вход по аккаунту


Demography life history and social structure in Propithecus diadema edwardsi from 1986Ц2000 in Ranomafana National Park Madagascar.

код для вставкиСкачать
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
lemurs; life table; mortality; fertility; population growth; group structure
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-
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:
Received 3 January 2002; accepted 19 December 2002.
DOI 10.1002/ajpa.10266
Published online 12 January 2004 in Wiley InterScience (www.
TABLE 1. Causes of death for females1
Age bracket
Unknown causes
All ages
5 (20.0%)
3 (23.1%)
4 (22.2%)
1 (14.3%)
3 (12.0%)
3 (23.1%)
3 (16.7%)
16 (64.0%)
6 (46.2%)
10 (55.6%)
6 (85.7%)
1 (4.0%)
1 (7.7%)
1 (5.6%)
0 (0%)
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.
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.
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
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
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.
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
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
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
4) The data had to exhibit a monotonic decline in
mortality rates in the early years, and a monotonic increase in older years.
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
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
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.
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
TABLE 2. Standard life table for Milne-Edwards’ sifakas1
Life expectancy
value (vx/v0)
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).
Fig. 4. Survivorship by age. Only 23.8% of females survive to
reproductive age. Straight line represents type II survivorship
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
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.
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).
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
Net reproductive
rate R0
Intrinsic rate
of increase r
Milne-Edwards’ sifaka
New World monkeys
Old World monkeys
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-
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
Offspring surviving
to reproduce4
14 ⫽ 0.238
12 ⫽ 0.510
15 ⫽ 0.798
Data from Gage (1998).
Represents survivorship (1x) at age of fertility onset. Data from Dyke et al. (1993) and Gage and Dyke (1993).
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.
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.
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-
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.
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.
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.
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.
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
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.
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
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
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.
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–
Без категории
Размер файла
149 Кб
edwards, parks, propithecus, 1986ц2000, history, madagascar, diadema, demographic, structure, social, ranomafana, national, life
Пожаловаться на содержимое документа