Journal of Tropical Ecology (2004) 20:247–258. Copyright © 2004 Cambridge University Press
DOI: 10.1017/S0266467403001275 Printed in the United Kingdom
Diurnal home range and roosting trees of a maternity colony of Pteropus
vampyrus natunae (Chiroptera: Pteropodidae) in Sedilu, Sarawak
Melvin Terry Gumal
National Parks and Wildlife Division, Sarawak Forest Department, Malaysia and Department of Anatomy, University of Cambridge, UK
(Accepted 11 April 2003)
Abstract: Field data on the roosting ecology of a maternity colony of Pteropus vampyrus natunae in a peat swamp forest in
Sarawak were collected for 19 mo. The colony moved between 10 roost sites and the diurnal home range (comprising
only day-roost sites) was about 1120 ha. Two movement phases were observed within the diurnal home range. The
ﬁrst phase consisted of continuous movement between roost sites each month (August 1998 to January 1999, and
August to November 1999), and this coincided with the months in which late pregnancies and pups dependent on
mothers were seen. The second phase had almost zero or restricted movement each month (February to July 1999,
and December 1999 to March 2000). Throughout the diurnal home range, the colony roosted in 20 families of trees
(27 genera and 31 species). The bats tended to roost in certain species and sizes of trees.
Key Words: ﬂying foxes, movements, peat swamp roost site, roost trees
Bats spend most of their lives in their roosts (Altringham
1996). By choosing a suitable roost, bats gain important
beneﬁts, such as better mating opportunities; improved
maternal care; increased social interactions and infor-
mation transfer; cheaper thermoregulation; reduced
commuting costs to foraging sites; and protection from
adverse weather and predators (Altringham 1996, Kunz
1982). In general, Pteropus species tend to roost in large
aggregations on exposed branches (Pierson & Rainey
The main threats facing colonial, tree-roosting Pteropus
species worldwide, are habitat loss to agriculture (Kunz &
Pierson 1994, Mickleburgh et al. 1992, Racey 1979)
and hunting at roosts (Kunz & Pierson 1994, Wiles
1992). Losses of critical roosting sites have been reported
for many species, among them, P. rodricensis in the
Mascarenes, P. livingstonii in the Comoros (Mickleburgh
et al. 1992, Pierson & Rainey 1992), and P. vampyrus in
Peninsular Malaysia (Davison 1992, Mohd-Azlan et al.
2001). With the loss of roosting sites, the population
may either decrease, especially for the most vulnerable
Correspondence address: 7, Jalan Ridgeway, Kuching 93200, Sarawak,
Malaysia. Email: melter firstname.lastname@example.org
species with highly specialized roosting requirements
(Kunz & Pierson 1994) or move to a different site
(Richards & Hall 1998). Hunting within the roosts
reduces the size of the colony and has led to the
abandonment of traditional maternity roost sites, as seen
in Pteropus marianus marianus (Wiles 1992).
In Sarawak, Pteropus vampyrus natunae was once con-
sidered widespread, forming large colonies throughout
the coastal lowland areas (Banks 1949, Medway 1977,
Payne et al. 1985). The population is currently con-
sidered threatened, and the species was upgraded to
a protected species in 1995 (Gumal 2001, Wildlife
Conservation Society & Sarawak Forest Department
1996), although the IUCN worldwide priority grade
for species action, lists it as not threatened in its
worldwide range (Mickleburgh et al. 1992). Currently,
empirical evidence points to over-hunting for the wildlife
trade (wild meat and Chinese traditional medicine)
and to sport hunting as the main causes for the
rapid decline of Pteropus vampyrus natunae in Sarawak
(Gumal et al. 1998). Over the next decade, however, the
proposed habitat loss from lowland peat swamp forests to
agricultural and forest plantations, will increasingly be a
concern, as all the existing maternity colonies are found in
these areas (Gumal 2001). Although roosting behaviour
and post-natal growth for captive Pteropus vampyrus
natunae have been partially documented (McCracken &
248 MELVIN TERRY GUMAL
Wilkinson 2000), their ﬁeld roosting behaviour has
largely been unreported.
The objective of this study was to investigate the
following questions: (1) Is there a pattern in the maternity
colony’s movements between the various roost sites? (2) Is
there a preference for certain sizes and species of roosting
trees, and does the general availability of suitable trees
inﬂuence the location of a maternity roost? Sedilu was
chosen as the study site because it was not threatened
with logging activities or large-scale land conversions to
oil palm or forest-tree plantations. Besides this, it was also
assumed that there would be control over illegal logging
and poaching of bats in the roost, as there were only
two main access routes to the area, both of which were
patrolled by staff from the Sarawak Forest Department.
The seasonal movements of the colony between roost
sites, a diurnal roosting home range for the colony,
sizes and species of roosting trees during the seasonal
movements, and the relative preference for sizes and
species of roost trees in relation to the availability of trees
in the forest are all described in this paper.
STUDY SITE AND SPECIES
The study was conducted in Sedilu Forest Reserve,
Sarawak, Malaysia (1◦24to 1◦26N and 110◦48to
110◦51E), from September 1998 to March 2000. The
reserve is formed on ﬂat alluvial plains and drained by the
Belanga, Sebangan, Tebelu and Sebuyau rivers, and is less
than 15 km from the coast. Although some of the lower
reaches of these rivers have been cleared and cultivated
with crops, such as coconuts, oil palm and rice, peat-
swamp forests still predominate as they make up 85.9%
of the land area within a 5-km radius of the maternity
roost sites (Sarawak Forest Department unpubl. data).
Pteropus vampyrus is the largest bat in the world, with a
recorded wingspan of about 1700 mm (Pierson & Rainey
1992). Adult forearms range between 185–200 mm
(Gumal 2001, Payne et al. 1985) and weights of up to
1100 g have been recorded (Payne et al. 1985). The
species can be found from southern Myanmar to Java in
Indonesia (Mickleburgh et al. 1992). Pteropus vampyrus
natunae is the subspecies found in Borneo, ranging up
to Natuna Islands, Balembangan and Banggi Islands
(Mickleburgh et al. 1992, Payne et al. 1985). In the
1970s, spectacular evening ﬂights of ﬂying foxes were
seen around major towns during the fruiting season, and
these animals were found throughout the lowland areas
(Payne et al. 1985). During the state-wide survey of roosts
between 1997 and 2000, only ﬁve maternity colonies
were found, i.e. at Patok Island, Loagan Bunut, Sarang,
Limbang and Sedilu (Gumal 2001). All the colonies
were located at remote and extremely inaccessible areas,
either in peat swamp or mangrove forests. Among all the
maximum of 15 000, estimated through dispersal counts
The following terms are used accordingly: roost trees refer
to the trees used by the bats for sleeping or resting during
the day; roost sites are areas of roost trees occupied by the
maternity colony, and diurnal home ranges encompass
all of diurnal roost sites for the maternity colony.
