Content uploaded by Weihong Ji
Author content
All content in this area was uploaded by Weihong Ji
Content may be subject to copyright.
45
WEIHONG, VEITCH and CRAIG: EFFICIENCY OF RODENT TRAPPING
New Zealand Journal of Ecology (1999) 23(1): 45-51 ©New Zealand Ecological Society
JI WEIHONG
1
, C.R. (DICK) VEITCH
2
and JOHN L. CRAIG
1
1
School of Environmental and Marine Sciences, University of Auckland, Private Bag 92019, Auckland, New Zealand.
E-mail: j.weihong@auckland.ac.nz
2
Department of Conservation, Private Bag 68-908, Newton, Auckland, New Zealand.
AN EVALUATION OF THE EFFICIENCY OF RODENT
TRAPPING METHODS: THE EFFECT OF TRAP
ARRANGEMENT, COVER TYPE, AND BAIT
__________________________________________________________________________________________________________________________________
Summary: Eradication of rodent species from some offshore islands has proved to be an effective means of
conserving native animal communities and restoring natural ecological processes on the islands. As methods
of eradication differ for different rodent species, a truthful monitoring method to detect species presence and
relative density is essential for a successful eradication programme. This study compared two spatial
arrangements (line vs. grid), 5 different baits (chocolate, cheese, soap, wax, oiled wood) and 3 cover types
(transparent plastic, wire netting, galvanised iron) on the detection of 2 species of rodents on Browns
(Motukorea) Island in June and August.
The two species of rodents present on the island were Norway rats (Rattus norvegicus) and mice (Mus
musculus). Trapping using conventional trapping lines and trapping grids was carried out in June and August,
respectively. The traps were set for 8 nights for both lines and grids. Trap lines caught 12.40 rats per 100
corrected trap-nights (100 ctn
-1
) and no mice; trap grids caught 5.3 rats 100 ctn
-1
and 0.2 mice 100 ctn
-1
. Trap
grids appeared to be better than trap lines for detecting the presence/absence of rodent species when two
species coexist and one appears subordinate to the other. On trap lines the trapping rate of rats was
consistently high for five of the first six nights. On trap grids the trapping rate was variable on all nights with
the first mice being caught on the third night. Three-night trapping sessions, conventionally used by most
researchers, should be reliable for testing the relative densities of numerically dominant species but may not
detect all rodent species present.
Of five different bait types tested for monitoring rodent presence, the preferred order was chocolate,
cheese, soap, wax and oiled wood. The efficiency of covers made of different materials, (galvanised iron,
plastic, and wire netting) was also tested. Wire netting covers had the highest trapping rate and galvanised
iron covers had the lowest. Three blackbirds (Turdus merula) were caught under wire netting covers,
indicating a risk to non-target organisms.
__________________________________________________________________________________________________________________________________
Keywords: Norway rats; Rattus norvegicus; mice; Mus musculus; trap line; trap grid; trap cover; bait.
Introduction
Introduced rodent species have been a major factor
in the reduction or extermination of native animal
populations in New Zealand (Atkinson, 1978;
Whitaker, 1978; Moors, 1983, Bremner, Butcher and
Patterson, 1984; McCallum, 1986; Miller and Miller,
1995), and continue to suppress or threaten native
fauna. Among the four rodent species introduced to
New Zealand, kiore (Rattus exulans Peale) is now
found mainly on offshore islands, while the other
three, Norway rats (R. norvegicus Berkenhout), ship
rats (R. rattus L.) and mice (Mus musculus L.) are
widely distributed on the mainland and many islands
(Clout, 1980; Efford, Karl and Moller, 1988;
Dowding and Murphy, 1994). Some islands have
only one rodent species present (Bettesworth, 1972;
Taylor and Thomas, 1989,1993), while others have
been invaded by more than one species (Hickson,
Moller and Garrick, 1986). Determining which
rodent species are present on an island in New
Zealand typically relies on the setting of a line of kill
traps with galvanised iron or mesh covers, often
alternating mouse and rat traps (Cunningham and
Moors, 1993), and is usually run for three nights.
