Effectiveness of pitfallldrift-fence systems for
sampling small ground-dwelling lizards and frogs
in southeastern Australian forests
Garry A. Webb'"
'State Forests of New South Wales, P.O. Box 100, Beecmft, New South Wales 21 19
¶Present address: Sumitorno Chemical Australia Pty Ltd, 501 V~ctoria
Chatswood, New South Wales 2067
The effectiveness of various pilfall trap designs and pitfalVdrifi fence systems for sampling
small ground-dwelling lizards and frogs in the forests of southeastern Australia was examined.
Pitfallldrifl fence systems employing long drifl fences were more affective than short or
no-fence systems in terms of the number of individuals and species caught, but were time
consuming and caused considerable habitat disturbance. Various pitfall trap designs were
compared. Generally, simple open-necked and funnel traps were more effective than traps with
shelter and shelterldrifi arm traps. It is suggested that in forested habitats, groups of individual
open necked pitfall traps or short-fence systems may be just as effective and no more costly
in installation time than systems employing long drifi fences and more complex bap types.
Key words: Survey techniques, Pit-falls, Lizards, Frogs.
INTRODUCTION amphibians in heavily forested habitats.
Relative installation times for the drift fence
Few studies have compared the relative configurations were also assessed.
effectiveness of diierent pitfall and pitfalVdrift
fence systems for catching vertebrates in
Australian ecosystems, despite their wide- METHODS
spread use in survey and ecological research Comparative studies were carried out at
(Pengilley 1972; Cockburn et al. 1979; Davidge several locations in Bondi State Forest
1979; Mather 1979; Hopper 1981; Webb (37"09'S, 149"OYE) near Bombala (New South
1981, 1983, 1985; Menkhorst 1982; Bennett Wales) (hereafter referred to as Bondi) and
et al. 1989). As several recent authors, e.g., at Glenorie (formerly Maroota State Forest)
Braithwaite (1983), Friend (1984) and Morton (33"33'S, 150°59'E) near Sydney (New South
et a l . (1988) have suggested, the choice of Wales) (hereafter referred to as Maroota).
pitfallidrift fence design and configuration Vegetation a t Bondi is tall open and open
has been largely a matter of logistics and eucalypt forest (sensu Specht et a l . 1974),
personal preference, pointing to a need comprising mainly Eucalyptus obliqua L'Herit.,
for quantitative comparison of techniques in Eucalyptus viminalis Labill. and Eucalyptus
different ecosystems. radiata Sieber ex DC., and savannah woodland
Comparisons have been made between comprising mainly Eucalyptus paucfZora Sieber
diierent pitfawdrift fence systems for sampling ex Spreng., Eucalyptus steUulnta Sieber ex DC.
populations of small ground vertebrates (small and Eucalyptus dal~ympleana Maiden. At
mammals, reptiles and amphibians) in Maroota, the vegetation is "sandstone complex"
tropical, arid and semi-arid Australian habitats dry open eucalypt forest (sensu Specht et al.
(Cockburn et al. 1979; Mather 1979; 1974) comprising Angaphma costata (Gaertn.)J .
Longmore and Lee 1981; Braithwaite 1983; Britt., Eucalyptus eximia Schauer, Eucalyptus
Friend 1984; Morton et al. 1988; Friend et al. g u m m i k a (Gaertn.) Hochr., Eucalyptus piperita
1989). While pitfall/drift fence trapping has Sm. and Eucalyptus punctata DC.
proven to be a successful method for these
relatively open habitats, the use of drift fences PiYalZ trap design
in the heavily forested areas of eastern
All trap designs are illustrated in Figure 1.
Australia may be logistically impractical due to
Five basic designs (OPEN, FUNNEL,
rough terrain and the often heavily littered
forest floor. SHELTER, DRIFT ARM and SHELTER1
DRIFT ARM) and two traD sizes (5 and 10
Between 1979 and 1981 four trials were Litres) were lsed in the fAur trials. Funnels
conducted in eucalypt forest in southeastern were fitted snugly into the opening of the
Australia to assess the relative effectiveness bucket and the narrow neck of the funnel was
of various pitfall trap designs and drift removed to leave an opening of approximately
fence configurations for sampling reptiles and 90 mm. Shelters consisted of 50 cm x 50 cm
I 18 Australian Zoologist 3 ( )
11 June 1999
(S5, S10) (A5)
F i p r r I. Pitfall trap designs used during this study. Trap codes used in Tables 1 4 are shown in brackets.
