The Role of Anthropic, Ecological, and Social Factors in Sleeping Site Choice by Long Tailed Macaques (Macaca fascicularis).
American Journal of Primatology 76:1140–1150 (2014)
RESEARCH ARTICLE
The Role of Anthropic, Ecological, and Social Factors in Sleeping Site Choice
by Long‐Tailed Macaques (Macaca fascicularis)
FANY BROTCORNE1,2*, CINDY MASLAROV1, I. NENGAH WANDIA3, AGUSTIN FUENTES4,
ROSELINE C. BEUDELS‐JAMAR2, AND MARIE‐CLAUDE HUYNEN1
1
Primatology Research Group, Behavioural Biology Unit, University of Liège, Liège, Belgium
2
Conservation Biology Unit, Education and Nature, Royal Belgian Institute of Natural Sciences, Brussels, Belgium
3
Primate Research Center, Universitas Udayana, Bali, Indonesia
4
Department of Anthropology, University of Notre Dame, Notre Dame, Indiana
When choosing their sleeping sites, primates make adaptive trade‐offs between various biotic and
abiotic constraints. In human‐modified environments, anthropic factors may play a role. We assessed
the influence of ecological (predation), social (intergroup competition), and anthropic (proximity to
human settlements) factors in sleeping site choice by long‐tailed macaques (Macaca fascicularis)
occupying a habitat at the interface of natural forests and human‐modified zones in Bali Barat National
Park, Indonesia. Over the course of 56 nights, we collected data relating to physical features of sleeping
trees, patterns of the use of sleeping sites within the home range, pre‐sleep behavior, diurnal ranging
patterns and availability of natural and human food. Overall, the macaques used 17 sleeping sites with
37 sleeping trees. When the monkeys slept in forest zones, they selected sleeping trees that had larger
trunks but were not significantly taller than surrounding trees. Though the macaques rarely re‐used
sleeping sites on consecutive nights, they frequently re‐used four sites over the study period. The group
favored sleeping within the core area of its home range, despite the occurrence of frequent agonistic
intergroup encounters there. Macaques preferentially selected sleeping trees located within or near
human‐modified zones, especially when human food was abundant and natural food was scarce. These
results partially support the hypothesis that long‐tailed macaques choose their sleeping sites to avoid
predation; proximity to human settlements appears to be the primary factor influencing sleeping site
choice in this primate species. Our results reflect the strong influence that anthropic factors have on
primates, which subsist in increasingly human‐dominated landscapes. Am. J. Primatol. 76:1140–1150,
2014. © 2014 Wiley Periodicals, Inc.
Key words:
sleeping site; human proximity; human food; predation avoidance; Bali Barat National
Park
INTRODUCTION
The choice of sleeping sites by diurnal primates
reflects diverse factors and constraints that are
mutually non‐exclusive and highly dependent on
context and the species in question. Predation
pressure, proximity of food, competition with conspecifics and physical comfort may influence the
choice of sleeping sites and sleeping trees by primates
[Anderson, 1998, 2000].
The predation avoidance hypothesis asserts that
primates have evolved anti‐predator strategies at
sleeping sites. To minimize detection by a predator,
they should select tall emergent trees [Bernard
et al., 2011; Fan & Jiang, 2008; Reichard, 1998] and
behave in a cryptic manner (e.g. moving silently and
rapidly) when entering sleeping trees [Heymann,
1995; Liu & Zhao, 2004; Smith et al., 2007]. Some
primates may prefer open‐canopy trees located along
© 2014 Wiley Periodicals, Inc.
riverbanks [Fittinghoff & Lindburg, 1980; Matsuda
et al., 2008] and connected to adjacent trees [Albert
et al., 2011]. These conditions improve their chances
of detecting approaching predators and facilitate
Contract grant sponsor: Belgian National Fund for Scientific
Research; contract grant sponsor: Fondation Belge de la
Vocation
Correspondence to: Fany Brotcorne, Primatology Research
Group, Behavioral Biology Unit, University of Liège, 22 Quai
Van Beneden, Liège 4020, Belgium. E‐mail: fbrotcorne@gmail.
com
Received 22 November 2013; revised 9 April 2014; revision
accepted 10 April 2014
DOI: 10.1002/ajp.22299
Published online 8 May 2014 in Wiley Online Library
(wileyonlinelibrary.com).
Sleeping Sites of Balinese Macaques / 1141
escape [Hankerson et al., 2007; Kurland, 1973].
Alternatively, sleeping trees with a crown not
touching the neighboring trees may offer protection
against arboreal mammalian predators [Barnett
et al., 2012]. Some researchers argued that a dense
canopy cover improves the concealment of primates
[Anderson, 2000; Liu & Zhao, 2004]. Physical barriers
such as large tree trunks, tall crowns, elevated first
branches, and the absence of lianas serve to impede
the access of predators into trees [Barnett et al., 2012;
Bernard et al., 2010; Di Bitetti et al., 2000; Ramakrishnan & Coss, 2001; Teichroeb et al., 2012;
Tenaza & Tilson, 1985]. As predators may memorize
the sleeping refuges of their prey [Emsens et al.,
2014], primates are expected to switch between
several alternate sleeping sites, thus, making them
less predictable for predators [Phoonjampa et al.,
2010; Reichard, 1998; Zhang, 1995]. Alternatively,
primates may re‐use the same sleeping sites repeatedly because suitable sites are rare [Duarte &
Young, 2011; Ramakrishnan & Coss, 2001; Tenaza
& Tilson, 1985], or because familiarity with sleeping
sites facilitates escape from predators [Di Bitetti
et al., 2000; Hankerson et al., 2007; Li et al., 2011].
According to the food proximity hypothesis,
primates select sleeping sites that maximize their
access to food patches and at the same time minimize
travel costs. Sleeping sites should be near the feeding
sites used prior to or just after the period of rest
[Chapman et al., 1989; Heymann, 1995; Li et al.,
2011; Smith et al., 2007; Teichroeb et al., 2012].
