Australian Journal of Ecology (1999) 24, 228–239

Bias in the effect of habitat structure on pitfall traps:

An experimental evaluation
BRETT A. MELBOURNE
Division of Botany and Zoology, The Australian National University, Canberra, ACT 0200, Australia
([email protected])

Abstract Habitat structure has been implicated as a source of bias for pitfall-trap data but most evidence is
observational or anecdotal. This study used an experimental approach to quantify biases due to habitat structure.
In a randomized block design, I manipulated native grassland to create three types of habitat structure and measured
pitfall-trap catches of grassland ants. Small patches of modified habitat were surrounded by otherwise unmodified
grassland with the assumption that population density remained unaffected by the modification and that the effects
observed were due to changes in trappability. I assessed magnitude, direction, predictability, and consistency of
bias for the following types of data: population abundance for single species, relative abundance among species,
species composition of assemblages, and species richness. The magnitude of the bias in population abundance
was large for most species. However, since the direction of the bias varied predictably with habitat structure, pitfall-
trap data can be used to judge differences in population abundance in some situations. The magnitude of the bias
in relative abundance was less than for abundance. However, there was inconsistency in the direction and magnitude
of bias among species. Thus, interpretation of relative abundance data in pitfall-trap studies may be compromised.
Species richness and species composition were biased by habitat structure but were affected significantly only when
the groundcover was very dense, suggesting a threshold effect of habitat structure. To help to interpret survey data,
pitfall-trap studies should routinely measure attributes of habitat structure and incorporate an experimental
component to characterize the bias.

Key words: biodiversity assessment, field experiment, Formicidae, population assessment, survey methods.

INTRODUCTION developed to factor out bias or to aid interpretation of

biased data (Niemelä et al. 1990; Spence & Niemelä
Pitfall traps are commonly used to address a question
1994; Digweed et al. 1995; Melbourne et al. 1997). In
of difference in population size or community struc-
this paper, I consider the biases that emerge when
ture in time or space. However, an effect of habitat
locations with different habitat structure are compared,
structure on trappability could result in biased data for
determine the extent to which different types of data
studies that compare locations with different habitat
are affected by bias and consider approaches to aid
structure, or when habitat structure changes in time.
interpretation of biased data.
Despite widespread recognition of this and other
The capture rate of a pitfall trap is a function of
potential sources of bias (reviews by Luff 1975;
population density and trappability. Trappability is here
Thomas & Sleeper 1977; Southwood 1978; Adis 1979;
defined as the probability of capture of an individual
Halsall & Wratten 1988; Topping & Sunderland 1992;
in the population. Differences in trappability in time
Spence & Niemelä 1994), pitfall traps continue to be
or space may lead to biased estimates for population-
used frequently in many areas of ecological research,
level data. One well-recognized influence on trappa-
for both vertebrates and invertebrates. This is perhaps
bility is the activity level of individuals in a population
because methods other than pitfall trapping are more
(Mitchell 1963). Indeed, pitfall-trap data are often said
difficult and expensive to implement, particularly for
to represent the ‘activity density’ of a population
surface-active invertebrates. Sometimes pitfall trapping
(Thiele 1977). Activity levels can be affected, inde-
may be the only feasible method, such as when popu-
pendently of population density, by a variety of factors
lation densities are low or when minimal impact
(Southwood 1978), particularly those that directly
methods are required for sensitive sites. Because of the
affect the physiological state or behaviour of the
advantages of pitfall trapping, approaches need to be
animal, such as weather (Whicker & Tracy 1987;
Honék 1988; Niemelä et al. 1989), leading to biased
Present address: CSIRO Wildlife and Ecology, Private Bag PO, estimates of population density. Community-level data
Wembley, WA 6014, Australia ([email protected]). share the same problems as population-level data but
Accepted for publication October 1998. in addition are affected by differences in trappability
 HA BI TAT S T R U CT U R E A N D P I T FA LL T R A P S 229