June 1999 Australian Zoologist 31(1) 119
masonite sheets supported approximately 5 cm trap type F10 (Fig 1 ) Pitfall traps were
above the trap by large pebbles. Drift arms checked daily during the period 3 January
were small lengths of 3 mm thick plastic to 29 January, 1979 and the animals released
sheeting, 20 cm long and 5 cm high. Four off-site. For the remainder of the study,
were placed at 90" to each other, supported traps were filled with a formalin solution and
by small metal pegs with one edge slightly the contents emptied at one week intervals.
overlapping the lip of the bucket. Where I
shelters were used in conjunction with drift Trial 2: Bondi (30 January-23 March, 1979).
arms, the shelter was supported by the four Two pitfall-drift fence systems were compared:
1. 4 m drlft fence with a funnel pitfall trap
(F10) (Fig. 1) at either end.
Trial 1: Bondi (3 January-23 March, 1979).
Three pitfall-drift fence systems were com- 2. 4 m drift fence with a shelter pitfall trap
pared. Four replicates of each of the three (S10) (Fig. 1) at either end.
systems and $our vacant plots used for
Two replicates of each were assigned one to
observation only, were assigned one to each
of the sixteen 50 m x 50 m grid cells in a each of four 50 m x 50 m plots in a 1 ha
4 ha u
mid. The followinp vltfallidrift fence
trapvine svstems were used:
Drift fences were of identical desipn to
(a) one 40 m drift fence centred in the middle the 4 m drift fences in Trial 1. pitfall-traps
contained formalin preservative and were
of the plot, aligned diagonally across the
emptied and refilled weekly during the period
plot with one pitfall trap beneath each
end, of the study.
@) one 4 m drift fence centred in the middle Trial 3: Maroota (29 September-23 November,
of the plot, and aligned diagonally across 1979). Four pitfall trap designs (F, S, SA
the plot with one pitfall trap beneath each and 0) (Fig. 1) and two trap sizes (10 L
end, and 5 L plastic buckets) were compared.
No drift-fences were used in this- trial.
(c) two single pitfall traps spaced 17.5 m apart
Six replicates of each of the eight trap
and aligned diagonally across the plot.
tvDes were randomly allocated one to each
Drift fences were 50 cm high blue polyfabric df' the 48 10 m X' 10 m grid cells in a
and secured vertically to the ground with 0.48 ha grid. Traps were sunk flush with
metal pegs. A shallow trench was dug to the ground and soil and litter replaced
accommodate the bottom of the drift fence around the rims of the traps. Traps were
and soil and litter replaced along each side part-filled with formalin preservative and
of the drift fence. Pitfall traps were funnel emptied weekly.
Tobk 1. Frogs and lizards captured with three pitfalvdrift fence trapping systems in Bondi State Forest. 0 rn, 4 rn and
40 m represent fence lengths. Only significant differences are indicated. Values in the same mw with the same letter
are not significantly different (P > 0.05). F-values and probabilities are pmvided where P < 0.1. Values for FROGSPEC,
LIZSPEC and ALLSPEC are mean number of species with total number of species far that treatment in brackets.
0m 4m 40 m
Limnodynortcr peronii 0 0 1
Prcudophtync bibmnii 1' '
1 8b F = 21.0, p < 0.01
Ranidcllo signifera 0 0 I
FROGTOT 1' 1
' loD F = 16.2. p < 0.01
FROGSPEC 0.25' (1) 0.25' (I) I .OOb(3) F = 4.5, p < 0.05
E&mpn*r hrntwolci 5 4 7
Lampropholis guichcnoti 0 1 0
Sapro~inincurmwttlino 0 0 1
Nannorcincus m c c q i 3
' 9' 24b F = 5.79, p < 0.05
Nivrosrincur c m r m g i II 4 16
Psrudcnoin cntncarkowii 0 0 2
8 50b F = 5.78, p < 0.05
LIZSPEC 2.25(3) 2.75 (4) 3.50 (5)
ALLTOT 20' 19' 5Sb F = 8.74, p < 0.01
ALLSPEC 2.5 (4) .