Competition with conspecific groups may also
influence the choice of sleeping site location within the
home range. Seeping in areas of exclusive use may
decrease the risk of intergroup encounters [Albert
et al., 2011; Li et al., 2011; Phoonjampa et al., 2010;
Smith et al., 2007; Von Hippel, 1998], while sleeping
near range boundaries may allow these boundaries or
nearby resources to be defended [Teichroeb et al.,
2012], depending on the species’ degree of territoriality [cf. risk hypothesis: Wrangham et al., 2007].
Additionally, factors such as parasite avoidance,
thermoregulation and the search for comfort, stability, and social contact in trees may play a role in the
choice of sleeping sites [Anderson, 1998; Di Bitetti
et al., 2000; Fan & Jiang, 2008; Liu & Zhao, 2004;
Zhang, 1995].
Primate habitat is becoming increasingly modified
by humans and anthropogenic change is having large
impacts on primate behavior and ecology [e.g. Fuentes
& Hockings, 2010; McKinney, 2009; Strum, 2010].
Predation pressure may be lessened near human
settlements as here predators have been persecuted by
humans and their populations reduced [Bishop
et al., 1981; Isbell & Young, 1993; Ramakrishnan &
Coss, 2001; Stanford, 2002]. Studying the choice of
sleeping sites in environments at the interface of forest
and human‐influenced habitats contributes to a better
understanding of how primates cope behaviorally with
human encroachment. In terms of predation risk, only
a few studies have analyzed sleeping site choice in
human‐dominated landscapes [Duarte & Young, 2011;
Ramakrishnan & Coss, 2001]. In this study, we
investigated the role of ecological, social and anthropic
factors in the sleeping site choice of a group of long‐
tailed macaques (Macaca fascicularis) living in a partly
human‐modified habitat within the Bali Barat National
Park (BBNP), Indonesia.
The geographical range of M. fascicularis extends
across southeast Asia from Bangladesh to the Sunda
Archipelago of Indonesia [Fooden, 1995]. Although in
the past the species occurred mainly in riverine
forests [Fittinghoff & Lindburg, 1980; Kurland,
1973], in more recent times its behavioral and
ecological flexibility (including a generalist, opportunistic feeding strategy) has enabled M. fascicularis to
colonize a wide variety of habitats [Fooden, 1995;
Gumert et al., 2011]. In particular, the species has
succeeded in exploiting fragmented forests and the
edges of disturbed habitats [Gumert et al., 2011].
Although the species has long been associated
with humans throughout its range [Fuentes &
Hockings, 2010], recently it has increasingly entered
into commensal relationships with humans, living in
close association with people, taking advantage of
human food to supplement its diet [Wheatley, 1999],
and even competing with humans for spatial and
dietary resources [Richard et al., 1989]. The island of
Bali has a particularly intense and widespread
human‐primate interface; long‐tailed macaques and
humans have coexisted and interacted for centuries
[Fuentes et al., 2005; Wheatley, 1999]. Fuentes
[2010] suggested that the long‐term sympatry between humans and M. fascicularis in Bali could have
resulted in intertwined ecologies, with humans
exerting selective pressures on the macaques.
The monkeys’ sleeping sites are one potential
point of influence. Among the factors potentially
influencing sleeping site choice, we focused on
predation and intergroup competition avoidance, as
well as proximity to human settlements and human
food. We investigated the following predictions: (i) If
avoidance of predators guides sleeping site choice,
macaques should use multiple sleeping sites and
alternate between them, and they should adopt
cryptic pre‐sleep behavior to decrease the risk of
detection by predators, (ii) Macaques could be safer
when sleeping in human‐modified zones (cf. definition in Methods) where natural predation pressure
was likely reduced. In this case, we expected a
habitat‐specific choice of sleeping trees, with macaques selecting taller trees with larger trunks only
when sleeping in forest zones, (iii) following the risk
hypothesis [Wrangham et al., 2007], we predicted
that the macaques should avoid sleeping in areas of
their home range where they frequently encountered
other groups, (iv) due to the long history of human‐
macaque interface on Bali, we predicted that
Am. J. Primatol.
1142 / Brotcorne et al.
anthropic factors should influence the habitat use
and sleeping site choice of the macaques. Particularly, we expected the macaques to use the human‐
modified zones and their surroundings because of the
food left by people. The preference for certain sleeping
site locations should also vary with food availability
within the home range. During the months of high
tourist activity, when the availability of human food
was substantial, we expected the macaques to sleep
more frequently near human‐modified zones.
METHODS
Study Site and Subjects
Bali Barat National Park (BBNP) is located in
the north‐western part of the island of Bali, Indonesia
(8°05’S–18°15’S, 114°25’E–114°56’E). The park covers an area of 19,002 ha comprising dry deciduous
monsoon forest, lowland rain forest, coastal forest,
and mangrove. The climate is monsoonal, with a dry
season lasting from May to September and a wet
season lasting from October to April; the annual
average rainfall is 1,198 mm. During the study
period, sunset time ranged between 18:08 and
18:33, and sunrise was between 06:23 and 06:34
(http://www.timeanddate.com).
We conducted this study in the western sector of
the park, an area of deciduous monsoon and coastal
forests located around the park headquarters,
which included a ranger station, tourist facilities,
a camping area, and a Hindu temple. Two roadways
crossed the study site (Fig. 1). Tourists visit the
park throughout the year, with peak numbers from
May to September (in 2013, monthly mean during
high tourist season ¼ 7,715 visitors vs. low tourist
season ¼ 2,423 visitors). Human‐derived food (hereafter referred to as human food), occasionally present
in human‐modified zones and alongside roadways,
mainly consisted of offerings (i.e. fruit or crackers)
placed in temples, or refuse and leftovers (i.e. rice and
fruits) in bins or scattered across the ground.