among species, which can arise either from differences In his main experiment, he clipped away a 30 cm radius
in activity levels (Honék 1988), or through trap area of grass from around three pitfall traps and
avoidance or escape behaviour (Luff 1975; Halsall & compared this to three traps in intact grass. He found
Wratten 1988; Topping 1993). Differences in the direc- that catches of some carabid beetles were higher in the
tion and magnitude of biases among species may cause traps that had grass clipped from around them. The
community-level data to vary in time or space, inde- result was attributed to the lower impedance of the
pendently of real changes in community structure. habitat surrounding cleared traps. Although there is
The limitations just described are not critical when anecdotal evidence consistent with Greenslade’s experi-
the question of interest is a relative difference in popu- mental findings (e.g. Wallin 1985; Honék 1988) his
lation density or community structure in time or space. study is apparently the only published experimental
Then, it is not necessary to know the absolute density investigation into the effect of habitat structure on
or structure; biased data (e.g. the activity density of the trappability. Since studies that compare locations with
‘pitfall-sampled’ community) are sufficient to address different habitat structure are common, more research
the question of ‘differences’. However, it must be is needed to identify limitations to the use of pitfall traps
assumed that biases are the same between the times or and to develop approaches to aid interpretation of
locations that are being compared. For comparisons in biased data. I conducted a field experiment to deter-
time, such as between seasons, this assumption is prob- mine the effect of habitat structure on trappability of
ably unreasonable at least because of changing weather grassland ants for the following types of data: popu-
conditions and animal behaviour. For comparisons in lation abundance for single species, relative abundance
space, this assumption is often reasonable provided that among species, species composition of assemblages,
weather conditions are not different between locations and species richness.
or can be averaged out over time (Baars 1979).
For studies that compare locations with different
habitat structure, the assumption that biases are the METHODS
same between the locations that are being compared is
critical. Trappability must not change with habitat
Study site and manipulations
structure. Here, habitat structure is defined as the
spatial configuration of the animal’s environment, The study site was located within a 20 hectare area of
particularly the environment immediately surrounding Themeda triandra dominated grassland (35°1294099E,
the trap. Examples of elements of habitat structure 149°0695099S), in the Australian Capital Territory
include the density of vegetation, the roughness of the (ACT), Australia. This grassland consisted of dense
soil surface, or the spatial arrangement of microland- T. triandra with a thick litter layer of dead and senescent
scape features. Habitat structure could affect trappa- plant material to a height of about 20 cm. The study
bility by affecting the movement behaviour of animals, site was a uniform area, 60 3 50 m, divided into 30
either through microclimatic effects or by a more direct 10 3 10 m plots (Fig. 1). Two pitfall traps were installed
response of individuals to the physical features of the 2.5 m apart at the centre of each plot. Pitfall traps were
habitat. Studies that compare locations with different similar to the design used by Margules (1993). Each
habitat structure are perhaps the most common. Some trap consisted of a plastic cup, 9 cm in diameter,
examples from the recent literature include forest frag- inserted into a polyvinyl chloride sleeve, sunk flush with
ments versus the surrounding landscape (Punttila et al. ground level. A 60 cm long, 5 cm high, galvanized iron
1994), grazed versus ungrazed plots (Dennis et al. drift fence was positioned across the centre line of the
1997), burnt versus unburnt sites (Greenberg et al. cup and a galvanized iron roof was positioned over the
1994; Cole et al. 1997), comparisons among agroeco- cup, flush with the drift fence. A preservative fluid, con-
systems (Cárcamo et al. 1995; Perfecto & Snelling sisting of 50% ethylene glycol with water, was used.
1995), or biodiversity studies that compare different Further details of the study site and pitfall traps are
habitat types (Abensperg-Traun et al. 1996; Gomez & given in Melbourne et al. (1997).
Anthony 1996; Hadden & Westbrooke 1996; Samways The experimental design included three levels of
et al. 1996). Sometimes comparisons made in time are habitat structure, which were created by modifying (or
coincidentally also comparisons between different not modifying) the structure of the habitat within an
habitat structures, because habitat structure may 80 cm radius of each pitfall trap. This radius was chosen
change over time. The effect of habitat structure on to be somewhat larger than the drift fence. The three
trappability, independent of its effect on population size levels of habitat structure encompassed a range of veg-
or community structure, could present a serious etation densities and were designed to match the struc-
problem when interpreting the results from these types ture of different native grassland communities found
of studies. in the ACT. For the most dense level (‘unmodified’)
The effect of habitat structure on trappability was no modification was made to the structure of the
first experimentally examined by Greenslade (1964). surrounding habitat. For the least dense level (‘cleared’)
2 30 B. A . MELBOURNE

all above-ground vegetation and litter was removed formed but remained closed for 10 days after the
from within an 80 cm radius of each pitfall trap. To manipulations were completed to allow disturbance
create a level of density intermediate between ‘unmodi- effects to dissipate. Traps were then opened for two
fied’ and ‘cleared’, only litter and senescent plant mat- weeks in autumn (5–20 April 1993).
erial was removed from within an 80 cm radius of each All ants caught were sorted to morphospecies
pitfall trap, such that inter-tussock spaces were created as described in Andersen & Reichel (1994).
(‘litter removed’). All debris was removed from Representative samples of morphospecies were identi-
‘cleared’ and ‘litter removed’ plots using a vacuum fied by S. O. Shattuck (pers. comm. 1996). Many
device. A randomized block design was used with 10 species could not be identified to named species with
blocks and 10 replicates for each level of habitat struc- confidence. These species were identified to species
ture (Fig. 1). Habitat structure treatments were groups where possible. In this paper, species groups are
randomly assigned to plots within blocks. In all indicated by the use of parentheses (International
analyses, data from the two pitfall traps in each plot Commission of Zoological Nomenclature 1985).
were summed and made up one replicate. Traps were Voucher specimens of each morphospecies have been
installed before experimental manipulations were per- deposited in the Australian National Insect Collection,
CSIRO Entomology, Canberra.

Experimental assumptions
Two assumptions were made in the experiment. First,
that the modifications to habitat structure made in the
area immediately surrounding the pitfall traps did not
affect the size of ant populations. Thus, any effects
observed would be due to a change in trappability
rather than to a change in the population size of a given
species. This assumption is based on the scale of the
manipulations. The size of each manipulation (the sum
of the two modified circular areas in each plot)
was 4 m2, or 4% of a 10 3 10 m plot (Fig. 1a). This
small area was not expected to alter substantially the
population size of any species, in the time frame of the
experiment. In other words, the true population size
at the scale of 10 3 10 m plots was assumed to be the
same, on average, among the different habitat structure
manipulations. A second assumption was that any
observed effects were due to differences in the structure
of the habitat rather than to other effects associated with
the manipulations, such as physical disturbance of the
habitat. Possible effects due to disturbance include dis-
ruption of ant nests and attraction of animals to the
disturbed area because of an increase in food items.
However, such initial disturbance effects associated
with installation of pitfall traps probably last for less
than one week for ants (Greenslade 1973), less than
two weeks for carabid beetles (Digweed et al. 1995) and
a few days for Collembola (Joose & Kapteijn 1968)
and some other invertebrate groups (Greenslade 1973).
I left a 10-day period, during which traps were closed,
between the habitat manipulations and the commence-
ment of trapping. This period was a trade off between
Fig. 1. Schematic representation of (a) experimental plots
allowing initial disturbance effects to dissipate and to
and (b) the experimental area. (a) Each plot consisted of a
10 3 10 m area with two pitfall traps positioned 2.5 m apart. minimize delayed effects, such as regrowth of vegetation
Habitat structure was modified (or not modified) in a cir- or population response to the new habitat.
cular area of radius 80 cm surrounding each trap. (b) Plots
were arranged in a randomized block design, with three plots Statistical analyses
to a block and treatments randomly assigned to plots within
blocks. Treatments were: U, unmodified; L, litter removed; Mixed model analyses of variance (ANOVA) were used
C, cleared. to model the effect of habitat structure on: (i) the abun-
 HA BI TAT S T R U CT U R E A N D P I T FA LL T R A P S 231