Capturing and radio-collaring of bats
Animals for radio-collaring were captured with single-ply
mist nets: 110 Denier, 5-mm meshes, four shelves, 4 m
in height and 20 m in length (Gajah Ltd). Bamboo poles
(20 m) were used as stakes. All the poles were trimmed to
20 m, as longer poles became too thin and often snapped
when the nets were stretched. Pulleys, tied to the top end
of the poles, were used to lift the nets into place. Two nets
were opened each night, from 18h30 to 05h30. Both nets
were checked every 3 h, and tended by two people each
Twenty animals were radio-collared, of which ten were
ﬁtted with large transmitters, and the other ten with
smaller transmitters (both transmitter types coded as
BE10-18 by Sirtrack Ltd). Stratiﬁed random-sampling
design was used as the ﬁrst ten animals that weighed more
than 550 g (hereafter known as adults) were ﬁtted with
the large transmitters. Eight males and two females were
ﬁtted with these large transmitters. The same sampling
design was used for ﬁtting the smaller transmitters on
animals that weighed between 330–550 g (hereafter
known as juveniles). Six males and four females were
ﬁtted with the small transmitters.
The large transmitters had a maximum battery
operational life of at least 10 mo. They were epoxy-
mounted on leather collars and the whole unit ﬁtted
around the bat’s neck with a plastic screw and nut.
These transmitter units (transmitter and collar) weighed
approximately 16 g, and were only ﬁtted on animals
heavier than 550 g to ensure that they were less than
3.0% of the animal’s weight, and thus, below the 5%
‘rule’ for transmitter mass for bats (Aldridge & Brigham
1988). The collars were designed to break after 1 y of use
(gross approximations by Sirtrack Ltd).
The transmitter units for juveniles were much smaller
and lighter, and weighed about 10 g each (about 3%
of the animal’s weight). Instead of leather, the smaller
collars made it possible to slip the transmitter unit
easily around their necks. The rubber collars came in
Diurnal home range and roosting trees of a maternity colony of Pteropus vampyrus natunae 249
Figure 1. Locations of trails, transects and plots at Sedilu, Sarawak. Dashed lines represent the 17-km survey trail. The trail was used for locating
roost sites. SF2 and SF7 are transects, and are highlighted as solid bars. Within the transects, all trees greater than 10 cm dbh were identiﬁed to
species level and measured for their dbh and height. R2 and R7 are plots, and shown as rectangles. All roost trees within the plots were identiﬁed to
species level and measured for their dbh and height. All bearings are approximate on the map.
two different sizes and were designed to break within
3 mo. These smaller transmitters had a maximum battery
operational life of 6 mo.
Bats were tracked using two portable receivers
(Telonics, Model TR4), each connected to a foldable,
hand-held three-element Yagi aerial (Sirtrack Ltd). All the
transmitters were conﬁgured to: low-power transmitters
with a pulse rate of 35 ppm; monopole whip antenna
pick-up distance of at least 5 km; duty cycle of 12 h
on, 36 h off; government-approved frequencies between
Locating roosts – walking surveys
A 17-km survey trail was cleared in the forest reserve,
and between August 1998 and January 1999 all the
maternity roost sites were found by walking the full length
of the trail (Figure 1), and listening for the calls of the
250 MELVIN TERRY GUMAL
animals. As calls were only audible from a range of about
50–75 m, roosts further away were not found. There was
no radio-tracking done during this period, as bats had yet
to be ﬁtted with radio-collars.
Tracking of radio-collared animals was carried out
after February 1999. Radio-tracking was only used to
determine the approximate location of a roost site. Within
50–75 m of the colony, the bats were audible to the
naked ear and the search thereafter, involved following
their calls. There were no attempts to ﬁnd the exact
locations of the radio-collared individuals or to map
the sizes of the roost as roosting animals often ﬂew
off at the sight of humans. Individual radio-collared
animals were either recorded as absent or present, but
searches would be made of missing or dropped radio-
collars, and this was done at the periphery of the roost
or in previous roosting sites. Once the colony was seen,
the exact location or coordinate was recorded using a
GPS (Garmin 12 XL). Radio-tracking of the bats was
carried out for 4 consecutive days per month (about the
same time each month), and the locations were recorded
for 5 min every 30 min (between 07h30–18h00), to
increase the accuracy of their estimated maternity colony
Locating roosts – helicopter surveys
Apart from determining the absence or presence of radio-
collared bats and looking for radio-collared, migrating
animals, Bell 206 Jet Ranger (four-seater) helicopters
were used to observe the approximate size (ha) of the
roost, and to ﬁnd out if the colony had maintained itself
as a single unit or as various discrete smaller units.
The roost location was recorded using the helicopter’s
GPS (Garmin 95). The helicopter circled the maternity
roosts, at 400 m asl, with a ﬂight speed of 60 km h−1.At
this height, the bats did not take ﬂight. Two aerials and
receivers were used during each helicopter survey. Both
aerials were hand-held inside the cabins, the ﬁrst in the
direction of the ﬂight path, the second at right angles to
the ﬂight path. Helicopter radio-tracking at Sedilu was
carried out twice a month between June to October 1999,
but only once in November 1999. After locating the radio-
collared animals, the helicopter circled the roost up to a
radius of 5 km in visual searches to determine if the colony
had formed smaller discrete units.
Constructing diurnal home ranges for the maternity colony
A home range often deﬁnes the ‘area used by an animal in
its normal activities of food gathering, mating and caring
for the young’ (White & Garrott 1990). In this paper
however, only part of the home range is discussed and
deﬁned, i.e. the diurnal roosting areas for the maternity
colony, hence the use of the term ‘diurnal home range’.