There were two non‐human primate species at
BBNP, the long‐tailed macaque (M. fascicularis) and
the ebony langur (Trachypithecus auratus) [Leca
et al., 2013; Wheatley et al., 1993]. Although we did
not quantify predator density in the study area, we
observed several potential predators of M. fascicularis: the reticulated python (Python reticulatus), the
water monitor lizard (Varanus salvator), diurnal and
nocturnal raptors such as changeable hawk‐eagle
(Nisaetus cirrhatus), Indian black eagle (Ictinaetus
malayensis), barred eagle‐owl (Bubo sumatranus), as
well as domestic dogs [Fam & Nijman, 2011;
Fooden, 1995; van Schaik & Mitrasetia, 1990]. Tigers
are known predators of M. fascicularis at other sites
[van Schaik et al., 1983], but the Bali tiger (Panthera
tigris balica) has been extinct since around 1940
[Whitten et al., 1996]. The leopard cat (Prionailurus
bengalensis), which has been reported as a potential
predator of macaques [Palombit, 1992], is the sole
wild felid that still occurred in the park (H.
Kesumanegara, personal communication). Among
the aforementioned predators, only the python, owl,
and leopard cat represented a nocturnal threat to
macaques.
We studied a group of long‐tailed macaques
with 24–26 members including 3 adult males, 7
adult females, 3–4 subadult males, 6 juveniles, and
Fig. 1. Study site in Bali Barat National Park (Indonesia) with an outline of the study group home range, showing two habitat types: forest
(natural zones with continuous tree canopy, 40 ha) and human zones (ranger station and grass patches with planted trees, 9 ha).
Am. J. Primatol.
Sleeping Sites of Balinese Macaques / 1143
4–6 infants. During the study, two births occurred
and one subadult male disappeared.
Data Collection
We carried out the study over a period of four full
months, from March to June 2011. We allocated the
first month to habituating the macaques to observers,
so the analyzed data set covers a 3‐month period. We
followed the macaques 5 days per week (mean:
15 days per month; range: 14–17 days) from sleeping
site (06:00) to sleeping site (18:30) when possible,
resulting in 530 hr of observation. We ended our
observations when visibility became too low for us to
see the monkeys. We devoted some additional
afternoons to locating sleeping site locations on
days when we had not followed the study group
(n ¼ 13 days, with 11 successful attempts). In
total, we recorded the location of sleeping sites for
56 nights.
We use the term sleeping tree to refer to trees in
which macaques stayed overnight; sleeping site to
refer to the circular plot delimited by a 20 m radius
around the edges of the crowns of the sleeping trees
used by the group (one sleeping site could be
composed of one or more sleeping trees occupied
simultaneously); control tree to refer to trees present
within the sleeping site but never used for sleeping;
re‐used sleeping site to refer to sites used more than
once; and favored sleeping site to refer to sites used
more than five times by the group over the course of
the study [Reichard, 1998].
For each sleeping site, we recorded the GPS
location of the plot center (using a handheld GPS
Garmin 60CSx; 10 m error), as well as its distance to
the nearest roadway and to any human settlement
(using a 3 point‐scale: 0–20 m; 21–50 m; >50 m). We
also distinguished between two habitat types: the
“forest zones,” consisting of natural forested zones
with continuous tree canopy, and the “human‐
modified zones” (hereafter also referred as to “human
zones”), consisting of roadways, the ranger station
with its associated buildings, and grass patches
with planted trees (Fig. 1) [Sha & Hanya, 2013].
For each sleeping tree (n ¼ 37), we recorded the
species (with the collaboration of a local botanist, H.
Kesumanegara), diameter at breast height (DBH)
and total tree height (visual estimate always made by
the same observer). We took the same measures for
control trees (n ¼ 138), which were nearby trees (DBH
>10 cm) selected to represent the four quadrants of a
sleeping site equally. Additionally, we noted the time
the macaques entered and left the sleeping site, the
GPS location and nature of encounters with other
primate groups during the day and at sleeping sites,
and the occurrence of vocalizations such as loud
collective contact calls and those associated with
intra‐group aggression during retirement to sleeping
sites on an ad libitum basis [Altmann, 1974].
During our daily follows of the macaques, we
recorded the group’s GPS location at 20‐min intervals
(n ¼ 1,230) to reconstruct its daily ranging patterns.
When possible, we took the GPS position from the
group’s center (rough estimate of the typical group
spread ¼ 30 m). When this was not possible, we
tolerated a maximum average distance of 10 m
from the group’s periphery. In addition, we used
30 sec scans at 2.5 min intervals (n ¼ 5,037)
[Altmann, 1974] to record the habitat type at the
group’s location (i.e. forests vs. human zones), the
distance between the group’s center and the nearest
human settlement or roadway (using a 3 point‐scale:
0–20 m, 21–50 m, >50 m), and the presence of human
food in bins or elsewhere within ca. 30 m of the
center of the group (this distance was a rough
estimate always made by the same observer, corresponding to the maximum beyond which visibility
was restricted).
To investigate whether preferences for sleeping
site locations related to changes in natural fruit
availability within the home range, we conducted a
tree phenological survey. Across the center of the
study area, we established a 1,200 m 20 m transect
running north‐south, along which we surveyed twice
a month the phenological status of all trees with DBH
10 cm (0: absence of fruit, 1: presence of ripe fruit, 2:
presence of sun‐dried fruit) [Newton, 2007]. This
transect crossed all the habitat types present in the
group’s home range.
Data Analysis
To assess whether sleeping trees had larger
DBHs and were taller than nearby control trees
within the sleeping site, we used nested ANOVAs
(tree‐type nested within sleeping site). We tested for
these differences in both habitat types, forest versus
human zones. Since the residuals of the models did
not conform to a Shapiro–Wilk normality test, the
DBH was log‐transformed and height was square‐
root‐transformed [Sokal & Rohlf, 1995]. We used
Levene’s test to confirm non‐deviance for homoscedasticity [Dytham, 2003]. To test for randomness of
sleeping site re‐use, we generated an expected
frequency distribution based on Poisson lambda
parameters [Sokal & Rohlf, 1995] that we compared
with the observed frequency distribution of the
sleeping site re‐use using a Kolmogorov–Smirnoff
test for goodness‐of‐fit [Day & Elwood, 1999; Phoonjampa et al., 2010]. The theoretical Poisson distribution was appropriate for the counts of re‐use of
sleeping sites, which are rare events [Agresti, 2007].