3 0 (5) 4.5(8) F = 3.90, p = 0.06
120 Australian Zodogisl 31(1) June 1999
Table 2. Frogs and reptiles captured by different pitfall trap designs in Bondi State Forest. Values in the same row with
the same letter are not significantly different (P < 0.05). F-values and probabilities are provided where P < 0.1. Values
for FROGSPEC, LIZSPEC and ALLSPEC are mean number of species with total number of species far that treatment
~onide'lla'si~n@ra 3 0
Litorio verrcauxii 3 0
FROGTOT 34' Ib F = 1089.0, p < 0.001
FROGSPEC 4.5' (6) 0.5b (1) F = 32.0, p < 0.05
Nnnnorcincw marcqi 2 0
Niveosrincus coventri 14 5
LIZTOT 16 5 F = 9.31, p = 0.09
LIZSPEC 2 1
ALLTOT 40' 6b F = 96.8, p < 0.01
ALLSPEC 6.0 (8) 1.5 (2) F = 16.20, p = 0.06
Trial 4: Bondi (5 January30 March, 1981). chosen to assign t h e name I bibronii rather
Four pitfall trap designs were compared. than I! dendyi to these animals.
Five replicates of each of the four trap The following acronyms are used in the text,
designs were randomly allocated one to each tables and figures: ALLTOT (total individuals
of twenty 5 m x 5 m grid cells in a 500 m2 for all species combined), ALLSPEC (total
grid. Pitfall traps were all 5 L traps containing number of species), LIZTOT (total number
formalin preservative and sunk flush with of individual lizards), LIZSPEC (total number
the ground. Four trap designs were included of lizard species), FROGTOT (total number of
(A, F, SA, S) (Fig. 1). Traps were checked individual frogs), FROGSPEC (total number of
and emptied weekly during this period. frog species).
One or two-way analysis of variance (PROC
ANOVA - SAS 1987) was used to compare RESULTS
different pitfall-drift fence systems and pitfall Trial I : Drift fence length
trap types in the four trials. Analyses were
also carried out on raw counts, and on Significantly m o r e Pseudophryne bibronii,
rank, logarithmic (log 1) and square root FROGTOT; Nannoscincus mccoyi, LIZTOT and
(SQRT 0.5) transformed data. Plots of ALLTOT were captured in pitfallldrift fence
raw and transformed count data and residual systems using long (40 m) rather than short
data were examined using the Wilks-Shapiro (4 m) or no (0 m) drift fences (Table 1). For
statistic and equality of variance was checked Niueoscincus coventryi, fewer individuals were
using both Bartletts test and Cochrans Q caught in 4 m drift-fence systems than in
test (Statistix v. 4.1, Analytical Software 1994). 40 m or 0 m systems but this difference was
The Log + 1 transformation provided the not expressed as a significant difference. The
best dataset for analyses. A mean separation mean score for FROGSPEC was significantly
procedure (Duncans multiple range test - higher for the 40 m system than the 4 m or
SAS 1987) was used to compare means for 0 m systems. This trend of increasing mean
trials with more than two treatments. The scores with fence length was also apparent
null hypothesis in each trial was that each for ALLSPEC and LIZSPEC but not significant
pitfall-drift fence system or pitfall trap type at the 0.05 level. For ALLSPEC, the difference
would capture equal numbers of animals and between fence lengths approached significance
species. (p = 0.06).