dance of the five most abundant ant species considered used to ordinate the sites based on the association
separately, and (ii) the number of species caught. Block values (Faith et al. 1987; Belbin 1991, 1994). A cut-
was included in the ANOVA model as a random effect. value of 0.63 was used for HMDS based on inspection
The abundance data for all species required log trans- of a histogram of the association values (Belbin 1994).
formation, which successfully stabilized the variance. The number of dimensions required to describe the
P-values were calculated from F-ratios to test the null data adequately was determined using the guidelines
hypothesis that there was no effect of habitat structure given in Clarke (1993) after inspecting the stress values
on pitfall-trap catches. There were no a priori hypo- from solutions in two to five dimensions obtained from
theses requiring comparisons between particular means the best of 10 random starts. The final solution for the
and thus a multiple comparison procedure was used. chosen number of dimensions was obtained from 100
A protected LSD (least significant difference) was random starts using a stopping rule of a maximum of
calculated from: t 3 SED, where t 5 critical value 50 iterations or a minimum stress difference of 0.005.
of student’s t-distribution with appropriate degrees of Solutions from non-metric and metric MDS also were
freedom at a 5 0.05 and SED 5 standard error of the inspected. Fourth, a permutation test (ANOSIM;
difference between means (Fisher 1935). Clarke 1993) was used to determine whether there was
Logistic regression (McCullagh & Nelder 1989) was a significant effect of habitat structure on species com-
used to model the effect of habitat structure on the position. The ANOSIM test described by Clarke
relative abundance of the five most abundant species. (1993) is based on rank similarities (non-metric
Logistic regression was used because ‘relative abun- ANOSIM) but the association values themselves
dance’ has an approximately binomial form, that is, the also can be used (metric ANOSIM; Belbin 1994). The
abundance of a given species is a proportion of the latter approach possibly has more statistical power.
abundance of all species combined. To estimate the Both approaches were used here. Following the test
change in deviance associated with habitat structure, for an overall effect, the test was repeated for pairs of
block was first added to the model followed by habitat the treatment level. All multivariate analyses were
structure. Block was included in the model whether it performed using PATN 3.3 (Belbin 1994).
was significant or not because it describes the (random)
variation associated with experimental blocks. Deviance
ratios and standard errors were calculated using the dis-
RESULTS
persion parameter set to the residual mean deviance.
P-values were calculated from deviance ratios to test Experimental manipulation of habitat structure in
the null hypothesis that there was no effect of habitat Themeda triandra grassland affected all types of data
structure on pitfall-trap catches. A common approach (abundance, relative abundance, species richness,
to analyzing relative-abundance data is ANOVA of species composition). In general, trappability increased
untransformed or arcsine-transformed percentages. as the structure of the habitat became more open. For
This approach gave similar means, identified the same convenience of presentation, pitfall-trap catches are
outliers and gave higher levels of statistical significance. here referred to using the same terminology as the
However, the ANOVA assumption of constant variance quantities they would normally be used to estimate.
was violated and attempts to stabilize the variance with Thus, for example, ‘abundance’ here implicitly means
various data transformations were not as successful as ‘the number of animals caught in pitfall traps’ since it
the use of logistic regression, which is more appropriate was only the number of animals caught in pitfall traps
for the distributional properties of relative-abundance that could be affected by the manipulations of habitat
data. Logistic regressions and ANOVAs were performed structure. The true population abundance at the scale
using the statistical package GENSTAT 5, release 3.1 of 10 3 10 m plots was assumed to be the same, on
(Numerical Algorithms Group, Oxford). average, among the different habitat structure classes.
The effect of habitat structure on species compo- In other words, the data were assumed to be affected
sition was investigated using multivariate techniques as (biased) not the ecological quantity itself. Note also that
follows. First, the species composition of each plot was the bias referred to here is in addition to other sources
compared with all other plots using the Czekanowski- of bias. It is assumed that the baseline data are
Sorensen measure of dissimilarity, which is one of a inherently biased by the factors discussed in the
class of measures that provide a robust estimate of eco- Introduction. The issue addressed here is whether
logical distance (Faith et al. 1987). Second, to detect additional bias is introduced, or the inherent bias
outliers, a cluster analysis was performed on the changes, with habitat structure.
association values using the UPGMA (unweighted pair
group arithmetic averaging) method (Sneath & Sokal
Population abundance
1973) with slight dilation (beta 5 – 0.1; Belbin 1994).
Outliers were then excluded from all other analyses. Abundance data were biased by habitat structure for
Third, hybrid multidimensional scaling (HMDS) was four of the five most abundant ant species (Fig. 2). For
2 32 B. A . MELBOURNE

all species for which there was a significant effect of Relative abundance among species
habitat structure, trappability increased as the habitat
Relative-abundance data were biased by habitat
became more open. However, comparisons of means
structure for three of the five most abundant species
indicated that there was variation among species in their
(Fig. 3). Indeed, the results for each species were idio-
responses to particular levels of habitat structure. The
syncratic; that is, for most species there was not a
abundance of Iridomyrmex (rufoniger) sp. A increased
monotonically ordered pattern with vegetation density
with each decrease in vegetation density, the abundance
and there was considerable variation among species. It
of Rhytidoponera (metallica) sp. increased significantly
appears also that the rank relative abundance and rank
only in the ‘cleared’ pitfall traps, and the abundance
abundance of species was affected (Table 1). For
of Pheidole sp. A and Paratrechina sp. increased to a
example, the relative abundance of I. (rufoniger) sp. A
similar extent in both the ‘litter removed’ and ‘cleared’
(about 17%) in the ‘litter removed’ pitfall traps was
pitfall traps.
lower than the relative abundance of Paratrechina sp.

