Demography life history and social structure in Propithecus diadema edwardsi from 1986Ц2000 in Ranomafana National Park Madagascar.код для вставкиСкачать
AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 125:61–72 (2004) Demography, Life History, and Social Structure in Propithecus diadema edwardsi From 1986 –2000 in Ranomafana National Park, Madagascar Sharon T. Pochron,* W. Troy Tucker, and Patricia C. Wright Department of Anthropology, State University of New York at Stony Brook, Stony Brook, New York 11794 KEY WORDS lemurs; life table; mortality; fertility; population growth; group structure ABSTRACT Prosimian lemurs differ fundamentally from anthropoid primates in many traits related to social structure. By exploring the demography of Milne-Edwards’ sifakas (Propithecus diadema edwardsi), and comparing it to other well-studied primates, we explore the effect of demographic and life-history factors on social structure. Speciﬁcally, we compare lemur survivorship and fertility patterns to two published composite models: one created for New World and another created for Old World monkeys. Using longitudinal data collected on individual Propithecus diadema edwardsi from four study groups from 1986 –2000 in Ranomafana National Park, Madagascar, we quantify 1) group composition, 2) birth seasonality, 3) interbirth interval, 4) life-table values, and 5) population growth estimates. The mortality, survivorship, and life-expectancy schedules indicate high infant and juvenile mortality. Fertility remains high until death. The intrinsic rate of increase and net reproductive rate indicate a shrinking population. We suggest that high mortality rather than low fertility causes the observed population decline. While sifaka survivorship closely resembles New World patterns, fertility resembles Old World patterns, i.e., like New World monkeys, few sifakas survive to reproductive age, and those that do, reproduce at a slow rate resembling the Old World pattern. This necessarily impacts social structure. An adult sifaka at the end of her lifespan will have one only daughter who survives to reproductive age, compared to 3.4 for New World or 2.7 for Old World monkeys. Demography limits the formation of large kin-based groups for sifakas, and survivorship and fertility patterns do not easily permit sifakas to form large same-sex family groups. Am J Phys Anthropol 125:61–72, 2004. © 2004 Wiley-Liss, Inc. In many traits related to social structure, prosimians differ fundamentally from anthropoid primates. These differences caused Wright (1999) and Kappeler (2000) to label lemurs as “puzzling” and “idiosyncratic.” Puzzling lemur traits include: male based sex ratios, low maximum group size, high infant mortality rates, photo-period-dependent limited estrus, female dominance over males, femalefemale aggression, and monomorphism in polygynous social systems. Speciﬁally, in the Milne-Edwards’ sifaka (Propithecus diadmema edwardsi) social structure differs dramatically relative to anthropoid primates. Baboon migration patterns, for example, can be neatly summarized: Males disperse from their natal group as they reach adulthood (Rasmussen, 1981; Smuts, 1985; Altmann et al., 1996). However, patterns of dispersal in the Milne-Edwards’ sifaka are less straightforward (unpublished data). Roughly 50% of males and females migrate, while the others remain in their natal group. Females tend to migrate prior to reproduction, while males move both as adolescents and after sexual maturity. Adults of both sexes, from adolescence onward, migrate as a result of targeted aggression (Wright, 1995, and unpublished data). Migration patterns are not the only difﬁcult-topredict aspect of sifaka social structure. While papionines live in polygamous multi-male/multi-female social groups, the Milne-Edwards’ sifaka cannot be so easily categorized. Propithecus diadema edwardsi demonstrate operational sex ratios consistent with harem, multi-male/multi-female, single-male/sin- © 2004 WILEY-LISS, INC. Grant sponsor: ANGAP; Grant sponsor: Université d’Antananarivo; Grant sponsor: Université de Fianarantsoa; Grant sponsor: ONE; Grant sponsor: DEF; Grant sponsor: MICET; Grant sponsor: Ministre de la Environnement; Grant sponsor: SUNY Stony Brook Research Foundation; Grant sponsor: ICTE; Grant sponsor: Wenner Gren Foundation for Anthropological Research; Grant sponsor: John D. and Catharine T. MacArthur Foundation; Grant sponsor: Liz Claiborne and Art Ortenberg Foundation; Grant sponsor: USAID UDLP Program; Grant sponsor: Earthwatch Institute; Grant sponsor: Douroucouli Foundation. *Correspondence to: Sharon T. Pochron, Department of Anthropology, State University of New York at Stony Brook, Stony Brook, NY 11794. E-mail: email@example.com Received 3 January 2002; accepted 19 December 2002. DOI 10.1002/ajpa.10266 Published online 12 January 2004 in Wiley InterScience (www. interscience.wiley.com). 62 S.T. POCHRON ET AL. TABLE 1. Causes of death for females1 Age bracket N Unknown causes Infanticide Fossa Hurricane All ages ⱕ1 ⱕ4 ⬎4 25 13 18 7 5 (20.0%) 3 (23.1%) 4 (22.2%) 1 (14.3%) 3 (12.0%) 3 (23.1%) 3 (16.7%) N/A 16 (64.0%) 6 (46.2%) 10 (55.6%) 6 (85.7%) 1 (4.0%) 1 (7.7%) 1 (5.6%) 0 (0%) 1 Only infants categorized as females contribute. gle-female pair and polyandrous breeding systems with equal frequencies. A group that uses a multimale/multifemale breeding system in one year may use a harem, monogamous, or polyandrous system during the next breeding season, as group compositions change. The type of social system into which an infant is born does not affect its survival (Pochron and Wright, 2003). Ecological factors no doubt play a role in demographic measures, but these factors are not the focus of this paper. Although both ecology and demography likely contribute to lemur idiosyncrasies, this paper examines the force of demography on lemur social structure. In this paper, we present a demographic analysis of 15 years of data on wild sifakas (Propithecus diadema edwardsi) in Ranomafana National Park, Madagascar. By exploring the demography of this species, and comparing it to well-studied primates, we aim to explore how demographic factors inﬂuence social structure. Our paper includes assessments of group compositions and birth seasonality. We also describe classic demographic measures, including 1) mortality, 2) survivorship, 3) fertility, 4) life expectancy, 5) reproductive value, 6) intrinsic rate of increase, and 7) net reproductive rate. In this paper, we compare the lemur pattern of survivorship and fertility to composites of Old and New World monkeys. Previous