The roost positions were entered into ArcView
3.2 (Environmental Systems Research Institute). Here,
the USGS (United States Geological Service) Animal
Movement Analysis extension programme (Hooge &
Eichenlaub 1997), was used to determine their diurnal
home ranges. All the coordinates were projected in
Universal Transverse Mercator (UTM), as the area
distortion was minimal within the zones (White & Garrott
1990). Distance calculations were in metric. Minimum
Area Polygons (MAPs) are used in this paper, as they
are the commonest method used to estimate home range,
allow for easy comparison with other home range studies
(Law & Anderson 2000, Robinson 1992), are simple
to use, ﬂexible in shape and easy to calculate (White &
Garrott 1990). The diurnal home range was constructed
by connecting the outer locations to form a convex
polygon, and then calculating the area of this polygon.
The disadvantage is that MAPs can be inﬂuenced by a
few outlying visits that can result in the range including
areas not visited (MacDonald et al. 1980). MAPs also do
not indicate the intensity of spatial use within the home
range and the technique is susceptible to bias, especially
with small sample sizes (Jennrich & Turner 1969). Large
sample sizes, taken at constant time intervals, offset this
bias (Solla et al. 1999). To determine the intensity of
spatial use (number of survey days spent by the colony
in an area), a grid was laid at Sedilu, and each quadrat
was 500 ×500 m commencing from the UTM coordinate
of 478,000; 160,000.
Sizes and species of roosting trees
Tree species, diameter at breast height (dbh) and heights
of the roosting trees were recorded in two plots (R2 and
R7), of 2 ha each (500 m ×40 m) (Figure 1). The plots
were established to examine the relationship of roost-tree
characteristics close to and away from the conﬂuence of
Belanga and Sebangan rivers and also to compare the
characteristics of roosting trees with all trees ≥10 cm
structure of the forest were examined in seven random
transects (500 m ×5 m), where all trees ≥10 cm dbh
were measured (dbh and height), and were identiﬁed to
the species level. The acronyms SF1 to SF7 were used
to identify these transects (Gumal 2001). The core of
the sampling plots R2 and R7, were transects SF2 and
SF7 respectively. The ﬂoristic structure and composition
of the peat swamp forest where the bats roosted were
homogeneous (Gumal 2001), and thus, it was assumed
that the two plots, R2 and R7, represented the sizes
and species of trees used by the bats for roosting in this
Diurnal home range and roosting trees of a maternity colony of Pteropus vampyrus natunae 251
Tree species and heights of roosting trees outside the
plots were also recorded. When the colony was found,
three trees were randomly selected. Selection was based
on a stratiﬁed random design, where the ﬁrst three trees
with animals seen were chosen, regardless of the height or
numbers of animals in each tree. Trees next to each other
were avoided. During the course of the day’s observations,
if all of the animals departed from one of the trees, another
tree would be randomly chosen. All tree species were
tagged, with samples removed to the Sarawak Forest
Department’s Herbarium for identiﬁcation by their staff.
Tree heights were estimated using a digital range ﬁnder.
Chi-square analyses were carried out on the
proportional representation between frequencies of roost-
tree species in the plots and the frequencies of the
same species in the transects (R2 and SF2, and R7 and
SF7). The analyses were based on the species common
to both the transects and plots, only included roost
trees ≥10 cm dbh, and excluded species with expected
frequencies of less than ﬁve.
Locations and movements of roost sites
At Sedilu, there was only one colony of Pteropus vampyrus
natunae observed during the 6 mo of helicopter radio-
tracking. During this period, the colony moved between
roost sites as a single unit. The mean circular area of
a roost site from the helicopter surveys was 100.8 ha
(SD =43.4 ha, n =11 surveys).
The colony used 10 roost sites during the study
period, but all of them were within the mixed swamp
forest (Figure 2a). The diurnal home range, which
encompassed all the various roosting sites, was 1120 ha.
Most of the roost sites were located along the Belanga
River. The site with the highest intensity of use, in terms of
the numbers of days roosted throughout the study period,
was in UTM quadrat 155,001–155,5000; 470,001–
478,500 (Figure 2a). This quadrat was over 3 km from
the conﬂuence of the Belanga and Sebangan rivers. The
colony roosted in this location for 31 (55.4%) out of the
total of 56 survey d (February 1999–March 2000). After
this quadrat, the intensity of use at the other locations
dropped sharply, as the next highest was 7 d (12.5%) at
UTM coordinate 155,001–156,000; 478,001–478,500.
This quadrat was adjacent to the previous, and was
2.5 km away from the conﬂuence of the Belanga and
The movement between roosts was not uniform each
month, but followed a pattern involving two distinct
phases, each lasting several months. Between September
1998 and January 1999, the colony frequently moved
between roost sites, sometimes on a daily basis. There was
minimal movement between February and July 1999,
followed by frequent movements between August and
November 1999, and then back to minimal movement
between December 1999 and March 2000 (Figure 2b).
Hereafter in this paper, the frequent monthly movement
pattern will be referred to as the continuous-movement
phase whereas the minimal movement pattern will be
known as restricted-movement phase. Analyses of the
roost movements are based on the movement phases as
Continuous-movement phase, September 1998 to January 1999.
Although roost sites were only located within 50–75 m of
the 17-km survey trail, it appeared that the colony moved
daily, and was conﬁned to either side of Belanga River
(bearing of 195◦) (Figure 2b). The northernmost roost
site was situated at the conﬂuence of the Belanga and
Sebangan rivers. The diurnal home range for the months
of November 1998 and January 1999 were 14 ha and
13 ha respectively (Figure 3).
In October 1998, the roost sites could not be found,
and in September and December, the roost sites were only
found in 2 of the 4 d of surveys. Due to the incomplete data
for September, October and December 1998, a diurnal
home range for the maternity colony for this phase could
not be quantiﬁed.
Restricted-movement phase, February to July 1999. During this
phase, all roost sites were conﬁned to a narrow area, with
the northernmost point being about 3.1 km south-east of
the Belanga and Sebangan River conﬂuence. The diurnal
home range during these 6 mo was 11.6 ha (Figure 2b).
The colony stayed at the same site in four of these months,
i.e. March, April, May and July.