Taxonomic treatment of lizards follows Trial 2: Shelter us funnel traps (with 4 m drift
Cogger (1992). The Pseudophryne sp. present fence)
in those trials in Bondi State Forest showed
only a limited degree of colour development Significantly fewer L. peronii, I bibronii,
in the groin and arm-pits. I have therefore FROGTOT, FROGSPEC and ALLTOT were
June 1999 Australian Zoolo~ist
LI'I LI'I 05'0 ~9-o ~9.1 0'2 66.7. 5'2 =!>ads lem!uv
SVS 6 el .O ( ~ 1 . 0 d)
= s.a '(09.0 = d) Z!S
' '(50.0 > d) u8!saa 5 9 0 Z P SI 8 S o108?nml o!alo~adn
0 0 0 0 0 0 0 I !?uoq?q #uOvdopn8y
zvs.S xo o
( ~ L ' = d) s ~ '(55.0 = d) ~ Z ! S' ( ~ 0 . 0> d) ~ + a a
~ B a q
a>ue>'j!u8!~ 5VS OIVS 5s 01s 53 Old 50 010 .%
adLj. d e q c
.sa>uaajj!p (50-0 > d) lue~'j!uB!s JOJ pap!no~d a e .uo!l.~edar ueaw .pap!nold a e aaxql 1Ie l o j sanlen-d '1.0 = d le
1ue>g!u8!r rem om1 aql jo uo!lJelalu! aql la az!r dell 'uS!sap dell alaqfi .s1q3exq u! luamleal, leql JOJ sa!>adr j a mqmnu [el01 ql!m sapads jo laqmnu ueam ale 3 3 d ~ l l pue v -?
33dSZ11 ' 3 3 d S 3 0 8 d JOJ ranlei\ '(su8!sap d e n jo sl!elap JOJ 1 axn8!d ass) ru>epaIl=qr = y~ 'mllaqs = s 'lauunj = 'uado = 0 -(sa~!q a w n p d e ~~ o (01 d o 5) laqrunu e pue
u!) ~j O,
sal!~daipue ~ 8 0 1 5 q q o ~
ad.41 d e n 103 mhualae e JO uo!,eu!quro= =!~amnueqdle ue se paleu8!sap a e radL1 ddeq. . s a l o j a w l s eloo.mw u! sad.41 dden flgl!d ,uaJajj!p u! p a r n d e ~ .€ 2
Table 4 . Frogs and Reptiles captured by different pitfall trap designs in Bondi State Forest. No significant
difference was found between trap types for any taxa. F-valuer and probabilities are provided where P < 0.1. Values
for FROGSPEC, LIZSPEC and ALLSPEC are mean number of species with total number of species for that treatment
Pitfall Trap Type
SA5 A5 S5 F5
Limodpnrtcs peronii 0 I 0 0
Prcudophtync bibrrmii 24 33 12 31
R n n h l l o rignijro 0 1 0 0
Litoria cwingii 0 0 1 0
Litorio vcrrcnu*ii 0 0 0 1
FROGTOT 24 35 13 32
FROGSPEC 1.0 (1) 1.2 (3) 1.0 (2) 1.2 (2)
Eulompnrs hentwolei 8 3 3 8
Lnmpmphollr pichenoti 1 3 1 I
Nonnorcincur mccoyi 0 3 0 I
Nivcascincur covenlryi 0 3 2 2
Pxudrmoia cntncartcouxii 4 7 3 6
Preudemoio spmceri 0 0 I 1
LIZTOT I3 19 10 19
LIZSPEC 1.2 (3) 2.2 (5) 1.8 (5) 2.0 (6)
ALLTOT 37 54
ALLSPEC 2.2 (4) 3.4 (8)
captured in 4 m pitfall/drift fence systems Trial 4: Pitfall hap types
with shelter traps (Trap type S10) than the
No significant differences were found between
same drift fence with funnel traps (Trap
trap types for a n y of the taxa examined
type F10) (Table 2). Mean counts for LIZTOT
(Table 4). For ALLSPEC, fewer species were
and ALLSPEC were also lower but not
caught in the shelterlarm traps (4 species)
significant at the 5% level of significance (both
than in the o t h e r three types (8, 7 and 8
P < 0.1). Fewer N. couentryi were captured in
species for the funnel, shelter and drift-arm
sheltered traps but this was not expressed as
a significant difference. traps respectively), although the difference
in mean species count was not significant.