papers (Gage and Dyke, 1988; Dyke et al., 1993; Gage and Dyke, 1993; Gage, 1998) summarized the literature on age-speciﬁc mortality and fertility in primates. They developed a model life table using statistical procedures to test for differences among populations, pooling all high-quality published information. They thus created composite models of survivorship and birth rate for Old World and New World monkeys, and we compare the lemur pattern to those composites. METHODS Study site Wright (1995) described the study site in detail. Ranomafana National Park, located at 21°16⬘ S latitude and 47°20⬘ E longitude in southeastern Madagascar, was created in 1991 and is classiﬁed as submontane rainforest. The park covers 43,500 ha (Wright, 1992) and ranges in altitude from 500 – 1,500 m (Wright, 1997). The rainfall varies from 2,300 – 4,000 mm per year, with most precipitation occurring between December–March (Overdorff, 1993; Hemingway, 1996). Average annual temperature is 21°C, with lowest temperatures from June– September (0 –12°C). The four study groups reside in the Talatakely Trail System (TTS), which was selectively logged from 1986 –1988. Subjects This sexually monomorphic species weighs 5– 6 kg on average (Glander et al., 1992; Wright, 1995; Pochron et al., 2002). Females give birth to only one infant at a time. P.C.W. and her team habituated four groups of sifakas, which have been observed regularly since habituation. Observations began in January 1986 for groups I and II, in January 1991 for group III, and January 1995 for group IV. Although this long-term project is ongoing, only data collected through December 2000 contribute to this paper. This yields a total of 46 group-years of observations (15 years each for groups I and II, 10 for group III, and 6 for group IV). Groups were followed for at least 5 simultaneous days per month using full-day, focal-animal follows. Births, deaths, and emigrations have been recorded continuously since the habituation of each group. Dates of births and deaths are accurate to within 1 week, on average. The ages of individuals born after 1985 are known, and toothwear provides estimates for those individuals born before 1986 or born in less-studied groups (Wright, 1995). Age classes of sifakas were assigned by year of birth or estimated year of birth. “Infant” refers to animals in their ﬁrst year, “juvenile” refers to animals aged 1–3, and “adult” refers to animals 4 years old or older. Wright (1995) detailed how subjects are darted, collared, and identiﬁed. Animals wear a collar and tag, and the combination provides that animal’s designation. Since the initiation of this ﬁeld site, 58 animals have been observed in groups I–IV. All individuals are recognizable, and more than half (N ⫽ 32) are collared. Noncollared animals include: 1) those who died before they reached their second birthday, 2) animals less than 2 years of age, and 3) adult emigrants who leave a group before being darted. The in-depth, long-term nature of this study allows high conﬁdence in distinguishing dead animals from those who migrated. Table 1 provides, for females, cause of death by age category. The three cases of infanticide were part of a targeted-aggression campaign by one female against another in which one female killed the infant of another. (So far, infanticidal males have killed only male infants (N ⫽ 3), who do not contribute to this data set.) Death by fossa (Cryptoprocta ferox) is assigned with DEMOGRAPHY OF PROPITHECUS DIADEMA EDWARDSI or without corpses (Wright, 1998). Fossas leave corpses bearing discarded intestines and deep tooth marks on long bones and/or the cranium. When fossas leave no corpses, evidence such as injured group mates, fossa footprints, scat, and sightings indicates predation (Wright, 1998). Female migration events are preceded by targeted aggression in this species (unpublished data), and unlike predation events, female/female aggression has not resulted in wounded adults (Wright, 1995). We have not observed animals under age 4 years migrate from our study groups, so infants and juveniles who disappear leaving no corpse are categorized as “dead.” We cannot assign cause of death for one adult female. We have ruled out migration because no targeted aggression took place before this female disappeared, and although she wore a collar, we never saw her again. Sex ratios To calculate group compositions, we assess the number of adults of both sexes in December, the peak breeding month. No migrations occur during the breeding season and no adult has yet died in December, so December provides the most stable month for counting the number of adults. Sexing Propithecus infants under age 6 months can be difﬁcult from behavioral observations because the clitoris and penis of animals under age 2 years are approximately the same size. After 6 months of age, males and female can be behaviorally distinguished, since males scent-mark using a chest gland, which females lack. Before 6 months of age, sexes can be determined physically. Once a corpse is recovered or a mother and dependent offspring are darted, females can be distinguished from males by a vaginal opening. Population-growth and life-table parameters Despite the 15-year depth of this project, which covers 46 breeding seasons, our sample size is small. These sifakas live in small groups, live many years, and reproduce slowly. To maximize sample size, we pooled females by age, regardless of their year of birth (Sade et al., 1976; Smith, 1982; Ha et al., 2001). All infants were pooled together, all 1-yearolds were pooled together, all 2-year-olds were pooled together, and so on, for each age. This means, for example, that infant mortality rate equals the proportion of infants born over the course of the study who die before their ﬁrst birthday. This technique, although standard in primatology, masks temporal variability in rates. In the example above, if infants die at a signiﬁcantly higher rate in one year relative to the others, that effect is reduced through averaging with other years. We also pooled groups to maximize sample size, and this masks variability between groups (Smith, 1982; Sade et al., 1976; Ha et al., 2001). Figure 1 shows the mortality data used both by year and after pooling by age. 63 For the pooled data, we calculated values of: 1) age-speciﬁc mortality (qx), 2) age-speciﬁc survivorship (lx), 3) age-speciﬁc fertility (mx), 4) life expectancy (Ex), and 5) reproductive value (vx/v0), using standard life-table methods (Smith, 1982; Pianka, 1983; Dyke et al., 1986; Gage and Dyke, 1988; Gotelli, 1998). To calculate