Fig. 2. The abundance in pitfall traps of five ant species in Fig. 3. The relative abundance in pitfall traps of five ant
response to experimental manipulation of habitat structure species in response to experimental manipulation of habitat
surrounding the trap. Results show mean abundance esti- structure surrounding the trap. Results show mean relative
mated using ANOVA for each species separately. P-values indi- abundance estimated using logistic regression for each
cate the level of significance for habitat structure. Black bars species separately. P-values indicate the level of significance
indicate the least significant difference for each species for habitat structure. Two-tiered error bars indicate the stan-
(LSD), which is equal to approximately two times the stan- dard error (inner tier) and 95% confidence interval (outer
dard error of the difference between means. Comparison of tier) for the mean. One outlier was excluded from the model
means using the LSD indicates the significance of the differ- for Iridomyrmex (rufoniger) sp. A (relative abundance 5 85%
ence between two means, within species, at a 5 0.05. in a ‘cleared’ plot) and also for Paratrechina sp. (relative
* 5 outlier, excluded from model (P 5 0.2 with outlier abundance 5 83% in an ‘unmodified’ plot). With outliers
included). included in the model, P 5 0.009 and P 5 0.3, respectively.
 HA BI TAT S T R U CT U R E A N D P I T FA LL T R A P S 233

(about 26%) but this pattern was reversed in the habitat structure became more open (Table 2). The
‘cleared’ pitfall traps (Fig. 3, Table 1). Also of note was main effect was for the most dense habitat
the pattern of the variance for relative abundance, from (‘unmodified’), in which frequency of occurrence was
large variance in the ‘unmodified’ pitfall traps to small very low for many species. Species that were
variance in the ‘cleared’ pitfall traps (Fig. 3). represented in the experiment by less than 10
individuals were caught only in the ‘litter removed’
and ‘cleared’ treatments in all but two cases (Table 2).
Frequency of occurrence
For most of the abundant species, the frequency
The frequency of occurrence for almost all of the of occurrence was higher in the modified plots than
19 species caught in the study increased as the the ‘unmodified’ plots and was either equal between

Table 1. The rank order of five ant species, from highest to lowest, for mean abundance and mean relative abundance of the
pitfall-trap catch in response to experimental manipulation of habitat structure

Abundance Relative abundance

Species U L C U L C

Iridomyrmex (rufoniger) sp. A 1 2 1 1 3 1

Pheidole sp. A 3 1 2 3 1 2
Rhytidoponera. (metallica) sp. 1 4 3 1 4 4
Paratrechina sp. 5 2 4 3 2 3
Notoncus (ectatommoides) sp. 3 5 5 5 5 5

Ranks were calculated from the means given in Fig. 2 and 3. Means were considered different when the larger mean was
1.10 times larger than the smaller mean, otherwise ranks were considered to be tied.
U, unmodified, L, litter removed, C, cleared.

Table 2. The 19 ant species caught in the experiment and the frequency of occurrence for each species in pitfall traps with dif-
ferent surrounding habitat structure

Species Code Frequency of occurrence Total

Unmodified Litter removed Cleared caught

Ponerinae
Heteroponera imbellis (Emery) GR03 0.1 0.7 0.3 20
Hypoponera sp. A GR04 0 0.5 0.6 35
Hypoponera sp. E GR06 0.1 0 0.5 8
Hypoponera sp. F GR07 0 0.1 0.1 2
Rhytidoponera. (metallica) sp. GR09 0.8 1.0 1.0 147
Myrmicinae
Meranoplus sp. GR11 0.1 0.3 0 8
Monomorium sp. C GR14 0 0.1 0 1
Pheidole sp. A GR18 0.4 0.9 1.0 313
Dolichoderinae
Iridomyrmex (rufoniger) sp. A GR29 0.8 1.0 1.0 361
Iridomyrmex (rufoniger) sp. B GR30 0 0.1 0 2
Iridomyrmex (bicknelli) sp. GR32 0 0.1 0 1
Iridomyrmex sp. A GR28 0 0 0.3 8
Iridomyrmex sp. C GR31 0 0.1 0.1 2
Ochetellus glaber (Mayr) GR26 0 0 0.1 1
Formicinae
Camponotus sp. GR36 0 0 0.1 1
Melophorus sp. A GR37 0 0.1 0.2 3
Melophorus sp. B GR38 0 0.1 0 1
Notoncus (ectatommoides) sp. GR40 0.4 0.5 0.6 48
Paratrechina sp. GR42 0.7 0.8 0.8 264

Frequency of occurrence was calculated from the number of plots in which the species was caught, divided by the total number
of plots (10) for each habitat structure class. The total number of individuals caught for each species also is given. Species have
been grouped by subfamily. Also shown are voucher codes that correspond to specimens submitted to the Australian National
Insect Collection, CSIRO Entomology, Canberra.
2 34 B. A . MELBOURNE

the ‘litter removed’ and ‘cleared’ plots or slightly higher

in the ‘cleared’ plots.

Species richness and composition

There was a significant effect of habitat structure on
species-richness data (P < 0.001, Fig. 4). Species
richness in the ‘litter removed’ and ‘cleared’ pitfall
traps was almost twice as high as in the ‘unmodified’
pitfall traps.
There was a significant effect of habitat structure on
species-composition data (non-metric ANOSIM,
P 5 0.01; metric ANOSIM, P 5 0.004), which is
reflected in the separation of habitat-structure types in
ordination space (Fig. 5). Pairwise comparisons indi-
cate that the difference in species composition was most
Fig. 4. The number of species caught in pitfall traps in significant for ‘unmodified’ compared to ‘cleared’ and
response to experimental manipulation of habitat structure ‘unmodified’ compared to ‘litter removed’ pitfall traps
surrounding the trap. Results show mean number of species (Table 3). However, only the ‘unmodified’ versus
caught estimated using ANOVA. There was a significant effect ‘cleared’ contrast using the metric test was significant
of habitat structure (P < 0.001). LSD stands for least signifi-
cant difference, which is equal to approximately two times
after Bonferonni correction (for an overall type I error
the standard error of the difference between means. rate of a 5 0.05, a9 5 0.017). There did not appear to
Comparison of means using the LSD indicates the signifi- be a significant difference between ‘litter removed’ and
cance of the difference between two means at a 5 0.05. U, ‘cleared’ pitfall traps (Table 3), however, these two
unmodified; L, litter removed; C, cleared. habitat-structure types were essentially separated in
ordination space (Fig. 5) suggesting some degree of
difference in species composition. Three dimensions
were required for HMDS to represent adequately
(stress 5 0.12) the composition of species after three
outlying plots (Fig. 6), all from the ‘unmodified’ class,
were removed. Non-metric and metric MDS gave
similar ordination solutions to HMDS. The presence
of the outliers indicates that, as well as being more
different to the other modified classes, there also was
greater variation in species composition within the
‘unmodified’ class. With outliers included, the effect
of habitat structure on species composition data also
was significant (non-metric ANOSIM, P < 0.001;
metric ANOSIM, P < 0.001).