Using ArcGIS 9.3.1 (ESRI) and the extension
Hawth’s Tool, we calculated the size of the home
range, both on a monthly and daily basis, with
the fixed Kernel method (95% kernels ¼ total home
range; 50% kernels ¼ core area). We assessed the
smoothing parameters href [Worton, 1989] with R
Am. J. Primatol.
1144 / Brotcorne et al.
3.0.2 and the package adehabitatHR. We compared
the daily home range sizes between months with a
Kruskall–Wallis test. To compare the locations of
sleeping sites and intergroup encounters within the
home range, we used Binomial tests with expected
values derived from the observed frequency of
macaques inside the core and peripheral areas,
which are, by definition, each used 50% of the time.
Using methods defined in previous studies
[Imaki et al., 2006; Sha et al., 2009], we established
a human proximity gradient within the home range
characterized by several buffer‐zones delimited by
their distance to any roads and/or human settlements, with the 3‐level scale used on the field: (i) 0–
20 m (including the human zones and their edges), (ii)
21–50 m (intermediary forest edge), and (iii) more
than 50 m (forest interior). We tested for random
buffer‐zone use by the macaques during their diurnal
activities and for sleeping using G‐tests for goodness‐
of‐fit with Williams’s correction [Sokal & Rohlf, 1995]
(with expected frequencies based on the surface
area of the buffers within the home range) and the
standardized residuals considered as significant
when the absolute value exceeded 2 [Agresti,
2007].
To analyze the variation in natural food availability, we used a fruit availability index (FAI) for
each tree species k and for each month m [Albert
et al., 2013]:
Month
FAIkm ¼ Dk Bk Pkm
where in Dk is the density of a given species k (number
of specimens on the phenology transect), Bk is the
mean basal area of species k, and Pk is the fruit
phenology score of species k in a given month m
(calculated as the percentage of specimens holding
ripe fruit on their crown). We used a Friedman test to
compare the FAI between months.
To analyze the effect of the human food (HF)
availability on the habitat type chosen for sleeping,
we assigned each observation day to a category of HF
availability (no/some HF vs. high HF), based on the
overall median of the daily proportion of scans when
HF was present around the macaque group (30 m
radius from the center of the group). We examined
variation in HF available among months using a
Kruskall–Wallis test. We used a G‐test of independence to evaluate whether selection of habitat type
for sleeping depended on month. Finally, we analyzed
the relationship between daily home range size and
daily HF availability with a Spearman correlation
test.
We performed the nested ANOVAs using R 3.0.2.
Other statistical tests were completed using STATISTICA 10.0 (two‐tailed statistics with a significance level set at 0.05). Data are presented as mean
values with standard deviations.
Am. J. Primatol.
Our study was approved by University of Liège,
Belgium, the Indonesian Ministry of Research and
Technology (#03B/TKPIPA/FRP/SM/III/2011), the
Indonesian Directorate of Forest Protection and
Nature Conservation (#SI.33/Set‐3/2011) and the
Bali Barat National Park (#S.308/BTNBB‐1/2011),
and using protocols developed by Universitas
Udayana Primate Research Center (UNUD‐PKP)
and The University of Notre Dame (USA) under
University of Notre Dame IACUC (#07‐001). This
research adhered to the American Society of Primatologists Principles for the Ethical Treatment of Non‐
Human Primates.
RESULTS
Physical Characteristics of Sleeping Trees
In total, macaques used 37 sleeping trees with a
mean DBH of 56.5 SD 38.4 cm and a mean total
height of 10.4 SD 3.3 m (Table I). In testing whether
the monkeys selected the tallest trees with the
largest trunks, we examined forest zones (82% of
the home range) and human zones (18% of the home
ranges) separately (Fig. 1). In forest habitat, the
average DBH of sleeping trees was significantly
larger than that of surrounding control trees (mST
¼ 66.5 SD 43.3 cm, n ¼ 23; mCT ¼ 47.8 SD 42.1 cm,
n ¼ 107; nested ANOVA: F13,104 ¼ 1.8, P < 0.05). In
contrast, in human zones, the average DBH of
sleeping trees was not different from that of controls
(mST ¼ 40.0 SD 21.1 cm, n ¼ 14; mCT ¼ 44.01 SD
23.5 cm, n ¼ 31; nested ANOVA: F4,37 ¼ 1.7, P ¼ 0.16).
The total height of sleeping and control trees did not
differ significantly, either in forest (mST ¼ 11.2 SD
3.1 cm; mCT ¼ 9.4 SD 3.6 cm; nested ANOVA:
F13,104 ¼ 1.0, P ¼ 0.37) or human zones (mST ¼ 9.2
SD 3.3 cm; mCT ¼ 10.3 SD 4.5 cm; nested ANOVA:
F4,37 ¼ 0.8, P ¼ 0.52). We can thus conclude that when
sleeping in forest zones, macaques selected trees with
larger trunks that were not significantly taller than
neighboring trees, while in human zones, sleeping
trees did not differ from neighboring trees in either
girth or height.
Macaques used nine species of tree for sleeping
(Table I). The most frequently used were Phyllanthus
emblica (n ¼ 11, 30%), Lannea coromandelica (n ¼ 8,
22%), and Acacia leucophloea (n ¼ 7, 19%).