age-speciﬁc fertility (mx), we halved the total number of offspring born in each year, regardless of sex, rather than using the total number of female offspring born. This method accommodates yearly variations in birth sex ratios of our small sample by assuming a 50:50 birth sex ratio. To assess the effect of left- and right-censoring of the data on the survival curve, we also calculated lx, using survival-analysis techniques (Cox and Oakes, 1984). A nonparametric (Kaplan-Meier) survival estimate was computed for the data, and diagnostic plots were produced for exponential, Weibull, and lognormal distributions. A Weibull distribution appeared most appropriate to the sifaka survival data, and was used to compute a survival curve in a subsequent parametric survival analysis that included information regarding both left- and rightcensoring. We graphed survivorship for analysis. One of three types of idealized survivorship curves can characterize the life history of a particular species, and these can be seen by plotting the logarithm of lx on the ordinate, and age on the abscissa (Gage, 1998; Gotelli, 1998). A type I curve has high survivorship during young and intermediate ages, and then a steep decline as individuals approach old age. Type I curves often describe human populations (Gage, 1998). The opposite curve, type III, depicts poor survivorship in young ages and improving survivorship with increasing age. New World primates (Gage, 1998) and many insects (Gotelli, 1998) demonstrate this pattern. If the semilogged survivorship curve forms a straight line (a type II curve), the probability of survival is constant across age (Gotelli, 1998). We also calculated the intrinsic rate of increase and the net reproductive rate. Both the intrinsic rate of increase and the net reproductive rate measure population growth. The intrinsic rate of increase (r), the per capita birth rate minus the per capita death rate, is computed iteratively from the Euler equation ⌺ lxmxe⫺rx ⫽ 1. A negative value indicates a declining population. The net reproductive rate (R0), the mean number of offspring produced per female over her lifetime, is calculated by multiplying each value of lx by the corresponding mx and summing the products across all ages (Gotelli, 1998). If R0 exceeds 1.0, females produce a net surplus of offspring each generation, and the population increases exponentially (Gotelli, 1998). Comparative primate demography Justification. In a series of papers, Gage (1998) and Dyke et al. (1993) created standard composite mortality and fertility schedules for Old and New World primates while assuming that within any 64 S.T. POCHRON ET AL. Fig. 1. Demographic data for female Milne-Edwards’ sifakas. The individual’s name runs along the x axis. Top: Data pooled by individual age. Bottom: Data organized by year of study. Pooled data were used to compute life-table parameters. Solid circles, deaths; open circles, migrations. broad taxonomic grouping, an underlying age-speciﬁc pattern of fertility and mortality exists. They also assumed that mortality and fertility rates observed in any one population represent a sample of underlying schedules to which the populations are subject; they argued that for many purposes, including cross-species comparisons, standardized estimates of the underlying schedule provide more reliable information than mortality and fertility rates computed directly from population data of one species (Dyke et al., 1993; Gage, 1998). Rather than selecting one or two Old World and New World primates for comparison to our sifaka data, we compare our data to these composite standardized models. Particularly for the mortality data, which are difﬁcult to measure in captive and wild groups relative to fertility (Gage, 1998), these composite standardized models provide a more representative and thus complete picture for comparison. As described below, the fertility data are less useful, both because they come from fewer species and because they come from captive data. However, we decided to include them because at worst, this is equivalent to comparing sifakas to one captive species. Such a comparison still provides interesting information, even if best used as suggestive. Mortality. With the ultimate goal of providing a method to compensate for inadequate primate mortality data, earlier researchers (Gage and Dyke, 1988, 1993; Dyke et al., 1993; Gage, 1998) conducted an extensive review of the published literature on primate demography, compiled a composite sample of mortality statistics, and used a hazard model to create a standard mortality schedule that statistically ﬁt Old World and New World monkeys as a group. Data included in the composite models were carefully selected. Before they included a published mortality schedule, data had to meet ﬁve criteria: 1) The information had to extend over the full lifespan of the species. 2) Age classes had to be deﬁned at regular intervals that were short enough to indicate changes in mortality rates throughout the species’ life history. 3) The data had to lack stochastic variation and/or gaps in age classes resulting from insufﬁcient numbers in age classes and small numbers of deaths. 4) The data had to exhibit a monotonic decline in mortality rates in the early years, and a monotonic increase in older years. DEMOGRAPHY OF PROPITHECUS DIADEMA EDWARDSI 5) The data had to exhibit an absence of bias resulting from poor reproduction and/or age misestimation. Using these criteria, for example, they honed 25 data sets down to eight for Old World primates (Gage and Dyke, 1988). These populations include four Macaca mulatta populations, two Macaca fuscata populations, and two populations of mixed Papio ssp. (mostly yellow baboons). One Macaca fuscata population was wild, another was provisioned, one M. mulatta population was provisioned, and the remaining populations were captive. The New World primates contributing to the composite model include: captive Callithrix jaccus, Leontopithecus rodalia, Saguinus fascicollis, and Saguinus oedipus. For captive and provisioned animals, research-related deaths were censored and not incorporated into the model life table estimates (Gage and Dyke, 1988; Gage and Dyke, 1993; Gage, 1998). Fertility. The fertility data used by Gage (1998) were not screened as well as the mortality data, and he therefore recommended that readers consider the fertility models preliminary. Gage (1995, 1998) based his fertility models on the assumption that all primates follow the senescent fertility pattern (characterized by a decline in fertility at the older ages), and used