Table 3. Pairwise comparisons of species composition

between the different habitat structure classes

Comparison P value
Non-metric Metric
Fig. 5. Scatterplot matrix of the 3-dimensional solution from
HMDS. The ordination represents the composition (pres- U vs L 0.041 0.026
ence/absence) of ant species caught in pitfall traps in U vs C 0.029 0.003
response to experimental manipulation of the habitat struc- L vs C 0.117 0.115
ture surrounding the trap. Each point represents one experi-
mental plot (i.e. one pair of pitfall traps). Axes are scaled P values were calculated using analysis of similarity
equally. The ordination illustrates the significant effect of (ANOSIM). Non-metric P values were calculated using the
habitat structure on species-composition data (non-metric ranks of the association values whereas metric P values were
ANOSIM, P 5 0.01) and demonstrates that the ‘litter calculated using the association. The values Czekanowski
removed’ and ‘cleared’ pitfall traps are more similar to each association measure was used.
other than to the ‘unmodified’ traps (Table 3). U, unmodified; L, litter removed; C, cleared.
 HA BI TAT S T R U CT U R E A N D P I T FA LL T R A P S 235

One possibility is that some easily measured trait (e.g.

body size) is related to the magnitude of the bias. For
the ants considered here, there is a pattern of increasing
bias with decreasing body size but there are too few
species to test this conjecture.
In the absence of a means to predict the magnitude
of the bias, pitfall-trap data could be used to judge
differences in population abundance when the direction
of the bias is known, albeit in a limited number of situ-
ations (Fig. 7). For grassland ants, there was a mono-
tonic increase in the bias as the habitat structure
became more open (Fig. 2). Thus, in a population
assessment study for grassland ants, it would be pos-
sible to determine whether population abundance was
lower in open compared to dense habitats but not vice
versa (Fig. 7) but it would not be possible to determine
accurately the magnitude of the difference. For the
grassland ants in this study, the direction of the bias
was the same for nearly all species (Fig. 2, Table 2),
which is a useful generalization. Melbourne et al.
(1997) used the guidelines given in Fig. 7 to interpret
Fig. 6. Cluster analysis of plots using UPGMA. The dend- the results from a pitfall-trap survey of five different
rogram represents the composition (presence/absence) of ant types of grassland for slugs and crickets. The response
species caught in pitfall traps in response to experimental of slugs and crickets to habitat structure was first deter-
manipulation of the habitat structure surrounding the trap. mined experimentally. Since habitat structure did not
Three outliers (*) were identified. Scale indicates dissimi- bias abundance data for slugs, the survey results were
larity. considered robust. In contrast, for crickets, bias in the
abundance (higher trappability in open grasslands)
DISCUSSION meant that only the direction of the difference in abun-
dance between habitats could be determined; crickets
were more abundant in one of the densest grasslands.
Bias and recommendations for assessment
Carabid beetles may not have the same consistency
procedures
in the direction of bias as ants. Greenslade (1964)
How seriously the interpretation of pitfall-trap data is observed a variety of responses to habitat structure for
compromised by bias depends on four characteristics carabid beetles, including higher trappability of
of the bias: (i) magnitude, (ii) direction, (iii) predict- predatory carabids in open plots, no bias for larger
ability, and (iv) for multispecies data, consistency of phytophagous species and lower trappability in open
magnitude and direction among species. Here, I plots for smaller phytophagous species. This suggests
emphasize studies where the question of interest is a that the results obtained for grassland ants cannot be
difference in population size or community structure generalized to other species. It will be necessary to
between different habitat types, rather than studies determine the bias experimentally for different organ-
attempting to obtain absolute estimates. I consider isms in different habitats before survey results can be
assessment procedures for four data types: population interpreted.
abundance for single species, relative abundance among The magnitude of the bias in relative abundance
species, species composition of assemblages, and (Fig. 3) was generally less than that for abundance but
species richness. inconsistency in the direction and magnitude of bias
For grassland ants, population-abundance data were among species (Fig. 3, Table 1), along with greater vari-
biased by habitat structure for most species and the ance, means that the interpretation of multispecies
magnitude of the bias was large, up to 10 times differ- relative-abundance data is more seriously compro-
ence in abundance between structural classes (Fig. 2). mised. To correct the bias for each species would rarely
Therefore, any attempt to compare population abun- be logistically feasible for large multispecies assem-
dance between different habitats would need to correct blages. Consequently, commonly used multispecies
for this bias. Correction would be difficult to achieve measures of community structure or diversity that
when the magnitude of the bias is different among include a relative abundance component will be diffi-
species, as it was for grassland ants. For this situation, cult to interpret when pitfall traps are used to compare
the magnitude of the bias would not be predictable locations with different habitat structure. These include
without experimental testing for the species of interest. the family of measures that reduce the species richness
2 36 B. A . MELBOURNE

and relative abundance information into a single index, Directly evaluating the bias in the relative-abundance
such as the Shannon–Weiner index and Pielou’s J9 data, as done here, provides a more powerful test of
(reviewed by Magurran 1988) as well as species the reliability of these measures.
abundance models such as the log normal distribution Species-richness and species-composition data were
and the logarithmic series (May 1975; Pielou 1975; biased by habitat structure but were seriously affected
Magurran 1988). Multivariate techniques such as only by the most dense habitat structure class (Figs
ordination and clustering that include this relative 4,5,6). This effect appears to be due to the effect of
abundance component also will be difficult to interpret, habitat structure on capture probabilities. Assuming
as will the ‘functional group’ approach to community that the frequency of occurrence data reflect the prob-
assessment (Greenslade & Halliday 1983; Andersen ability of capturing a species when it is actually pres-
1995). I did not directly test the measures and indices ent, there was a tendency for capture probabilities to
mentioned here because it makes no sense to do so. For increase as the habitat became more open but the main
most measures, species identity is not important so that effect of habitat structure on capture probabilities
an increase in the relative abundance of one species may involved the most dense habitat, in which frequency of
be offset by a decrease in another, leading to no change occurrence was very low for many species (Table 2).
in the index despite real changes in relative abundance. This resulted in species compositions for the dense
‘unmodified’ plots that were essentially a small and
random sample of the more common species present
at the experimental site. Thus, species richness esti-
mates were lower and species composition estimates
were different for the dense plots compared to the more