Continuous-movement phase, August to November 1999. The
colony shifted north, in a pattern similar to the movement
observed from September 1998 to January 1999. All of
the roost sites were on Belanga River except for a period
of 15 d (out of the total study period of 609 d, or one out of
56 radio-tracking survey d), where the roost site shifted
5.04 km eastward of the Belanga and Sebangan River
conﬂuence (Figure 2b). The diurnal home range during
this period was 891.8 ha (mean =97.5, SD =103.8 ha,
n=4 mo) (Figure 3). This continuous-movement phase
ceased in November 1999, as the restricted-movement
phase commenced in December 1999.
Restricted-movement phase, December 1999 to March 2000.
During these 4 mo, the roost movement reverted to the
pattern observed between February and July 1999, i.e.
conﬁned to a small area away from Sebangan River with
small movements each month. The diurnal home range
252 MELVIN TERRY GUMAL
Figure 2. (a) Grid at Sedilu, Sarawak, indicating the intensity of roost use by a maternity colony of Pteropus vampyrus natunae, from February
1999 to March 2000. Each grid is 500 m ×500 m. Only the grids used by the bats as roosts are highlighted. The numbers of days in which the
animals roosted within the grid are shown. (b) Diurnal home range of a single Pteropus vampyrus natunae maternity colony at Sedilu, Sarawak, from
September 1998 to March 2000. The Minimum Area Polygon of each movement phase can be distinguished by its separate shading pattern.
Diurnal home range and roosting trees of a maternity colony of Pteropus vampyrus natunae 253
Figure 3. Roost movements of Pteropus vampyrus natunae, showing the
total distance travelled per month and the diurnal home ranges per
month from November 1998 to March 2000. The diurnal home ranges
are shown as MAP (ha). MAPs and distances travelled were based on
the exact location of roost sites. Data for September and December 1998
are incomplete, as the colony could not be found for 2 days out of 4. The
colony was not found in October 1998. There were no radio-collared
animals between November 1998 and January 1999, and as such, exact
locations of roost sites further than 50–75 m from the 17-km survey
trail could not be found. Radio-tracking was used to locate roost sites
between February 1999 and March 2000. There was zero movement in
March, April, May, July and December 1999, and January and March
for this period was 25 ha (mean =5.3 ha, SD =10.6 ha,
n=4 mo) and the northernmost site was 1.74 km
south-east of the Belanga and Sebangan River conﬂuence
(Figure 2b). There were 3 mo during this period when
there was no movement of roost sites at all (Figure 3). It
is uncertain as to how long the roost sites were conﬁned
to this area, as the study ended in March 2000.
As seen from Figure 2b, there was some spatial
overlap between the roost sites for both restricted-
and continuous-movement phase diurnal home ranges.
There was also overlap between the roost sites for the
restricted-movement diurnal home range from February
to July 1999 and December 1999 to March 2000. The
two restricted-movement phases were characterized by
small diurnal home ranges, low movement frequency
between roost sites, and were also further inland, away
from Sebangan River. The two continuous-movement
phases were characterized by a high movement frequency
between roost sites and a diurnal home range that
stretched right up to the conﬂuence of the Sebangan
and Belanga rivers. There were differences in daily
distances moved between the four phases (Figure 4);
and also between the combined continuous-movement
vs. restricted-movement phases (Mann–Whitney U test,
U=68.0, P <0.01).
Roosting trees across the movement phases
Heights. The means and ranges in heights of the roosting
trees between the phases were quite similar, and there was
Figure 4. Daily movements between roost sites, by a maternity
colony of Pteropus vampyrus natunae at Sedilu, Sarawak. There were
differences in the daily distances moved (Kruskal–Wallis one-way
ANOVA, K =21.6, df =3, P<0.01, n1=6d,n
2=18 d, n3=12 d,
n4=12 d) (n1to n4refer to the four phases in chronological order).
The mean distances moved daily in the restricted-movement phases
were 100 m (SD =193 m, n2=18 d) and 270.8 m (SD =415 m,
n4=12 d), whereas larger means were observed for the continuous-
movement phases, i.e. 471.1 m (SD =228.7 m,n1=6 d) and 1370 m
(SD =1450 m, n3=12 d).The open boxes indicate means and the bars
represent 95% conﬁdence intervals for the mean.
no statistical difference in the mean height of the roosting
trees between the movement phases, indicating that the
bats frequently used certain sizes of trees as roosts.
Species. Throughout the diurnal home range, 20 tree
families (27 genera and 31 species, n =447 trees)
(Table 1) were used as roosts. The bats roosted in 14 tree
species in the restricted-movement phases of February
to July 1999, whereas they roosted in 11 tree species
(inclusive of an unknown species) in the restricted-
movement phase of December 1999 to February 2000
(Table 2). Twenty-three tree species were used as roosts in
the continuous-movement phase of August to November
In both of the restricted-movement phases (February
to July 1999 and December 1999 to February 2000), the
most common roosting trees were bare trees (defoliated
trees) (35% and 22% respectively). Due to the absence of
leaves, identiﬁcation was a problem. As such, these trees
were classed as bare trees, to differentiate them from those
that had leaves, but which could not be identiﬁed (coded as
unknown species). After the bare trees, the next three top-
roosting trees in these same restricted-movements phases
were Eugenia havilandii (29.8% and 15.8% respectively),
Shorea scaberrima (9.3% and 21.1% respectively) and
254 MELVIN TERRY GUMAL
Table 1. Family, species and number of individuals used by Pteropus
vampyrus natunae at Sedilu, Sarawak. This includes four species
(Gonystylus bancanus, Eusideroxylon zwageri, Barringtonia reticulata and
Azadirachta excelsa) recorded during the continuous-movement phase of
September 1998 to January 1999, but not in the other phases.