the Brass polynomial method of smoothing fertility. The Old World fertility schedule is based on data from captive Macaca mulatta, while the New World schedule is based on captive Callithrix jaccus, Saguinus fuscicollis, and Saguinus oedipus. These colonies emphasized breeding, so fertility was not intentionally limited (Gage, 1998). The net reproductive rate and the intrinsic rate of increase are calculated using the mortality and fertility models calculated above. Gage (1998) acknowledged that composite fertility models based on captive animals may reﬂect some combination of colony management issues and biological or social limits. We compare our Propithecus diadema edwardsi results to his as heuristic. RESULTS Group composition On average, during December, a group contains 1.44 adult females and 1.46 adult males for an operational sex ratio of 1.0:1.0. Multimale/multifemale groups (N ⫽ 11) consistently include two adult males and two adult females. Harem males (N ⫽ 9) have access to 2.2 ⫾ 0.4 females on average; polyandrous females (N ⫽ 13) have access to 2.3 ⫾ 0.5 males on average. Twelve groups included one adult male and one adult female. Groups have 0.6 ⫾ 0.7 infants and 1.1 ⫾ 1.1 juveniles. Wright (1995) reported a ratio of males:females:infants:juveniles of 1.0:1.2:0.6:1:3. Results based on additional data yield a ratio of 1.0:1.0:0.4:0.7, which has fewer juveniles than previously reported. 65 Birth sex ratio and seasonality Since 1986, 37 infants were born into groups I–IV, i.e., 20 female and 17 males (1.0:0.85). Birth timing suggests strong seasonality. Most infants (94.6%) were born in the austral winter, with the largest number born in June. In June, 73% (27/37) of infants were born, 13.5% (5/37) were born in May, and 10.8% (4/37) were born in July. One infant was born out of season, on September 24. Life table values Mortality. The life table (Table 2 and Fig. 2) shows that mortality (qx) is highest for infants and remains high until age 5, after which it drops to 0 until age 9. Half of the newborn females fail to survive to 1 year of age (Table 2), and an ANOVA shows that mortality does not differ signiﬁcantly by year (N ⫽ 70 births; DF ⫽ 14, 55; F ⫽ 1.28; P ⫽ 0.25; see Fig. 3). Most females reproduce for the ﬁrst time at age 4, although only 23.8% of females survive to that age. Taken together, these ﬁgures show that mortality is highest among newborns and juveniles. Survivorship. Survivorship values (lx) in Table 2 show that 23.8% of females survive to age 4, the age of ﬁrst reproduction. As shown in Figure 4, Propithecus diadema edwardsi demonstrates a combination of type II and type III curves. From birth until age 4, the population shows decreasing survivorship with age, following a type III curve. From age 9 until death, the probability of survival holds constant across age, following a type II curve. The overall pattern indicates high infant and juvenile mortality. Figure 5 shows the survivorship curves generated from the life table (marked line) and from the parametric survival analysis (plain line). The survival analysis accounts for both right- and left-censoring; however, it requires that a parametric distribution be ﬁt to the data. Comparing the survival curve generated from the survival analysis to the survival curve generated directly from the data shows that disregarding data censoring may result in overestimates of mortality rates up to age 5, and underestimates of adult mortality beyond age 7, as expected (Wood et al., 1992). Additionally, some of the discrepancy between methods is due to the need to ﬁt an artiﬁcial function (i.e., a Weibull distribution) to the data in order to control for the censored data. This is a problem with nonetiological approaches to hazard modeling (cf. Wood, 1994, p. 85– 86). Interbirth interval. We calculated the interbirth interval for 25 births. The average interbirth interval is 1.56 ⫾ 0.58 years. In 12 cases, the interbirth interval is 1 year; in 11 cases, it is 2 years; and in one case, the interbirth interval is 3 years. Variation in survival of the previous infant causes a bimodal distribution in interbirth interval. Survival of an infant to the mother’s next mating period explains signiﬁcant variation in interbirth interval (R2 ⫽ 0.35; t ⫽ ⫺3.5, P ⫽ 0.0019); in other words, the 66 S.T. POCHRON ET AL. TABLE 2. Standard life table for Milne-Edwards’ sifakas1 Age Obs 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Total 20 10 7 6 6 7 7 7 7 7 5 4 4 3 3 2 3 2 2 2 1 115 Mortality (qx) Survivorship (Ix) Fertility (mx) Life expectancy (Ex) Reproductive value (vx/v0) 0.5000 0.2000 0.2857 0.1667 0.1667 0.0000 0.0000 0.0000 0.0000 0.1429 0.2000 0.2500 0.2500 0.0000 0.3333 0.0000 0.0000 0.0000 0.0000 0.5000 0.0000 1 0.5000 0.4000 0.2857 0.2381 0.1984 0.1984 0.1984 0.1984 0.1984 0.1701 0.1361 0.1020 0.0765 0.0765 0.0510 0.0510 0.0510 0.0510 0.0510 0.0255 0.0000 0.0000 0.0000 0.0000 0.1667 0.0000 0.3571 0.1429 0.2857 0.3571 0.3000 0.3750 0.3750 0.1667 0.3333 0.2500 0.2500 0.5000 0.5000 0.0000 0.0000 4.3595 4.2577 6.5154 6.8943 8.2520 8.7024 9.2429 8.2429 7.2429 6.2429 5.2429 4.9500 4.9375 5.2500 5.6667 4.6667 5.5000 4.5000 3.5000 2.5000 1.5000 1.0000 0.9999 1.8791 2.2073 2.9038 3.2745 3.5045 3.2931 2.7589 2.4583 2.0416 1.8467 1.8167 1.8064 1.7934 1.5287 1.6849 1.3483 1.0321 0.5000 0.0000 0.0000 1 Only data from females were used. To maximize sample size, we pooled animals by age, regardless of year of birth. Obs represents number of animals at risk in each age class. “Obs” do not equal absolute number of lemurs observed (N ⫽ 58), because as a lemur increases with age, she is represented in increasing number of age classes. Fig. 2. Mortality by age, from data in Table 2. U-shaped pattern shows very high infant mortality, high juvenile mortality, undetectable adult mortality to age 9, and increasing adult mortality thereafter. Upper ages have small sample sizes (n ⬍ 5 by age 10). interbirth interval for mothers with surviving infants is signiﬁcantly greater than the interbirth interval for mothers with infants that died before the next mating period. If an infant survives until the mother’s next mating period, the average interbirth interval is 1.8 ⫾ 0.54 years (N ⫽ 16). If the infant dies, the average interbirth interval is 1.1 ⫾ 0.33 years (N ⫽ 9). Fertility. Age-speciﬁc fertility rates (mx) for females show an onset of fertility at age 4 and an increase at age 6. Fertility remains high until death (Fig. 6 and Table 2). Because fertility stays constant after age 6, it appears unaffected by age or other factors leading to mortality. Life expectancy. Life