Effects of selective logging on the microhabitat

Original article
Effects
of
selective
of
logging
non-volant
tropical
on
small
lowland
the
microhabitat-use
mammals
in
a Bornean
mixed-dipterocarp
Henry
patterns
forest
Bernard
Institute for Tropical Biology and Conservation, (JniversitiMalaysia
Locked
Bag No. 2073, 88999
Sabah,
Kota Kinabalii, Sabah, Malaysia
Abstract
To
assess the effects of habitat disturbance caused by selective logging, the microhabitat-use patterns of six
understorey species of small (<1 kg), non-volant mammals
in unlogged forests were compared
forests that had been selectively logged 13 and 25 years previously. The
with those in adjacent
study was conducted
in a lowland mixed-
dipterocarp forest at Tab inWildlife Reserve, located in eastern Sabah, Malaysian Borneo. Twenty-three
microhabitat
variables mainly related to vegetation structure, but also including other habitat features, were used to characterise
trap-stations set along twenty 200-m
lines per site).The
logged-un
long trap-lines equally distributed among
trapping-sites represented two primary and two logged forests.These
logged forest treatments located at two independent
analysis revealed significant separation between
appeared
logged
were arranged as pairs of
sampling locations. Canonical
discriminant function
trap-slations both occupied and non-occupied
the primary and logged forest pair sites at both locations. However,
small mammals
four trapping-sites (i.e.,five trap-
microhabitat-use patterns in primary compared
by small mammals
to logged forests.In general, the small mammals
to be able to utilisedifferent sets of microhabitats in different habitat types. Captures
in both primary and
forests were generally positively associated with higher density of fallen logs, twigs and rock piles,number
tree stumps and vegetation cover at the understorey and canopy
seemed
to suggest that habitat utilisationby small mammals
in
there was no consistent trend observed in the
levels.Wet
sitestend to be avoided. These
of
variables
is closely linked to food or foraging areas and shelter
or refuge sites,with probably a strong influence for predator avoidance. Tabin
small mammals
are able to persist
following the firstcut in logged forest indicating that although in general, logged forest cannot replace primary forest,
selectively logged forests may
stillconstitute an important component
for the preservation of small mammal
species
diversity.
Key
words: Small mammals,
microhabitat-use patterns, effects of selective logging, North Borneo
Introduction
of small, isolated primary
of developed
Owing
to
anthropogenic
the
extensive
and
timber extraction for export, many
east Asia are now
running
forest (Danielsen
and
1997). A
rapid
pace
of
forest change, particularly as a result of
large amount
Heegard,
of
countries in South-
out of sites with primary
1995;
Whitmore,
the remaining
tropical
forest landscape in this region occurs as a patchwork
1997; Whitmore,
many
forest remnants
in a "sea"
land (Corlett and Turner, 1997; Lynam,
1997; Laidlaw,
2000). Although
studies have generally documented
of derived habitats as compared
the inferiority
to pristine forests for
sustaining biological diversity, recent studies have
indicated that derived habitats may
role. Taking
stillhave a valuable
into consideration the level of damage
and regeneration time since the lastlogging, selectively
logged
forestsin the tropics have been found to support
has been selectivelylogged and may, therefore, to some
quite a significantproportion of original forest faunas,
extent be stilldominated
especially when
This study formed
stillremain
(Holloway
large tracts of undisturbed
forests
within or close to the logged-over
et al.,1992; Lambert, 1992; Dahaban
1993; Daniel sen and Heegard,
1996; Laurance
examined
et al.,
terrestrialsmall mammals
and Bierregaard, 1997; Bernard
and
the responses of North-Bornean
establishment
family.
part of a research project that
areas
1995; Grieser Johns,
by the dipterocarp
non-volant
to selective logging and the
of large agricultural crop plantations
(Bernard, in prep.).
Diun, 1999; Robinson and Robinson, 1999; Vasconcelos
et al.,2000; Willott et al, 2000; Laidlaw, 2000). In
general, although
Materials
and
methods
not equal to the original forests
in terms of species composition,
usually high in moderately
species richness is
Study
disturbed forest,such as in
area
Tabin Wildlife
Reserve
(5 °15'-5°10'N, 118
°
selectively logged forests, however, specialist species
30' ―118°45'E) is near-rectangular in shape, measures
tend to decline (Bernard, in prep.).