Fig. 7. Possible interpretations of abundance data in pitfall-

trap surveys for different patterns of bias caused by habitat
structure. The first column shows the pattern of bias (or
trappability) in relation to habitat structure. This might, for
example, be determined experimentally. For simplicity, the
relationship between the bias (log measured abundance
minus log real abundance) and habitat structure is shown as
a linear function. The second column shows the number of
animals caught in a hypothetical field survey of sites with dif-
ferent habitat structure. The third column shows the possible
interpretation of field-survey data (i.e. the true population
abundance) given the bias but assuming that only the direc-
tion of the bias (the sign of the slope in column one) is known.
Shading indicates that the slope of the line may vary. It is
only possible to determine the difference in the true abun-
dance between survey sites with different habitat structure
(column three) when the survey results (column two) have
a different pattern to the bias (a, b, e, f). Further, in these
cases it is only possible to determine the direction of the
difference in the true abundance, not the magnitude of the
difference (i.e. in the third column, the sign of the slope may
be determined but not the gradient). When the survey results
have the same pattern as the bias, any interpretation is
possible (i.e. in the third column, neither the sign nor the
gradient of the slope may be determined; c, d), in which case
no confident conclusions can be drawn. Note also that, when
bias is present, ‘no difference between survey sites’ is a
possible interpretation only when the survey result has the
same pattern as the bias (i.e. in the third column, the slope
can only equal zero in c and d). It follows that, since no
confident conclusions can be drawn for (c) and (d), one can
never be confident of a ‘no difference’ conclusion when bias
due to habitat structure is present. When there is no bias (not
shown here) it is possible to determine both the direction and
the magnitude of the difference in abundance. The habitat
structure axis is arbitrary but could, for example, represent
a gradient from dense to open habitats.
 HA BI TAT S T R U CT U R E A N D P I T FA LL T R A P S 237

open plots (Figs 4,5,6). There was less difference in fre- that habitats with a more complex structure have more
quency of occurrence between the more open habitat surface area available for animals to move around on,
classes (‘litter removed’ and ‘cleared’; Table 2), which hence, the number of pitfall traps per unit area is effec-
was reflected by the greater similarity in species com- tively reduced. For example, in dense vegetation,
position and species richness estimates (Figs 4,5). This animals may move around not only on the ground but
pattern, along with the sharp transitions observed for also above the ground. Thus, while population or
abundance data (three species out of five, Fig. 2) sug- community measurements are traditionally calculated
gests that there is a threshold effect of vegetation den- as if the surface of the ground was two dimensional,
sity on trappability. It would be useful to conduct animals perceive the surface area of the habitat in
further studies where habitat structure is manipulated three dimensions. The importance of this effect will
across a greater range of densities to elucidate better be related to the scale at which the organism perceives
the functional relationship between vegetation density its environment so that, for example, grass will add
and trappability. In conclusion, pitfall traps can be significant surface area for a small invertebrate like
used to assess differences in species richness, species an ant but little surface area for a reptile or small
composition, and frequency of occurrence provided mammal.
that differences in habitat structure do not span such Two hypotheses have been suggested to explain why
thresholds. changes in movement behaviour occur: (i) because of
Modifications and additions to the pitfall trapping changes in microclimate (Honék 1988) and (ii) because
method would help to interpret survey data. First, field of the response of the species to the physical layout of
surveys could incorporate an experimental component the habitat (Greenslade 1964; Crist et al. 1992). In the
to characterize the bias. For example, trapping could first hypothesis, several effects of habitat structure on
use both cleared and uncleared traps at the same the microclimate near the ground are expected,
location. By plotting the ratio of cleared to unmodified including changes to solar radiation, humidity and
traps against habitat structure, the direction of the bias temperature (Geiger 1965). The potential effect of
and shape of the bias function (i.e. Fig. 7 column 1) increasing temperature is usually regarded as resulting
could be determined, including the detection of from faster movement or more movement activity over-
thresholds. However, consideration needs to be given all (Honék 1988). While temperature clearly plays a
to the size of manipulated areas and the length of time part in pitfall-trap catches at the larger scale in relation
required for disturbances associated with manipulations to weather (Honék 1988; Niemelä et al. 1989), few
to dissipate, both of which need to be determined studies have considered the effect of temperature in
experimentally for different taxa and different habitats. relation to the microclimate resulting from habitat
Another consideration is that inherent biases due to structure. Honék (1988) and Wallin (1985) considered
trap design (e.g. trap size, drift-fence length, type of higher catches of some species on bare ground exposed
preservative) have the potential to interact with habitat to the sun to be due to the thermophilic nature of the
structure. Thus, it might be useful to vary both trap species but their evidence is anecdotal. In any case, the
design and habitat structure. Second, attributes of effect of habitat structure on temperatures near the
habitat structure should be measured at each location. ground will be complex and will depend on the time
Standard suites of measurements need to be deter- of day and the season (Geiger 1965). For example, at
mined for different taxa but could include, for example, night, temperatures on bare ground could be lower than
vegetation density, vegetation height, microtopography a habitat with dense ground cover because of greater
(surface roughness), and temperature near the ground. convective and radiative losses. Thus, effects of habitat
It is unlikely that any one method (e.g. clearing alone) structure due to temperature and other microclimatic
would provide a standard protocol but experimental factors will depend on the diurnal and seasonal timing
manipulations, further discussed below, in conjunction of activity for a given species.
with measurement of habitat structure, could provide The second hypothesis is that changes in movement
enough information to determine important properties behaviour will result because of the response of species
of the bias. to the physical layout of the habitat. This includes
effects on the shape of movement pathways, including
effects on tortuosity and net displacement, which will
Mechanisms and future experiments
affect the susceptibility of the animal to capture (Crist
It will be necessary to consider mechanisms so that & Wiens 1995) and will be dependent on the spatial
better generalisations can be made about how trappa- scale at which the animal perceives habitat structure
bility will be affected by habitat structure for which taxa (Crist et al. 1992; Wiens et al. 1995). Speed of move-
and in what way. The trappability of animals changes ment also may be affected due to the impedance of the
with habitat structure, either because of a dilution effect habitat (Greenslade 1964). Morrill et al. (1990) found
or because of effects on the movement behaviour of that speed of movement was an important determinant
animals. The dilution hypothesis stems from the fact of pitfall-trap efficiency for some carabids and tene-
2 38 B. A . MELBOURNE