Patterns of Use and Re‐use of Sleeping Sites
Over 56 nights, the macaques used 17 sleeping
sites and 37 sleeping trees (Table I). The group used
10 (59%) of the sleeping sites repeatedly (range: 2–9
times), but 4 of these sites accounted for 66% of the
total nights, and the most favored site was re‐used
nine times (Fig. 2). The other seven sleeping sites
were each used only once. Sleeping sites were re‐used
an average of 2.29 times (n ¼ 17). Based on this mean
Sleeping Sites of Balinese Macaques / 1145
TABLE I. Physical Characteristics (Mean SD) and Habitat Types of the Sleeping Trees Used by Long‐Tailed
Macaques at BBNP
SS
Habitat
type
Human
settlement
proximity
No ST
Tree species
Family
DBH (cm)
Height (m)
01
02
03
04
05
Forest
Forest
Forest
Forest
Forest
RESEARCH ARTICLE
The Role of Anthropic, Ecological, and Social Factors in Sleeping Site Choice
by Long‐Tailed Macaques (Macaca fascicularis)
FANY BROTCORNE1,2*, CINDY MASLAROV1, I. NENGAH WANDIA3, AGUSTIN FUENTES4,
ROSELINE C. BEUDELS‐JAMAR2, AND MARIE‐CLAUDE HUYNEN1
1
Primatology Research Group, Behavioural Biology Unit, University of Liège, Liège, Belgium
2
Conservation Biology Unit, Education and Nature, Royal Belgian Institute of Natural Sciences, Brussels, Belgium
3
Primate Research Center, Universitas Udayana, Bali, Indonesia
4
Department of Anthropology, University of Notre Dame, Notre Dame, Indiana
When choosing their sleeping sites, primates make adaptive trade‐offs between various biotic and
abiotic constraints. In human‐modified environments, anthropic factors may play a role. We assessed
the influence of ecological (predation), social (intergroup competition), and anthropic (proximity to
human settlements) factors in sleeping site choice by long‐tailed macaques (Macaca fascicularis)
occupying a habitat at the interface of natural forests and human‐modified zones in Bali Barat National
Park, Indonesia. Over the course of 56 nights, we collected data relating to physical features of sleeping
trees, patterns of the use of sleeping sites within the home range, pre‐sleep behavior, diurnal ranging
patterns and availability of natural and human food. Overall, the macaques used 17 sleeping sites with
37 sleeping trees. When the monkeys slept in forest zones, they selected sleeping trees that had larger
trunks but were not significantly taller than surrounding trees. Though the macaques rarely re‐used
sleeping sites on consecutive nights, they frequently re‐used four sites over the study period. The group
favored sleeping within the core area of its home range, despite the occurrence of frequent agonistic
intergroup encounters there. Macaques preferentially selected sleeping trees located within or near
human‐modified zones, especially when human food was abundant and natural food was scarce. These
results partially support the hypothesis that long‐tailed macaques choose their sleeping sites to avoid
predation; proximity to human settlements appears to be the primary factor influencing sleeping site
choice in this primate species. Our results reflect the strong influence that anthropic factors have on
primates, which subsist in increasingly human‐dominated landscapes. Am. J. Primatol. 76:1140–1150,
2014. © 2014 Wiley Periodicals, Inc.
Key words:
sleeping site; human proximity; human food; predation avoidance; Bali Barat National
Park
INTRODUCTION
The choice of sleeping sites by diurnal primates
reflects diverse factors and constraints that are
mutually non‐exclusive and highly dependent on
context and the species in question. Predation
pressure, proximity of food, competition with conspecifics and physical comfort may influence the
choice of sleeping sites and sleeping trees by primates
[Anderson, 1998, 2000].
The predation avoidance hypothesis asserts that
primates have evolved anti‐predator strategies at
sleeping sites. To minimize detection by a predator,
they should select tall emergent trees [Bernard
et al., 2011; Fan & Jiang, 2008; Reichard, 1998] and
behave in a cryptic manner (e.g. moving silently and
rapidly) when entering sleeping trees [Heymann,
1995; Liu & Zhao, 2004; Smith et al., 2007]. Some
primates may prefer open‐canopy trees located along
© 2014 Wiley Periodicals, Inc.
riverbanks [Fittinghoff & Lindburg, 1980; Matsuda
et al., 2008] and connected to adjacent trees [Albert
et al., 2011]. These conditions improve their chances
of detecting approaching predators and facilitate
Contract grant sponsor: Belgian National Fund for Scientific
Research; contract grant sponsor: Fondation Belge de la
Vocation
Correspondence to: Fany Brotcorne, Primatology Research
Group, Behavioral Biology Unit, University of Liège, 22 Quai
Van Beneden, Liège 4020, Belgium. E‐mail: fbrotcorne@gmail.
com
Received 22 November 2013; revised 9 April 2014; revision
accepted 10 April 2014
DOI: 10.1002/ajp.22299
Published online 8 May 2014 in Wiley Online Library
(wileyonlinelibrary.com).
Sleeping Sites of Balinese Macaques / 1141
escape [Hankerson et al., 2007; Kurland, 1973].
Alternatively, sleeping trees with a crown not
touching the neighboring trees may offer protection
against arboreal mammalian predators [Barnett
et al., 2012]. Some researchers argued that a dense
canopy cover improves the concealment of primates
[Anderson, 2000; Liu & Zhao, 2004]. Physical barriers
such as large tree trunks, tall crowns, elevated first
branches, and the absence of lianas serve to impede
the access of predators into trees [Barnett et al., 2012;
Bernard et al., 2010; Di Bitetti et al., 2000; Ramakrishnan & Coss, 2001; Teichroeb et al., 2012;
Tenaza & Tilson, 1985]. As predators may memorize
the sleeping refuges of their prey [Emsens et al.,
2014], primates are expected to switch between
several alternate sleeping sites, thus, making them
less predictable for predators [Phoonjampa et al.,
2010; Reichard, 1998; Zhang, 1995]. Alternatively,
primates may re‐use the same sleeping sites repeatedly because suitable sites are rare [Duarte &
Young, 2011; Ramakrishnan & Coss, 2001; Tenaza
& Tilson, 1985], or because familiarity with sleeping
sites facilitates escape from predators [Di Bitetti
et al., 2000; Hankerson et al., 2007; Li et al., 2011].
According to the food proximity hypothesis,
primates select sleeping sites that maximize their
access to food patches and at the same time minimize
travel costs. Sleeping sites should be near the feeding
sites used prior to or just after the period of rest
[Chapman et al., 1989; Heymann, 1995; Li et al.,
2011; Smith et al., 2007; Teichroeb et al., 2012].