120,521 ha and is covered mainly with lowland mixed-
Since derived
habitats increasingly cover
larger areas, and primary
undergoing
reduction and fragmentation,
that in the future the survival of many
faunas must depend
more
much
forests are correspondingly
it appears
tropical forest
heavily on the management
dipterocarp
Reserve
forest. The
comprises
water swamp
forest and a small amount
Topography
undulating
with some
between
includes some
peak is 570 m. Mean
habitats. In
this respect, it is
mm
species are able to adapt to significant changes
and May-June.
requirements
Knowledge
in
of the habitat
of various groups of wildlife is important
of mangroves.
steep hills. Elevations
100 to 300 m
pivotal to address the issue of whether tropical forest
their natural environment.
above
Mean
October-February
daily minimum
and maximum
are 22 °C and 32 °C, respectively. Many
parts of the Reserve
were
selectively logged
the 1960's to 1980's. When
conservation efforts to be concentrated where suitable
ceased in 1989, only the central part, known
habitats are present.
"Core
Compared
to regions
like the Neo-tropics
and
papers concerning the effects of human-induced
habitat
disturbance on various groups of tropical forest fauna
from
Southeast Asia (Cuaron, 2000; Sodhi
and
Liow,
Area"
of "Virgin
highest
annual precipitation is 3,100
and usually occurs between
temperatures
range
sea level. The
to their conservation in altered environments, to enable
Australian tropics, there are generally few published
brackish
in most parts of the Reserve is moderately
of derived habitats, ideally on a landscape scale that
primary
north-eastern corner of the
riverine forest, primary
(8,616 ha) and
Jungle
during
the last logging activities
seven
Reserves"
small
as the
patches
(totalling 2,018
ha)
contiguous with Tabin remain undisturbed (Sale, 1994).
As
a result of this treatment, the landscape
is composed
of a mosaic
at Tabin
of different-aged secondary
forest habitats (ranging
from
11 to 40 years after
2000). In this paper I present the results of a study
logging) patched with isolated primary forest remnants.
on the effects of tropical rain forest disturbance, in
Information
the form of selective logging, on the microhabitat-use
described
patterns of small (< 1 kg) non-volant mammal
species.
surrounding the Reserve has been vigorously converted
mixed-
to agricultural crop plantations since the early 1960's.
dipterocarp forest of Tabin Wildlife Reserve
located
At present Tabin is almost entirely surrounded by large
in the Malaysian
Borneo
oil palm
The study was carried out in a tropicallowland
State of Sabah, Northeast
(Figure 1). Specifically, I describe the microhabitat-use
of the vegetation structure in Tabin
elsewhere
(Bernard,
plantations. The
in prep.). The
is
land
age of the oil palms range
from recent plantings to more than twenty years.
patterns of the non-volant, understorey, small mammal
community
in relatively undisturbed
forests and in
forests that had been selectively logged 13 and 25 years
previously. This
was carried out by comparing
microhabitat configurations between
occupied
and non-occupied
the
trap-stations both
by the small mammals
at
both forest types.
after the original forest
sampling
Sampling
was
designed
to provide
quantitative
estimates of several aspects of the small mammal
community
structure in three different forest types:
primary lowland
logged
The term "secondary forest" is used in this paper to
refer to forest that has grown
Mammal
mixed-dipterocarp
lowland mixed-dipterocarp
plantation. Owing
forest,selectively
forest and oil palm
to the extremely
captures in the oil palm
low
number
of
plantation, only data from the
Figure
1. Map
of Sabah (Malaysian
Borneo) in relation to Southeast Asia, and Tabin Wildlife Reserve in eastern Sabah, indicating
locations of trapping-sitesat Location 1 (Core Area) and Location 2 (Lipad).
firsttwo forest types and that relatesto small mammal
oil palm plantations(Figure 1). Owing
species most
separation,the two sampling locationswere presumed
frequently captured
are analysed
and
presented.
Sampling
to be independent of each other. Two
was
approximately
confined to two
20 km
locations, located
apart, within
the
Tab
inarea.
sites(
1 km
to the wide
trapping-
apart),arranged as coupled primary-
Area, central
logged forest treatments, were established at each
sampling location.Paired siteswere chosen to have
part of Tabin, while Location 2 was situated at Lipad
generally similar topography and elevation(100-250
on the western
m a.s.l).
The following abbreviationswere used; PF1
Location 1 was
situated at the Core
side of Tabin
near
to the border of
Table 1. Twenty-three
microhabitat variables recorded at each trap station at Tabin Wildlife Reserve. All
variables were assessed in a 2.5 m radius around each trap station.
* Rank scores:
0―Not evident;
1―Rare; 2―Occasional;3―Frequent; 4―Abundant.
and LF1 representing
logged
trapping sites in primary
forestsin Location 1; PF2 and LF2
and
representing
on cotton balls (Bernard, 2003). Similar
were
trapping sitesin primary and logged forestsin Location
period. Animals
2. LF1
was
toe clipping before release.
logged
in 1974. No
selectively logged in 1986 while LF2 was
it is quite evident from the remaining stumps
that the number
the study
caught were individually marked
by
detailed information is available
on the logging intensityof the logged forest treatments.