brionids, and Andersen (1983) and Greenslade (1973) the seasonal tropics of Australia’s Northern Territory. J. Aust.
have observed that ants that move fast are especially Entomol. Soc. 33, 153–8.
prone to capture in pitfall traps. However, the effect of Baars M. A. (1979) Catches in pitfall traps in relation to mean
densities of carabid beetles. Oecologia 41, 25–46.
speed of movement appears to be variable among
Belbin L. (1991) Semi-strong hybrid scaling, a new ordination
species since Halsall & Wratten (1988) could find no algorithm. J. Veg. Sci. 2, 491–6.
effect of speed of movement on pitfall-trap catches of Belbin L. (1994) PATN, Pattern Analysis Package: Technical
carabids. Reference. CSIRO Division of Wildlife & Ecology, Canberra.
In summary, the effect of habitat structure on Cárcamo H. A., Niemelä J. K. & Spence J. R. (1995) Farming
trappability could involve effects of dilution, micro- and ground beetles: effects of agronomic practice on popu-
climate and response of the species to the physical lations and community structure. Can. Entomol. 127,
123–40.
layout of the habitat, acting independently or in
Clarke K. R. (1993) Non-parametric multivariate analyses of
concert. Well-designed experiments will be needed to changes in community structure. Aust. J. Ecol. 18, 117–43.
determine the most important mechanisms for different Cole E. C., McComb W. C., Newton M., Chambers C. L. &
taxa. Examples include simultaneous manipulation of Leeming J. P. (1997) Response of amphibians to clearcutting,
habitat structure and microclimate, simultaneous burning, and glyphosate application in the Oregon coast
manipulation of surface area and other attributes of range. J. Wildl. Manag. 61, 656–64.
Crist T. O., Guertin D. S., Wiens J. A. & Milne B. T. (1992)
habitat structure (to distinguish between the dilution
Animal movement in heterogeneous landscapes: an experi-
hypothesis and other mechanisms), and variation in the ment with Eleodes beetles in shortgrass prairie. Funct. Ecol.
size and shape of cleared areas (to determine the role 6, 536–44.
of diffusion processes). Such experiments may allow Crist T. O. & Wiens J. A. (1995) Individual movements and esti-
better generalisations to be made about the effect of mation of population size in darkling beetles (Coleoptera:
habitat structure on trappability for different taxa and Tenebrionidae). J. Anim. Ecol. 64, 733–46.
could point to modifications that would improve the Dennis P., Young M. R., Howard C. L. & Gordon I. J. (1997)
The response of epigeal beetles (Col.: Carabidae,
pitfall trapping method.
Staphylinidae) to varied grazing regimes on upland Nardus
stricta grasslands. J. Appl. Ecol. 34, 433–43.
Digweed S. C., Currie C. R., Cárcamo H. A. & Spence J. R.
ACKNOWLEDGEMENTS (1995) Digging out the ‘digging-in effect’ of pitfall traps:
Influences of depletion and disturbance on catches of
This research was funded by the ACT Parks and
ground beetles (Coleoptera: Carabidae). Pedobiologia 39,
Conservation Service with funds from the 561–76.
Commonwealth Endangered Species Program. Keith Faith D. P., Minchin P. R. & Belbin L. (1987) Compositional
Williams, Sarah Sharp and Kruno Kucolic provided dissimilarity as a robust measure of ecological distance.
much essential information for the choice of the study Vegetatio 69, 57–68.
site. Kendi Davies and Eric Melbourne provided assis- Fisher R. A. (1935) The Design of Experiments, 1st edn. Oliver
tance with field and laboratory work. I am grateful to and Boyd, Edinburgh.
Geiger R. (1965) The Climate near the Ground, 4th edn. Harvard
Steve Shattuck for assisting with the identification of
University Press, Cambridge, Massachusetts.
the ants, to Ross Cunningham for advice on aspects of Gomez D. M. & Anthony R. G. (1996) Amphibian and reptile
the statistical analysis, and to Julian Ash, Kendi abundance in riparian and upslope areas of five forest types
Davies, Penny Gullan, Peter Cranston and Ian Oliver in western Oregon. Northwest Sci. 70, 109–19.
for comments that greatly improved the manuscript. Greenberg C. H., Neary D. G. & Harris L. D. (1994) Effect of
high-intensity wildfire and silvicultural treatments on
reptile communities in sand-pine scrub. Conserv. Biol. 8,
1047–57.
REFERENCES
Greenslade P. J. M. (1964) Pitfall trapping as a method for study-
Abensperg-Traun M., Smith G. T., Arnold G. W. & ing populations of Carabidae (Coleoptera). J. Anim. Ecol. 33,
Steven D. E. (1996) The effects of habitat fragmentation and 301–10.
livestock-grazing on animal communities in remnants of gim- Greenslade P. J. M. (1973) Sampling ants with pitfall traps:
let Eucalyptus salubris woodland in the Western Australian digging-in effects. Insectes Soc. 20, 343–53.
wheatbelt. 1. Arthropods. J. Appl. Ecol. 33, 1281–301. Greenslade P. J. M. & Halliday R. B. (1983) Colony dispersion
Adis J. (1979) Problems of interpreting arthropod sampling with and relationships of meat ants Iridomyrmex purpureus and
pitfall traps. Zool. Anz. 202, 177–84. allies in an arid locality in South Australia. Insectes Soc. 30,
Andersen A. N. (1983) Species diversity and temporal distribu- 82–99.
tion of ants in the semi-arid mallee region of northwestern Hadden S. A. & Westbrooke M. E. (1996) Habitat relationships
Victoria. Aust. J. Ecol. 8, 127–37. of the herpetofauna of remnant buloke woodlands of the
Andersen A. N. (1995) A classification of Australian ant com- Wimmera plains, Victoria. Wildl. Res. 23, 363–72.
munities, based on functional groups which parallel plant Halsall N. B. & Wratten S. D. (1988) The efficiency of pitfall
life-forms in relation to stress and disturbance. J. Biogeogr. trapping for polyphagous predatory Carabidae. Ecol. Entomol.
22, 15–29. 13, 293–9.
Andersen A. N. & Reichel H. (1994) The ant (Hymenoptera, Honék A. (1988) The effect of crop density and microclimate on
Formicidae) fauna of Holmes Jungle, a rainforest patch in pitfall trap catches of Carabidae, Staphylinidae (Coleoptera),
 HA BI TAT S T R U CT U R E A N D P I T FA LL T R A P S 239