T h e difference i n counts for ALLTOT was
Trial 3: Pitfall trap types
marginally significant ( p = 0.07) and can
Significant differences between pitfall trap be attributable to the difference between the
designs were found for Helieoporus australiacus, open type traps (funnel and arm) and the
Uperoleia laeuigata, FROGTOT, FROGSPEC, straight shelter t r a p .
ALLTOT and ALLSPEC (Table 3). All H.
australiacus were caught in funnel type traps DISCUSSION
(F5, F10). For U . leauigata, FROGTOT and
FROGSPEC shelter traps (S5, S10) were Drift fence length
clearly inferior to all other trap designs. Trap
T h e presence a n d length of drift fences
design did not significantly influence captures
have been shown t o be significant factors in
for any lizard taxa. LIZSPEC showed a trend
towards higher counts in open type traps (05, the capture efficiency of pitfall/drift fence
0 1 0 , F5, F10) than shelter type traps (S5, systems in arid and semi-arid Australia
(Morton et al. 1988; Friend et al. 1989)
S 10, SA5, SA10). For ALLTOT and ALLSPEC,
and elsewhere (Campbell and Christman
open traps (05, 0 1 0 ) were superior to shelter-
1982; Vogt and H i n e 1982; Bury and Corn
arm traps (SA5, SA10).
1987). However, the effectiveness of such
Trap size did not significantly affect capture systems in heavily forested habitats in
rates of any taxa. While there was no signifi- Australia has n o t been demonstrated. In
cant quantitative differences between trap the forests of t h e United States, short (ca.
sizes for ALLSPEC there was a qualitative < 5 m) fences provide effective sampling
difference in that only the larger traps of herpetofaunal communities and are
caught the larger lizard species, Amphibolu~us preferred over longer fences because of ease
diemensis, Pygopus lepidopw, Eulamprus tenuis of installation (Vogt and Hine 1982; Bury
and Tiliqua scincoides. and Raphael 1983).
Australian Zoologist 31(1) 123
In this study pitfall-drift fence systems (1983) suggested using small drift fences.
with 40 m drift fences caught significantly However, the results of this study suggest
more of some taxa than did those systems that, in the eucalypt forests of southeastern
employing 4 m or no fences. Unfortunately Australia at least, groups of individual pitfall
a threshold fence length could not he traps may be just as efficient as drift fence
adequately determined from these data. systems of any length. In Bondi S.F. (Webb
In other studies this threshold has been 1991a) and the nearby Coolangubra State
estimated at >15 m (Vogt and Hine 1982), Forest (Webb 1985, 1991b), groups of pitfall
2.5-5.0 m (Bury and Corn 1987) and 7-10 m traps have been successful in capturing all or
(Friend et al. 1989). most of the small ground-dwelling vertebrate
fauna known from these areas as well as
Clearly, 40 m drift fences may be well in detecting the occasional rare, cryptic, large or
excess of the threshold length required to arboreal species.
adequately sample herpetofaunal communities
but it remains to be determined what
the optimum length would be. For some Pitfall trap size
species, the 4 m and 40 m fence systems Size of pitfall trap (within the range
were not substantially different. However, examined) appeared to have little effect on the
for the nocturnally active frog I bibmnii and
! number of individuals and species captured.
crepuscular lizard N. maccoyi, the 40 m fence This differs from Morton et al. (1988) and
system was significantly better than the 4 m Friend et al. (1989) as they caught significantly
fence system. This suggests that diurnally and more animals in larger rather than smaller
nocturnally active animals may respond to pitfall-traps. In those studies the response of
the artificial barrier in different ways. Possibly some composite taxa and individual species
the colour (blue) or physical structure of the was less clear, but large traps were generally
drift fence may act as a deterrent for diurnally more efficient for most taxa.
