ARTICLE IN PRESS
B I O L O G I CA L C O N S E RVAT I O N
x x x ( 2 0 0 8 ) x x x –x x x
available at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/biocon
Conservation value of forest fragments to Palaeotropical bats
Matthew J. Struebiga, Tigga Kingstonb, Akbar Zubaidc, Adura Mohd-Adnanc,
Stephen J. Rossitera,*
a
School of Biological and Chemical Sciences, Queen Mary, University of London, London E1 4NS, United Kingdom
Department of Biological Sciences, Texas Tech University, Lubbock, TX 79409-3131, United States
c
Faculty of Science and Technology, Universiti Kebangsaan Malaysia, 43600 UKM Bangi, Malaysia
b
A R T I C L E I N F O
A B S T R A C T
Article history:
Forested landscapes in Southeast Asia are becoming increasingly fragmented, making this
Received 18 March 2008
region a conservation and research priority. Despite its importance, few empirical studies
Received in revised form
of effects of fragmentation on biodiversity have been undertaken in the region, limiting
3 June 2008
our ability to inform land-use regimes at a time of increased pressure on forests. We esti-
Accepted 12 June 2008
mated the biodiversity value of forest fragments in peninsular Malaysia by studying fragmentation impacts on insectivorous bat species that vary in dependence of forest. We
sampled bats at seven continuous forest sites and 27 forest fragments, and tested the influ-
Keywords:
ence of fragment isolation and area on the abundance, species richness, diversity, compo-
Chiroptera
sition and nestedness of assemblages, and the abundance of the ten most common
Habitat fragmentation
species. Overall, isolation was a poor predictor of these variables. Conversely, forest area
Nestedness
was positively related with abundance and species richness of cavity/foliage-roosting bats,
Species–area relationship
but not for that of cave-roosting or edge/open space foraging species. The smallest of frag-
Isolation
ments (<150 ha) were more variable in species composition than larger fragments or con-
Malaysia
tinuous forest, and larger fragments retained substantial bat diversity, comparable to
Oil palm
continuous forest. Some fragments exhibited higher bat abundance and species richness
than continuous forest, though declines might occur in the future because of time lags
in the manifestation of fragmentation effects. Our findings suggest that fragments
>300 ha contribute substantially to landscape-level bat diversity, and that small fragments
also have some value. However, large tracts are needed to support rare, forest specialist
species and should be the conservation priority in landscape-level planning. Species that
roost in tree cavities or foliage may be more vulnerable to habitat fragmentation than those
that roost in caves.
Ó 2008 Elsevier Ltd. All rights reserved.
1.
Introduction
Habitat fragmentation is a major contributor to biodiversity
loss (Whitmore, 1997). Nowhere is this more dramatic than
in Southeast Asia, where tropical forests are becoming
increasingly disturbed and fragmented, and are rapidly being
lost to agriculture (Sodhi et al., 2007). Despite this, only a few
detailed studies of fragmentation have been conducted in the
region (e.g. Lynam and Billick, 1999; Pattanavibool and Dearden, 2002; Brühl et al., 2003; and Benedick et al., 2006). This
paucity of information hinders both our understanding
of the consequences of fragmentation for biodiversity in
* Corresponding author: Tel.: +44 2078827528.
E-mail addresses: m.struebig@qmul.ac.uk (M.J. Struebig), tigga.kingston@ttu.edu (T. Kingston), zubaid@ukm.my (A. Zubaid),
adura@pkrisc.cc.ukm.my (A. Mohd-Adnan), s.j.rossiter@qmul.ac.uk (S.J. Rossiter).
0006-3207/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.biocon.2008.06.009
Please cite this article in press as: Struebig, M.J. et al, Conservation value of forest fragments to Palaeotropical bats, Biol.
Conserv. (2008), doi:10.1016/j.biocon.2008.06.009
ARTICLE IN PRESS
2
B I O L O G I C A L C O N S E RVAT I O N
Southeast Asia, and our ability to advise stakeholders and
land owners on potential mitigation strategies (Meijaard and
Sheil, 2007).
Fragmentation reduces suitable habitat area and isolates
patches within a matrix of modified habitat. Island biogeography theory (MacArthur and Wilson, 1967) predicts that smaller, more isolated fragments support smaller populations, and
fewer species than are supported by larger or less isolated
fragments. However, a recent synthesis has reported that,
while fragment area is a significant predictor of species richness in most studies, the effects of isolation remain ambiguous (Watling and Donnelly, 2006). Species exhibit variable
responses to fragmentation. These responses are inconsistent
across taxonomic groups, but are more commonly correlated
with population size and fluctuation, disturbance sensitivity,
matrix use, biogeographic position and rarity (Henle et al.,
2004). Area-dependent declines in abundance and species
richness can result in predictable local patterns of species
extinction, with depauperate smaller fragments harbouring
nested subsets of assemblages found in larger fragments
(Wright et al., 1998). Thus, to evaluate the long-term conservation value of forest fragments, it is essential to describe
species diversity patterns and to elucidate the mechanisms
that produce them.
Bats constitute the second most species-rich order of
mammals (Wilson and Reeder, 2005) and up to half of mammal species in tropical forests (Findley, 1993). In recent decades, bat populations have experienced global declines, a
trend linked to extensive, recent habitat loss (Mickleburgh
et al., 2002). In Southeast Asia, 20% of bat species are predicted to become extinct by 2100 (Lane et al., 2006). Nonetheless, bats are frequently overlooked in biodiversity
assessments and fragmentation research, possibly because
they are widely perceived to be at low risk of extinction due
to their ability to fly.
The perception of bats as low priority subjects for conservation research may be overly optimistic because these animals exhibit combinations of traits that may increase their
sensitivity to habitat loss and disturbance. Ecomorphological
factors such as wing shape (Schnitzler and Kalko, 2001),
behaviours including coloniality and strong site fidelity (Miller-Butterworth et al., 2003) and slow rates of reproduction
(Barclay and Harder, 2003) constrain their ecological flexibility. In addition, bats are dependent on the availability of suitable roosting sites. Consequently, populations of species that
roost in trees (in hollows and cavities of standing trees, under
fallen trees and logs), may be adversely impacted by fragmentation via the direct loss of rare roosting sites in fragments
(Schulze et al., 2000) or changes to roost suitability resulting
from edge effects (e.g. disturbance levels and microclimate
changes; Laurance et al., 2002). These edge effects also have
the potential to influence the persistence of bats that roost
in foliage (under modified or unmodified leaves), because
these types of roost are more exposed to abiotic conditions
and disturbance. Similarly, fragmentation can also lead to
the separation of cave roosts from foraging habitat (fragments). However, given the natural patchy distribution of
caves, and that they typically support large populations of
bats, communal cave roosting species are likely to have been
selected for greater vagility over evolutionary time. As a re-
x x x ( 2 0 0 8 ) x x x –x x x
sult, these species typically commute great distances to and
from foraging grounds each night (Altringham, 1999), and so
may be better adapted to persist in a fragmented landscape
than their tree and foliage-roosting counterparts.
To date, fragmentation studies of tropical bats have focused almost exclusively on assemblages in the Neotropics
(but see Law et al., 1999 for a study in Australia), where bat
assemblages in forests are dominated by members of the
family Phyllostomidae. These studies have suggested that
bat species richness may be largely unaffected by fragmentation (Schulze et al., 2000; Estrada and Coates-Estrada, 2002;
Pineda et al., 2005; Faria, 2006; Bernard and Fenton, 2007),
though subtle impacts on assemblage structure have been detected (Cosson et al., 1999; Gorresen and Willig, 2004). However, meaningful comparisons between studies are
complicated by differences in sampling effort and fragmentation history, as well as variation in potential determinants of
assemblage structure in fragments, such as the degree of contrast between fragments and the matrix (low contrast, Estrada and Coates-Estrada, 2002; Pineda et al., 2005; Faria, 2006;
versus high contrast, Cosson et al., 1999; Bernard and Fenton,
2007), and the availability and size of forest patches in a landscape (Gorresen and Willig, 2004).
Palaeotropical bat assemblages are dominated by members of the families Rhinolophidae and Hipposideridae, and
the Vespertilionidae subfamilies Kerivoulinae and Murininae. These species are not present in the Neotropics, and
many of them are typically highly adapted for foraging in
the clutter of the forest interior (‘narrow-space’ ensemble,
sensu Schnitzler and Kalko, 2001). Consequently, these species may be more sensitive to forest loss and exhibit greater
avoidance of disturbed and open habitats than Neotropical
bats (reviewed in Kingston et al., 2003). Because of this
dependence on forest, we expect these species to be adversely affected by deforestation and other forest disturbance events (Lane et al., 2006).
We determined the conservation value of forest fragments
in the Palaeotropics by using bats as a focal animal group, and
quantifying abundance, species richness, diversity, assemblage composition and nestedness in a fragmented landscape
in central peninsular Malaysia. We focused on the narrowspace ensemble of insectivorous species due to their predicted vulnerability and because they can be readily captured
in forests using a single, standardised sampling technique.
Landscapes in Malaysia have undergone major changes over
the last century as forests have been rapidly cleared for timber, urbanisation and plantation agriculture (KathirithambyWells, 2005). Increasing demand for plantation products such
as oil palm (Elaeis guineensis) places pressures on land owners
to increase yields; one way to do this is to increase production
area by clearing forest remnants. Hence, we sought to inform
these management decisions regarding the value of such
remnants at a time of increased pressure on remaining forest
habitats.