However,
bait mixtures
used at all trapping-sites throughout
of trees extracted at both logged
forests are comparable
Microhabitat
variables
Twenty-three
microhabitat
variables relating to
vegetation structure, and other habitat features, were
(i.e.,
ca 8-15 trees per ha) (per.
observ.). Selective logging in Sabah typically involves
used
to quantify the heterogeneity
(Table 1). Information
of trap-stations
on all microhabitat
variables
felling and extraction of large (usually > 120 cm girth)
was collected during the firsttrapping period of the
commercially
respective sampling
valuable tree species (Marsh
1992). Although
than 45%
and Greer,
only selected trees are felled, more
were
largely those identified in previous studies as
good
descriptors
by logging (Aiken and
Leigh, 1992)
Small mammals
were sampled
per
month
for five consecutive
nights
November
between
2000 in Location 1 and between
2000 in Location 2. Trapping
smaller
due
to access
October
-
February -
difficulty. At
live cage traps (17 by 30 by 10 cm)
were set at trap-stationslocated at 20 m
five 200 m
each
mature
that
intervalsalong
long trap-lines.Baits used were a mixture
of cut ripe local variety bananas,
coconut
of understorey
species (e.g. Kemper
and Bell, 1985;
1997). Changes
in
these variables are, therefore, expected to influence the
patterns of microhabitat usage and local distributionof
the small mammals.
efforts in Location
trapping-site the animals were live-trapped using baited
wire-mesh
small mammal
of microhabitats
Patterson, et al., 1990; Lynam,
days and
1 was
variables selected
of the residual stands (smaller trees < 90 cm
girth) are also killed or damaged
May
locations. The
pieces of roasted
kernel and peanut butter suspended
Statistical analysis
Prior to analysis all variables were
as necessary to correct for non-normal
to improve
methods
normality. The
were
continuous
used,
transformed
distribution or
following transformation
logarithmic
(base
10)
for
variables, square root for count variables
and arc sine for variables that were
recorded
as
Table
2. Standardised
canonical
discriminant
function
coefficients of 23 microhabitat variables with function 1 (Fl)
and function 2 (F2)
Figure
2. Scatter plot of four trapping-sites based
on 23
microhabitat variables analysed using canonical discriminant
function analysis [ PF1
= trap-stations in primary forest of
Location 1; LF1 = trap-stationsin logged forest of Location 1;
PF2
= trap-stationsin primary forest of Location 2; LF2 = trap-
stations in logged forest of Location 2].
percentages. Except
for variables measured
orders, a value of 1.0 or 0.5 was
with zero scores before making
It was assumed
added
in rank-
to variables
the transformations.
that the probability of a capture of an
individual at a particular trap-station is not influenced
by previous and/or following individual captures, i.e.,
(using CDF
mammal
analysis). In
species were treated as the grouping variable,
while all 23 microhabitat
No
significant differences
configurations between
found
conducted
for data collected during the initialphase
of this project showed
that captures are independent
events (Bernard, 2003).
To describe
the small
the
mammals
in
logged
forests, canonical
(CDF)
analysis was
primary
patterns
as compared
discriminant
performed.
Before
eliminate
highly
running
the
tested for
correlated
variables from subsequent analyses. Maximum
correlations did not exceeded
to
function
analysis, the discriminating variables were
multicollinearity to
of
variable
r = 0.80, therefore, all
variables were retained.
Separate
CDF
and
analyses
1
Location
occupied
trap stations were
were
carried
2. Occupied
out
and
for
non-
treated as the grouping
variable, while all 23 microhabitat variables gathered
at all trap-stations were
variables. Prior
microhabitat
to
treated as the discriminating
running
configurations
species at each
location were
the
analysis,
the
of the small mammal
analysed
separately
small mammal
= 92, P < 0.085). Likewise,
in the
microhabitat
microhabitat
species were
no significant differences
configurations
species were
= 111.095, df
found
between
in Location
small
2 (Wilks'
A = 0.463, x1 = 57.345, df = 69, P < 0.8411). With
these findings, at a particular sampling
habitat type, the microhabitat
location and
variables data at all
capture-stations of all species were
combined.
