BIRTH SEASONALITY AND INTERBIRTH INTERVALS IN FREE-RANGING FORMOSAN MACAQUES, MACACA CYCLOPIS, AT MT. LONGEVITY, TAIWAN

 

                                                                                       Published in Primates 42:15-25 (2001)

MINNA J HSU, National Sun Yat-sen University
GOVINDASAMY AGORAMOORTHY, National Pingtung University of Science & Technology

JIN-FU LIN, Shri-Pu Junior High School

ABSTRACT.  The birth season of Formosan macaque, M. cyclopis during our study started in February and ended in August with a peak in the second half of April and the first half of May.  The average birth rate was 82% ± 21 for 114 females with four years of breeding records.  Our study reports that a time span of one year between births can be considered as the norm for the wild M. cyclopis.  Of the 288 inter-birth intervals (IBI), 88.5% showed a one-year interval (mean 364 ± SD 29 days); 11% showed two-year interval (727 ± 36 days); and 1% (2 females) had 3-year interval (range 1030-1040 days).  The IBI for females that had infant loss within 6 months of life were the shortest.  But there was no significant difference from that of females that had stillbirth (p>0.9) and infant that survived for first 6 months of life (p>0.06).  However, among 255 cases of 1-year IBI, stillbirth or following infant loss within 6 months of life did significantly shorten IBI for 10 days (F 1, 253 = 5.74, p<0.05).

Key words: reproduction, birth rate, conception, infant mortality, Macaca cyclopis

Running head: Birth interval and seasonality in free-ranging Macaca cyclopis

INTRODUCTION

In primates, prolonged lactation prevents pregnancy mainly by inhibiting the return of ovulation.  Postpartum endocrine changes in apes and monkeys appear to be similar to those found in humans (FRAWLEY et al., 1983; NADLER et al., 1981). Lactation infertility was reported to be a mechanism selected to achieve adequate spacing between successive births and hence to avoid maternal overburdening.  This mechanism is also affected by maternal nutrition and hence allows mothers to space births flexibly according to their capacity for feeding infants (FRISCH, 1978).  In recent years, the reproductive parameters of primates stimulated interest among zoologists to evaluate the interpretations of sexual selection theory (FEDIGAN, 1983; HARCOURT, 1987; DIXSON, 1998) and birth seasonality (COZZOLINO et al. 1992, BITETTI & JANSON, 2000).

Among the 19 extant species of the Macaca genus, M. cyclopis, is one of the least known.  It is endemic to the island of Taiwan (area 36,000 km²).  It is listed as Vulnerable in the IUCN Red List of Threatened Animals (1996) and protected by Taiwan's Wildlife Conservation Law (WCL, 1989).  It is distributed in a wide range of elevation (30-3300 m) occupying a variety of habitats.  Previous studies for more than a decade on this species were focused on field surveys to estimate the status and distribution (DIEN, 1985; MASUI et al., 1986; TANAKA, 1986; KAWAMURA et al. 1991; LEE & LIN, 1991, 1995).  Reproductive parameters such as sexual skin and vaginal desquamation changes during the menstrual cycle, time of ovulation and gestation period for M. cyclopis have been thoroughly studied in a laboratory condition by PENG et al. (1973).  They indicated that the birth season was from March to June, based on 48 wild caught pregnant monkeys that gave birth in the laboratory from 1965-1969.  Long-term data on the reproductive parameters for M. cyclopis in the field are limited to a single monthly census and observations of two troops with a maximum size of 29 individuals (WU & LIN, 1992, 1993). 

We have been monitoring several habituated troops of free-ranging Formosan macaques that live in the lowland rainforest habitat at Mt. Longevity, southern Taiwan since July 1993 to record data on ecology and population dynamics.  In addition to the occurrence of twins (HSU et al., 2000), we examine for the first time the birth seasonality and duration of inter-birth intervals (IBIs) in more than a hundred female Formosan macaques from 16 troops over a period of four years.  We also assess correlates of maternal age on births, and the influence of lactation status on females' reproductive activities.

MATERIALS AND METHODS

STUDY SITE

Mt. Longevity is located in Kaohsiung City adjacent to the Taiwan Strait.  It is about 5 km long and 2.5 km wide, an area of about 1116 ha with a peak of 354m (22°39'N, 120°15'E).  A topographic field map was used to estimate the total area used by all macaque troops that includes several uplifted coral reefs, undulating hillocks and valleys.  Due to long-term effects of water penetration in hills, strange shapes of stones and steep valleys have been formed and these combined with natural lowland rainforest habitat provide safe roosting areas for the macaques.  The floras of Mt. Longevity include 209 species in 72 families and 164 genera.  The habitat and succession of plant communities follow the pattern of distinct wet and dry season that is similar to the Kenting national park (WU & LIN, 1992).  According to the records of the Central Weather Bureau of Kaohsiung, average annual precipitation (1996-1999) was 2106 (SD ± 716) mm.  Rainfall was concentrated from June to August as the wet season with monthly average above 360 mm (Fig 1).  The dry season started from October and ended in March with monthly average rainfalls below or near 40 mm.  The average monthly temperature was the lowest in January (19.8 ℃) and highest in July (29.2 ℃).

DATA COLLECTION AND ANALYSIS

A field study to investigate the population of Formosan macaques that inhabit Mt. Longevity began in July 1993.  Data presented in this paper are limited to the systematic census of 7-16 habituated macaque troops on a weekly basis between 1 December 1994 and 31 December 1999.  During the February-August period of each year, adult females in these troops were monitored for 4-days per week to reliably record birth data.  Identification of individuals was based on their natural marks and body characteristics using video and photographic documentation.  Kin relations and age of study animals were known from long-term genealogical record compiled by the authors (unpublished data).  About 50% of troops (eight troops) receive small amount of food from people on an irregular basis (usually on weekends and holidays) while other troops still remain wild with minimal contact from tourists.  Among the later, two troops did not have observations long enough to calculate their IBI.  In general, tourists have been discouraged to feed monkeys.

We used scan sampling and ad libitum sampling (ALTMANN, 1974) to record data on mating behaviors during troop encounters.  Gestation length was calculated between the date of birth and the last successful copulation seen, and we limited the range of 142-175 days according to PENG et al. (1973) to avoid wrong calculation because of missing mating behaviors.  However, for some females without accurate copulation records, conception time was backdated from birth, assuming a gestation length of 162 days (PENG et al. 1973).  Conception rates were calculated as the proportion of females that gave birth to a surviving infant (in each month) during the previous birth season who gave birth during the next breeding season.  Average conception rates were calculated from 1996 to 1998.  For average IBIs of multiparous or primiparous females, we first calculated the mean IBI per female, and then calculated the average IBI per category.

We used various test programs from SAS (1989).  The effect of food provisioning or troop identity on average IBI of females was tested through ANOVA.  Waller-Duncan tests were also used for the average IBI of 14 troops.  A series of Spearman rank correlation coefficients was calculated to investigate the correlation of precipitation with monthly number of birth lags in time (0-12 months).  Various General Linear Models were used to test associations between successful rearing of infants and maternal condition, age or troop identity.  Number of live births per year was calculated as number of infants that survive till weaning divided by the number of years of observation on the respective focal female.  The number of infant deaths per year was calculated as the number of infants that died before weaning divided by the number of years of data for the respective focal female.  In both cases, the number of live births per year or the number of infant deaths per year was used as the dependent variables, and the age of mother and troop type as the independent variables.

RESULTS

SYNCHRONY AND BIRTH SEASON

Infant births in Formosan Macaques at Mt. Longevity were highly synchronized.  The birth season started in February and ended in August with a peak during mid-April to mid-May (Fig. 1, 2A, 2B).  A total of 475 births were recorded from 211 females in 16 troops.  Among them, five pairs were twins.  Majority (91.5%) of infants were born over a period of 2.5 months, between April 1 and June 15, with a peak of 52% births between April 16 and May 15.  The average birth rate was 82% ± 21 per year per female for 114 females with four years breeding records.

The average gestation length was 163 (SD ± 8) days (n=98), and a majority of females conceived during November.  Among 141 females, copulation occurred in an average of 175 ± 26 days (range 87 to 253 days) after their previous date of birth.  However, in 118 females, an average of 41 ± 27 days were counted between their first mating and actual dates of conception.  Although mating activities of Formosan macaques started in early September and mostly ended in late January annually, we found the conception period extended from September till next February.

Monthly precipitation had significant and positive correlation with number of births at 10-11 months of time lags (Fig. 3) and conceptions with 4-5 months' lag.  Those were the only significant correlation consistent throughout the study period.  The maximum Spearman's rank correlation coefficient occurred at ten (1996, 1998 and 1999, n=12, p<0.005) or eleven (1997, p<0.001) months lag of birth in time.