Eliminating rodents from islands in New
Zealand is now common practice, with 58
eradications successfully completed, 26 awaiting
confirmation of success and only two recorded as
failures (Veitch, 1995). The method of eradication
varies for different rodent species (Veitch and Bell,
1990). As a consequence, correct assessment of the
presence, distribution and relative density of rodent
species is essential for successful eradication.
Mice were known to infest Browns Island prior
to the arrival of Norway rats in the late 1980s
(J. L. Craig and C. R. Veitch pers. obs.) although
their arrival date is not known. At the start of the
46 NEW ZEALAND JOURNAL OF ECOLOGY, VOL. 23, NO. 1, 1999
study it was not known if the invading Norway rats
had caused a decrease in or eliminated the mouse
population. Studies have shown that the presence of
larger, and presumably more dominant, rat species
may reduce the trappability of mice (Brown et al.,
1996). In this study, we set out to test different
methods of rodent capture and monitoring on
Browns Island where more than one rodent species
was believed to be present.
Methods
Browns Island, or Motukorea, is located in the
Hauraki Gulf about 10 km east of central Auckland
City, and has an area of 60 ha (Fig. 1). The northern
and eastern part of the island is dominated by a
volcanic cone rising to 68 m above sea level. The
southern and western part is flat and extends to
within 1600 m of the adjacent mainland.
The island has been grazed by cattle and sheep
for more than 100 years. Most of the cover is now
grassland dominated by kikuyu grass (Pennisetum
clandestinum)
1
except around the upper parts of the
volcanic cone where the grass Microlaena stipoides
is predominant. Patches of trees remain on the steep
north-western, northern and north-eastern sea cliffs.
These comprise a canopy dominated by pohutukawa
(Metrosideros excelsa) with a mixed understory of
mainly exotics such as Rhamnus alaternus and
Chrysantemoides monilifera and scattered natives
such as Melicytis ramiflorus, Pseudopanax lessonii
and Myoporum laetum. There are also small patches
of rocky areas covered by Muehlenbeckia australis
vines. Such areas were mainly distributed at the foot
of the volcanic cone. Esler (1974) recorded the
presence of 132 plant species, comprising 54 native
and 78 exotic species.
Rodent capture
For all the operations described in this paper two types
of trap were used: “Ezeset” wooden rat snap traps and
“Ezeset” wooden mouse snap traps. All traps were
baited with a rolled oats/peanut butter mixture. All
traps were checked and bait renewed daily.
Two different methods of trapping, (a) trap lines
and (b) trap grids, were used to assess the best way
of detecting the presence of rodents.
a) Trap lines
Eight trap lines were established in three different
habitats around the island (Table 1; Fig.1). Trap
lines were operated for a total of 464 trap-nights for
each trap type from 12 to 23 June 1995. Rat and
mouse traps were set alternately along each trap line
at 25-m intervals, giving a 50-m interval between
traps of the same type. Seven mouse traps and three
rat traps were in areas covered by Muehlenbeckia
australis vines. Details of each line are given in
Table 1. Trapping success is expressed as number of
captures per 100 corrected trap-nights (100ctn),
which accounts for the non-availability of sprung
traps (Nelson and Clark, 1973). The trapping results
for different habitats and covers were analysed with
Maximum Likelihood ANOVA (SAS Institute,
1990) since the data were not normally distributed.
b) Trap grids
Two trap grids of 11 x 11 trap sites covering a total
area of 12.25 ha, or 20.4% of the total area of the
island (Fig. 1), were established in early August
1995. Traps with wire-netting covers were placed
with a single mouse trap and rat/mouse pairs
alternately at 25-m intervals on every second row of
the grid, with single mouse traps at each site of the
alternate rows. This resulted in each grid containing
25 rat traps and 121 mouse traps. Sixteen mouse
traps and 4 rat traps were in areas covered by
Muehlenbeckia australis vines. The remainder of the
grid areas was grassland. Traps were operated for
eight days, resulting in a total of 1936 trap-nights for
mouse traps, and 400 trap-nights for rat traps.
Figure 1: Browns Island showing trap lines and trap grids.