Thus,
in this study the microhabitat-use patterns of the small
mammals
at a particular sampling location and habitat
type refers to the combined
of allof the small mammal
Owing
Location
in the
in Location 1 (Wilks'A = 0.167, ^
mammal
microhabitat-use
at the
treated as the
discriminating variables.
events. Test
of capture locations
variables recorded
respective species capture stations were
captures at a particular trap-station are independent
of independence
this initialanalysis, small
microhabitat-use patterns
species captured.
to the creation of natural tree-fallgaps and
the unexpectedly
movements
high frequency
of wild elephant
in the forest during the study period, the
microhabitat
structures at/or within the immediate
areas of 34 (or 17%
both Location
It was presumed
1 and
of the total 200) trap-stations,at
Location
2, changed
that the small mammals
their preferences or ranging
markedly.
might shift
patterns in relation to
the affected trap-stations.Hence,
these stations were
Table 3. Captures and summary
statistics
for six most frequently trapped small mammal
species
in primary (PF) and logged (LF) forests at two sampling locations at Tabin Wildlife Reserve
(#) ―No.
a ―No.
of observations or capture events
of trapsmultiply by no. of nights traps were set
b ―Percent
excluded
from
all analyses mentioned
All statisticalanalyses were conducted
Windows
(Norusis, 1994). A
in this paper.
with SPSS
for
no. of observations divided by no. of trap-nights
more
impression
when
compared
to the
forests in Location 1.
At both locations, the primary forest sites differed
probability of P < 0.05
from
was considered significantin all analyses.
disturbed
the logged
forest sites with the former
having
the higher density of small (10 cm gbh) and very large
Results
trees (120 cm gbh), and more
leaf litter,
while the latter
were characterised by the presence of logging tracks,
Differences
in
microhabitat
structure
between
relative
microhabitat
contribution
variables
of
each
of
(with all other
the
greater number
forest sites had
a
of stumps and twigs on the forest floor.
variables
present) to the power
of the discriminant functions are
given in Table 2. On
average the four trapping-sites
differed by location (Figure 2; along the firstcanonical
discriminant
climbers, and vegetation cover at low
heights. Additionally, the logged
trapping-sites
The
denser woody
axis, Fl) and by forest type (along the
Overall captures
Data for the six most
mammals
(accounting
for
frequently caught
> 95%
of
small
the overall
captures) in the forest habitats of Tabin are presented in
second canonical discriminant axis, F2) (Table 2). Both
Table 3. A
discriminant functions (Fl and F2) explained 46%
individuals of all six species were recorded at Location
40%
and
of the total variation in the data, respectively,and
totalof 119 captures comprising 65 different
1. Maxomys
surifer constituted the largest number
were significant(Test of functions 1 through 3: Wilks'
of captures
(59, 50%
A= 0.128, x2 = 313.257, df = 69, P< 0.0001; Test of
(29, 46%). Overall trap success for Location 1 was 11.9
functions 2 through 3: Wilks'A=
(calculated as the number
0.316, %2 = 175.733,
df = 44, P < 0.0001). The number
classified was 82%
of cases correctly
(136 of 166).
nights). One
of captures made
per 100 trap
hundred and fifty-onecaptures comprising
67 different
Irrespective of forest type, the firstdiscriminant
of the total) and individuals
individuals
recorded at Location 2. M.
of
five species
surifer was again the most
function indicates that Location 1 had greater canopy
abundant
cover, higher
more
%
was
Overall trap success for Location 2 was 7.6%.
obvious
density
signs
characterised by
of
pig
very
of
palm
climber
damage.
patchy
greater coverage of bare ground
and
Location
ground
in some
2
forests in Location
2
(80 captures, 26 individuals),representing 53
of the total captures and 39 %
of total individuals.
cover; with
parts, while
others had a higher density of shrubs and small sized
trees.The
were
also had
Small
mammals
primary
compared
microhabitat-use
to logged forest
greater
vegetation cover at low heights, thus, giving a generally
A. Location 1 (Core Area)
patterns
in
Table
4. Percentage of variationexplained by the firsttwo canonical discriminant
functions for Location 1 (Core Area) and Location 2 (Lipad)
- F3=Function 1through Function3
- F3 = Function2 through Function3
P<0.0001 *P<0.05
Figure
3. Scatter plots of capture and non-capture
stations in primary and logged forest trapping-sites at two different sampling
locations i.e.,(a) Location 1 (Core Area); and (b) Location 2 (Lipad). The
plots were based on 23 microhabitat variables analysed
using canonical discriminant function analysis.