INTER-BIRTH INTERVAL (IBI)

The frequency distribution of the total of 288 samples of IBIs from 144 females was separated into three clusters (Fig. 4).  However, 67 other females gave birth once during the 4-year study period and did not contribute any IBI samples.  The average length of these 288 intervals between successive births was 408 days.  The variability was large (standard deviation of 128 days and a range from 270 to 1040 days; Table 1) mainly due to a combination of 1-year, 2-year and 3-year intervals.  Of the 288 IBIs, a vast majority (88.5%) occurred in the range from 270 to 488 days and the average length of these 1-year intervals was 364 ± 29 days.  Of the remaining IBIs, 31 cases that accounted for 10.8% were 2-year intervals with an average of 727 ± 36 days (range 662 - 825 days).  However, two IBIs (0.7%) were 3-year interval with an average of 1035 ± 7 days (range 1030-1040 days).

We did not find any significant effect in terms of human contact or provision in shorting the IBI of females (F _{1, 142} = 0.25, p>0.62).  Troop identity also did not play a significant role on the average length of IBI (F _{13, 130} = 1.34, p>0.20).  However, we found significant difference in the shortest average IBI (333 ± 52 days from troop Aa, n=4) and the longest average IBI (542 ± 224 days from troop E, n=9), and both troops had frequent contact with human.

IBI AND PARTURITION

The IBI following the birth of the first infant of primiparous females was not significantly different from that after the birth of the second infant of multiparous females (F _{35, 105} = 1.14, p>0.59).  The majority of primiparous females gave birth at 5 years old (67%, n=79); 14% gave birth at 6 years old, 10% at 4 years old.  We also found 9% (n=7) primiparous females who gave birth when they were 7-9 years old. Among these primiparous females, 81.5% of them had one-year IBI with 365 ±19 days after their first births, and rest had two-year intervals or above.  The average IBI of primiparous females was 404 ± 112 days, while IBI of 106 multiparous females was 408 ± 105 days.  Maternal age was not positively correlated with mean IBI of multiparous females ((p>0.15), but old adult females (10⁺ years) had the tendency to give birth during early or late birth season.  Only 23 births occurred in Feb-March and July-August.  22 births of them were from multiparous females, especially those that lost infants in the previous year contributed 64% of early births.  None of the primiparous females gave birth before April.  The earliest birth was recorded on 17 February 1999, 150 days after her last recorded copulation.  Another old female delivered an infant on 28 February 1997.  Only two births were recorded from two old adults in the early August (Fig 2B), and the latest birth recorded was on 12 August 1998.

IBI AND INFANT DEATH

About 22% infants died within the first year of life, and 10% infants died within the first week of their life in which 66% were stillbirth.  The majority of deaths occurred within the first week of life (45.7%, n=105), and another 17% deaths occurred before the infants reached one-month old.  The peak of infant deaths occurred in May (mean 31.4% ± 12.2) and June (mean 18.6% ± 9.7).  We witnessed two accidental deaths when two male infants aged 42 and 59 days, slipped off from trees and landed on rocks during the rainy days of June.

Variation among these IBIs could not be significantly explained by maternal age (F_{1, 284} = 3.19, p>0.07), neither previous stillbirths, infant losses within 6 months of life, nor infants surviving to their first 6 months of life (F_{2, 284} = 1.90, p>0.15).  The average maternal age at birth for stillbirths was not significantly older than for live births (F_{1, 286} = 0.57, p>0.4, Table 1).  However, the variability among the number of infant survived per year for females was significantly explained by troops’ identity (F_{15, 204} = 3.53, p<0.001), but not related to maternal age (F_{1, 204} = 2.22, p>0.1).  The highest infant mortality rate occurred in troop Aa (mean 80.0% ± 27.4, n=5).

The IBIs for females that had infant loss within 6 months of life were the shortest, which did not differ from IBI following stillbirth (F_{1, 63} = 0.01, p>0.9) or from IBI following infant surviving for first 6 months of life (F_{1, 275} = 3.31, p>0.07).  However, among 255 cases of 1-year IBI, combining stillbirth or following infant loss within 6 months of life did significantly reduce IBI by 10 days (F _{1, 253}  = 5.74, p<0.05) compared to those following infant surviving for first 6 months of life (Table 2).  The average 1-year IBI of females that had stillbirth or infant loss within 6 months of life was 356 ± 35 days (n=61).

The conception rates of females with surviving infants were a function of the dates of their deliveries in the previous birth season (Table 3).  The conception rates were higher for females that gave birth to infants in the early part of the previous birth season than for females that had a late birth.  Moreover, the conception rates of females that gave birth in previous July and August were zero.  These results indicate that timing of the birth excluded the reproductive outcome of the mothers during the following year.

The females who came into estrus earlier were usually multiparous females without pregnancy or with stillbirth or dead infants in the previous year.  Only three females more than 10 years old were observed to mate before September but none of them conceived during that time.  Among them, the first one did not give birth in 1998, while the second had a stillbirth in April 22, but both were seen in estrous and mated on June 22 and July 29 1998 respectively.  But none delivered any infant during 1999.  The third one didn’t give birth during 1998, mated on Aug. 21 1998, but only delivered an infant on 22 July 1999 as a result of copulation in February.

DISCUSSION

The birth seasonality demonstrated in our results indicated that seasonal variation in environmental cues and social interaction was crucial to time the reproductive activity of the animals and provide the biological rhythm.  Births of M. cyclopis during our study are consistent with the wild caught females that were pregnant during their capture later delivered infants in the laboratory only between March and June (PENG et al. 1973).  Both indicated over 90% births within three month period, highly synchronized in a much narrow range than the previous field report (75%, WU & LIN, 1992).  The birth peak in our study occurred 10-11 months after peak precipitation, compared to 8-10 months lag in Kenting (WU & LIN, 1992).  Heavy rains cause infant deaths, and therefore it is crucial to deliver infants before the wet season.  Our study site has distinct dry and wet seasons and the birth peak of M. cyclopis coincides when the availability and diversity of natural food resources starts to increase in May, just before the rainy season (June to August) (Lin and Hsu, unpublished data).  On the other hand, successful conceptions or artificial insemination of socially isolated M. cyclopis (except during mating with an adult male for 5 days) could be obtained throughout the year with the exception of July in laboratory controlled environment (PENG et al. 1973).  But in a recent study from the New England Regional Primate Research Center, 41% of births in M. cyclopis were recorded within three months (March to May) from females that lived in social groups (PETTO et al., 1995).  Even without proximate factors for food availability and/or changes of photoperiod, M. cyclopis in captivity was reported to be the most strongly seasonal in breeding compare to M. mulatta, M. fascicularis and M. arctoides (PETTO et al., 1995).  Furthermore, shift in birth timing has been influenced by altitude with combination of food availability and daily temperature changes in M. thibetana (ZHAO 1994).  It would be interesting to examine the patterns of birth seasonality in M. cyclopis in different parts of Taiwan with variations in altitudes and/or raining patterns.

The average birth rate over the four birth seasons (83%) is similar to that of many of rhesus macaque troops (DRICKAMER, 1974; WOLFE, 1986) but higher than that of provisioned and wild troops of Japanese macaques (27-54%, TAKAHATA et al., 1998).  Our findings concluded that one year between births can be considered the norm for the wild M. cyclopis in contrast to 2 years for the closely related M. fuscata in the wild (TAKAHATA et al., 1998).  The mean IBI and percentage of one-year interval was very similar to M. sylvanus (408.7 days and 88.4%, PAUL & THOMMEN 1984).  The majority IBI of M. cyclopis (88.5%) fell into the 1-year interval category, but mean IBI of primiparous was similar to that of multiparous females, which was different with a previous report (WU & LIN, 1992).  However, in cases involving old adult females, the birth interval was longer, and fell into 3-year interval category or even longer.  Although the average IBI in our study was 13.4 months, shorter than that of 15.4 months of live birth in this species (in 8 females, WU & LIN, 1992), our study included stillbirths (7%), and deaths shortly after birth, which are easy to miss in most field studies.  According to HENDRIE et al. (1996), live birth rate was 83.4% in M. mulatta and prenatal mortality (abortion plus stillbirths) accounted for 13-22% in M. mulatta and 18% in M. fascicularis of confirmed pregnancy, which was similar to 16% prenatal mortality in M. cyclopis (PENG et al. 1973).  Therefore, IBI of live birth should be longer than IBI including stillbirth.  However, one adult female M. cyclopis in the wild was reported to produce an offspring each year for six successive years (WU & LIN, 1992).

Although few mating activities of Formosan macaques occurred outside of mating season (September to next February), females either went through spontaneous pregnancy losses or did not conceive during that time.  They might have undergone a phase of amenorrhea as reported in M. mulatta (KEVERNE & MICHAEL 1970) or postpartum sterility that lasts at least until the beginning of the next mating season.