Competition with conspecific groups may also
influence the choice of sleeping site location within the
home range. Seeping in areas of exclusive use may
decrease the risk of intergroup encounters [Albert
et al., 2011; Li et al., 2011; Phoonjampa et al., 2010;
Smith et al., 2007; Von Hippel, 1998], while sleeping
near range boundaries may allow these boundaries or
nearby resources to be defended [Teichroeb et al.,
2012], depending on the species’ degree of territoriality [cf. risk hypothesis: Wrangham et al., 2007].
Additionally, factors such as parasite avoidance,
thermoregulation and the search for comfort, stability, and social contact in trees may play a role in the
choice of sleeping sites [Anderson, 1998; Di Bitetti
et al., 2000; Fan & Jiang, 2008; Liu & Zhao, 2004;
Zhang, 1995].
Primate habitat is becoming increasingly modified
by humans and anthropogenic change is having large
impacts on primate behavior and ecology [e.g. Fuentes
& Hockings, 2010; McKinney, 2009; Strum, 2010].
Predation pressure may be lessened near human
settlements as here predators have been persecuted by
humans and their populations reduced [Bishop
et al., 1981; Isbell & Young, 1993; Ramakrishnan &
Coss, 2001; Stanford, 2002]. Studying the choice of
sleeping sites in environments at the interface of forest
and human‐influenced habitats contributes to a better
understanding of how primates cope behaviorally with
human encroachment. In terms of predation risk, only
a few studies have analyzed sleeping site choice in
human‐dominated landscapes [Duarte & Young, 2011;
Ramakrishnan & Coss, 2001]. In this study, we
investigated the role of ecological, social and anthropic
factors in the sleeping site choice of a group of long‐
tailed macaques (Macaca fascicularis) living in a partly
human‐modified habitat within the Bali Barat National
Park (BBNP), Indonesia.
The geographical range of M. fascicularis extends
across southeast Asia from Bangladesh to the Sunda
Archipelago of Indonesia [Fooden, 1995]. Although in
the past the species occurred mainly in riverine
forests [Fittinghoff & Lindburg, 1980; Kurland,
1973], in more recent times its behavioral and
ecological flexibility (including a generalist, opportunistic feeding strategy) has enabled M. fascicularis to
colonize a wide variety of habitats [Fooden, 1995;
Gumert et al., 2011]. In particular, the species has
succeeded in exploiting fragmented forests and the
edges of disturbed habitats [Gumert et al., 2011].
Although the species has long been associated
with humans throughout its range [Fuentes &
Hockings, 2010], recently it has increasingly entered
into commensal relationships with humans, living in
close association with people, taking advantage of
human food to supplement its diet [Wheatley, 1999],
and even competing with humans for spatial and
dietary resources [Richard et al., 1989]. The island of
Bali has a particularly intense and widespread
human‐primate interface; long‐tailed macaques and
humans have coexisted and interacted for centuries
[Fuentes et al., 2005; Wheatley, 1999]. Fuentes
[2010] suggested that the long‐term sympatry between humans and M. fascicularis in Bali could have
resulted in intertwined ecologies, with humans
exerting selective pressures on the macaques.
The monkeys’ sleeping sites are one potential
point of influence. Among the factors potentially
influencing sleeping site choice, we focused on
predation and intergroup competition avoidance, as
well as proximity to human settlements and human
food. We investigated the following predictions: (i) If
avoidance of predators guides sleeping site choice,
macaques should use multiple sleeping sites and
alternate between them, and they should adopt
cryptic pre‐sleep behavior to decrease the risk of
detection by predators, (ii) Macaques could be safer
when sleeping in human‐modified zones (cf. definition in Methods) where natural predation pressure
was likely reduced. In this case, we expected a
habitat‐specific choice of sleeping trees, with macaques selecting taller trees with larger trunks only
when sleeping in forest zones, (iii) following the risk
hypothesis [Wrangham et al., 2007], we predicted
that the macaques should avoid sleeping in areas of
their home range where they frequently encountered
other groups, (iv) due to the long history of human‐
macaque interface on Bali, we predicted that
Am. J. Primatol.
1142 / Brotcorne et al.
anthropic factors should influence the habitat use
and sleeping site choice of the macaques. Particularly, we expected the macaques to use the human‐
modified zones and their surroundings because of the
food left by people. The preference for certain sleeping
site locations should also vary with food availability
within the home range. During the months of high
tourist activity, when the availability of human food
was substantial, we expected the macaques to sleep
more frequently near human‐modified zones.
METHODS
Study Site and Subjects
Bali Barat National Park (BBNP) is located in
the north‐western part of the island of Bali, Indonesia
(8°05’S–18°15’S, 114°25’E–114°56’E). The park covers an area of 19,002 ha comprising dry deciduous
monsoon forest, lowland rain forest, coastal forest,
and mangrove. The climate is monsoonal, with a dry
season lasting from May to September and a wet
season lasting from October to April; the annual
average rainfall is 1,198 mm. During the study
period, sunset time ranged between 18:08 and
18:33, and sunrise was between 06:23 and 06:34
(http://www.timeanddate.com).
We conducted this study in the western sector of
the park, an area of deciduous monsoon and coastal
forests located around the park headquarters,
which included a ranger station, tourist facilities,
a camping area, and a Hindu temple. Two roadways
crossed the study site (Fig. 1). Tourists visit the
park throughout the year, with peak numbers from
May to September (in 2013, monthly mean during
high tourist season ¼ 7,715 visitors vs. low tourist
season ¼ 2,423 visitors). Human‐derived food (hereafter referred to as human food), occasionally present
in human‐modified zones and alongside roadways,
mainly consisted of offerings (i.e. fruit or crackers)
placed in temples, or refuse and leftovers (i.e. rice and
fruits) in bins or scattered across the ground.