0PF1
= NON-CAPTURE
1PF1 = CAPTURE
STATIONS
STATIONS
0LF1
= NON-CAPTURE
1LF1
= CAPTURE
0PF2
= NON-CAPTURE
1PF2
= CAPTURE
0LF2
= NON-CAPTURE
1LF2
= CAPTURE
STATIONS
STATIONS
in primary forest at Location 2;
in primary forest at Location 2;
STATIONS
STATIONS
in logged forest at Location 1;
in logged forest at Location 1;
STATIONS
STATIONS
in primary forest at Location 1;
in primary forest at Location 1;
in logged forest at Location 2;
in logged forest at Location 2.
The total number of captures of all species
combined did not differsignificantly
between PF1 and
significantly different (Wilks' X= 0.177, % 2 = 118.492,
LF1 (Chi-squared goodness of fittest:=
functions correctly classified76%
1, P = 0.169).The
1.891,df =
number of occupied trap-stations
df = 69, P < 0.0001). The
resulting discriminant
stations to group classes. Three
were extracted. The
(63 of 83) of the trapdiscriminant functions
percent of totalvariance explained
was 55 (i.e.,
PF1 = 28; LF1 = 27), while unoccupied
trap stationswas 28 (i.e.,
PF1 = 17, LF1 = 11).The
by the firsttwo functions combined
microhabitat configurations between occupied and
4). The
non-occupied
used and avoided
trap-stations at PF1 and LF1 were
was 90.9%
(Table
microhabitat characteristicsof the trap-stations
by the small mammals
in PF1 and
Table 5. Standardised canonical discriminant function coefficientsof 23 microhabitat
variables with function 1 (Fl) and function 2 (F2) for Location 1 (Core Area), and
Location 2 (Lipad), respectively.
Values in bold (the firsteight ranks) highlight variables having
LF1
are best explained by function 2 (see Figure 3(a)).
Function
1 appeared to explain the general differences
in habitat structure between
LF1
PF1 (primary forest) and
(logged forest).
Based
primary
the largest contribution to the power
of totalvariance explained by the firsttwo discriminant
functions
combined
in both
and logged forests at Location 1 appeared
to
PF2
and LF2
vegetation at low level (Figure 3(a) and Table 5). Sites
3(b)).
with higher coverage of rock piles were also preferred,
4). The
B. Location 2 (Lipad)
At Location 2, based on function 2, preferred sites
mammals
of captures of all species combined
was not significantly different between
PF2
and LF2
goodness of fittest:^2 = 3.227, df = 1, P
of occupied and non-occupied
stations was 63 (PF2 = 34; LF2
= 9), respectively.The
trap-
= 29) and 20 (PF2 =11;
differences in microhabitat
occupied
and non-occupied
in
primary (PF2) and logged (LF2) forests(Figure
to differ from
configurations between
(Table
characteristics between
are best explained by function 2. Function
by the small mammals
but wet sitesgenerally tended to be avoided.
LF2
95.7%
1 appeared to explain the differences in habitat structure
between
= 0.072). Number
was
differences in microhabitat
occur at sites with a higher density of fallen logs and
(Chi-squared
function 2
trap-stationsused and avoided by the small mammals
on function 2, the small mammals
Total number
of function 1 and
more
in the primary forest appeared
that in the logged
forest. The
small
in primary forest seemed
to prefer siteswith
bare ground, but with many
small trees (Table
5). Sites
with
higher coverage
of twigs and
rock
piles, and with dense vegetation at the middle
storey
also appeared
forest,
to be preferred. In
preferred sitesseemed
and canopy
seemed
the logged
to have dense vegetation at low
levels. Sites with many
tree stumps
also
to be preferred in thishabitat.
trap-stationsin PF2 and LF2 were significant(Wilks'A=
0.149, x2 = 130.487, df = 69, P < 0.0001). The
resulting
three discriminant functions correctly classified86%
of 83) of the trap-stationsto group classes.The
Discussion
(71
percent
In this study, M. surifer represented the majority
of observations and individuals captured. Therefore,
their prey by sight. By
the discussion
rock piles, fallen logs, tree stumps, dense ground level
mammals
of microhabitat
utilisation by small
collectively is very much
weighted
the preferences of this species. Nonetheless,
associating with structures like
towards
vegetation and canopy
it needs
be less visible and can, therefore, rapidly escape from
layering, small mammals
will
to be stressed that all species considered in this study
predators (Whitten, 1981; Lagos et al.,1995; Dickman,
also inhabit the forest understorey. Differences
1995).
the microhabitat-use
patterns between
insignificant suggesting
overlap. In
general, the species
1962; Lim, 1970; Langham,
it can be assumed
microhabitat
configurations
addressed in this study may
The
share a common
and fruits (Davis,
1983; Payne et al.,1985).