Females of many non-seasonal species may renew their sexual activity shortly after the death of the infant and thus reduce their length of birth interval (DIXSON, 1998).  It has been reported among seasonally breeding closely related macaques such as M. fuscata and M. mulatta, that females that lost their infants within six months of life had dramatically shorter IBIs compared to females with surviving infants (TAKAHATA et al., 1998; RAWLINS & KESSLER, 1986).  Infant loss has a great impact on shortening IBI for species with year-round breeding (ALTMANN, et al., 1978) and for seasonally breeding species with a normal IBI of two years (TANAKA et al., 1970; SCUCCHI, 1984), but not for species with IBI of one year (PAUL & THOMMEN 1984).  Although stillbirth, or infant loss within 6 months of life, shortened 1-year IBI on average for 10 days in this study, only giving birth in the late season (July and August) dramatically prevented females from conceiving in the coming mating season.

The occurrence of abortions and stillbirths among captive and recently captured macaques was reported to be high (VALERIO et al., 1969) and appeared to be associated with physical, social or environmental stress.  For example, wild-caught or imported M. cyclopis had 31% (22 out of 72) abortions or stillbirths which is higher than the 16% in laboratory mating colonies (PENG et al., 1973).  However, free-ranging M. mulatta at Cayo Santiago with constant food provision had an incidence of abortions or stillbirths of 3.8% (RAWLINS & KESSLER, 1986), which is relative to the other rates, about the same as that of 7% in the present study.  The shortest average IBI as well as the highest infant mortality rate had occurred in troop Aa, which was a new branch troop formed as a result of fission in May-June 1997.  Our study indicates that social stress associated with adult male replacements and/or fission might have influenced the prenatal and infant mortality that further reduced 1-year IBI of females.

Acknowledgements.  The field research on Formosan macaques at Mt. Longevity has been partially supported by the National Science Council through a research grant awarded to G. AGORAMOORTHY and M.J. HSU (NSC 88-2313-B-020-023).  We thank C. M. Crockett, R. I. M. Dunbar, H. Takahashi, and other anonymous reviewers for critically reviewing earlier drafts of the paper.

REFERENCES

ALTMANN, J. 1974. Observational study of behavior: sampling methods. Behaviour, 49: 227- 265.

ALTMANN, J.; ALTMANN, S.A; HAUSFATER, G. 1978. Primate infant's effect on mother's future reproduction. Science, 201: 1028-1029.

BITETTI, M. S.; JANSON, C. H. 2000. When will the stork arrive? Patterns of birth seasonality in Neotropical primates. Amer. J. Primatol., 50: 109-130.

COZZOLINO, R.; CORDISCHI, C.; SURELI, F ; SCUCCHI, S. 1992. Environmental temperature and reproductive seasonality in Japanese macaques (Macaca fuscata). Primates, 33: 329-336.

DIEN, Z. M. 1985. The Formosan rock-monkey. Quert. J. Taiwan Mus., 11: 345-348.

DIXSON, A. F. 1998. Primate Sexuality. Comparative Studies of the Prosimians, Monkeys, Apes, and Human Beings. Oxford University Press, London.

DRICKAMER, L. C. 1974. A ten-year summary of reproductive data for free-ranging Macaca mulatta. Folia Primatol., 21: 61-80.

FEDIGAN, L. M. 1983. Dominance and reproductive success in primates. Ybk. Phy. Anthropol., 26: 91-129.

FRAWLEY, L. S.; MULCHAHEY, J. J.; NEILL, J. D. 1983. Nursing induces a biphasic release of prolactin in rhesus monkeys. Endocrinol., 112: 558-561.

FRISCH, R. E., 1978. Population, food intake, and fertility. Science, 199: 22-30.

HARCOURT, A. H. 1987. Dominance and fertility among female primates. J. Zool. Lond., 213: 471-487.

HENDRIE, T. A.; PETERSON, P. E.;SHORT, J. J.; TARANTAL, A. F.; ROTHGARN, E.; HENDRIE, M. I.; HENDRICKX, A. G. 1996. Frequency of prenatal loss in a macaque breeding colony. Amer. J. Primatol., 40: 41-53.

HSU, M. J.; LIN, J.F.; AGORAMOORTHY, G. 2000. Occurrence of twins among wild troops of Formosan macaques, Macaca cyclopis at Mt. Longevity, Taiwan. Folia Primatol., 71:154-156.

IUCN. 1996. The 1996 IUCN Red List of Threatened Animals. Gland: IUCN.

KAWAMURA, S.; NORIKOSHI, N.; AZUMA, N. 1991. Observation of Formosan monkeys (Macaca cyclopis) in Taipingshan, Taiwan. In: Primatology Today, A. EHARA, T. KIMURA, O. TAKENAKA, M. IWAMOTO (Eds.), Elsevier, Amsterdam, pp. 97-100.

KEVERNE, E. B.; MICHAEL, R. P. 1970. Annual changes in the menstruation of rhesus monkey. J. Endocr., 48: 669-670.

LEE, L. L.; LIN, Y. S. 1991. Status of Formosan macaques in Taiwan. In: Primatology Today, A. EHARA, T. KIMURA, O. TAKENAKA, M. IWAMOTO (Eds.), Elsevier, Amsterdam, pp. 33-36.

LEE, L. L.; LIN, Y. S. 1995. Status of Formosan macaques in Taiwan. Primate Conser., 11: 18-20.

MASUI, K.; NARITA, Y. S.; TANAKA, S. 1986. Information on the distribution of Formosan monkeys (Macaca cyclopis). Primates, 27: 123-191.

NADLER, R. D.; GRAHAM, C. E.; COLLINS, D. C.; KLING, O. R. 1981. Postpartum amenorrhea and behavior of apes. In: Reproductive biology of the great apes, C. E. GRAHAM (ed.), Academic Press, London, pp. 69-81.

PAUL, A.;THOMMEN, D. 1984. Timing of birth, female reproductive success and infant sex ratio in semifree-ranging barbary macaques (Macaca sylvanus). Folia primatol., 42: 2-16.

PENG, M. T.; LAI, Y. L.; YANG, C. S.; CHIANG, H. S.; NEW, A. E.; CHANG, C. P. 1973. Reproductive parameters of the Taiwan monkey (Macaca cyclopis). Primates, 14: 201-213.

PETTO, A. J.; LA REAU-ALVES, M. N.; ELLISON, P. T.; ABBRUZZESE, M. C. 1995. Reproduction in captive Taiwan macaques (Macaca cyclopis) in comparison to other common macaque species. Zoo Biol., 14: 331-346.

RAWLINS, R. G.; KESSLER, M. J. 1986. The Cayo Santiago Macaques. History, Behavior and Biology. State University of New York Press, Albany.

SAS INSTITUTE. 1989.  SAS/FSP software, usage and reference : version 6. SAS Institute, Cary, NC. 451pp.

SCUCCHI, S. 1984. Interbirth intervals in a captive group of Japanese macaques. Folia Primatol., 42: 203-208.

TAKAHATA, Y.; SUZUKI, S.; OKAYASU, N.; TAKAJASHI, H.; YAMAGIWA, J.; IZAWA, K.; FURUICHI, T.; HILL, D. A.; MARUHASHI, T.; SAITO, C.; SATO, S.; SPRAGUE, D. 1998. Reproduction of wild Japanese macaque females of Yakushima and Kinkazan islands: a preliminary report. Primates, 39: 339-349.

TANAKA, S. 1986. Further note on the distribution problems of the Formosan monkey (Macaca cyclopis). Kyoto Univ. Overseas Res. Report of Studies on Nonhuman Primates, 5: 95-104.

TANAKA, T.; TOKUDA, K.; KOTERA, S. 1970. Effects of infant loss on the interbirth interval of Japanese monkeys. Primates, 11: 113-117.

VALERIO, D. A.; MILLER, R. L.; INNES, R. M.; COUTNEY, K. D.; PALLOTTA, A. J.; GUTTMACHER, R.M. 1969. Macaca mulatta, Management of a laboratory breeding colony. Academic Press, New York & London, pp. 59-78.

WCL. 1989. Republic of China's Wildlife Conservation Law I-3266. Taipei, Taiwan, ROC.

WOLFE, L. D. 1986. Reproductive biology of rhesus and Japanese macaques. Primates, 27: 95-102.

WU, H. Y.; LIN, Y. S. 1992. Life history variables of wild troops of Formosan macaques (Macaca cyclopis) in Kenting, Taiwan. Primates, 33: 85-97.