There were two non‐human primate species at
BBNP, the long‐tailed macaque (M. fascicularis) and
the ebony langur (Trachypithecus auratus) [Leca
et al., 2013; Wheatley et al., 1993]. Although we did
not quantify predator density in the study area, we
observed several potential predators of M. fascicularis: the reticulated python (Python reticulatus), the
water monitor lizard (Varanus salvator), diurnal and
nocturnal raptors such as changeable hawk‐eagle
(Nisaetus cirrhatus), Indian black eagle (Ictinaetus
malayensis), barred eagle‐owl (Bubo sumatranus), as
well as domestic dogs [Fam & Nijman, 2011;
Fooden, 1995; van Schaik & Mitrasetia, 1990]. Tigers
are known predators of M. fascicularis at other sites
[van Schaik et al., 1983], but the Bali tiger (Panthera
tigris balica) has been extinct since around 1940
[Whitten et al., 1996]. The leopard cat (Prionailurus
bengalensis), which has been reported as a potential
predator of macaques [Palombit, 1992], is the sole
wild felid that still occurred in the park (H.
Kesumanegara, personal communication). Among
the aforementioned predators, only the python, owl,
and leopard cat represented a nocturnal threat to
macaques.
We studied a group of long‐tailed macaques
with 24–26 members including 3 adult males, 7
adult females, 3–4 subadult males, 6 juveniles, and
Fig. 1. Study site in Bali Barat National Park (Indonesia) with an outline of the study group home range, showing two habitat types: forest
(natural zones with continuous tree canopy, 40 ha) and human zones (ranger station and grass patches with planted trees, 9 ha).
Am. J. Primatol.
Sleeping Sites of Balinese Macaques / 1143
4–6 infants. During the study, two births occurred
and one subadult male disappeared.
Data Collection
We carried out the study over a period of four full
months, from March to June 2011. We allocated the
first month to habituating the macaques to observers,
so the analyzed data set covers a 3‐month period. We
followed the macaques 5 days per week (mean:
15 days per month; range: 14–17 days) from sleeping
site (06:00) to sleeping site (18:30) when possible,
resulting in 530 hr of observation. We ended our
observations when visibility became too low for us to
see the monkeys. We devoted some additional
afternoons to locating sleeping site locations on
days when we had not followed the study group
(n ¼ 13 days, with 11 successful attempts). In
total, we recorded the location of sleeping sites for
56 nights.
We use the term sleeping tree to refer to trees in
which macaques stayed overnight; sleeping site to
refer to the circular plot delimited by a 20 m radius
around the edges of the crowns of the sleeping trees
used by the group (one sleeping site could be
composed of one or more sleeping trees occupied
simultaneously); control tree to refer to trees present
within the sleeping site but never used for sleeping;
re‐used sleeping site to refer to sites used more than
once; and favored sleeping site to refer to sites used
more than five times by the group over the course of
the study [Reichard, 1998].
For each sleeping site, we recorded the GPS
location of the plot center (using a handheld GPS
Garmin 60CSx; 10 m error), as well as its distance to
the nearest roadway and to any human settlement
(using a 3 point‐scale: 0–20 m; 21–50 m; >50 m). We
also distinguished between two habitat types: the
“forest zones,” consisting of natural forested zones
with continuous tree canopy, and the “human‐
modified zones” (hereafter also referred as to “human
zones”), consisting of roadways, the ranger station
with its associated buildings, and grass patches
with planted trees (Fig. 1) [Sha & Hanya, 2013].
For each sleeping tree (n ¼ 37), we recorded the
species (with the collaboration of a local botanist, H.
Kesumanegara), diameter at breast height (DBH)
and total tree height (visual estimate always made by
the same observer). We took the same measures for
control trees (n ¼ 138), which were nearby trees (DBH
>10 cm) selected to represent the four quadrants of a
sleeping site equally. Additionally, we noted the time
the macaques entered and left the sleeping site, the
GPS location and nature of encounters with other
primate groups during the day and at sleeping sites,
and the occurrence of vocalizations such as loud
collective contact calls and those associated with
intra‐group aggression during retirement to sleeping
sites on an ad libitum basis [Altmann, 1974].
During our daily follows of the macaques, we
recorded the group’s GPS location at 20‐min intervals
(n ¼ 1,230) to reconstruct its daily ranging patterns.
When possible, we took the GPS position from the
group’s center (rough estimate of the typical group
spread ¼ 30 m). When this was not possible, we
tolerated a maximum average distance of 10 m
from the group’s periphery. In addition, we used
30 sec scans at 2.5 min intervals (n ¼ 5,037)
[Altmann, 1974] to record the habitat type at the
group’s location (i.e. forests vs. human zones), the
distance between the group’s center and the nearest
human settlement or roadway (using a 3 point‐scale:
0–20 m, 21–50 m, >50 m), and the presence of human
food in bins or elsewhere within ca. 30 m of the
center of the group (this distance was a rough
estimate always made by the same observer, corresponding to the maximum beyond which visibility
was restricted).
To investigate whether preferences for sleeping
site locations related to changes in natural fruit
availability within the home range, we conducted a
tree phenological survey. Across the center of the
study area, we established a 1,200 m 20 m transect
running north‐south, along which we surveyed twice
a month the phenological status of all trees with DBH
10 cm (0: absence of fruit, 1: presence of ripe fruit, 2:
presence of sun‐dried fruit) [Newton, 2007]. This
transect crossed all the habitat types present in the
group’s home range.
Data Analysis
To assess whether sleeping trees had larger
DBHs and were taller than nearby control trees
within the sleeping site, we used nested ANOVAs
(tree‐type nested within sleeping site). We tested for
these differences in both habitat types, forest versus
human zones. Since the residuals of the models did
not conform to a Shapiro–Wilk normality test, the
DBH was log‐transformed and height was square‐
root‐transformed [Sokal & Rohlf, 1995]. We used
Levene’s test to confirm non‐deviance for homoscedasticity [Dytham, 2003]. To test for randomness of
sleeping site re‐use, we generated an expected
frequency distribution based on Poisson lambda
parameters [Sokal & Rohlf, 1995] that we compared
with the observed frequency distribution of the
sleeping site re‐use using a Kolmogorov–Smirnoff
test for goodness‐of‐fit [Day & Elwood, 1999; Phoonjampa et al., 2010]. The theoretical Poisson distribution was appropriate for the counts of re‐use of
sleeping sites, which are rare events [Agresti, 2007].