At Glenorie, the larger lizards, A. dkmemis,
When installation time is considered, the I lepidopus, E. tenuis and T scincoides were
value of long (40 m) drift fences is question- captured only in the 10 L pitfall traps. How-
able. T h e ratio of average installation time ever, this bias was not expressed as a
for the three pitfallidrift fence systems used significant difference. Similarly, Morton et al.
here was 3.3:1.7:1 (for 40 m, 4 m and Om (1988) and Friend et al. (1989) found that
systems respectively). Thus, three 0 m systems smaller pitfall traps did not detect some of
(i.e., six separate pitfall traps) would require the larger species and generally caught fewer
less installation time than a single 40 m fence of the larger species. In an earlier study in
system. Using a crude numerical extrapola- North America, Banta (1957) found that deep
tion, the former would provide higher counts pitfall traps were necessary to capture the
for most taxa. However, for both N. maccoyi larger lizards in his study area. Vogt and Hine
and I! bibronii the single 40 m fence would (1982) also found large pitfall traps (19 L)
still be more efficient. were more effective than small (4 L) traps for
T h e longer fence was more effective for snakes, lizards and frogs.
some measures of species number (FROG- Thus, for generally small species, size of
SPEC and ALLSPEC). However, whether pitfall trap may not have a significant bear-
additional 4 m fences or pitfall traps would ing on capture rate, but trap size may be
compensate for this difference remains to be important i n detecting larger species.
Large array trapping systems like those used Pitfall trap design
by Campbell and Christman (1982) and
Morton et al. (1988) or simple, long, drift
fences (Friend 1984; Friend et al. 1989, this Sheltered pitfall traps are useful in arid
study) may be impractical for use in heavily environments in that they act as attractants
forested areas as a resource inventory tool to animals seeking shelter from extreme
because they are limited by unsuitable terrain, temperatures (Banta 1957; Williams 1968;
abundant natural obstacles such as logs Parker 1971). However, Read (1985) has
and other debris and by installation time suggested t h a t sheltered pitfall traps are of
(Bury and Raphael 1983). Bury and Corn little use i n capturing small mammals in
(1987) also noted the considerable ground arid Australia. Braithwaite (1983), working
disturbance caused by installation of long drift in tropical forest and woodland of northern
fences and suggested that a settling-in period Australia, found that sheltered pitfall traps
of approximately 30 days was required. In were less efficient than pitfall traps with an
forested North America, Bury and Raphael overlying 4 m drift fence. However, this
124 Australian Zoologist 31(1)
difference may be attributable to channelling REFERENCES
of animals by the drift fence rather than Analytical Software, 1994. Stolislix 4.1. An inleractivc
the design of the pitfall trap. T h e data rtatirlicol analnyrir program for microcomputers.
presented here from trials 2 and 3 indicate Analytical Software: St Paul, USA.
that sheltered pitfall traps are generally less Banta, B. H., 1957. A simple trap for collecting desert
efficient than un-modified open-necked pitfall reptiles. Hcrpetologica 13: 174-76.
traps, catching fewer individuals and fewer
Bennett, A. F., Schulz. M., Lumsden, L. F., Robertson,
species. Morrill (1975) has previously noted P and Johnson, P. G., 1989. Pitfall trapping of
this effect for ground-dwelling insects. small mammals in temperate forests of southeastern
Australia. Aust. Mammal. 12: 37-39.
Funnel traps Braithwaite. R. W. 1983. A comparison of two pitfall trap
systems. Kc. Nal. 100: 163-66.
How et al. (1984) found that funnel traps
were more effective than open traps in Bury, R. B. and Corn, P S.. 1987. Evaluation of pitfall
catching small mammals but not lizards. trapping in northwestern forests: trap arrays with
drift fences. J. Wildl. Monag. 51: 112-19.
This is presumably related to the ability of
small mammals to jump out of the wider Bury, R. B. and Raphael, M. G., 1983. Inventory
aperture open traps. In this study, funnels methods for amphibians and reptiles. Pp. ??-?? in
Renewable Rexource Inuenfori~sfor Moniloring Changes
were no more effective than open-necked and Trendr ed by J . F. Bell and T Atterbury. Proc.
pitfall traps for any taxa except H. australiacus International C o n f . Aug. 15-19, 1983, Carvallis,
(trial 3 ) . Given that the funnel should act Oregon. College of Forestry: Oregon State
as a barrier to escape from the bucket a University, Carvallis, Oregon.
greater capture would be expected for larger Campbell, H. W a n d Christman, S. P., 1982. Field
species. This was not the case, suggesting techniques for herpetofaunal community analysis.
that the funnel may obstruct the entrance Pp. 193-200 in He~petological Communities ed by
N . J. Scott. US Fish. Wildl. Serv. Wildl. Res. Rep.
to the bucket by allowing purchase on the 13. 239 Pp.
sides, or by being too small in aperture to
allow access. Cockburn, A., Fleming, M. and Wainer, J., 1979.