We hypothesised that (1) species richness, abundance and
diversity of bats is lower in smaller or more isolated forest
fragments than in larger or less isolated fragments; (2) fragmentation effects are stronger in species that roost in tree
cavities and/or foliage than in more vagile species that roost
in caves and (3) assemblages in smaller or more isolated frag-
Please cite this article in press as: Struebig, M.J. et al, Conservation value of forest fragments to Palaeotropical bats, Biol.
Conserv. (2008), doi:10.1016/j.biocon.2008.06.009
ARTICLE IN PRESS
B I O L O G I CA L C O N S E RVAT I O N
ments represent nested subsets of those in larger or less isolated fragments.
2.
Methods
2.1.
Study landscape
The Krau landscape in central Pahang state (3°40 0 N, 102°10 0 E;
Fig. 1) represents 562,060 ha of land and is bounded by continuous forest to the north and west. Historically this landscape
has experienced little deforestation, but in recent years (1966–
2002), 39 % of the forest has been felled (DAPM, 2005). Today,
large blocks of undisturbed continuous forest remain protected as the Krau Wildlife Reserve and neighbouring Forest
Reserves, while 43% of land is covered by rubber (Hevea brasiliensis) and oil palm plantations, surrounding smaller forest
fragments.
The majority of natural vegetation in the Krau landscape is
lowland or hill dipterocarp forest, with associated dominant
tree species including Dipterocarpus cornutus, D. baudii, Hopea
sangal, Shorea acuminata and S. ovalis, or Anisoptera laevis, D.
grandiflorus, S. leprosula, S. cutisii and Vatica cuspidata respectively (Yusof and Sorenson, 2000). The annual 24-h mean temperature is 26 °C, and monthly precipitation typically exhibits
two periods of maximum rainfall between September and
December, and March and May, separated by two periods of
minimum rainfall (Yusof and Sorenson, 2000).
x x x ( 2 0 0 8 ) x x x –x x x
2.2.
3
Forests sampled and patch metrics
We sampled bats between May 2002 and June 2007 at 35 lowland forest sites. These comprised five undisturbed sites and
two disturbed sites within Krau Wildlife Reserve (S01-S07,
mean distance between sites 17 km); and 27 forest fragments
varying in size from 3 ha to 11,339 ha (F01-F27, Fig. 1). Fragments were identified from land-use maps (DAPM, 2005) verified by visual interpretation of 2002 Landsat ETM satellite
images; they represent the range of forest remnant sizes
and land-use histories in the landscape. Fragments were subject to ongoing disturbance: all exhibited evidence of logging
as well as hunting of wild pigs (Sus scrofa) and mouse deer
(Tragulus spp.).
We used ArcView version 3.2 to calculate forest area and
two measures of isolation widely used in fragmentation studies (Watling and Donnelly, 2006): the shortest Euclidean distance to nearest continuous forest, and the distance to the
nearest forest patch. All metrics were independent from each
other (Pearson’s r < 0.3; p > 0.3), and were log transformed to
approximate to normal distributions.
2.3.
Bat sampling
We minimised methodological heterogeneity and capture
biases (Kingston et al., 2003) by restricting sampling to insectivorous species that are readily captured in the forest under-
Fig. 1 – Locations of sampling sites in the Krau landscape, peninsular Malaysia, including those in continuous forest (S prefix)
and forest fragments (F prefix). Dark grey areas represent forest cover in 2002 and light grey areas represent additional forest
cover in 1966, both according to Malaysian Ministry of Agriculture maps. White areas consist of a plantation mosaic,
primarily oil palm and rubber, but also with substantial areas of durian (Durio spp.) and Acacia spp. Forest cover for all
peninsular Malaysia is shown in the inset, and black lines indicate state boundaries.
Please cite this article in press as: Struebig, M.J. et al, Conservation value of forest fragments to Palaeotropical bats, Biol.
Conserv. (2008), doi:10.1016/j.biocon.2008.06.009
ARTICLE IN PRESS
4
B I O L O G I C A L C O N S E RVAT I O N
storey, conducting fieldwork only in dry seasons, and avoiding
periods of heavy rain. Bats were captured using up to fifteen
four-bank harp traps positioned across flight paths (trails, logging skids, streams, or swamp beds) each night and then
moved to a new position the following day – hence one trap
set for one complete night constituted one harp trap night
(HTN), following Kingston et al. (2003).
Bats were collected and identified following the procedures of Kingston et al. (2006) and were marked either with
uniquely numbered forearm bands or wing biopsies, so that
recaptures could be recognised and excluded from analyses.
Individuals were released within 12 h at the capture point.
We classified species into three classes based on dispersal
capabilities inferred from wing morphology and roosting
ecology. Wing morphology was first used to define species
at the level of ensemble by distinguishing bats that forage
in narrow spaces (‘narrow-space’ bats, sensu Schnitzler and
Kalko, 2001; ‘Strategy I’ bats, sensu Kingston et al., 2003) from
those that primarily forage in edges or open spaces (‘Strategy
I’ and ‘Strategy III’ bats, sensu Kingston et al., 2003). Species in
the narrow-space ensemble were then further partitioned
based on their roosting ecology into two classes: (1) tree cavity/foliage-roosting species and (2) cave-roosting species
(including rock crevices).
2.4.
Sampling design
The 27 fragments varied substantially in isolation history, as
well as in distances to other fragments and to areas of karst
limestone, which hosted large populations of cave-roosting
bats that dominated assemblages. Because these factors were
likely to obscure the effects of fragmentation on bat assemblages, we analysed a subset of 15 fragments, each of which
was a minimum of 500 m from other fragments and 2 km
from karst sites, and had a sampling effort of at least 15
HTN. Isolation distances for these sites ranged from 2.1 to
11.0 km (mean 6.4 km) from continuous forest, and 0.6 to
2.3 km (mean 1.3 km) from other fragments. For comparisons
with undisturbed continuous forest, we classified fragments
by size, which ranged across three orders of magnitude: small
(mean 70 ha, range 31–102); medium (mean 353 ha, range 251–
443); and large (mean 5410 ha, range 2025–11 339). Data from
sites within each size class were then pooled to provide sufficient sample sizes for statistical analyses.
2.5.
Statistical analyses
A suite of analyses was designed to evaluate the effects of forest fragmentation on bat abundance, species richness, diversity, assemblage composition and nestedness. We used the
Simpson index as our measure of diversity because this measure is weighted toward common species, and allows for
examination of patterns of species dominance, or in its reciprocal form, species evenness. Because of the exploratory
nature of our study and extensive debate regarding the use
of adjustments for multiple tests in the ecological literature
(e.g. Roback and Askins, 2005), we report the exact p-values
for all analyses. We tested the influence of fragment area
and isolation on assemblage variables (total bat abundance,
x x x ( 2 0 0 8 ) x x x –x x x
observed and predicted species richness, and diversity), using
generalised linear models (GLMs) undertaken in the R-statistical package version 2.5.1 (http://www.r-project.org). This approach focussed on site-level species richness or total bat
abundance (i.e. standardised richness or abundance for all
species pooled in all traps at each site), and was also used
separately to test the influence of fragment metrics on the
abundance and species richness for each of the three classes
of bats (i.e. edge/open space foraging species, cave-roosting
narrow-space species, and tree cavity-foliage roosting narrow-space species). Observed species richness (Sobs) and reciprocal Simpson diversity (1/D, evenness) were derived from
sample-based rarefaction curves (Colwell, 2004), and species
richness was predicted at a standard number of individuals
(200) using the Shen multinomial model (Shen et al., 2003;
Chao and Shen, 2003-2005). Square root transformations were
used on abundances and predicted species richness to
approximate normal distributions without special treatment
of zeros (McCune and Grace, 2002).
Sample sizes were sufficient to warrant testing the responses in the abundance of the ten most common species
to fragment area and isolation. However, the low abundances
of these species at some sites limited our ability to detect
these responses using the GLM approach. Therefore, we
quantified the responses in species abundance to area and
both isolation metrics with a procedure that focused on
trap-level abundance of these species using a generalised linear mixed-effects model (GLMM) with Poisson error terms.
Modelling sites as random effects in GLMMs also controlled
for pseudoreplication within a site and accounted for variance attributable to particular sites. All GLMMs were undertaken in R with the lmer function from the lme4 package
(Bates, 2008). The three fragment metrics were modelled as
fixed effects, sites were modelled as random effects, and
the response variable was a species’ abundance in a trap (15
traps per site). p-values were generated from 10 000 Markov
chain Monte Carlo (MCMC) simulations using the languageR
package (Baayan, 2008).
Bat abundance, species richness and reciprocal Simpson
diversity (i.e. evenness) among different size fragments and
continuous forest were compared using non-parametric
Kruskal–Wallis tests with post-hoc pairwise Mann–Whitney
U tests. This procedure was also undertaken for the abundances of the ten species for which we had sufficient sample
sizes. Estimates of observed species richness and Simpson
diversity were partitioned into additive components within
sites (a); between sites of similar sizes (b1); and between sites
of different sizes (b2), using an individual-based randomisation procedure (Veech and Crist, 2007). Because additive partitioning resulted in a and b components being measured in the
same units (Crist et al., 2003), we could assess the relative
contributions of size classes of fragments to overall (c) insectivorous bat species richness and Simpson diversity (i.e. dominance) over the Krau landscape.