and Lycosidae (Araneae) in cereal fields. Pedobiologia 32, Perfecto I. & Snelling R. (1995) Biodiversity and the transfor-
233–42. mation of a tropical agroecosystem: ants in coffee plantations.
International Commission of Zoological Nomenclature (1985) Ecol. Applic. 5, 1084–97.
International Code of Zoological Nomenclature, 3rd edn. Pielou E. C. (1975) Ecological Diversity. Wiley, New York.
International Trust for Zoological Nomenclature in associ- Punttila P., Haila Y., Niemelä J. & Pajunen T. (1994) Ant
ation with British Museum (Natural History), London. communities in fragments of old-growth taiga and managed
Joose E. N. G. & Kapteijn J. M. (1968) Activity-stimulating surroundings. Ann. Zool. Fenn. 31, 131–44.
phenomena caused by field-disturbance in the use of pitfall Samways M. J., Caldwell P. M. & Osborn R. (1996) Ground-
traps. Oecologia 1, 385–92. living invertebrate assemblages in native, planted and
Luff M. L. (1975) Some features influencing the efficiency of invasive vegetation in South Africa. Agric. Ecosyst. Environ.
pitfall traps. Oecologia 19, 345–57. 59, 19–32.
Magurran A. E. (1988) Ecological Diversity and its Measurement. Sneath P. H. A. & Sokal R. R. (1973) Numerical Taxonomy.
Croom Helm, London. W. H. Freeman and Company, San Francisco.
Margules C. R. (1993) The Wog Wog habitat fragmentation Southwood T. R. E. (1978) Ecological Methods, 2nd edn.
experiment. Environ. Conserv. 19, 316–25. Chapman and Hall, London.
May R. M. (1975) Patterns of species abundance and diversity. Spence J. R. & Niemelä J. K. (1994) Sampling carabid assem-
In: Ecology and Evolution of Communities (eds M. L. Cody & blages with pitfall traps: the madness and the method. Can.
J. M. Diamond), pp. 81–120. Harvard University Press, Entomol. 126, 881–94.
Cambridge. Thiele H. U. (1977) Carabid Beetles in their Environments.
McCullagh P. & Nelder J. A. (1989) Generalised Linear Models, Springer Verlag, Berlin.
2nd edn. Chapman and Hall, London. Thomas D. B. & Sleeper E. L. (1977) The use of pitfall traps for
Melbourne B. A., Gullan P. J. & Su Y. N. (1997) Interpreting estimating the abundance of arthropods, with special refer-
data from pitfall-trap surveys: crickets and slugs in exotic and ence to the Tenebrionidae (Coleoptera). Ann. Entomol. Soc.
native grasslands of the Australian Capital Territory. Mem. Am. 70, 242–8.
Mus. Vict. 56, 361–7. Topping C. J. (1993) Behavioural responses of three linyphiid
Mitchell B. (1963) Ecology of two carabid beetles, Bembidium spiders to pitfall traps. Entomol. Exp. Appl. 68, 287–93.
lampros (Herbst) and Trechus quadristriatus (Schrank). II. Topping C. J. & Sunderland K. D. (1992) Limitations to the use
Studies on populations of adults in the field, with special of pitfall traps in ecological studies exemplified by a study
reference to the technique of pitfall trapping. J. Anim. Ecol. of spiders in a field of winter wheat. J. Appl. Ecol. 29, 485–91.
32, 377–92. Wallin H. (1985) Spatial and temporal distribution of some abun-
Morrill W. L., Lester D. G. & Wrona A. E. (1990) Factors dant carabid beetles (Coleoptera: Carabidae) in cereal fields
affecting the efficiency of pitfall traps for beetles (Coleoptera: and adjacent habitats. Pedobiologia 28, 19–34.
Carabidae and Tenebrionidae). J. Entomol. Sci. 25, 284–93. Whicker A. D. & Tracy C. R. (1987) Tenebrionid beetles in the
Niemelä J., Haila Y., Halma E., Pajunen T. & Punttila P. (1989) shortgrass prairie: daily and seasonal patterns of activity and
The annual activity cycle of carabid beetles in the southern temperature. Ecol. Entomol. 12, 97–108.
Finnish taiga. Ann. Zool. Fenn. 26, 35–41. Wiens J. A., Crist T. O., With K. A. & Milne B. T. (1995) Fractal
Niemelä J., Halme E. & Haila Y. (1990) Balancing sampling effort patterns of insect movement in microlandscape mosaics.
in pitfall trapping of carabid beetles. Entomol. Fenn. 1, 233–8. Ecology 76, 663–6.