______________________________________________________________
1
Botanic nomenclature follows Allan (1961) and Connor
and Edgar (1987).
47
WEIHONG, VEITCH and CRAIG: EFFICIENCY OF RODENT TRAPPING
The end of the trap line trial and the beginning
of the trap grid trial were separated by an interval of
just over one month so that we can assume little
direct interference between the two sampling
sessions.
Trap covers
Three types of trap cover were used: clear plastic,
galvanised iron and wire-netting. Covers were used
in sequence so that the first rat and first mouse traps
had clear plastic, the next rat and next mouse traps
had galvanised iron and the next rat and next mouse
had wire netting, and so on. Trap covers are
routinely used in poisoning operations to reduce the
risk of capturing non-target animals such as birds.
Bait preferences
Five bait types were tested to determine which was
most preferred and hence most useful for monitoring
the presence of rodents. These included cheese
(Anchor tasty), cooking chocolate, candle wax, soap
(Sunlight) and soybean oil on wood.
The baits were cut into approximately 1-cm
cubes and were held by U-shaped wires to the edge
of a 5 cm x 5 cm x 1 m wood plank. Ten cubes of
each type were positioned, with 1-cm gaps between
each cube, in a repeating order along the plank. Two
such bait sample blocks were set up and used at a
total of 10 different locations over five nights, from
15 to 19, June 1995. Baits were recorded as either
not touched, chewed or eaten. All baits were
replaced each day. The results were analysed using a
Maximum Likelihood ANOVA to detect any
preference for bait types.
Results
Trap lines
The 464 trap-nights for each trap type returned a
total capture of 51 rats and no mice. Full details of
the capture rates and numbers of sprung traps under
different cover types and in different habitats are
given in Table 2. Numbers of captures per 100 ctn
-1
were similar for five of the eight nights but lower
and varied considerably in the other three (Fig. 2,
2
=15.22; d.f.=8; P<0.05). The average number of
rats trapped was 12.40 (S.E.= 1.90) rats per 100 ctn.
There was no significant difference in trapping rates
of rats between the different habitats (
2
= 4.11;
d.f.=2; P>0.05) (Fig. 3).
Table 1: Numbers of rat and mouse traps under different cover type and in different habitat types on the trap lines.
__________________________________________________________________________________________________________________________________
Number of covers
_____________________________________________________________
Trap line Number of traps Plastic Galvanised iron Wire netting Nights
____________________________________________________________________________________
Habitat type number Rat Mouse Rat Mouse Rat Mouse Rat Mouse operated
__________________________________________________________________________________________________________________________________
Bush 2,3,6 17 17 6 6 5 5 6 6 8
Hill grassland 4,7 23 23 7 7 8 8 8 8 8
Flat grassland 1,5,8 18 18 6 6 6 6 6 6 8
Total traps 58 58 19 19 19 19 20 20
Total trap nights 464 464 152 152 152 152 160 160
__________________________________________________________________________________________________________________________________
Table 2: Numbers of rat captures and sprung traps under different covers and in different habitats along the trap lines.
__________________________________________________________________________________________________________________________________
Cover Habitat
__________________________________________________________ _____________________________________________________________
Plastic Galvanised iron Wire netting Bush Hill Flat grassland
Rats Sprung Rats Sprung Rats Sprung Rats Sprung Rats Sprung Rats Sprung
Night caught traps caughts traps caught traps caughts traps caught traps caughts traps
__________________________________________________________________________________________________________________________________
12221334 41121
23133213 22231
33221243 33212
41304350 43117
52312522 33034
63021514 04121
72102323 20023
80100211 11100
Total 16 13 10 14 25 19 20 19 17 8 14 19
__________________________________________________________________________________________________________________________________
48 NEW ZEALAND JOURNAL OF ECOLOGY, VOL. 23, NO. 1, 1999
Trap grids
The total capture on the two grids was 21 rats and 5
mice (Table 3). Capture rates were highest (close to
10.0 rats 100 ctn
-1
) on nights 2,3, and 7, whereas
capture rates on the other nights did not exceed 7.00
rats 100 ctn
-1
(Fig. 4). Mice were first caught on the
third night of trapping. All mice were caught in the
centre of the grid whereas rats were more frequently
caught on the other lines of the grid. The mice were
all caught in small rocky areas covered with
Figure 2: Mean (± S.E.) nightly rat capture rates100 ctn
-1
on the trap lines with all three habitats combined (n= 3).