Species abundance distributions within fragment size
classes and continuous forest were determined using standardised rank abundance (Whittaker) plots, on which the differences of species rank between the pooled assemblages
could be visually inspected. Differences in species abundances distributions between fragment size classes and con-
Please cite this article in press as: Struebig, M.J. et al, Conservation value of forest fragments to Palaeotropical bats, Biol.
Conserv. (2008), doi:10.1016/j.biocon.2008.06.009
ARTICLE IN PRESS
B I O L O G I CA L C O N S E RVAT I O N
tinuous forest were assessed using a v2 randomisation test in
Ecosim version 7 (Gotelli and Entsminger, 2007). Because this
test uses the same expected abundance values for both observed and simulated data (i.e. 1000 simulations) the results
were not sensitive to small expected values arising from rare
species.
To determine variation in the compositional structure of
bat assemblages among sites, we used non-metric multidimensional scaling (NMDS) with the Bray-Curtis dissimilarity
index. Bray-Curtis coefficients were based on species abundances, which were square-root transformed to compress
values of abundant species relative to those of rare species
without the need to adjust zeros (i.e. species absences)
(McCune and Grace, 2002). Ordinations were implemented
using the software PC-ORD version 5 (McCune and Mefford,
2006) with 500 iterations and 250 runs of both real and randomised data. Because assemblage data are often composed
largely of rare or absent species, removing some of these
species may enhance the detection of relationships between
composition and causal factors, such as fragment metrics
(McCune and Grace, 2002). Therefore, we performed several
ordinations starting with the inclusion of all species, and
then removed subsets of species based on ensemble or rarity. The final ordination was chosen based on reducing stress
from additional axes but also retaining enough species for
the ordination to remain biologically meaningful. A GLM
was used to evaluate whether forest metrics determined
the positions of forest sites in ordination space, and hence
the compositional differences of bat assemblages between
forest sites.
Finally we determined the extent to which bat assemblages were nested by calculating nestedness derived from
presence–absence matrices of species in fragments. We performed separate analyses using matrices of all bat species
and sub-matrices for each of the three classes of bat. The
resulting temperature metric T describes the level of ‘heat
disorder’, a measure of the distribution of unexpected presences and absences in a matrix. Maximum order, or perfect
nestedness, is indicated by a temperature of zero, and significance can be assessed by comparing the observed temperature to a null distribution based on MCMC simulations. We
used the binary matrix nestedness calculator (BINMATNEST,
Rodrı́guez-Gironés and Santamarı́a, 2006), an algorithm that
overcomes limitations of other calculators concerning the
reordering and packing of matrices, the definition of the isocline of perfect order, and the appropriateness of null models
used to assess significance. BINMATNEST provides three
alternative null models on which to assess significance, with
model 3 being the most conservative according to the
authors. We therefore used this model to evaluate significance, and based our p-values on 5000 simulated matrices.
To determine if the maximally nested matrix produced an
ecologically meaningful nested arrangement, in terms of forest fragmentation, the order of forest fragments in the maximally nested matrix was correlated (Spearman rank
coefficient) with forest fragments ordered by area or isolation,
as surrogates of fragmentation intensity. Hence, a significant
correlation coefficient would suggest a nested arrangement
that resulted from the fragmentation process (Rodrı́guez-Gironés and Santamarı́a, 2006).
x x x ( 2 0 0 8 ) x x x –x x x
3.
5
Results
We captured a total of 10 343 insectivorous bats of 46 species
from 1830 HTN over seven sites in continuous forest and 491
HTN over 27 fragments (Appendix 1). Of the 7488 individuals
captured in continuous forest, tree cavity/foliage-roosting
and cave-roosting narrow-space species represented a similar
proportion of all bats captured: 47% of individuals (21 species)
were tree cavity/foliage-roosting; 52%, (12 species) were caveroosting; and 1% (7 species) were edge/open space foragers.
Conversely, of the 2857 individuals captured in fragments,
the proportion of tree cavity/foliage-roosting species was lower (26% 17 species), while that of cave-roosting species (68% 12
species) and edge/open space foraging species (6% 9 species)
was higher.
Only six individuals were recaptured between sites, all of
which were cave-roosting species (Table 1). No species were
recorded in every fragment, but four species (Rhinolophus affinis, R. lepidus, R. trifoliatus and Murina suilla) were widespread
(present in >70% of fragments and all continuous forest sites).
Sixteen species were uniformly rare (< 1% of captures in both
continuous and fragment forest sites, Appendix 2) and six
(Coelops robinsoni, Hipposideros armiger, Harpiocephalus mordax,
Kerivoula krauensis, Murina rozendaali and Myotis siligorensis)
were only recorded in continuous forest. Three species captured in fragments (Hesperoptenus blanfordi, Hipposideros lylei
and Scotophilus kuhlii) were absent from our surveys in continuous forest, but have been recorded in that habitat by other
studies reviewed in Kingston et al. (2006).
3.1.
Patterns of abundance and assemblage composition
More bats were captured in larger fragments than smaller
fragments (Fig. 2; Table 2), but there was no response in total
or ensemble abundance based on either measure of fragment
isolation. Fragment area explained the majority of variation
in total bat abundance (72.7% Fig. 2a), with measures of isolation consistently removed from GLMs. When partitioning this
relationship by ensemble and roosting class, area explained
variation in abundance of tree cavity/foliage-roosting bats
(54.5% Fig. 2b), but not for cave-roosting bats (Fig. 2c) or
edge/open space foraging bats (Fig. 2d). There was no response to fragment isolation exhibited by any ensemble or
roosting class.
The total bat abundance of the smallest fragments was
significantly lower than continuous forest (Kruskal–Wallis
v2 = 11.78, p = 0.008; all pairwise Mann–Whitney comparisons,
U < 0.001, p = 0.009), but abundance in medium and large fragments was similar to continuous forest (U = 7.0–11.0, p > 0.05).
When considered by roosting class, fewer tree cavity/foliageroosting bats were captured in fragments of all size classes
compared to continuous forest sites (v2 = 10.73, p = 0.013; all
pairwise analyses, U > 1, p < 0.05). However, for cave-roosting
and edge/open space foraging bats no differences between
size classes were detected (v2 = 5.38, p = 0.15; and v2 = 5.01,
p = 0.17, respectively).
Species abundance distributions were unequal across fragment size classes (observed v2 = 913.17, simulated
v2 = 111.18 ± 13.08, p < 0.0001; Fig. 3). Kerivoula intermedia was
among the most dominant tree cavity/foliage-roosting spe-
Please cite this article in press as: Struebig, M.J. et al, Conservation value of forest fragments to Palaeotropical bats, Biol.
Conserv. (2008), doi:10.1016/j.biocon.2008.06.009
ARTICLE IN PRESS
6
B I O L O G I C A L C O N S E RVAT I O N
x x x ( 2 0 0 8 ) x x x –x x x
Table 1 – Individuals recaptured in forest fragments in the Krau landscape during this study, with the Euclidean distance
between sites of capture and recapture
Species
Hipposideros cervinus
Rhinolophus affinis
Rhinolophus stheno
Sex and
reproductive conditiona
#A
$ A PL
#A
$AP
$ A PL
#A
Identity band
number
MBCRU A6251
MBCRU A8593
MBCRU A5326
MBCRU B2183
THK 34227
MBCRU A6546
Site captured
Site recaptured
Distance between
sites (km)
S05
S05
S05
S01
F17
S05
F10
F12
F09
F08
F18
F10
12.5
10.7
10.4
4.9
1.9
10.9
a A, mature adult; PL, post-lactating; P, pregnant.
Fig. 2 – Relationships between insectivorous bat abundance (a–d) or species richness (e–h) and forest area, for ensembles of
bats with different roosting/foraging ecology. Black circles indicate forest fragments and grey circles indicate continuous
forest sites. Regression models performed using fragment sites only. Fragment F15 was identified as an outlier with a hyperabundance of cave-roosting species and so was excluded from regressions.
cies in continuous forest (rank = 2) and large fragments
(rank = 1), but was much rarer in medium (rank = 19) and
small fragments (rank = 20). In contrast, Rhinolophus affinis
was a rare cave-roosting species in continuous forest
(rank = 18), but was much more abundant in fragments (ranks
from large fragments to small = 3, 1, 3).
For species-level analyses of abundance significantly fewer
bats were captured in fragments compared to continuous forest for two tree cavity/foliage-roosting species – K. intermedia
(v2 = 9.11, p = 0.028) and K. papillosa (v2 = 11.99, p = 0.007).
These results were also supported by GLMM models, which
showed that more individuals of these species were captured
in traps set in larger fragments than in smaller fragments (K.
intermedia, p = 0.010; K. papillosa, p < 0.0001), and that both isolation metrics were not significant predictors of abundance.
Similarly, species abundance of two cave-roosting species,
Hipposideros cervinus and H. larvatus, was greater in larger fragments than smaller fragments (p = 0.005 and p = 0.014 respectively), but also increased with greater distance from
continuous forest (p = 0.030 and p = 0.032 respectively). The
six other common species (tree cavity/foliage roosting: Murina
suilla, Rhinolophus trifoliatus; and cave-roosting: H. bicolor 131,
H. bicolor 142, R. affinis, R. lepidus) exhibited no response in
abundance to any fragment metric (p > 0.1).
3.2.
Species richness and diversity
Smaller fragments typically supported fewer bat species than
larger fragments or continuous forest. Significant positive
relationships existed between species richness and log-area
using observed and predicted species richness (r2 = 0.309,
p = 0.01, Fig. 2e; and r2 = 0.324, p = 0.02 respectively). When ob-
Please cite this article in press as: Struebig, M.J. et al, Conservation value of forest fragments to Palaeotropical bats, Biol.