WU, H. Y.; LIN, Y. S. 1993. Seasonal variation of the activity and range use patterns of a wild troop of Formosan macaque in Kenting, Taiwan. Bull. Inst. Zool., 32: 242-252.

ZHAO, Q. K. 1994. Birth timing shift with altitude and its ecological implication for Macaca thibetana at Mt. Emei. Oecologia Montana, 3:2 4-26.

Authors' Names and Addresses: MINNA J HSU, Department of Biological Sciences, National Sun Yat-sen University, Kaohsiung, Taiwan, Republic of China; GOVINDASAMY AGORAMOORTHY, Department of Wildlife Conservation, National Pingtung University of Science and Technology, Taiwan, Republic of China, and S.M. Govindasamy Nayakkar Memorial Foundation, Thenpathy 609111, Tamilnadu State, India; and JIN-FU 

壽山之臺灣獼猴 (Macaca cyclopis) 社群大小與結構

 Troop Size and Structure in Free-ranging Formosan Macaque (Macaca cyclopis)at Mt. Longevity, Taiwan

Published in Zoological Studies 40:49-60 (2001)

Minna J. Hsu^1,* and Jin-Fu Lin²

¹Department of Biological Sciences, National Sun Yat-sen University, Kaohsiung, Taiwan 804, R.O.C.

² Shi-Pu Junior High School, Kaohsiung County, Taiwan 840, R.O.C.

Key words: Reproduction, Troop composition, Sex ratio, Male tenure, Fission.

Running head: Hsu and Lin - Troop Size and Structure in Macaca cyclopis

*To whom correspondence and reprint requests should be addressed.
Tel: 886-7-5252000 ext. 3623. Fax: 886-7-5253623.
E-mail: hsumin@mail.nsysu.edu.tw

Abstract
Minna J. Hsu and Jin-Fu Lin (2001) Troop size and structure in free-ranging Formosan macaque (Macaca cyclopis) at Mt. Longevity, Taiwan. Zoological Studies 40(0): 000-000.  Among the 19 extant species of the genus Macaca that are found in southern and eastern Asia as well as northwestern Africa, the Formosan macaque is one of the least known.  A long-term field study to investigate the population dynamics and social behavior of 7-16 troops of free-ranging Formosan macaques at Mt. Longevity, Taiwan has been conducted since July 1993.  Between December 1994 and December 1997, a systematic census was conducted on a biweekly basis to record data on the demography of Formosan macaques.  We used focal animal sampling and ad libitum sampling twice per week in respective troops to record data on social behavior including male replacement and fission processes.  The maximum density of macaques has been estimated as 26 individuals per km² in October 1997.  The average troop size was 26.1  ± 9.7 (n = 7) in January 1995 and it reached the highest level of 47.0 ± 21.2 (n = 13) in August 1997.  Two cases of fission were observed.  The branch troops, Ia and Aa, that were formed as a result of fission had the smallest size with 9 individuals in the beginning, while troop I had the largest size of 86 individuals.  Births were recorded mainly between April and June (97%) with a peak in mid-April to mid-May.  The annual average overall sex ratio was 1.06 ± 0.28, while the adult sex ratio (adult males to adult females) was 0.53 ± 0.12.  The average tenure length of alpha males was 16.8 ± 18.9 months (n = 34) and ranged from 1 wk to a maximum of 6 yr.  The average alpha male tenure in newly formed troops was significantly shorter (p < 0.01) than that in the remaining troops.  About 88% of alpha male changes occurred between October and February, which paralleled the peak and the end of the mating season, respectively.

　

Key words: Reproduction, Troop composition, Sex ratio, Male tenure, Fission.

INTRODUCTION

Non-human primates are known to exhibit a wide variety of social and grouping patterns (Eisenberg et al. 1972).  Changes in primate troop size and density have reflected short-term alterations in environmental conditions (Eisenberg 1979), epizootics (Collias and Southwick 1952), and natural disasters (Horwich and Johnson 1986).  Changes in population structure can also be long-term adaptations to altered environmental conditions such as deforestation and human predatory pressure (Freese et al. 1982).  Troop size and composition are the result of demographic processes that take years to develop, and short-term studies of relatively few groups may lead to misconceptions about primate social organization (Crockett 1985).  Understanding the adaptive significance of troop size and flexibility on a long-term basis is therefore fundamental for understanding primate social ecology (van Schaik 1999).  It is widely believed that predation and inter-troop resource competition as selective pressures favor group living and influence troop composition among non-human primates (Treves and Chapman 1996).

The genus Macaca is one of the most widespread primate groups in the world, but the Formosan macaque, M. cyclopis, is considered to be one of the least known.  Among the 19 extant species of the genus Macaca that are found in southern and eastern Asia as well as northwestern Africa (Fa and Lindburg, 1996), the Formosan macaque is endemic to the island of Taiwan (area 36000 km²).  Much previous research on this species was focused on field surveys and observations to estimate the population status and distribution (Dien 1985, Masui et al. 1986, Tanaka 1986, Kawamura et al. 1991, see Lee and Lin 1991 1995 for review).  Although it is distributed in a wide range of elevations (100-3300 m) occupying a variety of habitats (Masui et al. 1986, Lee and Lin 1991 1995), the density is relatively low, at around 0.1 group/km² in the southeastern coastal mountains (Masui et al. 1986).  Hunting pressure on M. cyclopis has been substantial with mass captures for the purpose of exports for medical or experimental use (Masui et al. 1986).  Even during 1986-1987, illegal hunting of this species was still reported (Lee and Lin 1995).  It is listed in the IUCN Red List of Threatened Animals (1996) as well as being protected under Taiwan's Wildlife Conservation Law (1989).

The social structure of macaques is generally characterized as often occurring as a large stable multimale-multifemale troop.  Masui et al. (1986) reported a sighting of a troop of M. cyclopis with more than 45 individuals (2 adult males and 13 adult females) in Taitung.  However, long-term data on group size and birth for M. cyclopis are limited to a single field study of 2 single-male-multifemale troops, with a troop size of 9-16 individuals (Wu and Lin 1992 1993).  Variation in grouping patterns has been found recently among species such as M. fuscata (Fukuda 1989), M. nemestrina (Oi 1990), and M. sylvanus (Menard and Vallet 1996).  Birth rate and immature survival rate have been differently predicted as a function of group size by models on the evolution of group living in primates (Wittenberger 1980, Wrangham 1980).  In an earlier survey of Mt. Longevity, Lin and Wang (1991) found 5 troops with a total of 76 Formosan macaques, but no details on troop composition or birth rate were described.  Nevertheless, most troops at Mt. Longevity have a multimale-multifemale structure, and this makes the population interesting for socio-ecological investigations including variations in demographic parameters.

Habituation of animals is often required in order to collect detailed demographic assessment of a non-human primate species.  We have been monitoring 7-16 habituated troops of free-ranging Formosan macaques that live in lowland rainforest habitat at Mt. Longevity, southern Taiwan since July 1993.  In this paper, we present 3 yr of systematic data on troop structures and dynamics of Formosan macaques.  Details are also given on troop size, composition, density, sex ratio, fission, dispersal, and alpha male tenures in order to compare the population and social dynamics data of Formosan macaques with other well-studied species of the genus Macaca.

MATERIALS AND METHODS

Description of Mt. Longevity study site

     Mt. Longevity is located in Kaohsiung City adjacent to the Taiwan Strait.  It is about 5 km long and 2.5 km wide with a height of 354 m and has an area of about 1116 ha.  This mountain is isolated from other nearby mountains and forests due to the development of the city over the last 40-50 yr.  The terrain is dominated by uplifted coral reefs, and the total surface area of Mt. Longevity is estimated to be 35 km² due to the complex topography (Lin and Wang 1991).   We used a topographic field map to calculate the total area used by all macaque troops, which includes several uplifted coral reefs, and undulating hillocks and valleys.  The northern part of Mt. Longevity is still a restricted military base, and researchers are not permitted to enter the area.  Thus the study area excluding the military area covers about 25 km².  Due to the long-term effects of water penetration in the hills, strange shapes of stones and steep valleys have been formed, and these combined with natural lowland rainforest habitat provide safe roosting areas for the macaques.  The flora of Mt. Longevity includes 209 species in 164 genera and 72 families (Lin and Wang 1991).  About 5% of the plants are ferns, which grow in the few wet gorges of the forest floor, or on trees.  The dominant tree species include Ficus septica, F. wightiana, and F. caulocarpa, which also constitute a major food source for the macaques.  Shrubs of Severinia buxifolia and Lantana camara and vines of Bauhinia championii are widely distributed.  Introduced plants include Acacia confusa, Leucaena glauca, and Delonix regia; agricultural fruit trees such as Euphoria longana, Mangifera indica, and Annona squamosa are also found.  The habitat and succession of plant communities follow the pattern of distinct wet and dry seasons that is similar to that in the nearby Kenting National Park (Wu and Lin 1992).  The dry season begins in October and ends in April (Fig. 1).  The average annual rainfall between 1995-1997 was about 1453 ± 576 mm concentrated from May to September as the wet season with monthly averages near or above 100 mm (Fig. 1).  Moreover, from November to February, the average monthly rain falls below 35 mm.  The average monthly temperature was lowest in February (18.6 ℃) and highest in July (29.9 ℃, Fig. 1).