Accordingly,
community
species were
a high level of microhabitat
diet of insects, other arthropods
the microhabitat-use
in
that the combined
of the small mam
ma Is
represent, very generally,
patterns
of small mammal
Comparisons
sampling
the patterns
microhabitat-
forests at the two
locations in this study did not reveal any
the Core Area
there is some
evidence
did not change
selective logging, results from
Li pad (Location
forest were
forest. This
different from
shows
generally are resilient and capable
may
very few
(August,
investigators have
actually demonstrated
clear associations with specific microhabitat variables,
especially in a
primary
multivariate approach.
and regenerating lowland
forest at Pasoh Forest Reserve
showed
that most
small mammal
captures occurred
and Bell, 1985). In
captures were
found
particular,
to be positively
for the fact that small mammal
proportions.
To assess the impact
comparing
results from
possible that observed
of habitat disturbance
two
differences may
sites.To
overcome
area has been suggested as a more
with flooding (Kemper
and Bell, 1985). Leaf
litterand
especially for
ground-dwelling
species (Davis, 1962; Kemper
and
semi
a rocky habitat, fallen logs and tree stumps
specific requirements
provide
for reproduction, such as refuge
sites for nesting and rearing of young
Tomblin
arboreal
and Bell, 1985), whereas
(Davis, 1962;
and Alder, 1998).
(Wu
et al., 1996; Heydon
opportunity
is, however,
1993). The
primary
present study were
factors,other
and logged sampling
sites in the
than those related to the effects of
logging, are believed to be minimal. Although,
a
there
general difference in the vegetation
structure between
the two sampling
locations in this
the primary and logged
of small mammal
species in
forest treatments at a particular location appeared to be
forest sites. These
include
related only to the differences occurring as a result of
density of fallen logs and twigs, coverage of rock piles
and wet areas, number
et al.,
differences in environmental
study, the differences between
unlogged
desirable approach
and Bulloh, 1997). Such
the presence or absence
and
conducted at the same
very rare (Dahaban
by the present study as significant in determining
logged
this problem,
paired and each pair was located
close to the other. Hence,
was indeed
Several microhabitat variables have been identified
such as local
and soil substrate, or other pre-disturbance
differences between
a "before and after" experiment
rotting logs have been associated with food abundance
be due to the
heterogeneity of the physical environment
topography
by
different sites, it is
emergents
correlated
species
present in the logged forest sites,although in different
correlated with the density of rotting logs, leaf litter,
and seedlings, and negatively
to
capability
that were recorded in the primary forest sites were also
study in
dipterocarp
in Peninsular Malaysia
small mammal
on drier sites (Kemper
A
mixed
account
that of
of adapting
the tropics are many
and particularlycomplex
2)
that small mammals
changes in their natural environments. This
and Bell, 1985). In the Asian tropics,
with
indicated that the microhabitats of the small mammals
of local habitat usage and distribution of organisms in
1983; Kemper
at
(Location 1) that the microhabitat-use
patterns of the small mammals
the logged
ecological factors that determine
and logged
consistent trend. While
in the primary
of the forest understorey.
of the small mammal
use patterns in primary
of tree stumps
and vegetation
habitat change due to selective logging.
Malcolm
(1997) remarked
that relative to other
density at low heights (1-5 m), middle heights (> 5 m)
taxa, small mammals
and at canopy
succcssional series,or other ecosystems with relatively
comparable
level. In general these variables are
to that identified by
(1985), which
Kemper
and
Bell
are related to foraging and refuge sites,
are adapted
high disturbance regimes
habitats.Although
implies that compared
Many
small mammals
terrestrial and
aerial, locate
and hence more
early
secondary
difficultto envisage, this hypothesis
with probably a strong influence for predator avoidance.
predators, both
to overcome
may
with
other
faunal
have experienced
groups,
a significantly
different suite of habitats during their evolutionary
history (Malcolm,
was
1997). Although
specifically for Amazonian
same
may
Malcolm's
small mammals,
be true for small mammals
of Southeast Asia. Given
remark
the
in tropicalforests
communities.
Bernard,
mammal
that, in the case of Tabin, in
Parks
Bernard,
forests is undoubtedly
however,
species diversity in logged
high. It should be emphasised,
that these forests had
once. Furthermore,
been
logged
only
the relatively long period that
species of North Borneo
based on line-
trapping with wire-mesh live cage traps.Sabah
addition to the primary forest patches, the prospect of
sustaining small mammal
Ecology, 64: 1495-1507.
H. (2003) Bait preferences of some small
Nature Journal, 6: 27-44.
H. and
Diun,
avifauna of Gunung
Sabah, Malaysia.