expectancy peaks at age 5, at which age females can expect to live an additional 9 years (Fig. 7). Life expectancy stays constant from ages 8 –15, after which it declines. Fig. 3. Infant mortality by year. In each year, percent of female infants that die within their ﬁrst year. Reproductive value. Reproductive value peaks for 5-year-old females. A 5-year-old female who dies before her sixth birthday needs to produce 3.5 offspring in order to leave as many offspring as she would if she had instead survived to age 6. As Figure 8 shows, reproductive value remains constant for females between ages 8 –15, and it declines sharply thereafter to 0 for females aged 19 –20. Population Growth The intrinsic rate of increase is ⫺0.065, indicating a shrinking population. The net reproductive rate is 0.52. Since R0 is less than 1.0, the population is not replacing itself (Gotelli, 1998). DEMOGRAPHY OF PROPITHECUS DIADEMA EDWARDSI Fig. 4. Survivorship by age. Only 23.8% of females survive to reproductive age. Straight line represents type II survivorship curve. Fig. 5. Survivorship by age. Marked line shows survival curve generated from data using life-table methods. Plain line shows survival curve generated from survival analysis ﬁtting a parametric (Weibull) distribution, and controlling for censored data. 67 Fig. 7. Life expectancy by age, from data in Table 2. Additional years of life expected for females surviving to each age. Fig. 8. Reproductive value by age, from data in Table 2. This represents number of offspring a female dying before reaching age x ⫹ 1 would have to produce at age x in order to leave as many descendents as if she had instead survived at age x ⫹ 1 and enjoyed average survivorship and fertility thereafter. Comparative primate demography Fig. 6. Fertility (offspring per female aged x) by age x, from data in Table 2. This represents number of female births per year to females of each age. Fertility appears unaffected by age or other factors leading to mortality. Relationship between fertility and mortality Age-speciﬁc fertility and age-speciﬁc mortality are not correlated in this population (r ⫽ ⫺0.25; P ⫽ 0.25); fertility does not predict mortality. Sifaka survivorship very closely resembles the curve for New World monkeys in both shape and magnitude (Fig. 9). Both curves drop quickly and then decrease gradually. Conversely, the curve for Old World monkeys declines gradually throughout their lifetime. For the fertility curve (Fig. 10), however, these sifakas exhibit an Old World rather than New World pattern. Fertility for both Milne-Edwards’ sifakas and Old World monkeys increases gradually and then gradually declines, unlike the sharp early peak seen in New World primate fertility. Table 3 provides a comparison of population growth-rate estimates. Sifakas demonstrate the weakest population growth pattern. Table 4 provides the estimates of the number of female offspring surviving to reproductive age. 68 S.T. POCHRON ET AL. TABLE 3. Population growth estimates for the Milne-Edwards’ sifaka, composite New World monkeys, and composite Old World monkeys Fig. 9. Survivorship (lx) curves for Milne-Edwards’ sifaka and composite Old and New World monkeys. Composite data come from Gage and Dyke (1993) and Dyke et al. (1993); data for sifakas come from Table 2. Fig. 10. Fertility (mx) curves for Milne-Edwards’ sifaka and composite Old and New World monkeys. Composite data come from Gage (1998). To smooth sifaka fertility curve, we regressed age and fertility from Table 2 and used quadratic regression curve. Relationship is highly signiﬁcant (R2 ⫽ 0.57; N ⫽ 25; F ⫽ 14.31; P ⫽ 0.001). DISCUSSION Population growth and life history The negative intrinsic rate of increase and low net reproductive rate both indicate a decline in number for Propithecus diadema edwardsi. Because net reproductive rate adjusts for the mortality schedule, we can suggest that mortality rather than fertility causes its low value (Gotelli, 1998). The mortality schedule itself shows highest levels among newborns and high levels in juveniles; it is undetectable Model Net reproductive rate R0 Intrinsic rate of increase r Milne-Edwards’ sifaka New World monkeys Old World monkeys 0.520 0.679 1.902 ⫺0.065 ⫺0.068 0.056 for sifakas aged 5– 8. The survivorship schedule is constant across age, with the exception of the youngest members, who have poor survivorship. Also, the life-expectancy calculations show that if a sifaka lives to age 5, she can expect to live 9 more years. These conclusions need to be tempered somewhat due to the censored nature of the data, which results in some overestimation of juvenile mortality and adult survivorship (Fig. 5). However, the estimates of infant mortality are not affected by censoring, and parametric survival functions do not map well to the extremely low young-adult mortality rates observed, lending credence to the high juvenile mortality and adult survivorship values obtained from the lifetable analysis. Experimental ﬁtting of nonparametric hazard functions (Kaplan-Meier), which allow for right-censored data, shows lower juvenile mortality than predicted by the parametric model, and higher adult survivorship than predicted by analyses ignoring censoring, and it conﬁrms the nonmonotonicity of the survival function (unpublished results). Mortality and fertility appear to be independent in this population. Low birth rates can contribute to population decline, but this does not describe sifaka demography. In Propithecus diadema edwardsi, fertility does not signiﬁcantly vary with age, even for the oldest females in the study. Reproductive value is high from the onset of reproduction until age 15, supporting the statement that fertility issues are not responsible for population decline. While ecological factors play a role in demographic measures, we do not focus on them in this paper. The lack of correlation between age-speciﬁc fertility and age-speciﬁc mortality indicates that mortality at most ages is unlikely to be caused by a lack of resources or limited range size. However, if limited resources cause poor maternal nutrition, a lack of resources could result in high infant mortality. Moreover, the high mortality rates among newborns and juveniles indicate that predation, infanticide, or catastrophe are likely leading to the observed population decline. In small groups of primates, predation can have an overwhelming impact at the local level (Isbell, 1990). This pattern ﬁts the observation that predation by fossa (Cryptoprocta ferox) causes 64% of all female deaths (Table 1). Lemur demography As Overdorff et al. (1999) reported, current information concerning prosimian life history is based primarily on four Malagasy species from three gen- 69 DEMOGRAPHY OF PROPITHECUS DIADEMA EDWARDSI TABLE 4. Estimates of number of female offspring surviving to reproductive age and parameters used to create estimates for Milne-Edwards’ sifaka, composite New World monkeys, and composite Old World monkeys Milne-Edwards’ sifaka New World monkeys Old World monkeys Average age of fertility onset (years)1 Fertility onset2 Lifetime fertility3 Offspring surviving to reproduce4 4 2 5 14 ⫽ 0.238 12 ⫽ 0.510 15 ⫽ 0.798 4.36 6.62 3.34 1.04 3.38 2.67 1 Data from Gage (1998). Represents survivorship (1x) at age of fertility onset. Data from Dyke et al. (1993) and Gage and Dyke (1993). 