Family Species Number
Annonaceae Mezzettia leptopoda
(Hk. f. & Th.) King 1
Annonaceae Xylopia coriifolia Ridl. 2
Aquifoliaceae Ilex hypoglauca Loes. 2
Bombaceae Durio carinatus Mast. 2
Celastraceae Lophopetalum multinervium Ridl. 11
Crypteronicaceae Dactylocladus stenostachya Oliv. 6
Dipterocarpaceae Dryobalanops rappa Becc. 80
Dipterocarpaceae Shorea platycarpa Heim 3
Dipterocarpaceae Shorea scaberrima Burck 54
Ebenaceae Diospyros pseudomalabarica Bakh. 7
Euphorbiaceae Baccaurea bracteata Muell.-Arg. 1
Euphorbiaceae Neoscortechinia kingii
Pax. & K. Hoffm. 2
Gonystylaceae Gonystylus bancanus (Miq.) Kurz 9
Guttiferae Calophyllum fragrans Ridl. 1
Guttiferae Garcinia beccarii Pierre 1
Icacinaceae Platea latifolia Bl. 1
Lauraceae Eusideroxylon zwageri
Teijsm. & Binn. 1
Lauraceae Litsea accedens Boerl. 3
Lecythidaceae Barringtonia reticulata Miq. 1
Leguminosae Copaifera palustris de Wit 4
Meliaceae Azadirachta excelsa (Jack) Jacobs 2
Myrtaceae Eugenia cerina Hend. 1
Myrtaceae E. havilandii Merr. 75
Myrtaceae E. palembanica Merr. 3
Myrtaceae E. zeylanica Wright 3
Rosaceae Parastemon urophyllus A.DC. 4
Rutaceae Euodia nervosa K. & V. 2
Rutaceae Tetractomia parviﬂora Ridl. 2
Sapindaceae Nephelium maingayi Hiern 1
Sapotaceae Ganua motleyana Pierre ex Dubard 54
Sapotaceae Palaquium walsuraefolium
Pierre ex Dubard 14
Bare trees 85
Ganua motleyana (8.7% and 15.8% respectively). The bats
also roosted in bare trees in the continuous-movement
phase (August to November 1999), but the most common
roosting tree during this phase was Ganua motleyana
Roost selection in relation to tree size and the ﬂoristic
composition of the forest
There were no differences in dbh and heights of roosting
trees between the sample plots (R2 and R7), and also dbh
and heights of all trees ≥10 cm dbh between transects
(SF2 and SF7). Both the variances of the dbh and heights
of roost trees in R2 and R7 were large as the bats roosted
in trees of varying sizes.
Statistical comparisons were then made between the
mean dbh and heights of the roosting trees with that
of the mean dbh and height of trees in the area, by
excluding all roosting trees <10 cm dbh (21 out of a
total of 234 trees) from the analyses. The comparisons
were made between R2 and SF2, and R7 and SF7. In
all the analyses, Pteropus vampyrus natunae selected roost
treesthat were signiﬁcantly larger in both girth and height
than the average tree size recorded in transects SF2 and
SF7 (Table 3), indicating that the bats tended to use larger
trees as roosts.
The animals also tended to use certain species of trees
as roosts. There was a signiﬁcant difference between
the observed values in R2 and the expected values in
SF2, (χ2=29.8, df =4, P <0.001), and also in R7 and
SF7 (χ2=31.2, df =4, P <0.001). Several tree species
were over-represented as roost-trees at the plots, i.e.
Dryobalanops rappa and Lophopetalum multinervium at
R2, and Ganua motleyana at both R2 and R7. Under-
representation was also seen at R2, as Eugenia havilandii
was seldom used as roosts. At R7, Copaifera palustris,
Neoscortechinia kingii and Palaquium walsuraefolium were
rarely used as roosts, even though these were the three
most common tree species there.
Historically, Pteropus vampyrus natunae was reported as
strongly colonial and returned to the same area in the
nipah (Nypa fruticans) and mangrove forests each year
(Banks 1949, Lim 1966). There were, however, no
reports of this subspecies continuously occupying one
area, or moving between various roost sites in that area
throughout the year, although at least three Pteropus
species exhibited such continuous use of a maternity
roost area, i.e. P. tonganus in American Samoa (Brooke
et al. 2000, Grant & Bannack 1999), P.poliocephalus (Eby
1991, Parry-Jones 1987, Parry-Jones & Augee 2001)
and P. conspicillatus (Richards 1990), both in Australia.
Fidelity to a diurnal home area, rather than to a
speciﬁc roost, also appears to be a common characteristic
among: foliage-roosting bats, e.g. Artibeus lituratus and
Vampyrodes caraccioli (Kunz 1982), and Cynopterus sphinx
(Storz et al. 2000); tree roosting bats, e.g. Eptesicus fucus
and Lasionycteris noctivagans (Vonhof & Barclay 1996);
and some tree-hollow-roosting bats, e.g. Vespadelus
pumilus (Law & Anderson 2000). All these species make
frequent movements between roosts, and in the case of A.
lituratus and V. caracciolo, they seldom use the same roost
for more than 2 d consecutively (Morrison 1980).
The restricted- and continuous-movement phases were
also observed in three other maternity roost sites in
Sarawak (Gumal 2001). In Sedilu, the continuous-
movement phase coincided with the months in which
Diurnal home range and roosting trees of a maternity colony of Pteropus vampyrus natunae 255
Table 2. Roost tree species selected by a maternity colony of Pteropus vampyrus natunae during three movement phases at Sedilu, Sarawak.
Restricted-movement phase Continuous-movement phase Restricted-movement phase
February to July 99 August to November 99 December 99 to February 2000
No. No. No.
Total 161 Total 87 Total 76
Species 14 Species 23 Species 10
Bare trees 56 Bare trees 12 Bare trees 17
Eugenia havilandii 48 Ganua motleyana 17 Shorea scaberrima 16
Shorea scaberrima 15 Dryobalanops rappa 13 Ganua motleyana 12
Ganua motleyana 14 Shorea scaberrima 9Eugenia havilandii 12
Palaquium walsuraefolium 11 Eugenia havilandii 7 Unknown 5
E. palembanica 3Lophopetalum multinervium 4Dryobalanops rappa 4
Copaifera palustris 2Parastemon urophyllus 4Durio carinatus 2
Dryobalanops rappa 2Copaifera palustris 2Dactylocladus stenostachys 2
Gonystylus bancanus 2Gonystylus bancanus 2P. walsuraefolium 2
Shorea platycarpa 2Diospyros pseudomalabarica 2Euodia nervosa 2
Tetractomia parviﬂora 2E. zeylanica 2D. pseudomalabarica 1
Dactylocladus stenostachys 1Calophyllum fragrans 1Xylopia coriifolia 1
Ilex hypoglauca 1Garcinia beccarii 1
Mezzetia leptopoda 1Neoscortechinia kingii 1
Neoscortechinia kingii 1Palaquium walsuraefolium 1
Xylopia coriifolia 1
Baccaurea bracteata 1
Dactylocladus stenostachys 1
E. cerina 1
Ilex hypoglauca 1
Litsea accedens 1
Nephelium maingayi 1
Platea latifolia 1
S. platycarpa 1
late pregnancies and pups dependent on mothers were
seen (Gumal 2001). Short-term movement between
roosts, as seen during the continuous-movement phase,
is potentially costly and, as such, should be balanced by
the beneﬁts associated with moving (Lewis 1995). There
are at least ﬁve beneﬁts of roost lability: (1) bats switch
roosts in response to human disturbance; (2) commuting
distance to foraging areas is minimized; (3) unfavourable
conditions in the roost are avoided; (4) predation is
reduced; and (5) bats escape from large populations of
ectoparasites (Lewis 1995).