P. (1999) A checklist of the
Rara
Forest Reserve, Tawau,
Sabah
Parks Nature
Journal, 2:
45-57.
Corlett, R.T.
and
Turner,
I.M.
(1997) Long-term
has elapsed since the last logging activitiestook place
survival in tropicalforest remnants in Singapore and
(i.e.,13 to 25 years) may
Hong
forests to recover much
structure.On
have
allowed
the logged
of their original vegetation
whether the small mammals
may
be able
Kong.
Management,
and
Communities,
is unknown.
London,
(1997), however, remarked that
repeated logging, especially at short intervals,is likely
to be more detrimental than a single episode of logging.
While
the present
study
has
shown
that logged
over forest can stillplay a vital role in biodiversity
conservation, many
species especially interior primary
W.L. and Bierregaard,
R.O. (eds.), Tropical Forest Remnants:
to thrive in this habitat after a second or third logging
Whitmore
In, Laurance,
Cuaron,
Conservation
Univ. Chicago
of
Ecology,
Fragmented
Press, Chicago
and
pp. 333-345.
A.D. (2000) A global perspective on habitat
disturbance
and
tropical rainforest mammals.
Conserv. Biol, 14: 1574-1579.
Dahaban,
Z., Nordin,
Immediate
M.N.
and Bennett, E.L. (1993)
effectson wildlife of selectivelogging in
forest specialist are stillreliant on primary forests for
a hilldipterocarp forest in Sarawak:
their long term survival. Thus, derived habitats should
Edwards,
D.S., Booth, W.E. and Choy, S.C. (eds.),
be maintained only in addition to, and not in place of,
Tropical
Rainforest
areas of primary forest.
Kluwer
In,
― Current Issues,
Acad. Pub., Dordrecht, pp. 341-346.
Danielsen,
Acknowledgements
Research
mammals.
logging
F. and
and
Heegard,
M.
(1995) Impact
plantation development
of
on species
diversity:A case study from Sumatra. In, Sandbukt,
This research was conducted
"collaboration on
Malaysia
Sabah
DANCED.
and Danish
Additional
Universities" funded
Sabah
through
research grant (UMS
Fjeldsa, Men
a short term
reviewers commented
E. Bignell and
on the drafts
of this paper. Januarius Gobilik provided the map
Tabin. Special
by
No. 18/99). Jon
no Schilthuizen, David
two anonymous
Universiti
financial support was provided
by Universiti Malaysia
fundamental
as part of the project
biodiversity between
thanks go to Maryati Mohamed,
Borchsenius
and Axel D. Poulsen
support and
interest in the project. Permission
of
Finn
for their constant
to
conduct the study at Tabin Wildlife Reserve was kindly
granted by the Wildlife Department
of Sabah.
0. (ed.),Management
of Tropical Forests: Towards
an Integrated Perspective, Ctr. Dev. &
En v.,Univ.
Oslo, Oslo, pp. 73-92.
Davis, D.I). (1962) Mammals
of the lowland rain-forest
of North Borneo. Bull. Natn. MusSingapore,
no.
31: 1-129.
Dickman,
C.R. (1995) Diets and habitat preferences of
three species of crocidurine shrews in arid southern
Africa. Jour. Zool. (London), 237: 499-514.
Grieser Johns, A. (1996) Bird population persistence in
Sabahan
logging concessions. Bio. Conserv., 75: 3-
10.
Heydon,
M.J.
and
densities in a
Bulloh,
P. (1997) Mousdeer
tropical rainforest: The impact
of
selective logging. Jour. App. Ecoi, 34: 484-496.
References
Holloway,
J. D., Kirk-Spriggs,
(1992) The response
Aiken, S.R. and
Leigh, C.H. (1992) Vanishing
Forests'.The Ecological
Claredon
August,
Rain
Transition in Malaysia.
Press, Oxford, 194 p.
P.V. (1983) The
role of habitat complexity
and heterogeneity in structuring tropical mammal
of some
A.H. and
Khen,
C.V.
rain forest insect
groups to logging and conversion to plantation.In,
Marshall, A.G. and Swaine, M.D.
Rain Forest Disturbance
(eds.). Tropical
and Recovery, The Royal
Society, London, pp. 425-436.
Kemper,
C. and Bell, D.T. (1985) Small mammals
and
habitat structure in lowland rain forest of Peninsular
Malaysia. Jour. Trap. Ecol., 1: 5-22.
Lagos,
V.O.,
Contreras,
Gutierrez, J.R. and
L.C.,
L.P.,
along an elevational transect in temperate rainforests
A
with a Neotropical rodent. Oikos,
of Chile. Jour. Mammal.,
Payne,
R.K. (2000) Effects of habitat disturbance and
protected areas on mammals
of Peninsular Malaysia.
for Bornean
lowland
Marshall, A.G. and Swaine,
Rain Forest Disturbance
M.D.