Conserv. (2008), doi:10.1016/j.biocon.2008.06.009
ARTICLE IN PRESS
B I O L O G I CA L C O N S E RVAT I O N
7
x x x ( 2 0 0 8 ) x x x –x x x
Table 2 – Landscape metric and insectivorous bat assemblage characteristics for fragment and continuous forest sites used
for analyses
Forest class and site
Landscape metrics
Area (ha)
a
Assemblage characteristics
b
Isolation (km)
Nearest forest (km)
N
c
Sobsd
S200e
1/Df
Small fragments
F03 RTP Lembah Klau
F09 Paya Parit
F11 Ulu Rugan
F22 Desa Bakti
F25 Jambu Rias
100
31
122
107
32
7.4
5.0
8.8
5.6
4.6
2.3
0.6
1.7
2.1
1.2
39
41
23
42
20
10
11
5
12
7
11.8
27.3
8.9
17.0
7.4
4.7
6.3
3.2
6.1
6.3
Medium fragments
F06 Klau Kecil
F08 Paya Luas
F10 Hutan Kerdau
F14 Kampung Lebu
F15 Rumpun Makmur
443
353
319
400
251
3.7
2.1
8.1
7.4
4.6
1.4
1.0
0.6
2.3
1.8
63
43
71
77
157
17
11
14
17
19
19.1
12.7
19.8
18.9
20.4
10.7
6.8
5.4
8.2
7.0
Large fragments
F01 Kemasul 1
F02 Kemasul 2
F21 Belungu
F23 Klau Besar
F24 Jengka
2883
11 339
5225
5581
2025
7.7
7.5
11.0
5.5
7.6
0.6
1.2
1.8
0.7
0.6
65
113
104
91
112
15
17
14
17
13
19.8
18.3
18.9
19.5
14.5
4.2
8.8
4.2
9.1
4.9
Continuous forest
S01 Kuala Lompat
S02 Lubuk Baung
S03 Kuala Serloh
S04 Kuala Gandah
S05 Jenderak Selatan
137
137
137
137
137
–
–
–
–
–
67
66
49
61
162
16
13
11
16
12
20.0
16.6
11.9
19.4
12.5
10.0
7.9
7.4
8.0
3.5
a
b
c
d
e
f
000
000
000
000
000
–
–
–
–
–
The shortest straight-line distance to continuous forest.
The shortest straight-line distance to the nearest forest fragment.
Number of individuals captured in 15 harp traps set on trails in a site.
Number of observed species.
Predicted number of species at the 200 individual abundance level using the model proposed by Shen et al. (2003).
Reciprocal Simpson index – higher values indicate a more diverse assemblage with even species abundances.
served species richness was considered separately for each
ensemble and roosting class, only tree cavity/foliage-roosting
bats exhibited a positive response to log-area (Fig. 2f–h). No
significant relationships existed between reciprocal Simpson
diversity (i.e. evenness) and any of the fragment metrics. Fewer species were recorded in small fragments than continuous
forest (v2 = 10.73, p = 0.013; U = 2.00, p = 0.027), but fragment
size classes were similar in terms of predicted species richness (v2 = 7.20, p = 0.066), or reciprocal Simpson diversity
(v2 = 4.45, p = 0.217).
Additive partitioning revealed that species richness of bats
within sites, between sites of similar size, and between sites
of different size, contributed almost equal proportions to
the overall insectivorous bat species richness (a = 36.8%;
b1 = 35.7%; b2 = 27.7%). However, the majority of species dominance, as measured by Simpson diversity, was attributed to
within sites (a = 89.7% b1 = 6.9%; b2 = 3.3%) suggesting that
sites were highly dominated by common species.
based on assemblages with all species, or with rare species
excluded, had higher stress (20.0) and so were less reliable,
but showed a similar pattern. The final ordination represented 82% of variation in dissimilarity, and showed that
small fragments (<150 ha) were atypical of the pattern of
assemblage structure at other sites based on their wide scatter on the ordination plot (Fig. 4). Hence, bat species composition in small fragments was more variable than that in larger
fragments or continuous forest. Forest area was the sole predictor of assemblage composition described by NMDS axis 1,
which represented the majority of variation in dissimilarity
(r2 = 0.430, p = 0.001). There was no such relationship between
the scores of other axes with area or isolation. Hence, larger
sites, including those in continuous forest and large fragments, were more similar to each other in terms of bat species composition than smaller sites.
3.3.
When all bat species were considered in the analysis assemblages were significantly nested (observed T = 25.29, expected
T = 44.77 ± 4.04, p < 0.001), and the rank order of forest fragments in the maximally nested matrix was strongly negatively correlated with forest fragments ordered by area
(rs = 0.639, p = 0.001). Tree cavity/foliage roosting bats were
Bat assemblage composition
The final NMDS ordination of species dissimilarity among
sites consisted of two axes and was based on a matrix that included all bats except for the edge/open space foraging
ensemble (27 species, stress = 16.7, (Fig. 4)). Other ordinations
3.4.
Nestedness of species assemblages
Please cite this article in press as: Struebig, M.J. et al, Conservation value of forest fragments to Palaeotropical bats, Biol.
Conserv. (2008), doi:10.1016/j.biocon.2008.06.009
ARTICLE IN PRESS
8
B I O L O G I C A L C O N S E RVAT I O N
x x x ( 2 0 0 8 ) x x x –x x x
Fig. 3 – Rank abundance (Whittaker) plots for insectivorous bats in three size classes of forest and continuous forest in the
Krau landscape. Species are ranked according to the abundance of each species (n) and the total abundance of all species for
each forest class (N). Species codes are in Appendix 2.
Fig. 4 – Nonmetric multimensional scaling (NMDS)
ordination for Bray-Curtis species dissimilarity of
insectivorous bat assemblages in 15 forest fragments and
five continuous forest sites in the Krau landscape. A 2dimensional ordination that excluded edge/open space
species was the best solution (stress = 16.7), and
represented the majority of species (27) and variance in
dissimilarity (82%). Points are scaled to log transformed
forest area, the sole significant predictor of assemblage
composition. Grey points represent sites in continuous
forest and black points represent forest fragments.
also significantly nested when assessed separately (observed
T = 20.00, expected T = 36.81 ± 5.23, p < 0.001), and the rank order of fragments exhibited a similar negative relationship
with fragment area (rs = 0.693, p = 0.001). Cave-roosting bats
exhibited nested subsets, but were not as strongly nested as
other groups of bats (observed T = 17.22, expected
T = 34.06 ± 6.54, p < 0.01), and no significant correlation between fragment rank and area was evident (rs = 0.261,
p = 0.174). No nested pattern was evident for edge/open space
foraging bats (observed T = 14.69, expected T = 21.20 ± 6.71,
p = 0.216), and no correlation was observed between the
nested rank order of fragments and either measure of isolation for any of the matrices (p > 0.1). Hence, nested analyses
suggested that bat assemblages in smaller fragments were
subsets of those in larger fragments, that this was more evident for tree cavity/foliage-roosting and cave-roosting narrow-space bats, and that fragment area rather than
isolation played a causal role in this nested structure.
4.
Discussion
We recorded diverse insectivorous bat assemblages and
found evidence that fragmentation has negative effects on
bat abundance, species richness and assemblage composition. Overall bats responded to changes in fragment area,
but not isolation, and assemblages in small fragments were
nested subsets of those in large fragments. Moreover, tree
Please cite this article in press as: Struebig, M.J. et al, Conservation value of forest fragments to Palaeotropical bats, Biol.
Conserv. (2008), doi:10.1016/j.biocon.2008.06.009
ARTICLE IN PRESS
B I O L O G I CA L C O N S E RVAT I O N
cavity/foliage-roosting species appeared more susceptible to
fragmentation than cave-roosting species or edge/open foraging species.
We found that fragment area, but not fragment isolation,
influenced assemblage-level abundance, species richness,
composition and nestedness rankings of bat species, in agreement with the majority of fragmentation studies (Watling and
Donnelly, 2006). At the level of individual species, we also
found that the abundance of four species (Hipposideros cervinus, H. larvatus, K. intermedia and K. papillosa) responded positively to increases in fragment area; however, contrary to
expectations, the abundance of two of these (H. cervinus and
H. larvatus) also increased with greater isolation distance from
continuous forest. Given that both of these species roost in
caves, we suggest that this response was an artefact of the
distribution of caves; several large cave systems are known
in the east and west of the Krau landscape, but only a few
small caves are known near the continuous forest sites, and
none of these support large bat populations. Thus it is likely
that, at least for some cave-roosting species, their roosts are
simply of sufficient distance from continuous forest for bats
not to be recorded in great numbers at the sites we studied.
This notion is also supported by the greater abundance rankings of the cave-roosting species H. larvatus, Rhinolophus affinis
and R. lepidus in fragments compared to continuous forest
(Fig. 3).