The ﬁrst two beneﬁts were not apparent in Sedilu,
as human disturbance was eliminated at this remote
roost site and the movements did not decrease the
distances to the foraging areas. Even though most
Pteropus species are seen roosting in bare branches
(Pierson & Rainey 1992), there are reports that some
species prefer roosting in thick foliage as it protects them
from the heat of the sun (Advani 1982, Brooke et al.
2000). Avoidance of heat cannot be an explanation for
the regular movement in the continuous phase, as the
animals in heavily foliaged trees received protection from
Table 3. Roost selection by Pteropus vampyrus natunae, in a peat swamp forest in Sedilu, Sarawak. Heights and girths of
roost trees in two, 500 ×40-m forest plots (R2 and R7) in relation to two, 500 m×5 m forest transects (SF2 and SF7)
are compared. Only trees ≥10 cm dbh are included in the comparisons.
Variables, sampling plots and
transects – all trees >10 cm dbh Mean SD N Test statistic
R2 27.0 13.2 141 T’=6.3, df =273.7, P <0.001
SF2 transect 18.08 11.1 156
R7 25.70 14.7 72 T’ =3.3, df =104.8, P =0.001
SF7 transect 19.30 11.9 209
R2 15.80 7.9 141 T’ =5.4, df =226.3, P <0.001
SF2 transect 11.70 4.8 156
R7 16.3 10.31 72 T’ =3.6, df =85.7, P =0.001
SF7 transect 11.68 5.9 209
256 MELVIN TERRY GUMAL
the sun, regardless of whether there were continuous
Predation may have a signiﬁcant selective pressure on
behaviour of bats (Lewis 1995), and roost movement is
reported to be effective at deterring predators that are
attracted to the smell of frequently used roosts or sit-
and-wait predators that learn the location of a roost
(Fenton et al. 1994, L. Hall, pers. comm.). There were
four species of raptors seen at the maternity roost site, i.e.
white-bellied sea eagle (Haliaeetus leucogaster), brahminy
kite (Haliastur indus), black eagle (Ictinaetus malayensis)
and one unknown, probably changeable hawk eagle
(Spizaetus cirrhatus) (Gumal 2001). White-bellied sea
seen or heard almost daily (Gumal 2001). Raptors were
seen soaring above the roost, perching on trees close to
roosting animals, chasing the bats on the wing and diving
into the roosting trees. There were three conﬁrmed kills of
Pteropus vampyrus natunae by the white-bellied sea eagle
(Gumal 2001). On all three occasions, the raptor caught
the animal in the roosting tree and then carried its prey off
to another perch. Due to the small sample sizes, statistical
comparisons could not be made to verify the effect of
the movements phases on the predator activity (Gumal
According to Brooke et al. (2000), the potential reason
for small-scale roost movements in Pteropus tonganus was
to reduce infestations of ectoparasites on the animals.
There were at least two types of ectoparasites found
on Pteropus vampyrus natunae, i.e. ﬂies and mites. Flies
(families Diopsidae and Tabanidae) were seen hovering
around the bats, and the roosting animals would be
seen ﬂicking them off their bodies, whereas Dermacentor
australis (mite) had its nymphal stage on the animals.
Ectoparasites from the families Laelapidae, Nycteribiidae
and Spinturnicidae have also been reported for Pteropus
vampyrus malaccensis (Beck 1971). Roost lability before
the eggs or pupae develop may be an effective means
of curtailing ectoparasite reproduction (Lewis 1995).
Lability may be especially advantageous for lactating
females, because hairless juveniles are particularly prone
to parasitism (Watkins 1972). The continuous phases
which occurred during late pregnancy and the birth
season (Gumal 2001), could thus be a response to protect
the young from the parasites. If this were the case, it
could then follow that the restricted-movement phases
existed, because the older pups were stronger and more
able to withstand the ectoparasites, and the reason for
situating this phase in the most remote part of Belanga
River was that it offered the pups greater protection from
villages close to the conﬂuence of Sebangan and Belanga
Rivers (Gumal 2001). To help conﬁrm the reasons for the
movement phases, it is thus suggested that the parasite
loads on bats and their predation by raptors over the two
phases be examined.
Across the restricted- and continuous-movement
phases, the bats tended to be found in similar sizes of larger
trees. The preference for larger trees is obvious as the
mean sizes of the roost trees were signiﬁcantly larger
than the mean found in the surrounding peat swamp.
Pteropus species have been found to use various sizes of
roost trees, e.g. P. giganteus roosted in trees with a dbh
greater than 7 cm (Goyal et al. 1993), P. voeltzkowi used
trees with dbh ranging from 10 cm to 19 m (Entwistle
& Corp 1997), and P. tonganus used trees that ranged in
height from 3.6 to 30.5 m (Brooke et al. 2000). In all these
examples, however, comparisons were not made between
the mean size of roost trees and that of the surrounding
The bats also appeared to prefer certain species as
roost trees. A preference for certain tree species as roosts
has been reported for P.giganteus as it used Eucalyptus
spp. throughout the year (Goyal et al. 1993). The use
of these trees was higher than the frequency mean for
the species in that habitat. Such preferences were not
found for P. tonganus, which used the most common
species for roosting (Brooke et al. 2000). Although Ficus
spp. and Casuarina spp. are two of the most commonly
quoted roosting trees for the genus Pteropus (Pierson &
Rainey 1992), the bats did not use these trees as roosts.