(eds.), Tropical
Sodhi, N.S.
W. L. and Bierregaard,
Conservation
Lim,
R.O. (eds.) (1997)
of Fragmented
A.J. (1997) Rapid
diversity in monsoon
Communities.
and London, 616 p.
W.L.
and
Communities,
London,
Malcolm,
Conservation
Univ. Chicago
J.R.
In, Laurance,
Biomass
in Amazonian
and
Conservation
Univ. Chicago
C.W.
two
(1998) Differences in
morphologically
similar
79: 953-961.
of ants to selective logging of a
forest.Jour. App. Ecol., 37: 508-
T.C. (1997) Tropical forest disturbance,
and
species
loss. In, Laurance,
forest fragments
W.L. and Bierregaard, R.O. (eds.),Tropical Forest
and Bierregaard,
Remnants:
of
Ecology,
Press, Chicago
Chicago
and
Ecology, Management,
of Fragmented
Fragmented
of
and
Whitten,
diversity
of
Communities,
and Conservation
Univ. Chicago
Press,
and London, pp. 3-12.
J.E.J. (1981) Ecological separation of three
Island, Indonesia. Jour. Zool. (London),
forest fragments.
Fragmented
Press, Chicago
and
Willott, S. J., Lim,
S.L. (2000)
Communities,
London,
Valley.
Phil. Trans. Roy. Soc. London, Ser. B, 335: 331-339.
Norusis, M. J. (1994) SPSS
G.S. and Sutton,
on the
rainforest. Conserv. Biol.,
14: 1055-1065.
Wu,
Sabah, Malaysia: An introduction to Danum
D.C., Compton,
Effects of selective logging
butterflies of a Bornean
pp.
and Greer, A.G. (1992) Forest land-use in
193: 405-
420.
Ecology, Management,
207-221.
Marsh,
Asia.
514.
disappearance
Bierregaard, R.O. (eds.),
Tropical Forest Remnants:
and
Improving
diurnal squirrels in tropical rainforest on Siberut
(1997)
W.L.
(2000)
H. L., Vilhena, j.M.S and Caliri, G.J.A.
central Amazonian
pp. 222-240.
small mammals
L.H.
research in Southeast
Alder, G.H.
Responses
Whitmore,
R.O. (eds.), Tropical Forest Remnants:
Management,
Liow,
biology
D.C. and
Vasconcelos,
(2000)
51: 730-740.
evergreen
Plan for Tabin Wildlife
tropical forest rodents. Jour. Mammal.,
decline of small mammal
in Thailand. In, Laurance,
and
habitat use between
B.L. (1970) Distribution,relative abundance, food
Malaysia. Jour. Mammal.,
S.K. (1999) Effects
Conserv. Biol.,14: 1211-1212.
habits and parasite of giant rats (Rattus) in west
Lynam,
Robinson,
landscape. Conserv. Biol.,13: 58-66.
conservation
Ecology, Management,
Press, Chicago
and
Reserve. Sabah Wildlife Department, Kota Kinabalu,
Tomblin,
Tropical Forest Remnants:
W.R.
77 p.
tree shrew, T. glis in peninsula Malaysia. Jour. Zool.
Univ. Chicago
of Borneo. The Sabah
fragmented
and Recovery, The Royal
(London), 197: 323-344.
and
Field Guide to the Mammals
Sale, J.B. (1994) Management
N.P.E. (1983) The ecology of the common
Laurance,
Phillipps, K. (1985) A
Robinson,
of selective
forest birds. In,
Society, London, pp. 443 ―451.
Langham,
71: 620-633.
and
of selective logging on forest bird populations in a
F.R. (1992) The consequence
logging
J., Francis, C.M.
Society, Kota Kinabalu, 332 p.
Conserv. Biol, 14: 1639-1648.
Lambert,
P.L. and Lang, B.K. (1990)
Quantitative habitat associations of small mammals
Jaksic, F.M. (1995) Effects of
74: 259-264.
Laidlaw,
Inc.,Chicago, 577 p.
Patterson, B.D., Meserve,
Meserve,
predation risk on space use by small mammals:
field experiment
SPSS
D.L., Luo, J. and Fox, B.J. (1996) A comparison
of ground-dwelling
small mammal
communities
in
primary and secondary tropicalrainforests in China.
Jour. Trop. Ecol., 12: 215-230.
Professional Statistics6.1.
Received: June 10, 2003
Accepted: November
26, 2003