decreased as a function of troop size (without infants) increases. This indicates that the annual growth rate of large-sized troops decreased through low immigration and possibly high emigration but not necessarily through high infant mortality, which is not consistent with the predation (van Schaik 1983). However, if the proportion of adult females exceeds 28% of troop size, the troop will suffer higher infant mortality. More field studies are needed to investigate the birth rate differences among females of various social ranks to further test the predation — intra-group female competition hypothesis. Acknowledgments The field research was partially funded by the Republic of China’s Council of Agriculture and National Science Council (NSC88-2313-B-020-023, NSC94-2311-B-110-003) through research grants awarded to M.J. Hsu and G. Agoramoorthy. 16 M.J. Hsu, J.F. Lin and G. Agoramoorthy References Agoramoorthy G. 2002. Formosan macaques in crisis at Mt. Longevity, Taiwan. IPPL News: 12-13. Berman C.M., Rasmussen K.L.R. & Suomi S.J. 1997. Group size, infant development and social networks in free-ranging rhesus monkeys. Animal Behaviour 53: 405-421. Burton F.D. & Chan L. 1996. Behavior of mixed species groups of macaques, pp. 389-412. In: Fa J.E. & Lindburg D.G., Edits. Evolution and ecology of macaque societies. Cambridge: Cambridge University Press. Caldecott J.O. 1986. An ecological and behavioural study of the pig-tailed macaque. Contribution to primatology. Vol. 21. Basel: S. Krger. Chapman C.A., Wrangham R.W. & Chapman L.J. 1995. Ecological constraints on group size. An analysis of spider monkeys and chimpanzee subgroups. Behavioral Ecology and Sociobiology 36: 59-70. Ciani A.C. 1984. A case of infanticide in free-ranging group of rhesus monkeys (Macaca mulatta) in the Jackoo forest, Simla, India. Primates 25: 372-377. Dunbar R.I.M. 1988. Primate social system. London: Croom Helm. Hill R.A., Lycett J.E. & Dunbar R.I.M. 2000. Ecological and social determinants of birth intervals in baboons. Behavioral Ecology 5: 560-564. Hsu M.J. & Agoramoorthy G. 1997. Wildlife conservation in Taiwan. Conservation Biology 11: 834-836. Hsu M.J., Agoramoorthy G. & Lin J.F. 2001. Birth seasonality and interbirth intervals in freeranging Formosan macaques, Macaca cyclopis, at Mt. Longevity, Taiwan. Primates 42: 15-25. Hsu M.J. & Lin J.F. 2001. Troop size and structure in free-ranging Formosan macaques, Macaca cyclopis at Mt. Longevity, Taiwan. Zoological Studies 40: 49-60. Hsu M.J., Lin J.F., Chen L.M. & Agoramoorthy G. 2002. Copulation calls in male Formosan macaques: Honest signals of male quality? Folia Primatologica 73: 220-223. Hsu M.J., Moore J., Lin J.F. & Agoramoorthy G. 2000. High incidence of supernumerary nipples and twins in wild Formosan macaques at Mt. Longevity, Taiwan. American Journal of Primatology 52: 199-205. IUCN 2003. 2003 IUCN Red List of Threatened Animals. Gland: IUCN. Janson C.H. & Goldsmith M.L. 1995. Predicting group size in primates: foraging costs and predation risks. Behavioral Ecology 6: 326-336. Jiang H. 1988. The rhesus monkey (Macaca mulatta) population dynamic and distribution at Nanwan peninsula of Hainan Island. Acta Ecologica Sinica 8: 86-94. Kawanaka K. 1973. Intertroop relationships among Japanese monkeys. Primates 14: 113-159. Koyama N., Takahata Y., Huffman M. A., Norikoshi K. & Suzuki H. 1992. Reproductive parameters of female Japanese macaques. Thirty years data from the Arashiyama troops, Japan. Primates 33: 33-47. Kumar A. 1995. Birth rate and survival in relation to group size in the lion-tailed macaque, Macaca silenus. Primates 36: 1-9. Lefebvre D., Menard N. & Pierre J S. 2003. Modelling the influence of demographic parameters on group structure in social species with dispersal asymmetry and group fission. Behavioral Ecology and Sociobiology 53: 402-410. Lehner P.N. 1996. Handbook of ethological methods. Cambridge: Cambridge University Press. Liu G.C. 1997. Monkey-watching at Chichi dashan area. Nature Conservation 18: 55-61. 