Using ArcGIS 9.3.1 (ESRI) and the extension
Hawth’s Tool, we calculated the size of the home
range, both on a monthly and daily basis, with
the fixed Kernel method (95% kernels ¼ total home
range; 50% kernels ¼ core area). We assessed the
smoothing parameters href [Worton, 1989] with R
Am. J. Primatol.
1144 / Brotcorne et al.
3.0.2 and the package adehabitatHR. We compared
the daily home range sizes between months with a
Kruskall–Wallis test. To compare the locations of
sleeping sites and intergroup encounters within the
home range, we used Binomial tests with expected
values derived from the observed frequency of
macaques inside the core and peripheral areas,
which are, by definition, each used 50% of the time.
Using methods defined in previous studies
[Imaki et al., 2006; Sha et al., 2009], we established
a human proximity gradient within the home range
characterized by several buffer‐zones delimited by
their distance to any roads and/or human settlements, with the 3‐level scale used on the field: (i) 0–
20 m (including the human zones and their edges), (ii)
21–50 m (intermediary forest edge), and (iii) more
than 50 m (forest interior). We tested for random
buffer‐zone use by the macaques during their diurnal
activities and for sleeping using G‐tests for goodness‐
of‐fit with Williams’s correction [Sokal & Rohlf, 1995]
(with expected frequencies based on the surface
area of the buffers within the home range) and the
standardized residuals considered as significant
when the absolute value exceeded 2 [Agresti,
2007].
To analyze the variation in natural food availability, we used a fruit availability index (FAI) for
each tree species k and for each month m [Albert
et al., 2013]:
Month
FAIkm ¼ Dk Bk Pkm
where in Dk is the density of a given species k (number
of specimens on the phenology transect), Bk is the
mean basal area of species k, and Pk is the fruit
phenology score of species k in a given month m
(calculated as the percentage of specimens holding
ripe fruit on their crown). We used a Friedman test to
compare the FAI between months.
To analyze the effect of the human food (HF)
availability on the habitat type chosen for sleeping,
we assigned each observation day to a category of HF
availability (no/some HF vs. high HF), based on the
overall median of the daily proportion of scans when
HF was present around the macaque group (30 m
radius from the center of the group). We examined
variation in HF available among months using a
Kruskall–Wallis test. We used a G‐test of independence to evaluate whether selection of habitat type
for sleeping depended on month. Finally, we analyzed
the relationship between daily home range size and
daily HF availability with a Spearman correlation
test.
We performed the nested ANOVAs using R 3.0.2.
Other statistical tests were completed using STATISTICA 10.0 (two‐tailed statistics with a significance level set at 0.05). Data are presented as mean
values with standard deviations.
Am. J. Primatol.
Our study was approved by University of Liège,
Belgium, the Indonesian Ministry of Research and
Technology (#03B/TKPIPA/FRP/SM/III/2011), the
Indonesian Directorate of Forest Protection and
Nature Conservation (#SI.33/Set‐3/2011) and the
Bali Barat National Park (#S.308/BTNBB‐1/2011),
and using protocols developed by Universitas
Udayana Primate Research Center (UNUD‐PKP)
and The University of Notre Dame (USA) under
University of Notre Dame IACUC (#07‐001). This
research adhered to the American Society of Primatologists Principles for the Ethical Treatment of Non‐
Human Primates.
RESULTS
Physical Characteristics of Sleeping Trees
In total, macaques used 37 sleeping trees with a
mean DBH of 56.5 SD 38.4 cm and a mean total
height of 10.4 SD 3.3 m (Table I). In testing whether
the monkeys selected the tallest trees with the
largest trunks, we examined forest zones (82% of
the home range) and human zones (18% of the home
ranges) separately (Fig. 1). In forest habitat, the
average DBH of sleeping trees was significantly
larger than that of surrounding control trees (mST
¼ 66.5 SD 43.3 cm, n ¼ 23; mCT ¼ 47.8 SD 42.1 cm,
n ¼ 107; nested ANOVA: F13,104 ¼ 1.8, P < 0.05). In
contrast, in human zones, the average DBH of
sleeping trees was not different from that of controls
(mST ¼ 40.0 SD 21.1 cm, n ¼ 14; mCT ¼ 44.01 SD
23.5 cm, n ¼ 31; nested ANOVA: F4,37 ¼ 1.7, P ¼ 0.16).
The total height of sleeping and control trees did not
differ significantly, either in forest (mST ¼ 11.2 SD
3.1 cm; mCT ¼ 9.4 SD 3.6 cm; nested ANOVA:
F13,104 ¼ 1.0, P ¼ 0.37) or human zones (mST ¼ 9.2
SD 3.3 cm; mCT ¼ 10.3 SD 4.5 cm; nested ANOVA:
F4,37 ¼ 0.8, P ¼ 0.52). We can thus conclude that when
sleeping in forest zones, macaques selected trees with
larger trunks that were not significantly taller than
neighboring trees, while in human zones, sleeping
trees did not differ from neighboring trees in either
girth or height.
Macaques used nine species of tree for sleeping
(Table I). The most frequently used were Phyllanthus
emblica (n ¼ 11, 30%), Lannea coromandelica (n ¼ 8,
22%), and Acacia leucophloea (n ¼ 7, 19%).
Patterns of Use and Re‐use of Sleeping Sites
Over 56 nights, the macaques used 17 sleeping
sites and 37 sleeping trees (Table I). The group used
10 (59%) of the sleeping sites repeatedly (range: 2–9
times), but 4 of these sites accounted for 66% of the
total nights, and the most favored site was re‐used
nine times (Fig. 2). The other seven sleeping sites
were each used only once. Sleeping sites were re‐used
an average of 2.29 times (n ¼ 17). Based on this mean
Sleeping Sites of Balinese Macaques / 1145
TABLE I. Physical Characteristics (Mean SD) and Habitat Types of the Sleeping Trees Used by Long‐Tailed
Macaques at BBNP
SS
Habitat
type
Human
settlement
proximity
No ST
Tree species
Family
DBH (cm)
Height (m)
01
02
03
04
05
Forest
Forest
Forest
Forest
Forest