3 Summation of fertility (mx) over lifetime from Gage (1998). Represents number of female offspring average female will have over course of life, assuming she lives a complete lifespan. 4 This is product of survivorship at age of fertility onset and lifetime fertility. It estimates number of female (or male) offspring that live to reproductive age to be born to a female who lives a complete lifespan. 2 era: Eulemur fulvus rufus, Lemur catta, Propithecus verreauxi, and P. diadema edwardsi. Consistent with most other lemurs, P. d. edwardsi reproduces for the ﬁrst time at age 4 years and has an average interbirth interval of 18 months. The average for females with surviving infants is signiﬁcantly longer (1.8 years) than for females with infants who died (1.1 years). The age at ﬁrst reproduction is 4 years for Eulemur fulvus rufus (Overdorff et al., 1999), 2– 4 years for Lemur catta (Sussman, 1991; Koyama et al., 2001), and 2.5–5 years for P. verreauxi (Richard et al., 1991). The mean interbirth interval is 18.6 months for E. f. rufus (Overdorff et al., 1999) and 18 months for P. verreauxi (Richard et al., 1991). Lemur catta have a shorter interbirth interval than other lemurs, at 13–14.4 months (Koyama et al., 2001; Sussman, 1991). The age-speciﬁc fertility of Propithecus diadema edwardsi, like that of Lemur catta (Koyama et al., 2001), shows nearly constant fertility across age. This contrasts with the age-speciﬁc fertility of P. verreauxi, which climbs after age at ﬁrst reproduction and then declines (Richard et al., 1991), and probably of Eulemur fulvus rufus, which demonstrates an increasing reproductive value with age (Overdorff et al., 1999). The strong birth seasonality (73% of births occur in June) shown by Propithecus diadema edwardsi is not unusual for lemurs or prosimians (van Horn and Eaton, 1979; Pereira and Weiss, 1991; Kappeler, 1997). Each species of Malagasy lemur shows strict breeding seasonality, with a limited estrous period for each female in each year (Wright, 1999). Nor is the length of birth season demonstrated by P. d. edwardsi unusual for lemurs. Richard et al. (1991) report that the annual birth season for P. verreauxi lasts about 1 month (from mid-June to mid-July); Overdorff et al. (1999) reported a 1-month birth season (September) for Eulemur fulvus rufus; Koyama et al. (2001) reported that 82% of Lemur catta infants in Berenty are born in September, and Sussman (1991) reported a 5-week birth period for L. catta in Beza Mahafaly. These lemur species each show high infant mortality, consistent with Propithecus diadema edwardsi. Eulemur fulvus rufus has an infant mortality rate of 41% (Overdorff et al., 1999). P. verreauxi infant mortality rates are 53–70%. Sussman (1991) reported an infant mortality rate of 30 –52% for Lemur catta, while Koyama et al. (2001) reported a rate of 37.7% in Berenty, a site that is occasionally provisioned. These numbers are consistent with our value of 50%. In short, P. d. edwardsi appears to demonstrate normal lemur demographic patterns. Comparative primate demography While population growth estimates indicate a shrinking population for Propithecus diadema edwardsi, and life-history parameters indicate that the decline derives from mortality rather than fertility issues, a comparison to other primate life histories might help place this lemur’s pattern into a primate perspective. Dyke et al. (1993), Gage and Dyke (1993), and Gage (1998) provided composite models of survivorship and birth rates for New World and Old World monkeys to which we compared our sifaka data. They also provided composite net reproductive rates and intrinsic rates of increase. Figures 9 and 10 imply that sifakas demonstrate the most acute primate pattern. Although New World monkeys have low survivorship at early ages, they demonstrate early and high reproduction rates. And although Old World monkeys delay and spread out their fertility, they have high relative survivorship throughout their lifespan, even at the youngest ages. Sifakas survive poorly, especially at young ages, and their fertility peaks slowly and at a low level. Few female sifakas (23.8%) survive to reproductive age, and those that do, reproduce at a slow rate more similar to Old World than New World monkeys. Table 3 shows that sifakas demonstrate the weakest population growth pattern. These ﬁndings are consistent with those of Richard et al. (2002), who reported that for Propithecus verreauxi verreauxi, females stood out in a comparative allometric analysis (using 16 wild and unprovisioned primate populations), not only because sifakas took relatively longer than any other primate to begin reproduction, but also because they were the only folivorous primate in the sample to begin reproduction later, rather than earlier, than average. Following Stearns (1992), Richard et al. (2002) argued that for sifakas, it pays to reduce reproduc- 70 S.T. POCHRON ET AL. tive effort in order to live longer and reproduce more times, sampling a larger number of reproductive conditions and increasing the number of offspring born into good conditions. Records from the Duke University Primate Center show that captive lemurs live a remarkably long time for small-bodied primates. A male Propithecus verreauxi verreauxi lived 30.5 years, a female Varecia variegata variegata lived 36.0 years, and a female Eulemur fulvus lived approximately 32 years. Such longevity would allow lemurs to bear offspring into a variety of environmental conditions, as suggested by Richard et al. (2002). This may be a particularly beneﬁcial strategy for animals reproducing in Madagascar’s harsh and unpredictable environment (Overdorff et al., 1999; Wright, 1999; Pochron and Wright, 2002; Richard et al., 2002). The possibility that the pattern of high pre-reproductive mortality coupled with slow and late reproduction is common to other island