Census and observation plan

     The study area was thoroughly searched for Formosan macaque troops along forest trails, footpaths, and transects using methods described by Hsu and Agoramoorthy (1996).  Interviews with local people were conducted during the initial stages of the census to locate monkey troops and to collect preliminary data on the population size since July 1993.  The existence of a few troops has been known since February 1989 (Table 1).  Some of these troops were also recorded in an earlier survey conducted by Lin and Wang (1991).  Systematic census of 7-16 troops was conducted on a biweekly basis between December 1994 and December 1997 (Table 1).  The roosting sites of most study troops were marked on a field map (Fig. 2).  Identification of individuals was based on their natural marks and body characteristics using video and photographic documentation (National Research Council 1981).  We did not attempt to capture monkeys for marking purposes.  Body weights of selected individuals from habituated troops were measured by using peanuts to attract the animal to stand or sit on a body scale.

Individuals were classified into broad age-sex classes based on direct measurement of chronological age with known birth year or based on physical characteristics and body size according to US National Research Council (1981).  Chronological ages of individuals born after the onset of our study were known and were estimated for those individuals born earlier.  Most females give birth to their first infants at around 5 yr of age.  Subadult males were between the ages of 5 and 6 yr, and their secondary sexual characteristics had not fully developed as in adult males.  Other individuals aged between 2 and 4 yr were considered to be juveniles.  Infants of both sexes were typically nursing, and being cared for by their mothers when less than 1 yr old.  We used focal animal sampling and scan sampling (Altmann 1974) to record data on social behavior including male replacement and fission processes twice per week in respective troops.

Data calculation and analysis

The annual growth rate was calculated using troop sizes in consecutive years for each troop when compared with the troop size of the previous year (December).  The average annual growth rate was calculated from the annual growth rate for each of the 7 troops identified in the early part of the study for each year from 1994 to 1997.  Population density was calculated as the maximum number of monkeys counted within the study period divided by the study area (25 km²).  Solitary males were often associated or interacted with particular troops and were seen mating with some peripheral females during the breeding season.  Therefore they were included within a troop size count to get a better estimation of adult sex ratio (adult males/adult females) instead of being calculated separately.  The average adult sex ratio was calculated monthly for all available social troops.  The overall sex ratio for each troop was calculated monthly as the number of males (adult, subadult, juvenile, and infants combined) divided by the number of females within a troop during the monthly census period.  The annual overall sex ratio for each troop was the average of overall sex ratio calculated each year for 1995, 1996, and 1997.

Then we calculated the mean annual overall sex ratio for each troop and we used this data set to get an average of the entire population during the study period.  The tenure length of an alpha male was calculated as the duration of time that a male maintained this status in a troop.

All statistical analyses were conducted with Statistical Analysis System software (SAS Institute 1989).  All mean values are presented ± 1 standard errors.  The significance of troop size and the number of adult females on the rate of troop growth was tested using analysis of variance (ANOVA).  Pearson correlation coefficients were used to test the relations of average troop growth to the average number of adult females and average troop size (with or without infants).  Various regression analyses were conducted to obtain the best model to estimate the rate of troop growth.  Spearman rank correlation coefficients were calculated for troop size without infants, number of adult females, and numbers of births and deaths of each troop.  The effects of troop size (without infants) and number of adult females on dependent variables, such as the number of surviving infants, number of births, birth rate, and infant mortality, were tested using General Linear Models.  The nonparametric Wilcoxon test was used to compare average tenure length of alpha males in newly formed troops to that of troops with longer histories (> 2 yr).

RESULTS

Density and troop size

Troop sizes of M. cyclopis at Mt. Longevity had a wide range from 9 to 86 with a maximum average troop size of around 47 individuals in August 1997 (SE 21.2, n = 13).  The troop size increased during our study period (Table 1, Fig. 3) and the maximum density was estimated as 26 individuals/km² in October 1997.  The maximum number of individuals recorded was 652 among 15 troops in October 1997.  Their troop composition and age class distribution are shown in Table 2.  Initially, our study troops included 7 troops (171 individuals, Table 1) in December 1994, and later 9 new troops were found and habituated including 2 troops (Aa and Ia) that formed as a result of fission from troops A and I (Table 1).  Two of our study troops changed their original home ranges and disappeared from our study site, one temporarily while the other never reappeared.  Troop Q disappeared on 26 June 1996 and was never seen again, while troop N disappeared on 9 June 1996 but reappeared on 5 October 1997.

The average size of 7 troops of Formosan macaques in December 1994 was 24.4 ± 10.1 (Table 1), and it reached the highest at 47.1 ± 16.9 in August 1997 but decreased again to 44.4 ± 16.5 in December 1997 (Fig. 4).  The average annual growth rate of these 7 troops was 24.4% per year.  This annual growth rate actually decreased from 29.0% (± 16.7 SE) in 1994/1995 to 25.4% (± 9.0 SE) in 1995/1996 to 18.7% (± 16.9 SE) in 1996/1997.  In addition, the average annual growth rates of troops D and E with smaller troop sizes were higher than that of the other 5 medium- to large-sized troops (Fig. 5).  In 11 troops that had been monitored for more than 2 yr (Table 1), the average annual growth rates were negatively correlated with average number of adult females (r = -0.68, p < 0.05) and average troop sizes excluding infants (r = -0.61, p < 0.05).  This trend remained even when we excluded troops A and I (that underwent fission) from the analysis.  However, the best regression model (R² = 0.466, F_1,9  = 7.86, p < 0.05) for estimating the annual growth rate (%) was 46.6 - (2.2114 x the number of adult females).

 Troop Aa was the smallest with 9 individuals in December 1997 (Table 1), while troop I reached a record high of 86 individuals in August 1997.  All troops had a multi-male structure with the exception of the initial troop formation of Ia and Aa.  These 2 new troops had single males associated with adult females during the initial stages, but became multi-male troops after 7 and 16 wk, respectively.

Infant birth and survival

Births were recorded between February and June annually, and 85% of the births occurred within 2 months (April 15 to June 15, Fig. 3).  Most successful matings with sperm ejaculation were observed from September until the following February, which indicated a distinct mating season.  The total birth rate over 3 birth seasons was 0.7 infants per adult female per year.  Birth rates increased from 0.5 infants per adult female per year in 1995 to 0.7 in 1996 and 0.8 in 1997 if all adult females in all troops were combined.  However, the infant birth rate (birth per adult females) did not change significantly (F_1,11 = 1.774, p > 0.20) with troop size excluding infants (Fig. 6A) and also did not correlate with the number of adult females (F_1,11 = 1.078, p > 0.32).  A total of 235 births were recorded during 1995-1997, of which 2 pairs were twins.  Both sets of twins survived for over 3 yr and were still alive at the time of writing.  Most infants (97%) were born between April and June with the exception of 1 in February and 5 in March.  Thirty-nine females were observed to deliver their 1st infants when they were about 4 yr old.

The number of surviving infants was correlated with the number of births (Spearman correlation coefficient, r =0.89, p < 0.001), and the number of adult females and troop size without infants (p < 0.001).  The number of infant deaths, however, did not correlate with any of these variables.  The average mortality rate for infants within 6 mo of birth was 22.58% (± 2.95 SE) in 1996-1997.  When troop Aa was excluded from the data set, the average infant mortality did not correlate linearly with average troop size (excluding infants) for 12 troops (F_1,10 = 0.07, p > 0.79, Fig. 6B).  The percentage of infants within a troop was relatively constant and did not correlate with troop size (p > 0.05) when troop Aa was excluded.  Troop Aa was small and newly formed from fission, and in that troop, 2 infants had died before September 1997 (Table 2).

Sex ratio

The average overall sex ratio was approximately 1:1 (1.06 ± 0.28, n = 16, range 0.63 to 1.67), and the average adult sex ratio was close to 0.53 (± 0.12, n = 16, range 0.30 to 0.71).  Solitary adult males accounted for 5% of the entire population, and they were seen interacting with social troops especially during mating season (between September and January) which slightly increased the adult sex ratio (Fig. 3).

The proportions of adult females in troops did not significantly decrease (p > 0.05) when troops' sizes increased from 20 to 85.  However, the smallest troop, Aa, had the highest percentage of adult females (45.5%), while the largest troop, I, had the lowest percentage of adult females (21.3%, Table 2).