Our inability to detect a clear effect of fragment isolation
on bat assemblage structure might reflect the comparatively
limited distribution in values of isolation distance compared
to area in the study (see Watling and Donnelly, 2006). However, we also suspect that the lack of an impact of isolation
is real, and is attributable to several aspects of the study area
and focal species. First, the Krau landscape is characterised
by a low level of matrix contrast. The structural contrast between fragments and matrix determines the extent to which
animals can move across fragment boundaries, and is highly
dependent on a species’ vagility and its perception of habitat
(see Ewers and Didham, 2006). In an example of extreme matrix contrast, Meyer and Kalko (in press) studied land-bridge
islands in Panama and found that island (fragment) isolation,
rather than area, was linked to patterns of nestedness in bat
assemblages. In Krau, the more hospitable matrix consisting
of plantations and village gardens is likely to be more easily
traversed by bats. Indeed, the tolerance of a species to different habitats defines its effective isolation, which might differ
from that described by Euclidean distances, and so further
complicate our ability to detect isolation effects (Ricketts,
2001).
The extent to which isolation impacts assemblage structure in fragments will also be influenced by the history of
the landscape. Fragmentation is usually an ongoing process,
and although this process began in the Krau landscape ca.
50 years ago, it has been much more recently that most
changes have occurred. In addition, the rate of fragmentation
has been much faster in some areas than others. Thus, there
might be a delay in the realisation of isolation effects, with
current patterns dominated by the effects of area, which will
reflect both the prey base available to insectivorous bats, as
well as viable roosting opportunities for tree cavity/foliage
roosting species.
x x x ( 2 0 0 8 ) x x x –x x x
9
Perhaps the most obvious explanation for a lack of isolation effect is that the bats studied are sufficiently vagile to
cover distances between the fragments. However, we advocate caution in drawing this conclusion for several reasons,
not least because while empirical studies demonstrate that
some of the bat species recorded are highly mobile, the vagility of several groups appears to be more limited. Indeed, edge/
open space foraging bats, together with several species in our
cave-roosting class, may be able to commute between forest
patches, as well as utilise matrix habitats. The relatively
small, long and narrow wings characteristic of these species
(Kingston et al., 2003) result in high wing loading and high aspect ratios, which have been linked to fast energy-efficient
flight (Norberg and Rayner, 1987). Radio-tracking studies have
shown that cave-roosting rhinolophid and hipposiderid species can commute several kilometres in a single night (e.g.
H. speoris, Pavey et al., 2001; R. hipposideros, Bontadina et al.,
2002), and recapture data from our study (Table 1) confirm
that dispersal distances can exceed the distances between
some fragments. In contrast, bat species in our tree cavity/foliage roosting class appear to be more restricted to areas
around available roosts in forest fragments. These species
are characterised by low wing loading and low aspect ratios
(Kingston et al., 2003), associated with slow, manoeuvrable
but energetically expensive flight that is suited to clutter but
poorly adapted to long distances (Norberg and Rayner, 1987).
This prediction is well supported by banding records from
the Krau Wildlife Reserve; of 3900 bat recaptures, none of
the recapture distances for tree cavity/foliage roosting species
exceeded 1 km (Sujarno-Kudus, 2006). Moreover, radiotracking studies of four of these species (Kerivoula papillosa, Hipposideros ridleyi, Rhinolophus sedulus and R. trifoliatus) have
revealed that home ranges are limited to < 100 ha and do
not extend beyond forest boundaries (Allen, 2005; Fletcher,
2006). In light of such empirical evidence, we speculate that
the combined effects of low matrix contrast, heterogeneous
fragmentation rates and variation in bat species vagility are
likely to have ameliorated the impacts of isolation for at least
some of our study species.
Area-dependent relationships with species richness, nestedness and assemblage structure suggest that in our study
area large tracts of forest are needed to conserve intact bat
assemblages. Nonetheless, the greatest differences in assemblage structure were among the smallest fragments, and
many of the medium- and large-sized fragments (>300 ha)
retained substantial bat diversity, in come cases equalling
or even exceeding that of continuous forest sites. In fact,
additive partitioning revealed that almost a third of bat species richness at the landscape level was generated by diversity between sites of different size classes (i.e. b2), which
was a similar contribution to that of species richness from
within sites (i.e. a). Although this suggests that smaller fragments have substantial value for bat diversity when considered together at the landscape level, again, there are several
reasons to be cautious. First, most sites were highly dominated by common species that were relatively mobile; the
greatest component of Simpson diversity (i.e. species dominance) was attributed to individual sites, and species abundance distributions in fragments indicated that dominant
bats were frequently cave-roosting species. Second, species
Please cite this article in press as: Struebig, M.J. et al, Conservation value of forest fragments to Palaeotropical bats, Biol.
Conserv. (2008), doi:10.1016/j.biocon.2008.06.009
ARTICLE IN PRESS
10
B I O L O G I C A L C O N S E RVAT I O N
predicted to be vagile contributed most to the differences in
assemblage composition between small fragments. In particular, there was no nested pattern for edge/open space foraging bats, the nestedness of cave-roosting bats could not be
predicted by fragment area, and some small fragments were
seen to host edge/open space foraging species that were not
recorded elsewhere during the study (Hesperotenus blanfordi
and Scotophilus kuhlii). Third, our analyses have not fully accounted for the rare specialist bat species known to occur
in peninsular Malaysia, which are likely to be at a greater
risk of extinction than common generalists (Davies et al.,
2004). Although 11 of the 46 species we captured are IUCN
red-listed (Appendix 2), these were typically found in larger
fragments and continuous forest. Despite a large cumulative
sampling effort over the landscape, six species were only
found in continuous forest; three of these (Coelops robinsoni,
Harpiocephalus mordax and Murina rozendaali) are red-listed,
and one (Kerivoula krauensis) has only recently been described, is currently considered endemic to Krau Wildlife Reserve, and has not yet been assessed by the IUCN. Finally,
because of the recent history of fragmentation in some parts
of the Krau landscape, crowding effects are likely to be substantial over the short term (see Ewers and Didham, 2006).
Hence, many small and isolated fragments may still owe
an extinction debt (Tilman et al., 1994), at least for the few
cavity/foliage roosting species that currently persist in them,
and the long-term responses of bat assemblages to fragmentation in Malaysia may yet be realised.
This study represents one of the first to date of bats and
fragmentation in the Palaeotropics, with the vast majority of
previous tropical fragmentation research having been undertaken in the Neotropics. Comparisons between our results
and those of other studies of bats and fragmentation are
not only complicated by the fundamental differences of bat
assemblage composition between the Neo- and Palaeotropics,
but also by fragmentation history. Studies in naturally fragmented landscapes in the Neotropics suggest that fragmentation has had limited impact on bat assemblages, with species
richness and composition remaining similar between fragments and continuous forest (Montiel et al., 2006; Bernard
and Fenton, 2007). However, in historical examples of fragmentation, processes are likely to have selected for species
traits that confer resistance to habitat change (Balmford,
1996), and hence patterns in these landscapes do not necessarily predict those that arise from more rapid human-induced fragmentation. In some cases, subtle impacts on
assemblage structure have often been detected in these situations, and researchers have suggested characteristics that
may influence the resilience of bat species to fragmentation.
In Guatemala, for example, the most abundant bats in fragments were found to be typically large frugivores (Schulze
et al., 2000), while in French Guiana they were large frugivores
that were also canopy specialists (Cosson et al., 1999). In a
study of insectivorous bats in Australia, resilient species appeared to be fast flying, poorly manoeuvrable species (Law
et al., 1999)ßand in a detailed study of land-bridge islands in
Panama, edge-sensitivity was suggested to be the key influence on vulnerability (Meyer et al., 2008). However, single
traits are often poor predictors of species sensitivity to frag-
x x x ( 2 0 0 8 ) x x x –x x x
mentation, and profiles describing groups of traits that may
act synergistically may be more accurate (Davies et al., 2004;
Henle et al., 2004). In this regard, our study suggests that multiple traits correlated with roosting ecology (e.g. vagility, population size, foraging behaviour, see Altringham, 1999) might
have important roles in determining the differential responses of bat species to fragmentation.
4.1.
Conservation implications
The ability of bats to fly calls into question whether these animals are a poor model group to infer the impacts of land-use
change, or on which to base landscape management policies.
Despite this, few empirical studies have compared the responses of bats to land-use changes with different taxa within the same landscape. The exceptions, based exclusively in
the Neotropics, have demonstrated that different animal
groups vary in their response to land-use change, and that
the response of bat assemblages is not shared by other taxa,
which typically are more heavily affected (Pineda et al.,
2005; Barlow et al., 2007; Gardner et al., 2008). In fact, Amazonian bat assemblages were similar amongst secondary forests
and plantations, a response that exhibited the poorest congruence with other groups of vertebrates, invertebrates and
plants (Gardner et al., 2008), and which was related to their
high vagility (Barlow et al., 2007). However, this finding might
have arisen because analyses were conducted on all bat species at the assemblage level. Our study suggests that, in the
Palaeotropics at least, not all bat species are as mobile as
might be perceived, and that partitioning assemblage analyses based on foraging or roosting strategies may improve
our ability to detect responses to land-use changes. Studies
of bats and other animal groups in the same disturbed Palaeotropical landscapes are needed to elucidate how the responses of tree cavity/foliage roosting bats compare to those
of other taxa.
Our study supports the view that larger fragments contain
more species, with assemblages resembling those in ‘intact’
natural habitats. Therefore, conservation strategies in Palaeotropical landscapes should favour large areas of forest, at
least for conserving bat populations. Large fragments in the
Krau landscape are currently managed for timber production
as part of the Permanent Forest Estate, echoing trends elsewhere in Malaysia; they are therefore likely to remain in the
landscape given their size and economic value. Small and
medium sized fragments, however, are typically afforded
low conservation and economic status and their long-term
fate rests in the hands of plantation managers, local landowners and the state government. Our study suggests that
fragments > 300 ha can support considerable bat diversity. In
addition, although small fragments do not appear to support
the rare, specialist species of most conservation concern, they
contribute substantially to landscape-level bat diversity, and
may also facilitate the movements of some species across
managed landscapes. By comparison, a preliminary study of
bat diversity in rubber and oil palm plantations suggests that
these plantation types host a much more depauperate bat
fauna compared to forest (Danielsen and Heegaard, 1995).