T h e comparative effectiveness of drift fence pitfall
trapping and conventional cage trapping of verte-
Dnft arm traps brates in the Big Desert, northwestern Victoria. Ec.
Nat. 96: 92-95.
The use of small drift arms to increase
the coverage of the pitfall traps was also no Cogger, H. G., 1986. Repfiles and Amphibians ofAuslralia.
Revised ed. Reed: Sydney.
more efficient than open-necked traps (funnel
traps - trial 4). When combined with shelter Davidge, C., 1979. A census of a community of small
lids, these drift arms were also less effective terrestrial vertebrates. Avst J . Ecol. 4 : 165-70.
than open-necked traps but provided slightly Friend, G. R., 1984. Relative efficiency of two pitfall-drift
improved counts over straight shelter traps fence systems for sampling small vertebrates. A w f .
for a number of taxa (trials 3 and 4). Zool. 21: 423-33.
Friend, G. R.. Smith, G. T., Mitchell, D. S. and Dickman,
C. R., 1989. Influence of pitfall and drift-fence
CONCLUSIONS design o n capture rates of small vertebrates in semi-
arid habitats of Western Australia. Aust. Wild. Res.
The data presented here, and that of others, 16: 1-10.
suggests that, for forested areas in south-
eastern Australia, groups of simple, open- Hopper, S. D., 1981. A p i t trap s u r v q of small mammals,
lizards a n d frogs o n the Two Peoples Bay Native
necked, large pitfall traps may be as, or more, Reserve. Dept. Firh. Wildl. West. A w l . Repl. No. 4 3 .
efficient than some of the more complex Pp. 1-21.
systems examined to date. However, long drift
How. R. A.. H u m o h r i e s . W F. and Dell.. ,. 1984.
fences may be more efficient for nocturnal ~
Vertebrate burbeys ~n seo~~i-ar,d
~ ~ ~
animals than shorter fences o r individual Pp. 193-96 in S u r q .Mtlho& /or .\'o:urr (.'dn,rnal>un
pitfall traps. Further experimentation on LI. I rd b) K Myer,, C . K . Marguler m d I . hfusto
drift fence length and density and spacing CSlRO Div. Water a n d Land ~erourcer: Canberra.
of pitfall traps would be useful, particularly Longmore, R. and Lee, P., 1981. Some observations on
for nocturnal or crepuscular species. techniques for assessing the effects of fire on reptile
populations in S t u r t National Park. A w t . J. Hcrp.
Mather, P. B., 1979. An examination of the reptile fauna
I thank the Forestry Commission of New of Wyperfeld National Park using pitfall trapping.
South Wales for their permission to conduct Kc. Not. 96: 98-101.
these trials and providing research facilities. Menkhorst, P W., 1982. Pitfall trapping of reptiles in the
Rod Kavanagh, Marcia Lambert and Peter Big Desert, Victoria. Kc. Not. 99: 66-70.
Lind kindly made constructive comments on Morrill, W. L.. 1975. Plastic pitfall trap. Enuimn. Enl.
earlier drafts. 4: 596.
Australian Zoologist 31(1) 125
Morton, S. R., Gillam, M. W., Jones, K. R. and Fleming, Webb, C. A,, 1981. Geographical distribution of
M. R., 1988. Relative efficiency of different pit-trap reptiles and amphibians in the southern Eden
systems for sampling reptiles in spinifex grasslands. forestry region. Fol: Comm. NSW Unpubl. Rep. No.
Aust. Wildl. Rcs. 15: 571-77. 783. Pp. 1-115.