biota is worth further investigation. We note that the comparison of the sifaka pattern to composite Old and New World patterns provides a heuristic that helps place sifakas into a primate perspective. While the curves portrayed from composite life-history models characterize no single population, a comparison of survivorship between composite primate species provides interesting information. As a whole, neither the Old nor New World standard model life table describes the Milne-Edwards’ sifaka, but components of the composite models ﬁt remarkably well. This ﬁnding begs the question: how does lemur demography inﬂuence social structure in ways that contrast patterns seen in Old and New World monkeys? To answer this question, we examine possibilities available to sifakas concerning social structure, particularly group membership and migration patterns. For instance, if lemurs found themselves in an environment that favored philopatry by either sex, would they have the same options available to them as, say, chimpanzees, baboons, or squirrel monkeys? In other words, if lemurs experience appropriate selection pressure, would their demography permit them to form core groups of related females (or males) while the males (or females) dispersed? If female bondedness (or male bondedness) provided lemurs with an advantage in their past environments, we might expect to see either male primary dispersal or female primary dispersal. We might also expect to see a clear preference for one mating system over another (van Noordwijk and van Schaik, 1999). Instead, some male sifakas migrate and some female sifakas migrate (unpublished data), and they have no preference for any mating system. Although we have not looked for environmental correlates with migration patterns, environmental variation has not been able to explain variation in mating systems (Pochron and Wright, 2003). Demography may offer an avenue for explanation. For female bondedness to occur, an adult female must have daughters of breeding age available to her. For male bondedness to occur, an adult male must have adult sons available to him. The life-history data (see Results) show the calculation of how many daughters of breeding age a female sifaka can expect to have while she herself is breeding, and compare that number to composite New and Old World monkeys. As Table 4 shows, an adult sifaka at the end of her lifespan will have only one daughter who survives to reproductive age. Composite New World patterns show that an adult female will have 3.4 daughters available to her, and Old World females have 2.7. Sifakas would need to increase their lifetime fertility to 12.6 offspring (from 4.36) or increase their survivorship to 0.688 (from 0.238) to gain three adult female offspring reproducing at the same time as their mother, a level consistent with both New and Old World monkeys. Female (or male) bondedness could potentially arise via siblings rather than mothers and daughters (or fathers and sons). However, since a mother sifaka at the end of her lifespan has only one daughter of breeding age, while Old and New World monkeys have about three, sibships of either sex are highly unlikely for sifakas; they are not unlikely for monkeys. For sifakas, high infant mortality rates preclude bondedness of same-sex siblings. These comparisons indicate that, regardless of pressure for female bondedness (or male bondedness), for these sifakas, forming large kin-based groups is difﬁcult due to demography. Survivorship and fertility patterns do not easily permit sifakas to form large same-sex family groups. Life history and demography may force lemurs into opportunistic social structures and away from sex-speciﬁc stable patterns of bonding. High pre-reproductive mortality, coupled with slow and late reproduction, may prune the populations in such a way as to leave no options save ﬂuid social and mating systems. CONCLUSIONS The population decline suggested by this paper should be taken with some caution. Few primate populations show numerical stability (Dunbar, 1988). Without human interference, many primate populations would likely show alternating periods of decline and growth (Mittermeier and Cheney, 1986; Sussman, 1991). Despite indicators of population decline, alternating periods of decline and growth may best describe Propithecus diadema edwardsi. Over time, groups seem to grow and shrink, oscillating between 2–5 animals and back again. Furthermore, no animal born since the onset of this study has died of old age. Because these animals are long-lived, we have seen many infants born, but few animals complete their lifespan. This means that we have oversampled the young animals and undersampled the old. Since mortality lies primarily in the young animals, its strength may be overstated. DEMOGRAPHY OF PROPITHECUS DIADEMA EDWARDSI Lastly, despite the time depth of this project, and despite the fact that multiple groups were studied each year, only 58 animals have been observed over the duration of this study. Within age/sex classes, sample size quickly diminishes, as can be seen in Table 2. As in any sampling situation, size is an important aspect of data quality. Numbers of vital events vary by chance from year to year, and no matter how carefully a population is studied, this variability cannot be eliminated. The extent to which stochastic variability can inﬂuence measurements of underlying mortality rates is inversely proportional to population size (Gage and Dyke, 1988), which is demonstrably small for Propithecus diadema edwardsi. These factors together may result in an overly pessimistic measure of population decline. It remains to be seen whether this population decline will continue into the future. ACKNOWLEDGMENTS We acknowledge the support and assistance of ANGAP, the Université d’Antananarivo, the Université de Fianarantsoa, ONE, DEF, MICET, and the Ministre de la Environnement. The SUNY Stony Brook Research Foundation and ICTE are thanked for their support. We give special thanks to our ICTE RNP research assistants: the late Georges Rakotonirina, Raymond Ratsimbazafy, and Rémi Rakotosoa. The Wenner Gren Foundation for Anthropological Research, the John D. and Catharine T. MacArthur Foundation, the Liz Claiborne and Art Ortenberg Foundation, the USAID UDLP Program, the Earthwatch Institute, and the Douroucouli Foundation are thanked for their assistance with funding. Also, we thank Diane Doran for her very helpful comments on the manuscript, Jukka Jurnvall for his helpful comments, and Wayne Linklater for some ideas that we tested here. We also thank two anonymous reviewers