With demonstrated predation impacts on arthropod popula-
Please cite this article in press as: Struebig, M.J. et al, Conservation value of forest fragments to Palaeotropical bats, Biol.
Conserv. (2008), doi:10.1016/j.biocon.2008.06.009
ARTICLE IN PRESS
B I O L O G I CA L C O N S E RVAT I O N
tions in agricultural areas (Williams-Guillén et al., 2008),
insectivorous bats have an ecosystem value that could benefit
plantation managers. Hence, protecting large tracts of forest,
while retaining some forest fragments in plantations should
form an integral part of landscape planning, and has the potential to both benefit plantation management and bat
conservation.
Acknowledgements
We are grateful to Christoph Meyer and an anonymous reviewer for critical suggestions that greatly improved the
manuscript. Thanks to the Economic Planning Unit of the
Malaysian Government for granting us permission to conduct bat research in Malaysia, and the Malaysian Department of Wildlife and National Parks (DWNP), the Pahang
11
x x x ( 2 0 0 8 ) x x x –x x x
State Forestry Department, the Federal Land Development
Authority (FELDA), and numerous private landowners for
allowing us access to research sites. Thanks also to Paul
_
Banks, Monika Bozek,
Christine Fletcher, Joanne Kelly, LeeSim Lim, Juliana Senawi, Rakhmad Sujarno Kudus, Anthony
Turner and Zamiza Zainal for assistance with fieldwork, and
to Richard Nichols and Philippa Lincoln for statistical advice.
Research in fragments was funded by a PhD studentship
awarded to MJS from the Natural Environment Research
Council UK, and a grant from Bat Conservation International/US Forest Service. Research in Krau Wildlife Reserve
was supported by grants to TK from Lubee Bat Conservancy,
National Science Foundation (NSF # 0108384, DEB & East Asia
and Pacific Program), Earthwatch Institute, and National
Geographic (Committee for Research & Exploration; Conservation Trust).
Appendix 1
Fragment and continuous forest sites visited in the Krau landscape, peninsular Malaysia, between May 2002 and June 2007, with
a summary of bat survey results
Site name
Surrounding
land-usea
Area (ha)
Isolation
(km)b
Nearest
forest (km)c
A, O
A, O
O, R
O
A
R, O, G
O
O, R, G
R, O, G
R, O, G
O, R
C, O
O
O, G
R, G
C, G, R
R, G, O
R, G
R, G
O
O, R
A, P
O, R
O, R
O, R
O, R
R
F
F
F
F
2883
11 339
100
1838
551
443
1356
353
31
319
122
161
44
400
160
93
32
115
100
300
5225
107
5581
2025
32
35
3
137 000
137 000
137 000
137 000
7.7
7.5
7.4
2.5
18.1
3.7
12.3
2.1
5.0
8.1
8.8
6.9
3.0
7.4
4.6
5.5
6.3
6.6
13.4
14.7
11.0
5.6
5.5
7.6
4.6
3.6
10.8
–
–
–
–
0.6
1.2
2.3
1.9
1.2
1.4
0.6
1.0
0.6
0.6
1.7
0.4
1.3
2.3
1.8
0.6
0.5
0.3
0.7
0.7
1.8
2.1
0.7
0.6
1.2
1.1
0.6
–
–
–
–
Trap nights
Nd
Sobse
F, Fragment; S, Continuous forest (KWR)
F01
F02
F03
F04h
F05
F06
F07g
F08
F09
F10
F11
F12h
F13f
F14h
F15
F16
F17
F18
F19f,h
F20g
F21
F22
F23h
F24
F25
F26
F27
S01
S02
S03
S04
Kemasul 1: 3°23 0 N, 102°11 0 E
Kemasul 2: 3°26 0 N, 102°08 0 E
RTP Lembah Klau: 3°42 0 N, 101°58 0 E
FELDA Jenderak: 3°37 0 N, 102°19 0 E
Bukit Besar: 3°22 0 N,102°15 0 E
Klau Kecil: 3°47 0 N, 101°53 0 E
Gunung Senyum: 3°41 0 N, 102°27 0 E
Paya Luas: 3°42 0 N, 102°19 0 E
Paya Parit: 3°41 0 N, 102°23 0 E
Hutan Kerdau: 3°39 0 N, 102°25 0 E
Ulu Rugan: 3°36 0 N, 102°20 0 E
Dato’ Shariff: 3°40 0 N, 102°23 0 E
Kampung Gun: 3°33 0 N, 101°58 0 E
Kampung Lebu: 3°38 0 N, 101°56 0 E
Rumpun Makmur: 3°43 0 N, 102°23 0 E
Tebing Tinggi: 3°51 0 N, 102°23 0 E
Bukit Dinding: 3°49 0 N, 102°24 0 E
Bukit Ketupat: 3°48 0 N, 102°24 0 E
Paya Perak: 3°36 0 N, 102°26 0 E
Batu Sawar: 3°39 0 N, 102°28 0 E
Belungu: 3°44 0 N, 102°33 0 E
Desa Bakti: 3°48 0 N, 102°28 0 E
Klau Besar: 3°75 0 N, 101°89 0 E
Jengka: 3°59 0 N, 102°47 0 E
Jambu Rias: 3°45 0 N, 102°10 0 E
Karak: 3°41 0 N, 102°05 0 E
Tasek Chatin: 3°47 0 N, 102°35 0 E
Kuala Lompat: 3°43 0 N, 102°17 0 E
Lubuk Baung: 3°43 0 N, 102°13 0 E
Kuala Serloh: 3°40 0 N, 102°10 0 E
Kuala Gandah: 3°36 0 N, 102°09 0 E
28
137
40
220
27
90
27
97
14
18
38
120
15
483
16
44
15
43
29
93
15
23
14
30
9
11
21
85
14
219
16
33
22
75
17
42
14
35
13
202
22
107
16
43
17
358
19
105
16
20
8
122
5
2
330
998
356
1491
355
614
356
1194
(continued on next
16
19
12
14
5
20
15
11
11
16
5
10
5
17
20
10
19
13
9
8
16
12
19
13
7
12
2
29
28
30
25
page)
Please cite this article in press as: Struebig, M.J. et al, Conservation value of forest fragments to Palaeotropical bats, Biol.
Conserv. (2008), doi:10.1016/j.biocon.2008.06.009
ARTICLE IN PRESS
12
B I O L O G I C A L C O N S E RVAT I O N
x x x ( 2 0 0 8 ) x x x –x x x
Appendix 1 (continued)
Site name
Surrounding
land-usea
Area (ha)
Isolation
(km)b
Nearest
forest (km)c
Trap nights
Nd
F
F, G
F, G
137 000
137 000
137 000
–
–
–
–
–
–
370
39
24
1766
821
604
Sobse
F, Fragment; S, Continuous forest (KWR)
S05
S06
S07
Jenderak Selatan: 3°38 0 N, 102°17 0 E
Lembah Klau: 3°42 0 N, 102°03 0 E
Perlok: 3°49 0 N, 102°13 0 E
27
15
17
a Land-use surrounding the site in order of increasing area. A, Acacia plantation; O, oil palm plantation; P, pine plantation; R, rubber plantation;
C, cleared land; F, forest; G, mixed gardens.
b The nearest straight-line distance to continuous forest.
c The straight-line distance to the nearest forest fragment.
d Total number of individuals captured at a site, including recaptures from other sites, but excluding those within a site.
e Observed species richness for all insectivorous bat species.
f Sites in which surveys were influenced by heavy rain.
g The Gunung Senyum fragment (F07) contains an outcrop of karst limestone with very large abundances of cave-roosting bats, which skewed
analyses for both this fragment and the nearest neighbour Batu Sawar (F20). Hence, these fragments were excluded from subsequent
analyses on these grounds.
h Fragments of reduced area since 2002 due to recent or current forest clearance. Area estimates are corrected based on observations on the
ground and inspection of local forest maps if available.
Appendix 2
Bat species surveyed in the Krau landscape during this study
FAMILY/Taxon
Species
code
Red list
statusa
Landscape
distributionb
Ensemblec
No. continuous
sites occupied(Nmax = 7)d
No. fragments
occupied(Nmax= 27)d
MEGADERMATIDAE
Megaderma spasma
Msp
R
T
5
1
NYCTERIDAE
Nycteris tragata
Ntr
R
T
7
7
EMBALLONURIDAE
Emballonura monticola
Emo
R
E
5
2
RHINOLOPHIDAE
Rhinolophus affinis
Rhinolophus lepiduse
Rhinolophus luctus
Rhinolophus macrotis
Rhinolophus robinsoni
Rhinolophus sedulus
Rhinolophus stheno
Rhinolophus trifoliatus
Raf
Rle
Rlu
Rma
Rro
Rse
Rst
Rtr
W
W
R
R
R
C
C
T
T
C
T
C
T
7
7
3
1
4
5
7
6
22
22
6
2
2
13
17
23
HIPPOSIDERIDAE
Coelops robinsoni
Hipposideros armiger
Hipposideros bicolor 131f
Hipposideros bicolor 142f
Hipposideros cervinus
Hipposideros cineraceus
Hipposideros diadema
Hipposideros doriaee
Hipposideros galeritus
Cro
Har
Hb31
Hb42
Hce
Hci
Hdi
Hdo
Hga
T
C
C
C
C
C
C
T
C
4
1
7
7
7
4
7
4
1
0
0
18
13
19
5
17
5
3
NT
W
NT
A
A
R
NT
R
R
Please cite this article in press as: Struebig, M.J. et al, Conservation value of forest fragments to Palaeotropical bats, Biol.