Table 3: The number of trap nights, sprung traps and
uncorrected captures of rats and mice on both trapping
grids combined.
______________________________________________________________
Trap nights Sprung traps Captures
Day Rat Mouse Rat Mouse Rat Mouse
______________________________________________________________
150242 6 8 1 0
250242 7 18 4 0
350242 10 15 4 2
450242 8 18 2 1
550242 5 19 2 1
650242 5 10 1 0
750242 3 6 4 1
850242 4 5 3 0
Total 400 1936 48 99 21 5
______________________________________________________________
Figure 3: Mean (±S.E.) rat capture rates 100 ctn
-1
of eight
days (n = 8) along the trap lines in different habitats.
Muehlenbeckia australis. The average capture rates
were 6.09 rats 100 ctn
-1
(S.E. = 1.08) and 0.2 mice
100 ctn
-1
(S.E.= 0.08).
Trap covers
Rat trapping success varied significantly with the
different cover types used on the trap lines (
2
=
7.47;d.f. = 2; P<0.05). The descending order of
success for the different covers was wire-netting,
clear plastic and galvanised iron (Fig. 5). Wire
netting covers had a significantly higher catch rate
than galvanised iron covers (S=47.5, P<0.05,
Wilcoxon’s rank sum test). The only non-target
animals caught were three blackbirds (Turdus
merula L.), all caught under netting covers.
Figure 4: Mean (±S.E.) captures rates100 ctn
-1
of rats and
mice on the two grids (n= 2).
Figure 5: Mean (±S.E.) rat capture rates 100 ctn
-1
of eight
days under different cover types, (n = 8).
Baits
All five bait types showed some feeding by rodents.
Some cubes were completely eaten while others were
chewed (Fig. 6). The descending order of preference
(chewed + eaten) was cheese, chocolate, soap, wax
and oiled wood (
2
=117; d.f.=4; P<0.001).
49
WEIHONG, VEITCH and CRAIG: EFFICIENCY OF RODENT TRAPPING
Discussion
Rats were trapped in similar numbers in each of the
three different habitat types on the island. The
forested habitats were small and rats probably
moved to and from the adjacent grasslands. The
mice were all caught in two small rocky areas
covered by dense Muehlenbeckia australis vines,
which probably provided sheltered sites inaccessible
to Norway rats. The presence of mice in dense
ground cover microhabitats in areas containing
another dominant rodent species has been reported
before (Brown et al., 1996). During 928 trap-nights
along trap lines in June, with equal numbers of traps
set for rats and mice, 56 rats but no mice were
caught. On Motuihe Island a reverse situation
occurred where rats were not caught on trap lines,
while the trapping rate for mice was high (4.45 mice
100 ctn
-1
; Stubbs, 1996). Subsequent trapping on
Motuihe Island confirmed the presence of rats (C.
R.Veitch, unpubl. data). On Browns Island, both rats
and mice were caught in two trapping grids operated
a month later, during August. One possible
explanation for catching mice during the grid session
is that a trapping grid, which is likely to remove a
large proportion of the dominant species from the
centre of the grid, is more effective than the
traditional trap line for determining the presence of a
subordinate rodent species, particularly when the
subordinate species is present in much smaller
numbers (Brown et al., 1996). However, in the
present study, mice began to appear in the traps
before the rat trapping rates had decreased.
Alternative explanations for catching mice on
grids but not on lines are firstly, that although there
was a higher proportion of mouse traps in areas with
Muehlenbeckia australis on trap lines (12%) than on
trap grids (7%), trap grids are more likely to cover a
higher proportion of such microhabitats. Secondly,
mouse abundance and /or trappability changed more
between the two trapping periods than did that of
rats. However, in two previous New Zealand studies
on mouse population ecology by kill trapping, in
Woodhill State Forest near Auckland (Badan, 1979)
and Mana Island near Wellington (Efford et al.,
1988), mouse numbers started to increase after a low
in summer and reach a peak in late autumn and early
winter, declining afterwards. This suggests that, in
our study, when the trapping grids were operated in
August (late winter), mice numbers were unlikely to
be higher than when the trap lines were set in June
(early winter). Thirdly, small difference in average
nightly mouse trapping rates between the two trials
(0 mice 100 ctn
-1
vs. 0.2 mice 100 ctn
-1
) may have
been an artefact of the limited trapping effort.