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. Masui K., Narita Y.S. & Tanaka S. 1986. Information on the distribution of Formosan monkeys (Macaca cyclopis). Primates 27: 123-191. Menard N. & Vallet D. 1996. Demography and ecology of Barbary macaques (Macaca sylvanus) in two different habitats, pp. 95-106. In: Fa J.E. & Lindburg D.G., Edits. Evolution and ecology of macaque societies. Cambridge: Cambridge University Press. Effects of group size in wild Formosan macaques 17 National Research Council 1981. The Techniques for the Study of Primate Population Ecology. Washington, DC: National Academy Press. Peng M.T, Lai Y.L., Yang C.S. & Chiang H.S. 1973. Formosan monkey (Macaca cyclopis). Present status in Taiwan and its reproductive biology. Experimental Animal 22: 447-451. Pride R.E. 2005. Optimal group size and seasonal stress in ring-tailed lemurs (Lemur catta). Behavioral Ecology 16: 550-560. SAS Institute Inc. 2000. SAS/ETS software: changes and enhancements, release 8.1. Cary, NC, U.S.A. Soltis J., Thomsen R., Matsubayashi K. & Takaenaka O. 2000. Infanticide by resident males and female counter-strategies in wild Japanese macaques (Macaca fuscata). Behavioral Ecology and Sociobiology 48: 195-202. Southwick C.H. & Siddiqi M.R. 1988. Partial recovery and a new population estimate of rhesus monkey populations in India. American Journal of Primatology 16: 187-197. Southwick C.H., Richie T., Taylor H., Teas J. & Siddiqi F. 1980. Rhesus monkeys in India. Patterns of growth, decline, natural regulation, pp. 151-170. In: Cohen M. et al., Edits. Biosocial mechanisms of population regulation. New Haven: Yale University Press. Sterck E.H.M., Watts D.P. & van Schaik C.P. 1997. The evolution of female social relationships in nonhuman primates. Behavioral Ecology and Sociobiology 41: 291-309. Sugiura H., Saito C., Sato S., Agetsuma N., Takahashi H., Tanaka T., Furuichi T. & Takahata Y. 2000.Variation in intergroup encounters in two populations of Japanese macaques. International Journal of Primatology 21: 519-535. Sugiyama Y. & Ohsawa H. 1982. Population dynamics of Japanese monkeys with special reference to the effect of artificial feeding. Folia Primatologica 39: 238-263. Takahata Y., Koyama N., Ichino S. & Miyamoto N. 2005. Inter and within-troop competition of female ring-tail lemurs: a preliminary report. African Study Monographs 26: 1-14. Takahata Y., Suzuki S., Okayasu N., Takahashi H., Yamagiwa J., Izawa K., Agetsuma N., Hill D., Saito C., Sato S., Tanaka T. & Sprague D. 1998a. Does troop size of wild Japanese macaques influence birth rate and infant mortality in the absence of predators? Primates 39: 245-251. Takahata Y., Suzuki S., Okayasu N., Takahashi H., Yamagiwa J., Izawa K., Furuichi T., Hill D., Maruhashi T., Saito C., Sato S. & Sprague D. 1998b. Reproduction of wild Japanese macaque females of Yakushima and Kinkazan islands. a preliminary report. Primates 39: 339-349. Takasaki H. 1984. A model for relating troop size and home range area in a primate species. Primates 25: 22-27. Treves A. & Chapman C.A. 1996. Conspecific threat, predation avoidance and resource defense. implications for group living in langurs. Behavioural Ecology and Sociobiology 39: 45-53. van Schaik C.P. 1983. Why diurnal primates living in groups. Behaviour 87: 120-144. Vessey S.H. 1968. Interactions between free-ranging groups of rhesus monkeys. Folia Primatologica 8: 228-239. Walter H. & Leith H. 1967. Klimadiagramm-Weltatlas. Jena: VEB Gustav Fischer Verlag. Wildlife Conservation Law 1989. Republic of China’s Wildlife Conservation Law I-3266. Taipe, Taiwan: ROC. Wolfe L.D. 1986. Reproductive biology of rhesus and Japanese macaques. Primates 27: 95- 102. Wrangham R.W. 1980. An ecological model of female-bonded primate groups. Behaviour 75: 262-300. Wu H.Y. & Lin Y.S. 1992. Life history variables of wild troops of Formosan macaques (Macaca