Conserv. (2008), doi:10.1016/j.biocon.2008.06.009
ARTICLE IN PRESS
B I O L O G I CA L C O N S E RVAT I O N
13
x x x ( 2 0 0 8 ) x x x –x x x
Appendix 2 (continued)
FAMILY/Taxon
Hipposideros larvatus
Hipposideros lylei
Hipposideros ridleyi
VESPERTILIONIDAE
Glischropus tylopus
Harpiocephalus mordax
Hesperoptenus blanfordi
Kerivoula hardwickii
Kerivoula intermedia
Kerivoula krauensise
Kerivoula minuta
Kerivoula papillosa
Kerivoula pellucida
Miniopterus medius/schreibersiig
Murina aenea
Murina cyclotis
Murina rozendaali
Murina suilla
Myotis ater
Myotis horsefieldi
Myotis ridleyi
Myotis siligorensis
Phoniscus atrox
Phoniscus jagorii
Scotophilus kuhlii
Tylopus pachypus
Tylopus robustula
Species Red list Landscape Ensemblec
No. continuous
No. fragments
code
statusa distributionb
sites occupied(Nmax = 7)d occupied(Nmax= 27)d
Hla
Hly
Hri
Gty
Hmo
Hbl
Kha
Kin
Kkr
Kmi
Kpa
Kpe
Msc
Mae
Mcy
Mro
Msu
Mat
Mho
Mri
Msi
Pat
Pja
Sku
Tpa
Tro
NT
VU
F
VU
A
F
R
NT
nc
NT
NT
VU
VU
NT
A
R
A
W
R
A
R
R
F
R
R
C
C
T
3
0
5
13
1
7
E
T
E
T
T
T
T
T
T
E
T
T
T
T
E
E
E
E
T
T
E
E
E
0
2
0
2
7
4
1
7
7
0
3
7
4
6
1
3
5
1
6
1
0
1
1
8
0
1
1
13
0
1
19
15
6
1
4
0
23
8
1
4
0
8
2
1
2
0
a IUCN red list status following review by the Southeast Asian Mammal Databank (http://www.ieaitaly.org/samd/. Accessed 31 March 2008): NT,
Near Threatened; VU, Vulnerable; nc, not yet classified.
b Landscape distribution of species based on presence and abundance at sites in continuous and fragmented forest. W, widespread, present in
at least 70% of fragments, and all continuous forest sites; R, rare, comprise < 1% of combined captures from all sites; A, absent from
fragments; and F, absent from continuous forest.
c T, cavity/foliage-roosting narrow-space species; C, cave-roosting narrow-space species; E, edge/open space foraging species.
d The number of sites occupied does not control for the different trapping effort applied to each site. Therefore absence at a site does not
necessarily reflect true absence.
e Kingston et al. (2003) included R. lepidus as R. refulgens, H. doriae as H. sabanus, and K. krauensis as K.sp. We follow the updated nomenclature in:
Simmons (2005) Chiroptera, In: Wilson DE, Reeder DM (eds) Mammal species of the World: a taxonomic and geographic reference. John
Hopkins University Press, Baltimore, pp 312-529, and Francis, C.M., Kingston, T., Zubaid, A., 2007. A new species of Kerivoula (Chiroptera:
Vespertilionidae) from peninsular Malaysia. Acta Chiropterologica 9, 1–12.
f Hipposideros ‘bicolor’ comprises two phonic types with mean echolocation call frequencies of 131 kHz and 142 kHz. They differ genetically
enough to warrant treatment as separate taxa and can be distinguished in the hand by subtle differences in noseleaf shape, forearm length
and tibia length (Kingston et al., 2006).
g Miniopterus individuals captured could not be reliably identified by external measurements to medius or schreibersii. The red list status is Least
Concern for M. medius, and Near Threatened for M. schreibersii.
R E F E R E N C E S
Allen, D. 2005. Foraging ecology of Kerivoula papillosa. MSc thesis,
University of East Anglia, Norwich.
Altringham, J.D., 1999. Bats: Biology and Behaviour. Oxford
University Press, Oxford.
Baayan, R.H., 2008. The languageR package for R: data sets and
functions with ‘‘Analyzing linguistic data: A practical
introduction to statistics’’, version 0.92, <http://cran.rproject.org/web/packages/languageR/index.html>.
Balmford, A., 1996. Extinction filters and current resilience: the
significance of past selection pressures for conservation
biology. Trends in Ecology and Evolution 11, 193–196.
Barclay, R.M.R., Harder, L.D., 2003. Life histories of bats: life in the
slow lane. In: Kunz, T.H., Brock-Fenton, M. (Eds.), Bat Ecology.
University of Chicago Press, Chicago, pp. 209–256.
Barlow, J., Gardner, T.A., Araujo, I.S., Avila-Pires, T.C., Bonaldo,
A.B., Costa, J.E., Esposito, M.C., Ferreira, L.V., Hawes, M.I.M.,
Hernandez, M.I.M., Hoogmoed, M.S., Leite, R.N., Lo-Man-Hung,
N.F., Malcolm, J.R., Martins, M.B., Mestre, L.A.M., MirandaSontos, R., Nunes-Gutjahr, A.L., Overal, W.L., Parry, L., Peters,
Please cite this article in press as: Struebig, M.J. et al, Conservation value of forest fragments to Palaeotropical bats, Biol.
Conserv. (2008), doi:10.1016/j.biocon.2008.06.009
ARTICLE IN PRESS
14
B I O L O G I C A L C O N S E RVAT I O N
S.L., Ribeiro-Junior, M.A., da Silva, M.N.F., da Silva Motta, C.,
Peres, C.A., 2007. Quantifying the biodiversity value of tropical
primary, secondary and plantation forest. Proceedings of the
National Academy of Sciences 104, 18555–18560.
Bates, D., 2008. The lmer package for R: linear mixed-effects
models using S4 classes, version 0.99875-9, http://cran.rproject.org/web/packages/lme4/index.html.
Benedick, S., Hill, J.K., Mustaffa, N., Chey, V.K., Maryati, M., Searle,
J.B., Schilthuizen, M., Hamer, K.C., 2006. Impacts of rain forest
fragmentation on butterflies in northern Borneo: species
richness, turnover and the value of small fragments. Journal of
Applied Ecology 43, 967–977.
Bernard, E., Fenton, M.B., 2007. Bats in a fragmented landscape:
species composition, diversity and habitat interactions in
savannas of Santarém, Central Amazonia, Brazil. Biological
Conservation 134, 332–343.
Bontadina, F., Schofield, H., Naef-Daenzer, B., 2002. Radio-tracking
reveals that lesser horseshoe bats (Rhinolophus hipposideros)
forage in woodland. Journal of Zoology 258, 281–290.
Brühl, C.A., Eltz, T., Linsenmair, K., 2003. Size does matter – effects
of tropical rainforest fragmentation on the leaf litter ant
community in Sabah, Malaysia. Biodiversity and Conservation
12, 1371–1389.
Chao, A., Shen, T.J., 2003-2005. Program SPADE (Species Prediction
And Diversity Estimation). Program and user’s guide published
at <http://chao.stat.nthu.edu.tw>.
Colwell, R.K., 2004. EstimateS: statistical estimation of species
richness and shared species from samples. Version 7.0.
Available at <http://purl.oclc.org/estimates> (accessed July
2007).
Cosson, J.F., Pons, J.M., Masson, D., 1999. Effects of forest
fragmentation on frugivorous and nectarivorous bats in
French Guiana. Journal of Tropical Ecology 15, 515–534.
Crist, T.O., Veech, J.A., Gering, J.C., Summerville, K.S., 2003.
Partitioning species diversity across landscapes and regions: a
hierarchical analysis of a, b and c diversity. American
Naturalist 162, 734–743.
Danielsen, F., Heegaard, M., 1995. Impact of logging and
plantation development on species diversity: a case study
from Sumatra. In: Sandbukt, Ø. (Ed.), Management of Tropical
Forests: Towards an Integrated Perspective. University of Oslo,
Oslo, pp. 73–92.
DAPM, 2005. Industrial crops and herbs statistics, Malaysia 2003,
second ed. Department of Agriculture Peninsular Malaysia,
Putrajaya.
Davies, K.F., Margules, C.R., Lawrence, J.F., 2004. A synergistic
effect puts rare, specialized species at greater risk of
extinction. Ecology 85, 265–271.
Estrada, A., Coates-Estrada, R., 2002. Bats in continuous forest,
forest fragments and in an agricultural mosaic habitat-island
at Los Tuxtlas, Mexico. Biological Conservation 103, 237–245.
Ewers, R.M., Didham, R.K., 2006. Confounding factors in the
detection of species responses to habitat fragmentation.
Biological Reviews 81, 117–142.
Faria, D., 2006. Phyllostomid bats of a fragmented landscape in
the north-eastern Atlantic forest, Brazil. Journal of Tropical
Ecology 22, 531–542.