Parker. W S., 1971. Influence of trap cover type on
. Webb, G. A,, 1983. Diet in a herpetofiunal community
pitfall trapping of Lizards. Hcrp. RN. 3: 94. on the Hawkerbury sandstone formation in the
Sydney region. Hcrpctofauna 14: 87-91.
PengiUey, R. K., 1972. Systematic relationships and
ecology of some lygosomine lizards from south- Webb, G. A,, 1985. Habitat use and activity patterns in
eastern Ausualia. 2 Volr. Unpubl. Ph.D. thesis, ANU, same southeastern Australian stinks. Pp. 23-30 in
Canberra. Biology of Awfrnlarinn Fmgr and Repfilcs ed by G.
Grigg, R. Shine and H. Ehmann. Surrey Beatty &
Read, D. G., 1985. Notes on capture techniques for small Sons: Chipping Norton, New South Wales.
mammals of the arid zone.Aust. Zoo[. 21: 545-50.
Webh, G. A,, 1991a. A survey of the reptiles and
SAS, 1987. SLnfistkal Annlyris System. SAS Institute: Galy, amphibians of Bondi State Forest and surrounding
North Carolina, USA. areas, near Bambala, New South Wales. Awt. Zool.
Specht, R. L., Roe, E. M. and Boughton. V H., .
1974. Conservation of major plant communities in Webh, G. A., 1991b. The effects of logging on
Australia and Papua New Guinea. Awl. J . Bot. Suppl. populations of small ground-dwelling vertebrates in
SF. No. 7. Pp. 1-667. montane eucalypt forest in south-eastern N w South
Wales. Unpubl. M.Sc. thesis, Australian National
Vogt, R. C. and Hine, R. L., 1982. Evaluation University, Dept Forestly, Canberra.
of techniques for assessment of amphibian and
reptile populations in Wisconsin. Pp. 201-17 in Williams. S. C., 1968. Methods of sampling scorpion
Herpetological Cmrnuniti8s ed by N. J. Sott, Jr. US I
populations. Proc. G $ A r d . Sci. 4th Series 36:
Fish. and Wildl. Service, Wildl. Res. Rep. No. 13. 221-30.
Ethel Mary Read Research Grants
In 1985 the Council of the Royal Zoological In 1998, eight $600 grants were awarded. These
Society of New South Wales moved to create a were:
small research grants fund targeting projects in Andrew Baird uames Cook University) - The
Australasian zoology. A scheme based on multiple
small grants was devised, the monetary principal .length of t h e larval phase of corals.
Belinda Cooke (University of Technology,
Sydney) - Behavioural and ecophysiological
for the fund came from a bequest to the Society
from Mrs Ethel Mary Read, a long-time supporter
of the Society. After much deliberation, it was
decided that t h e target group For these grants
.studies on t h e Regent Honeyeater.
Emma Cronin (University of Adelaide) -
Energetics of respiration and development of the
should be students or new workers who had yet to cephalopods Sepia apama and Sepwteuthu awtralir.
prove their expertise in zoological research. It was Matthew Crowther (Sydney University) -
Evolution, systematics and ecology of Antechinus
felt that established researchers had other avenues
stunrti-Antechinw flauipes complex.
for funding whereas novices in zoology were very
Sharon Downes (Sydney University) - Using
limited in sources of financial assistance. This
behavioural assays to investigate the mechanics
granting system would therefore be catering for of predator-prey coevolution in reptiles.
short-term projects which may be pilot studies o r Natasha Lebas (University of Western Australia)
research into unproven areas of study. Established - T h e m a t i n g system of the dragon lizard
researchers were not excluded From applying for Ctenophow o m t w .
funds but the small slze of the grants hmited Elizabeth Tasker (Sydney University) - Effect
their use, such as for bridging finance or to cover of management-induced differences in habitat
short-falls in larger projects. T h e first call for complexity on small mammal community
applications for t h e Ethel Mary Read Research StNCtUre.
Grants was announced in January 1986. This grant Dimitrios Zabaras (University of Western Sydney)
scheme is now well known. New grant applications - Chemical communication in insects of
are announced annually and the review committee economic significance.
receives applications from students and junior Arthur White
researchers in all parts of Australia. Council Member
126 Australian Zoologist 31(1)