and Clark Spenser Larsen for their extraordinarily helpful comments that made this paper much richer. LITERATURE CITED Altmann J, Alberts SC, Haines SA, Dubach J, Muruthi P, Coote T, Geffen E, Cheeseman DJ, Mututua RS, Saiyalel SN, Wayne RK, Lacy RC, Bruford MW. 1996. Behavior predicts genetic structure in a wild primate group. Proc Natl Acad Sci USA 93:5797–5801. Cox DR, Oakes D. 1984. Analysis of survival data. London: Chapman and Hall. Dunbar RIM. 1984. Reproductive decisions. Princeton: Princeton University Press. Dunbar RIM. 1988. Primate social systems. New York: Comstock. Dyke B, Gage TB, Mamelka PM, Goy RW, Stone WH. 1986. A demographic analysis of the Wisconsin Region Primate Center rhesus colony, 1962–1982. Am J Primatol 10:257–269. Dyke B, Gage TB, Ballou JD, Petto AJ, Tardif SD, Williams LE. 1993. Model life tables of the smaller New World monkeys. Am J Primatol 29:269 –285. Gage TB. 1998. The comparative demography of primates: with some comments on the evolution of life histories. Annu Rev Anthropol 27:197–221. Gage TB, Dyke B. 1988. Model life tables for the larger Old World monkeys. Am J Primatol 16:305–320. 71 Gage TB, Dyke B. 1993. Model life tables for the larger Old World monkeys: a revision. Am J Primatol 29:287–290. Glander KE, Wright PC, Daniels PS, Merenlender AM. 1992. Morphometrics and testicle size of rain forest lemur species from southeastern Madagascar. J Hum Evol 22:1–17. Gotelli NJ. 1998. A primer of ecology. Sunderland, MA: Sinauer Associates, Inc. Ha JC, Robinette RL, Sacket GP. 2000. Demographic analysis of the Washington Regional Primate Research Center pigtailed macaques colony, 1967–1996. Am J Primatol 52:187–198. Hemingway CA. 1996. Morphology and phenology of seeds and whole fruit eaten by Milne-Edwards sifaka, Propithecus diadema edwardsi, in Ranomafana National Park. Int J Primatol 17:637– 659. Isbell LA. 1990. Sudden short-term increase in mortality of vervet monkeys (Cercopithecus aethiops) due to leopard predation in Amboseli National Park, Kenya. Am J Primatol 21:41–52 Kappeler P. 1997. Intrasexual selection and testis size in strepsirhine primates. Behav Ecol 8:10 –19. Kappeler P. 2000. Causes and consequences on unusual sex ratios among lemurs. In: Kappeler P, editor. Primate males. Cambridge: Cambridge University Press. p 55– 63. Koyama N, Nakamichi M, Oda R, Takahata Y. 2001. A ten-year summary of reproductive parameters for ring-tailed lemurs at Berenty, Madagascar. Primates 42:1–14. Mittermeier RA, Cheney DL. 1986 Conservation of primates and their habitats. In: Smuts BB, Cheney DL, Seyfarth RM, Wrangham RW, Struhsaker TT, editors. Primate societies. Chicago: University of Chicago Press. p 477– 490. Overdorff DJ. 1993. Similarities, differences, and seasonal patterns in the diets of Eulemur rubriventer and E. fulvus rufus in the Ranomafana National Park, Madagascar. Int J Primatol 14:721–753. Overdorff DJ, Merenlender AM, Talata P, Telo A, Forward ZA. 1999. Life history of Eulemur fulvus rufus from 1988 –1998 in southeastern Madagascar. Am J Phys Anthropol 108:295–310. Pereira ME, Weis ML. 1991. Female mate choice, male migration, and the threat of infanticide in ring-tailed lemurs. Behav Ecol Sociobiol 28:141–152. Pianka ER. 1983. Evolutionary ecology. New York: Harper and Row. Pochron ST, Wright PC. 2002. Dynamics of testes size compensates for variation in male body size. Evol Ecol Res 4:577–585. Pochron ST, Wright PC. In review. Predicting the number of males: a long-term view of the operational sex ratios from Propithecus diadema edwardsi in Ranomafana National Park, Madagascar (1986 –2000). Behav Ecol Sociobiol. Pochron ST, Wright PC, Schaentzler E, Ippolito M, Rakotonirina G, Ratsimbazafy R, Rakotosoa R. 2002. Effect of season and age on the gonadosomatic index of Milne-Edwards sifakas (Propithecus diadema edwardsi) in Ranomafana National Park, Madagascar. Int J Primatol 23. Rasmussen DR. 1981. Communities of baboon troops (P. cynocephalus) in Mikumi National Park, Tanzania. Folia Primatol (Basel) 36:232–242. Richard AF, Rakotomanga P, Schwartz M. 1991. Demography of Propithecus verreauxi at Bezy Mahafaly, Madagascar: sex ratio, survival and fertility. Am J Phys Anthropol 84:307–322. Richard AF, Dewar RE, Schwartz M, Ratsirarson J. 2002. Life in the slow lane? Demography and life histories of male and female sifaka (Propithecus verreauxi verreauxi). J Zool Lond 256:421– 436. Rowe N. 1996. The pictorial guide to the living primates. East Hampton: Pogonias Press. Sade DS, Cushing K, Cushing P, Duaif J, Figueroa A, Kaplan JR, Lauer C, Rhodes D, Schneider J. 1976. Population dynamics in relation to social structure on Cayo Santiago. Yrbk Phys Anthropol 20:253–262. Smith DG. 1982. A comparison of the demographic structure and growth of free-ranging and captive groups of rhesus monkeys (Macaca mulatta). Primates 23:24 –30. Smuts BB. 1985. Sex and friendship in baboons. New York: Aldine de Gruyter. 72 S.T. POCHRON ET AL. Stearns SC. 1992. The evolution of life histories. New York: Oxford University Press. Sussman RW. 1991. Demography and social organization of freeranging Lemur catta in the Beza Mahafaly Reserve, Madagascar. Am J Phys Anthropol 84:43–58. Van Horn RN, Eaton GG. 1979. Reproductive physiology and behavior in prosimians. In: Doyle GA, Martin RD, editors. The study of prosimian behavior. New York: Academic Press. p 79–122. Van Noordwijk MA, van Schaik CP. 1999. The effects of dominance rank and group size on female lifetime reproductive success in wild long-tailed macaques (Macaca fascicularis) Primates 40:105–130. Wood JW. 1994. Dynamics of human reproduction: biology, biometry, demography. New York: Aldine de Gruyter. Wood JW, Holman DJ, Weiss KM, Buchanan AV, Lefor B. 1992. Hazards models for human population biology. Yrbk Phys Anthropol 35:43– 87. Wright PC. 1992. Primate ecology, rainforest conservation and economic development: building a national park in Madagascar. Evol Anthropol 1:25–33 Wright PC. 1995. Demography and life history of free-ranging Propithecus diadema edwardsi in Ranomafana National Park, Madagascar. Int J Primatol 16:835– 854. Wright PC. 1997. The future of biodiversity in Madagascar: a view from Ranomafana National Park. In: Goodman SG, Patterson BD, editors. Natural change and human impact in Madagascar. Washington, DC: Smithsonian Institution Press. p 381– 405. Wright PC. 1998. Impact of predation risk on the behaviour of Propithecus diadema edwardsi in the rain forest of Madagascar. Behaviour 135:483–512. Wright PC. 1999. Lemur traits and Madagascar ecology: coping with an island environment. Yrbk Phys Anthropol 43:31– 72.