Findley, J.S., 1993. Bats: a community perspective. Cambridge
University Press, Cambridge.
Fletcher, C.D. 2006. Roosting ecology of insectivorous bats in Krau
Wildlife Reserve, Pahang, Peninsular Malaysia. Ph.D. Thesis,
Universiti Kebangsaan Malaysia, Bangi.
Gardner, T.A., Barlow, J., Araujo, I.S., Avila-Pires, T.C., Bonaldo,
A.B., Costa, J.E., Esposito, M.C., Ferreira, L.V., Hawes, M.I.M.,
Hernandez, M.I.M., Hoogmoed, M.S., Leite, R.N., Lo-Man-Hung,
N.F., Malcolm, J.R., Martins, M.B., Mestre, L.A.M., MirandaSontos, R., Overal, W.L., Parry, L., Peters, S.L., Ribeiro-Junior,
M.A., da Silva, M.N.F., da Silva Motta, C., Peres, C.A., 2008. The
x x x ( 2 0 0 8 ) x x x –x x x
cost effectiveness of biodiversity surveys in tropical forests.
Ecology Letters 11, 139–150.
Gorresen, P.M., Willig, M.R., 2004. Landscape responses of bats to
habitat fragmentation in Atlantic forest of Paraguay. Journal of
Mammalogy 85, 688–697.
Gotelli, N.J., Entsminger, G.L. 2007. EcoSim: null models software
for ecology. Version 7. Acquired Intelligence Inc. & Kesey-Bear.
Jericho, <http://garyentsminger.com/ecosim.htm>.
Henle, K., Davies, K.F., Kleyer, M., Margules, C., Settele, J., 2004.
Predictors of species sensitivity to fragmentation. Biodiversity
and Conservation 13, 207–251.
Kathirithamby-Wells, J., 2005. Nature and nations: forests and
development in peninsular Malaysia. Nias Press, Copenhagen.
Kingston, T., Francis, C.M., Zubaid, A., Kunz, T.H., 2003. Species
richness in an insectivorous bat assemblage from Malaysia.
Journal of Tropical Ecology 19, 67–79.
Kingston, T., Boo Liat, L., Zubaid, A., 2006. Bats of Krau Wildlife
Reserve. Universiti Kebangsaan Malaysia Press, Bangi.
Lane, D.J.W., Kingston, T., Lee, B.P.Y-H., 2006. Dramatic decline in
bat species richness in Singapore, with implications for
Southeast Asia. Biological Conservation 131, 584–593.
Laurance, W.F., Lovejoy, T.E., Vasconcelos, H.L., Bruna, E.M.,
Didham, R.K., Stouffer, P.C., Gascon, C., Bierregaard, R.O.,
Laurance, S.G., Sampaio, E., 2002. Ecosystem decay of
Amazonian forest fragments: a 22-year investigation.
Conservation Biology 16, 605–618.
Law, B.S., Anderson, J., Chidel, M., 1999. Bat communities in a
fragmented forest landscape on the south-west slopes of New
South Wales, Australia. Biological Conservation 88, 333–345.
Lynam, A.J., Billick, I., 1999. Differential responses of small
mammals to fragmentation in a Thailand tropical forest.
Biological Conservation 91, 191–200.
MacArthur, R.H., Wilson, E.O., 1967. The Theory of Island
Biogeography. Princeton University Press, Princeton.
McCune, B., Grace, J.B., 2002. Analysis of Ecological Communities.
MJM Software Design, Oregon.
McCune, B., Mefford, M.J., 2006. Multivariate analysis of ecological
data. PCORD version 5. MJM Software Design, Oregon.
Meijaard, E., Sheil, D., 2007. Is wildlife research useful for
wildlife conservation in the tropics? A review for Borneo with
global implications. Biodiversity and Conservation 16, 3053–
3065.
Meyer, C.F.J., Kalko, K.M.V., 2008. Bat assemblages on Neotropical
land-bridge islands: nested subsets and null model analyses of
species co-occurrence patterns. Diversity and Distributions.
doi:10.1111/j.1472-4642.2007.00462.x.
Meyer, C.F.J., Frund, J., Pineda Lizano, W., Kalko, K.M.V., 2008.
Ecological correlates of vulnerability to fragmentation in
neotropical bats. Journal of Applied Ecology 45, 381–391.
Mickleburgh, S.P., Hutson, A.M., Racey, P.A., 2002. A review of the
global conservation status of bats. Oryx 36, 18–34.
Miller-Butterworth, C.M., Jacobs, D.S., Harley, E.H., 2003. Strong
population substructure is correlated with morphology and
ecology in a migratory bat. Nature 424, 187–191.
Montiel, S., Estrada, A., Leon, P., 2006. Bat assemblages in a
naturally fragmented ecosystem in the Yucatan Peninsula,
Mexico: species richness, diversity and spatio-temporal
dynamics. Journal of Tropical Ecology 22, 267–276.
Norberg, U.M., Rayner, J.M.V., 1987. Ecological morphology and
flight in bats (Mammalia; Chiroptera): wing adaptations, flight
performance, foraging strategy and echolocation.
Philosophical Transactions of the Royal Society of London
Series B Biological Sciences 316, 335–427.
Pattanavibool, A., Dearden, P., 2002. Fragmentation and wildlife in
montane evergreen forests, northern Thailand. Biological
Conservation 107, 155–164.
Pavey, C.R., Grunwald, J-E., Neuweiler, G., 2001. Foraging habitat
and echolocation behaviour of Scheider’s leafnosed bat,
Please cite this article in press as: Struebig, M.J. et al, Conservation value of forest fragments to Palaeotropical bats, Biol.
Conserv. (2008), doi:10.1016/j.biocon.2008.06.009
ARTICLE IN PRESS
B I O L O G I CA L C O N S E RVAT I O N
Hipposideros speoris, in a vegetation mosaic of Sri Lanka.
Behavioral Ecology and Sociobiology 50, 209–218.
Pineda, E., Moreno, C., Escobar, F., Halffter, G., 2005. Frog, bat and
dung beetle diversity in the cloud forest and coffee
agrostsystems of Veracruz, Mexico. Conservation Biology 19,
400–410.
Ricketts, T.H., 2001. The matrix matters: effective isolation in
fragmented landscapes. American Naturalist 158, 87–99.
Roback, P.J., Askins, R.A., 2005. Judicious use of multiple
hypothesis tests. Conservation Biology 19, 261–267.
Rodrı́guez-Gironés, M.A., Santamarı́a, L., 2006. A new algorithm to
calculate the nestedness temperature of presence–absence
matrices. Journal of Biogeography 33, 924–935.
Schnitzler, H.-U., Kalko, K.M.V., 2001. Echolocation by insecteating bats. BioScience 51, 557–569.
Schulze, M.D., Seavy, N.E., Whitacre, D.F., 2000. A comparison of
the phyllostomid bat assemblages in undisturbed neotropical
forest and in forest fragments of a slash-and-burn farming
mosaic in Peten, Guatemala. Biotropica 32, 174–184.
Shen, T.J., Chao, A., Lin, J.F., 2003. Predicting the number of new
species in further taxonomic sampling. Ecology 84, 798–804.
Simmons, N., 2005. Chiroptera. In: Wilson, D.E., Reeder, D.M.
(Eds.), Mammal species of the World: a taxonomic and
geographic reference, third ed. John Hopkins University Press,
Baltimore, pp. 312–529.
Sodhi, N.S., Brook, B.W., Bradshaw, C.J.A., 2007. Tropical
Conservation Biology. Blackwell Publishing, Oxford.
x x x ( 2 0 0 8 ) x x x –x x x
15
Sujarno-Kudus, R., 2006. Temporal variation of insectivorous bat
assemblages at Krau, Malaysia. M.Sc. thesis, Universiti
Kebangsaan Malaysia, Bangi.
Tilman, D., May, R.M., Lehman, C.L., Nowak, M.A., 1994. Habitat
destruction and the extinction debt. Nature 371, 65–66.
Veech, J.A., Crist, T.O., 2007. PARTITION: software for hierarchical
additive partitioning of species diversity, Version 2.0 <http://
www.users.muohio.edu/cristto/partition.htm>.
Watling, J.I., Donnelly, M.A., 2006. Fragments as islands: a
synthesis of faunal responses to habitat patchiness.
Conservation Biology 20, 1016–1025.
Whitmore, T.C., 1997. Rainforest disturbance, disappearance and
species loss. In: Laurance, W.F., Bierregaard, R.O. (Eds.),
Tropical Forest Remnants: Ecology, Management and
Conservation of Fragmented Communities. University of
Chicago Press, Chicago, pp. 3–12.
Williams-Guillén, K., Perfecto, I., Vandemeer, J., 2008. Bats limit
insects in a neotropical agroforestry system. Science 320, 70.
Wilson, D.E., Reeder, D.M., 2005. Mammal species of the World: a
taxonomic and geographic reference, third ed. John Hopkins
University Press, Baltimore.
Wright, D.H., Patterson, B.D., Mikkelson, G.M., Cutler, A., Atmar,
W., 1998. A comparative analysis of nested subset patterns of
species composition. Oecologia 113, 1–20.
Yusof, E., Sorenson, K.W., 2000. Krau Wildlife Reserve: protected
area management experiences. Journal of Wildlife and Parks
18, 3–13.
Please cite this article in press as: Struebig, M.J. et al, Conservation value of forest fragments to Palaeotropical bats, Biol.
Conserv. (2008), doi:10.1016/j.biocon.2008.06.009