Troop fission

Two cases of fission were observed during 1996-1997.

Case 1. On 3 December 1995 troop I had 72 individuals, and its home range was on the west slope of  Mt. Longevity where visitors seldom hike.  On 3 February 1996, the gamma male (AM1) was found with 3 low-ranking adult females and 5 juveniles from troop I, and it was later named as a new troop (Ia).  Three of the juveniles (2 males and 1 female) were 1 yr old, and the other 2 males were 2 yr old.  The adult sex ratio (adult males to adult females) of troop I before fission was 0.21, and it changed to 0.19 after fission, whereas the adult sex ratio of the newly formed troop Ia was 0.33.  Troop Ia was a 1-male troop for 7 wk before becoming a multi-male troop.  Five months later, another 3 low-ranking adult females with 2 newborn infants emigrated from troop I to the newly formed troop Ia.  Two more juvenile males joined the troop on 21 September 1996.  Even 1 yr after the formation of troop Ia (on 7 February 1997), an adult female along with a juvenile male from troop I was observed to immigrate into troop Ia.  During the same period of time, 3 adult solitary males also immigrated on separate occasions into troop I.  Thus the adult sex ratio of troop I increased to 0.42, which was very similar to the ratio of troop Ia (0.43).

Case 2.  Prior to fission in April 1997, troop A had 42 individuals including 12 adult females.  Three low-ranking adult females were seen 200 m away from the main troop A on 16 April 1997.  Three weeks later, an adult female with a 2-d-old infant accompanied by 3 male juveniles joined the trio.  Although 2 peripheral males (AM1 and AM2) from troop A followed the 4 females, none was accepted as a leader since the alpha female (AF1) led the others.  She also repelled the 2 males when they approached her.  On 23 August 1997, 2 other adult males (AM3 and AM4) were seen with the females.  A week later, AM3 was established as the alpha male of the newly formed troop Aa.

Dispersal of males and females

Males stayed in their natal troops until 5 to 6 yr of age (n = 54) and dispersed either to become solitary or to immigrate into other troops.  However, most juvenile males (3 to 4 yr old) usually moved gradually from the core area to the peripheral part of their natal troops.  We did not observe any natal troop males engaging in true sexual mating with adult females in their natal troops.  In the majority of the observed inter-male encounters (n = 31), solitary adult male invaders fought with the dominant male leader of social troops to access estrous females, often resulting in injuries.  On the other hand, subadult males and old adult males (n = 39 in 11 troops) did not engage in aggressive encounters with alpha males in social troops.  Subadult males and newly immigrated adult males were frequently observed in the peripheral part of social troops forming coalitions that ranged from 2-14 individuals.  However, alpha males usually displaced these males and forced them to the periphery of the troops.  The average tenure length of alpha males was 16.8 ± 18.9 months (n = 34) with a range from 1 wk to a maximum of 6 yr.  Average alpha male tenure (3.1 mo, n = 7) in troops that were newly formed as a result of fission, was significantly shorter (p < 0.01) than average tenure (20.4 mo, n = 24) in the rest of the troops.  About 88% of changes of alpha males occurred during the peak of the mating season.  The majority of adult females stayed in their natal troops, with the exception of troop fission.  Sometimes subordinate females were mobile between main troops and newly formed troops.  In one case, an adult female was observed to move between main troop A and new troop Aa on 4 occasions before her disappearance from the study area.  Although female transfer is rare compared to male transfer in Formosan macaques, 6 solitary adult females successfully immigrated into other troops during our study.

DISCUSSION

     Wild troops of Formosan macaques have been in existence in Mt. Longevity for many centuries, and the earliest record was cited on a 17th century Dutch colonial map with the name Apenberg, meaning Apes Hill (Lin and Wang 1991).  Most of Mt. Longevity has been a restricted military base since the Japanese occupation of Taiwan for more than half a century and has only been partially opened to the public since 1989.  The protection of the natural forests at Mt. Longevity by the military has left the majority of the flora little disturbed over the years.  The forest also provides sufficient food resources for the monkeys, and in addition, some troops with home ranges close to trails irregularly receive some food items from visitors.  In an earlier survey of Mt. Longevity, Lin and Wang (1991) found 5 groups of Formosan macaques ranging from 12 to 19 individuals per group, and they were named as troops A, B, C, F, and K in our study.  These troops and troop E were habituated earlier than the rest of the troops.  The other troops were afraid of humans when we first encountered them.  Hunting pressure is thought to be the major cause for small troop size and low density of M. cyclopis in the wild (Masui et al. 1986, Wu and Lin 1992) before hunting was banned in 1989.  The newly added study troops, solitary male immigration into social troops, and the reduction of hunting pressure were contributing factors for the increase in population size of monkeys at Mt. Longevity.

The widely variable range of troop size recorded in this study is consistent with other macaque species such as M. nemestrina (Caldecott 1986), M. sylvanus (Menard and Vallet 1996), M. mulatta (Southwick and Siddiqi 1988), and M. fascicularis (Kyes 1993).  The average troop size in this study is comparable to a previous report of a troop of 45⁺ individuals (Masui et al. 1986).  Although it was reported that troop size could exceed 100 (Lee and Lin, 1995), the authors were unable to verify is large a size.  However, a social troop of M. cyclopis seldom exceeds 80 individuals and is not as large as reported for the provisioned troops of rhesus macaques on Cayo Santiago (Berman et al. 1997) and in Kowloon (Burton and Chen 1996) or for Japanese macaques (Furuya 1969).

 The ratios of infants to adult females (crude birth rate) appeared to be constant regardless of troop size (excluding infants).  This situation differed with Japanese macaques (Takahata et. al. 1998a) and lion-tailed macaques (Kumar 1995).  We found that a relatively larger troop (larger than 60 excluding infants) had an advantage in inter-troop competition with a higher percentage of winning troop interactions (Lin and Hsu unpubl. observ.).  Nevertheless, in a small-sized troop, in which few separate matrilineal groups of females occur, intra-troop competition between different matrilineal females may be reduced, but their infants might be extremely vulnerable during the alpha male changes and inter-troop interactions.

Troop size of M. cyclopis on Mt. Longevity did not continue to grow, and some underwent fission even though their troop size was less than 50.  Two new troops were formed in this study as a result of fission.  Similar cases of fission have been observed among M. fuscata (Maruhashi and Takasaki 1996, Furuya 1969), M. mulatta (Rawlins and Kessler 1986), M. sylvanus (Menard and Vallet 1996), M. sinica (Dittus 1988), M. silenus (Kumar 1995), and M. cyclopis (Wu and Lin 1992).  However, troop sizes before fission in M. cyclopis were much smaller than those reported for M. fuscata (100-650, Furuya 1969).  We observed relatively low-ranking females splitting from main troops to form new troops.  Furthermore, low-ranking adult females were seen joining one of the newly formed troops even several months after the establishment of the troop.  This suggests that intra-troop competition among females for resources or position might have been the proximate cause of troop fission (Wrangham 1980).  In addition, sexual advantage in attaining 2nd rank may have been the incentive for a male to form the new troop, Ia.  The costs and benefits of troop fission may vary between individuals, and the phenomenon needs future quantitative analysis.

Although the troop sizes of M. cyclopis at Mt. Longevity ranged widely from 9 to 86, the average troop size did not change much whether calculated either from the original 7 troops or from 13 troops in August 1997.  In both cases, the maximum average troop size was around 47 individuals.  Although the average troop size of these 7 troops increased from 24.4 (1994) to 42.9 (1997), the average annual growth rates actually decreased from 29.0% (1995) to 18.7% (1997).  An annual troop size increase of 25.2% was previously reported for this species at Kenting (Wu and Lin 1992), which is close to our findings at Mt. Longevity.  The annual growth rates of small-sized but socially stable troops, such as troops D and E in our study, were higher than that of large-sized troops.  The rate of growth of a troop is a decreasing function of the number of adult females in the troop.  This indicates that annual grow rates of large-sized troops decrease possibly through lower immigration, higher emigration (including fission) or higher mortality of juveniles and/or infants.  The 2 troops at Kenting reported by Wu and Lin (1992) were relatively small with 9 to 16 individuals, which are closer to the smallest troop size recorded during this study.  Unfortunately, the high growth rate shown at Kenting was not sustained; these 2 troops disappeared from their home ranges after incidents of illegal hunting.

Hunting pressure and intra-troop competition appears to play major roles in limiting troop size and population size increases in M. cyclopis at Mt. Longevity.  The threat of illegal hunting still continues despite the fact that this species has enjoyed legal protected status for over a decade.  About 5% of our study individuals managed to escape from hunters’ traps with severe wounds or broken limbs (unpubl. data).