In fact, there is also no lowland swamp Casuarinaceae in
As stated previously, some Pteropus species prefer
roosting in thick foliage as this protects them from the
heat of the sun. At Sedilu however, it is not known
why certain species of trees were used as roosts, as data
on variables like foliage density and tree form were not
recorded. It was also not possible to determine whether
the bats preferred sturdier species, as strength properties
of the trees (e.g. compressive, shearing and bending)
were not known. Such documentation is only available
for some commercial timber species like Dryobalanops
rappa and Shorea scaberrima (Sarawak Forest Department
unpublished information), whereas most of the roost trees
were considered as non-commercial.
Similarly, it is also uncertain why larger trees were
chosen as roosts. Larger trees would be a better choice,
if they provide extra roosting space, and therefore,
more animals in each tree, leading to better protection
from predators – the selﬁsh herd (Altringham 1996,
Hamilton 1971). Larger trees could also mean easier
‘free-fall take-off’ for these large bats (Kingdon 1974).
Unfortunately, however, there were insufﬁcient sightings
of kills by predators (Gumal 2001), as well as problems
with qualifying ease of take-off at different sized roost trees.
From ﬂoristic composition and tree structure alone
(trees ≥10 cm, dbh), the maternity roost sites did not
have to be centred in Belanga River, as the forests on the
banks of Sebangan River were extremely similar (Gumal
2001). Since the animals were not restricted in their
Diurnal home range and roosting trees of a maternity colony of Pteropus vampyrus natunae 257
mobility, the ﬂoristic structure and composition of the
forest on their own, cannot be the variables determining
the locations of the roosting sites. Possible reasons why
the bats roosted at this remote area could be historical
philopatry, as well as retreating to this less-accessible site
to avoid the historical mass hunting at the more accessible
roosting areas (Gumal 2001). Such roost retreats to avoid
hunters is not uncommon as it has been documented
for Pteropus vampyrus edulis (Goodwin 1979), Pteropus
marianus pelewensis (Wiles et al. 1997) and Pteropus
mariannus yapensis (Falanruw 1988).
The bats had restricted- and continuous-movement
phases in their diurnal home range, and both phases were
concentrated at Belanga River, i.e. at the most remote site
in Sedilu, even though the forest throughout the area is
shown to be largely homogeneous. The mean diameter
and height of roost trees tended to be larger than that
found in the surrounding forest. The bats preferentially
roosted in certain species of trees.
This study was ﬁnanced by the Sarawak Forest Depart-
ment (Malaysia) and Wildlife Conservation Society. WWF
Malaysia is thanked as it provided for funds for the write-
up. I would also like to acknowledge the constructive
criticisms and advice on earlier drafts by David Chivers,
Kim McConkey, Elizabeth Bennett and Christina Yin.
This paper also beneﬁted from the comments of Robert
Hodgkison and two anonymous reviewers. Jegung
Suka and Runi Sylvester from Sarawak Forest Depart-
ment’s Botany Unit identiﬁed all the tagged plants.
Tabanidae and Diopsidae ﬂies were identiﬁed by Paulus
Medeng (Sarawak Forest Department’s Entomology
Unit) and mite specimens by Lim Boo Liat (Federal
Department of Wildlife and National Parks). Lihon ak
Singga and Johnson ak Panggi are thanked for helping
me clear trails in the peat swamp. Finally thank you
to Katherine and Emily. Approval for the research was
given under SBC-RA-0041-MTG.
ADVANI, R. 1982. Distribution and status of Chiroptera species in
Rajasthan, India. S¨
ALDRIDGE, H. D. J. N. & BRIGHAM, R. M. 1988. Load carrying and
maneuverability in an insectivorous bat: a test of the 5% ‘rule’ of
radio-telemetry. Journal of Mammalogy 69:379–382.
ALTRINGHAM, J. D. 1996. Bats: biology and behaviour. Oxford University
Press, Oxford. 262 pp.
BANKS, E. 1949. Bornean mammals. Kuching Press, Kuching. 88 pp.
BECK, A. J. 1971. A survey of ectoparasites in West Malaysia. Journal of
Medical Entomology 8:147–152.
BROOKE, A. P., SOLEK, C. & TUALAULELEI, A. 2000. Roosting
behaviour of colonial and solitary ﬂying foxes in American Samoa
(Chiroptera: Pteropodidae). Biotropica 32:338–350.
DAVISON, G. W. H. 1992. P.v. malaccensis. Pp. 142–143 in
Mickleburgh,P.,Hutson,A.M.&Racey,P.A.(eds).Old world fruit
bats: an action plan for their conservation. IUCN, Gland.
EBY, P. 1991. Seasonal movements of grey-headed ﬂying-foxes Pteropus
poliocephalus (Chiroptera: Pteropodidae) from two maternity camps
in northern New South Wales. Wildlife Research 18:547–559.
ENTWISTLE, A. & CORP, N. 1997. Status and distribution of the Pemba
ﬂying fox Pteropus voeltzkowi. Oryx 31:135–142.
FALANRUW, M. W. C. 1988. On the status, reproductive biology and
management of fruit bats of Yap. Micronesica 21:39–51.
FENTON, M. B., RAUTENBACH, I. L., SMITH, S. E., SWANEPOEL, C. M.,
GROSELL, J. & JAARSVELD, J. V. 1994. Raptors and bats: threats and
opportunities. Animal Behaviour 48:9–18.
GOODWIN, R. E. 1979. Bats of Timor: systematics and ecology. Bulletin
of the American Museum of Natural History 163:73–122.
GOYAL, S. P., SALE, J. B. & GUPTA, A. 1993. Roost preference,
population and food habits of the Indian ﬂying fox Pteropus giganteus
at Behran Dun. Bat Research News 34:19.
GRANT, G. S. & BANNACK, S. A. 1999. Harem structure and
reproductive behaviour of Pteropus tonganus in American Samoa.
Australian Mammalogy 21:111–120.
GUMAL, M. T. 2001. Ecology and conservation of a fruit bat in
Sarawak, Malaysia. Ph.D. thesis. University of Cambridge, England.