The effects of food provisioning on birth rate, growth rate, and troop size in our study are not significant.  Although the Japanese macaque birth rate was high in provisioned troops (0.54-0.59) compared to that of wild troops (0.27-0.35, Sugiyama and Ohsawa 1982, Koyama et al. 1992, Takahata et al. 1998b), we found that the birth rate was constant across all troops regardless of human provisioning and interaction.  About 50% of troops (8 troops) receive food from people on an irregular basis (usually on weekends and holidays), but it is difficult to access the exact amount of food provided by tourists.  Other troops still remain wild with minimal contact with tourists.  Troop I, the largest group with 72 individuals was wild and was not habituated when we first made contact in 1995.  Even after fission in 1996, troop I remains the largest known troop with minimal contact with people.

The birth rate (0.67 infants per adult female per year) and overall annual growth rate (24.4%) in our study were actually less than those of non-provisioned troops reported previously in a similar eco-region (25.2% and 0.8, prospectively, Wu and Lin 1992).  Masui et al. (1986) indicated that the ratio of infants to adult females (0.77) in a group of Formosan macaques was slightly higher than that of wild Japanese macaques (Takasaki and Masui 1984).  Therefore, the birth rate of Formosan macaques is more similar to that of the rhesus macaque (0.82, Wolfe 1986) than the wild or provisioned Japanese macaque (0.27-0.59, Sugiyama and Ohsawa 1982, Wolfe 1986, Takahata et al. 1998b).

M. mulatta is considered the possible ancestor for the island species such as M. cyclopis and M. fuscata (Hoelzer and Melnick 1996).  We found striking similarities in the breeding patterns of these species.  Seasonality in breeding in our study is consistent with a previous field report (Wu and Lin 1992) and captive study (Petto et al. 1995) for M. cyclopis, as well as for M. mulatta (Lindburg 1987) and M. fuscata (Kawai et al. 1967, Nigi 1976).  Moreover, a laboratory study on M. cyclopis reported that when male-female pairs were kept isolated (not in a social troop), they were able to breed throughout the year (Peng et al. 1973a, b).  But when several individual females were kept in a social group, they showed a distinct seasonal breeding pattern that was similar to the wild situation (Petto et al. 1995).  This suggests that social factors may play a crucial role in the breeding seasonality of M. cyclopis, but this needs to be confirmed by further physiological and behavioral studies.  Our study indicates that synchronization of birth is more restricted to a narrow range as 97% of births occurred between April and June, whereas Wu and Lin (1992) reported that 75% occurred in this period of time.  Although the seasonality of birth reached a peak in mid-April to mid-May for M. cyclopis, mounting and mating were observed throughout the mating season.  The ages of females who delivered their first infants at 4 and 5 yr old coincided with previous reports for this species, both in captivity (Peng et al. 1973a,b) and in the wild (Wu and Lin 1992).  This was similar to results for M. mulatta (Rawlins and Kessler 1986, Wolfe 1986), but 1-2 yr earlier than wild M. fuscata (Takahata et al. 1998b).

     We found that the annual average overall sex ratio was roughly balanced when all troops were combined, but the adult sex ratio within social troops was frequently not balanced.  This was mainly due to male-male competition resulting in the eviction of other adult males entering some social troops.  All troops in our study had a multi-male troop structure with 2 exceptions in which new troops were formed as a result of fission, and they had either a single adult male (troop Ia) or no adult male (troop Aa) for a short period of time.  However, these troops were not socially stable and became vulnerable to male invasions.  This was indicated by the shorter tenure length of alpha males in newly formed troops than those of troops with longer histories (> 2 yr).  It is also interesting to note that the adult sex ratio of main troop I was similar to that of the branch troop Ia after a year.  On average, no sexual advantage for either males or females accrued to 1 particular group.

Solitary adult males were often seen around social troops prior to immigration into social troops.  But they often engaged in aggressive interactions during the mating season (September to February) with troop alpha males and received severe injuries.  Similarly, wounding and death of males during troop takeovers have been reported among Tibetan macaques (Zhao 1996).  Moreover, subadult males and old adult males were seen to be peripheralized by the dominant alpha male, and the peripheralized low ranking males often formed coalitions within a social troop,  similar to that observed in M. mulatta (Southwick et al.1965), M. radiata (Sugiyama 1971), and M. fuscata (Imanishi 1957), respectively.  Age at natal dispersal was usually at about 5 yr, which is similar to that of M. fuscata (Sprague et. al. 1998).

Although a previous study found a range of alpha male tenure length of 5-28⁺ mo for 2 social troops (Wu and Lin 1992), we have recorded a wide range that varied from 1 wk to 6 yr with an average of 16.8 mo.  Most alpha male changes occurred during mating seasons, an indication of severe competition among dominant males (the troop alpha male versus intruders).  Thus troop alpha male takeovers in Formosan macaques are influenced by seasonal breeding in order to access sexually receptive females.  Similarly, alpha male changes and the appearance of extra-troop males were reported to be higher during mating seasons in species such as M. fuscata (Fukuda 1982, Sprague 1992), M. mulatta (Neville 1968, Drickamer and Vessy 1973), and M. sylvanus (Mehlman 1986).  On the other hand female Formosan macaques remained in their natal troops in most cases with a few exceptions, and thus M. cyclopis is considered female-bonded which is similar to other species in the genus Macaca.

Acknowledgments: The field research on Formosan macaques at Mt. Longevity was partially supported by the Council of Agriculture through a research grant awarded to M. J. Hsu.  We are especially grateful to G. Agoramoorthy for his expert advice and inspirational contribution to this study.  We thank F. D. Burton, C. Crockett, J. Moore, A. Treves and anonymous referees for their critical comments on earlier versions of the manuscript.

REFERENCES

Altmann J. 1974. Observational study of behavior: sampling methods. Behaviour 49: 227- 265.

Berman CM, KLR Rasmussen, SJ Suomi. 1997. Group size, infant development and social networks in free-ranging rhesus monkeys. Anim. Behav. 53: 405-421.

Burton FD, L Chan.1996. Behaviour of mixed species groups of macaques. In JM Fa, DG Lindburg eds. Evolution and ecology of macaque societies, Cambridge: Cambridge Univ. Press, pp. 389-412.

Caldecott JO. 1986. An ecological and behavioural study of the pig-tailed macaque. Contrib. Primatol. 21. Basel: S. Krger.

Collias NE, CH Southwick. 1952. A field study of population density and social organization of howling monkeys. Proc. Am. Philos. Soc. 96: 143-156.

Dien ZM. 1958. The Formosan rock-monkey. Quart. J. Taiwan Mus. 11: 345-348.

Dittus WPJ. 1988. Group fission among wild toque macaques as a consequence of female resource competition and environmental stress. Anim. Behav. 36: 1626-1645.

Drickamer LC, SH Vessey. 1973. Group changing in free-ranging male rhesus monkeys. Primates 15: 359-368.

Eisenberg JF. 1979. Habitat, economy and society: some correlation and hypotheses for neotropical primats. In I Bernstein, E Smith, eds. Primate ecology and human origins. New York: Garland Press, pp. 215-262.

Eisenberg JF, N Muchenhirn, R Rudran. 1972. The relation between ecology and social structure in primates. Science 176: 863-874.

Fa JE, DG Lindburg. 1996. Evolution and ecology of macaques societies. Cambridge: Cambridge Univ. Press.

Freese CH, PG Heltne, N Castro, G Whitesides. 1982. Patterns and determinants of monkey densities in Peru and Bolivia with notes on distribution. Intl. J. Primatol. 3: 53-90.

Fukuda F. 1982. Relationships between age and troop shifting in male Japanese monkeys. Jpn. J. Ecol. 32: 491-498.

Fukuda F. 1989. Habitual fission-fusion and social organization of the Hakone troop T of Japanese macaques in Kanagawa Prefecture, Japan. Intl. J. Primatol. 10: 419-439.

Furuya Y. 1969. On the fission of troops of Japanese monkeys. II General view of troop fission of Japanese monkeys. Primates 10: 47-69.

Hoelzer GA, DJ Melnick. 1996. Evolutionary relationship of the macaques. In JE Fa, DG Lindburg, eds. Evolution and ecology of the macaque societies. Cambridge: Cambridge Univ. Press, pp. 3-19.

Horwich RH, ED Johnson. 1986. Geographical distribution of the black howler (Alouatta pigra) in central America. Primates 27: 53-62.

Hsu MJ, G Agoramoorthy. 1996. Conservation status of primates in Trinidad, West Indies. Oryx 30: 285-291.

Imanishi K. 1957. Social behavior of Japanese monkeys, Macaca fuscata. Psycholgy 1: 47-54.