GUMAL, M. T., JAMAHARI, S., IRWAN, M., JANTAN-BRANDAH, C.,
KAMAL, M. & RAZAK-PAWI, A. 1998. The ecology and role of the
large ﬂying fox (Pteropus vampyrus) in Sarawakian rain forests –
1997 report. Hornbill 1:32–47.
HAMILTON, W. D. 1971. Geometry for the selﬁsh herd. Journal of
Theoretical Biology 31:295–311.
HOOGE, P. N. & EICHENLAUB, B. 1997. Animal movement extension
to Arcview version 1.1. Alaska Biological Science Centre. USGS.
Anchorage. 29 pp.
JENNRICH, R. L. & TURNER, F. B. 1969. Measurement of non-circular
home range. Journal of Theoretical Biology 22:227–237.
KINGDON, J. 1974. East African mammals, an atlas of evolution in Africa.
Volume II, Part A (insectivores and bats). Academic Press, London.
KUNZ, T. H. 1982. Roosting ecology of bats. Pp. 1–55 in Kunz, T. H.
(ed.). Ecology of bats. Plenum, New York.
KUNZ, T. H. & PIERSON, E. D. 1994. Bats of the world: an introduction.
Pp. 1–46 in Nowak, R. M. (ed.). Walker’s bats of the world. The Johns
Hopkins University Press, London.
LAW, B. & ANDERSON, J. 2000. Roost preferences and foraging ranges
of the eastern forest bat Vespadelus pumilus under two disturbance
histories in northern New South Wales, Australia. Austral Ecology
258 MELVIN TERRY GUMAL
LEWIS, S. E. 1995. Roost ﬁdelity of bats: a review. Journal of Mammalogy
LIM, B. L. 1966. Abundance and distribution of Malaysian bats in
different ecological habitats. Federation Museums Journal 11:61–76.
MACDONALD, D. W., BALL, F. G. & HOUGH, N. G. 1980. The evalu-
ation of home range size and conﬁguration using radio tracking data.
Pp. 405–425 in Amlaner, C. J. & MacDonald, D. W. (eds). A hand-
book on biotelemetry and radio tracking. Pergamon Press, Oxford.
MCCRACKEN, G. F. & WILKINSON, G. S. 2000. Bat mating systems.
Pp. 321–362 in Chrichton, E. G. & Krutzsch, P. H. (eds). Reproductive
biology of bats. Academic Press, London.
MEDWAY, Lord. 1977. Mammals of Borneo: ﬁeld keys and annotated
checklist. Monographs of the Malaysian Branch of the Royal Asiatic
MICKLEBURGH, S. P., HUTSON, A. M. & RACEY, P. A. 1992. Old
world fruit bats: an action plan for their conservation. IUCN, Gland.
MOHD–AZLAN, J., ZUBAID, A. & KUNZ, T. H. 2001. Distribution,
relative abundance, and conservation of the large ﬂying fox, Pteropus
vampyrus, in Peninsular Malaysia: a preliminary assessment. Acta
MORRISON, D. W. 1980. Efﬁciency of food utilisation by fruit bats.
PARRY-JONES, K. A. 1987. Pteropus poliocephalus (Chiroptera: Ptero-
podidae) in New South Wales. Australian Mammalogy 10:81–85.
PARRY-JONES, K. A. & AUGEE, M. L. 2001. Factors affecting the
occupation of a colony site in Sydney, New South Wales by the Grey-
headed Flying-fox, Pteropus poliocephalus (Pteropodidae). Austral
PAYNE, J., FRANCIS, C. M. & PHILLIPS, K. 1985. A ﬁeld guide to the
mammals of Borneo. The Sabah Society with WWF Malaysia, Kota
Kinabalu. 332 pp.
PIERSON, E. D. & RAINEY, W. E. 1992. The biology of ﬂying foxes of
the genus Pteropus: a review. Pp. 1–17 in Wilson, D. E. & Graham,
G. L. (eds). Paciﬁc Island ﬂying foxes: proceedings of an international
conservation conference. U.S. Fish and Wildlife Service Biological
Report No 90(23), Washington, DC.
RACEY, P. A. 1979. Two bats in the Seychelles. Oryx 15:148–152.
RICHARDS, G. C. 1990. The spectacled ﬂying-fox, Pteropus con-
spicillatus (Chiroptera: Pteropodidae), in north Queensland. 1. Roost
sites and distribution patterns. Australian Mammalogy 13:17–24.
RICHARDS, G. & HALL, L. 1998. Conservation biology of Australian
bats. Are recent advances solving our problems? Pp. 271–281 in
Kunz,T.&Racey,P.(eds).Bat biology and conservation. Smithsonian
Institution Press, London.
ROBINSON, M. F. 1992. The behavioural ecology of the serotine bat. Ph.D.
thesis. University of Cambridge, Cambridge.
SOLLA, S., BONDURIANSKY, R. & BROOKS, R. 1999. Eliminating
autocorrelation reduces biological relevance of home range
estimates. Journal of Animal Ecology 68:221–234.
STORZ, J. F., BALASINGH, J., NATHAN, P. T., EMMANUEL, K. & KUNZ,
T. H. 2000. Dispersion and site ﬁdelity in a tent-roosting population
of the short-nosed fruit bat (Cynopterus sphinx)insouthernIndia.
Journal of Tropical Ecology 16:117–131.
VONHOF, M. J. & BARCLAY, R. M. 1996. Roost-site selection and
roosting ecology of forest-dwelling bats in southern British Columbia.
Canadian Journal of Zoology 74:1797–1805.
WATKINS, L. C. 1972. Nycteceius humeralis. Mammalian Species 23:1–4.
WHITE, G. C. & GARROTT, R. A. 1990. Analysis of wildlife radio-tracking
data. Academic Press Inc., San Diego. 383 pp.
WILDLIFE CONSERVATION SOCIETY & SARAWAK FOREST
DEPARTMENT. 1996. A master plan for wildlife in Sarawak.
WCS, New York and Sarawak Forest Department, Kuching.
WILES, G. 1992. Pteropus tokudae. P. 136 in Mickleburgh, P., Hutson,
A. M. & Racey, P. A. (eds). Old world fruit bats. An action plan for their
conservation. IUCN, Gland.
WILES, G. J., ENBRING, J. & OTOBED, D. 1997. Abundance, biology and
human exploitation of bats in the Palau Islands. Journal of Zoology