IUCN. 1996. The 1996 IUCN red list of threatened animals. Gland Switzerland: International Union Conservation of Nature.

Kawai M, S Azumu, Kyoshiba. 1967. Ecological studies of reproduction in Japanese monkeys (Macaca fuscata). I. Problems of the birth season. Primates 8: 35-74.

Kawamura S, N Norikoshi, N Azuma. 1991. Observation of Formosan monkeys (Macaca cyclopis) in Taipingshan, Taiwan. In A Ehara, T Kimura, O Takenaka, M Iwamoto, eds. Primatology today. Amsterdam: Elsevier, pp. 97-100.

Koyama N, Y Takahata, MA Huffman, K Norikoshi, H Suzuki. 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.

Kyes RC. 1993. Survey of long-tailed macaques introduced onto Tinjil Island, Indonesia. Am. J. Primatol. 31: 77-83.

Lee LL, YS Lin. 1991. Status of Formosan macaques in Taiwan. In A Ehara, T Kimura, O Takenaka, M Iwamoto, eds. Primatology today. Amsterdam: Elsevier, pp. 33-36.

Lee LL, YS Lin. 1995. Status of Formosan macaques in Taiwan. Primate Conser. 11: 18-20.

Lin CD, JP Wang. 1991. A preliminary survey of Formosan macaque population and habitat utilization in Kaohsiung Chaishan. Taipei: Council of Agriculture, pp. 50. (in Chinese).

Lindburg DG. 1987. Seasonality of reproduction in primates. In G Mitchell, J Erwin, eds. Comparative primate biology. behavior, cognition and motivation. New York: Alan R Liss, pp. 167-218.

Maruhashi T, H Takasaki. 1996. Socio-ecological dynamics of Japanese macaque troop ranging. In JE Fa, DG Lindburg, eds. Evolution and ecology of the macaque societies. Cambridge: Cambridge Univ. Press, pp. 146-159.

Masui K, YS Narita, S Tanaka. 1986. Information on the distribution of Formosan monkeys (Macaca cyclopis). Primates 27: 123-191.

Mehiman P. 1986. Male intergroup mobility in a wild population of Barbary macaque (Macaca sylvanus), Ghomaran Rif Mountains, Morocco. Am. J. Primatol. 10: 67-81.

Menard N, D Vallet. 1996. Demography and ecology of Barbary macaques (Macaca sylvanus) in two different habitats. In JE Fa, DG Lindburg, eds. Evolution and ecology of the macaque societies. Cambridge: Cambridge Univ. Press, pp. 95-106.

National Research Council. 1981. The techniques for the study of primate population ecology. 233 pp. Washington, DC: National Academy Press.

Neville MK. 1968. Male leadership changes in a free-ranging troop of Indian rhesus monkeys (Macaca mulatta). Primates 9: 13-27.

Nigi H. 1976. Some aspects related to conception of the Japanese monkey (Macaca fuscata). Primates 17: 81-87.

Oi T. 1990. Population organization of wild pig-tail macaques (Macaca nemestrina nemestrina) in Sumatra. Primates 31: 15-31.

Peng MT, YL Lai, CS Yang, HS Chiang. 1973a. Fomosan monkey (Macaca cyclopis). Present status in Taiwan and its reproductive biology. Exp. Anim. 22: 447-451.

Peng MT, YL Lai, CS Yang, HS Chiang, AE New, CP Chang. 1973b. Reproductive parameters of the Taiwan monkey (Macaca cyclopis). Primates 14: 201-213.

Petto AJ, MN LaReau-Alves, PT Ellison, MC Abbruzzese. 1995. Reproduction in captive Taiwan macaques (Macaca cyclopis) in comparison to other common macaque species. Zoo Biol. 14: 331-346.

Rawlins RG, MJ Kessler. 1986. The Cayo Santiago macaques. History, behavior and biology. Albany: State Univ. of New York Press.

SAS Institute. 1989.  SAS/FSP software, usage and reference: version 6. Cary, NC. 451 pp.

Southwick CH, MR Siddiqi. 1988. Partial recovery and a new population estimate of rhesus monkey populations in India. Am. J. Primatol. 16: 187-197.

Sprague DS. 1992. Life history and male intertroop mobility among Japanese macaques (Macaca fuscata). Intl. J. Primatol. 13: 437-454.

Sprague DS, S Suzuki, H Takahashi, S Sato. 1998. Male life history in natural populations of Japanese macaques: migration, dominance rank, and troop participation of males in two habitats. Primates 39: 351-363.

Sugiyama Y. 1971. Characteristics of the social life of bonnet macaques (Macaca radiata). Primates 12: 247-266.

Sugiyama Y,  H Ohsawa. 1982. Population dynamics of Japanese monkeys with special reference to the effect of artificial feeding. Folia Primatol. 39: 238-263.

Takahata Y, S Suzuki, N Okayasu, H Takajashi, J Yamagiwa, K Izawa, N Agetsuma, D Hill, C Saito, S Sato, T Tanaka, D Sprague. 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, S Suzuki, N Okayasu, H Takajashi, J Yamagiwa, K Izawa, T Furuichi, DA Hill, T Maruhashi, C Saito, S Sato, D Sprague. 1998b. Reproduction of wild Japanese macaque females of Yakushima and Kinkazan Islands: a preliminary report. Primates 39: 339-349.

Takasaki H, K Masui. 1984. Troop composition data of wild Japanese macaques reviewed by multivariate methods. Primates 25: 308-318.

Tanaka S. 1986. Further note on the distribution problems of the Formosan monkey (Macaca cyclopis). Kyoto Univ. Res. Rep. 5: 95-104.

Treves A, CA Chapman. 1996. Conspecific threat, predation avoidance and resource defense: implications for group living in langurs. Behav. Ecol. Sociobiol. 39: 45-53.

van Schaik CP. 1999. The socioecology of fission-fusion sociality in orangutan. Primates 40: 69-86.

Wildlife Conservation Law. 1989. Republic of China's wildlife conservation law I-3266. Taipei, Taiwan, R.O.C.

Wittenberger JF. 1980. Group size and polygyny in social mammals. Am. Natural. 115: 197-221.

Wolfe LD. 1986. Reproductive biology of rhesus and Japanese macaques. Primates 27: 95-102.

Wrangham RW. 1980. An ecological model of female-bonded primate groups. Behaviour 75: 262-300.

Wu HY, YS Lin. 1992. Life history variables of wild troops of Formosan macaques (Macaca cyclopis) in Kenting, Taiwan. Primates 33: 85-97.

Wu HY, YS Lin. 1993. Seasonal variation of the activity and range use patterns of a wild troop of Formosan macaque in Kenting, Taiwan. Bull. Inst. Zool. 32: 242-252.

Zhao QK. 1996. Etho-ecology of Tibetan macaques at Mount Emei, China. In JE Fa, DG Lindburg, eds. Evolution and ecology of the macaque societies. Cambridge: Cambridge Univ. Press, pp. 263-2

壽山之臺灣獼猴 (Macaca cyclopis) 社群大小與結構

                        徐芝敏¹       林金福²
    獼猴屬(Macaca)主要分布於亞洲東部與南部及非洲的西北部。在世界上獼猴屬現存的19種物種 中，臺灣獼猴(M. cyclopis)是最不為人所知者。壽山自由活動之臺灣獼猴社群動態和社會行為的長期 野外研究，始於1993年7月。我們經由個體辨識，來確認猴群成員組成變化以及行為。從1994年12月 至1997年12月，每月至少二次，以有系統的方式，來調查社群成員的數目變動。在相關的社群中， 每週至少二次使用焦點取樣及隨機取樣的方式，來記錄社會行為，包括雄猴首領替代及分群的過程 。依壽山表面面積計算，壽山臺灣獼猴的密度最高可達每平方公里26隻。平均社群大小在1995年1月 為26.1 ± 9.7隻 (n = 7)，平均社群大小在1997年8月最高，可達47.0 ± 21.2 隻(n = 13)。研究期間曾觀察 到二次分群。最小的社群為剛分出之小旁支社群Ia 與Aa，僅有9名個體；最大的社群為社群I，曾達86名個體。每年嬰猴出生的日期皆很集中，主要從4月到6月(97 %)，且在4月中至5月中到達高峰。 社群平均性比率接近一對一，但成年雄猴數目僅約為成年雌猴數目之半。平均首領任期長度為16.8± 18.9 月(n = 34)，任期長度從一週到最長六年。平均首領任期長度在新形成的社群中，顯著的比其 他的社群平均首領任期長度短。大約88%首領任期的替換發生於每年10月至翌年2月間，即從交配季 節的高峰至季末。

關鍵詞：生殖、社群組成、性比例、雄性任期、分群。

¹國立中山大學生物科學系

²高雄縣立溪埔國中