The greatest catch along trap lines occurred
during all except night four of the first six nights of
trapping. This result suggested that the commonly
used three-night trapping session (Cunningham and
Moors, 1993) should allow an adequate measure of
relative rodent densities. However, it may not
demonstrate the presence of a less abundant rodent.
The highest trapping rate for rats within the trap
grids occurred on the second and third nights,
whereas the first mice were not caught until the third
night. Hence the standard three-night session may
not confirm the presence of small populations of a
second rodent (in this case mice).
Future trapping to determine the presence and
absence of rodents needs to take into account the
inter-relationships between species when setting
trapping regimes. The operation of traps, especially
in grids, for periods exceeding three days is
recommended as the most effective regime. In this
study, the operation of trap grids for periods
exceeding three days was better than trap lines for
detecting mice. Further conclusions cannot be drawn
from this study because of small sample sizes and
lack of replication. However, our results indicated
that the accuracy of existing information on the
distribution of rodents, based on trap lines run for
three days, could be questionable.
Traps set under three different trap covers
returned significantly different catches. Traps with
wire netting covers had a catch rate 2.5 times higher
than the ones with the traditionally used galvanised
iron covers. However, wire netting covers did not
deter blackbirds. This catch of non-target species,
potentially disastrous in areas with endangered
species, made clear recommendation regarding
optimal cover-type difficult.
At permanent bait stations, five different bait
types tested showed that Norway rats had clear
preferences. The natural longevity of each bait type
Figure 6: Preference of rodents for different baits
expressed as mean percentage of the bait of each type
present (±S.E.), (n = 10 for each bait type).
50 NEW ZEALAND JOURNAL OF ECOLOGY, VOL. 23, NO. 1, 1999
should also be considered before use in monitoring.
The most preferred bait types, cheese and chocolate,
decay or melt rapidly if not eaten by rats. Soap, the
next most preferred item, would be longer lasting if
kept dry in a bait station, but waterproof galvanised
iron covers reduced rat interaction with the bait
inside. Candle wax and oiled wood are the most
durable baits but were also the least favoured ones
in this study.
Although the study showed that the
conventional methods of rodent trapping and
monitoring may be improved, more experiments are
required before clear recommendations can be
made. We suggest that similar trials in different
areas, independent tests of efficiency of the different
trap covers, and independent tests of bait preference
would improve methods of rodent monitoring.
Acknowledgements
We thank Sandra Anderson for her assistance
throughout the study, and Athol Gardner, Sarah
Mackinnon, Ian Fraser, Berry Green and Renee
Grove for their help with the field work. We
appreciate the help from Dianne Brunton for advice
on statistical analysis and John Ogden and Ewen
Cameron for identification of plants. Kay
Clapperton, Colin O’Donnell, Josh Salter and an
anonymous referee provided considerable assistance
with earlier drafts.
References
Allan, H. H. 1961. Flora of New Zealand, Volume 1.
Government Printer, Wellington, N.Z. 1085 pp.
Atkinson, I.A.E. 1978. Evidence for effects of
rodents on the vertebrate wildlife of New
Zealand islands. In: Dingwall, P.R.; Atkinson,
I.A.E.; Hay, C. (Editors), The ecology and
control of rodents in New Zealand nature
reserves, pp. 7-30. Department of Lands and
Survey Information Series No 4, Wellington,
N.Z.
Badan, D. 1979 (unpublished). The ecology of mice
(Mus musculus L.) in two forests near
Auckland. MSc thesis, University of Auckland,
Auckland, N.Z. 175 pp.
Bettesworth, D.J. 1972 (unpublished). Aspects of the
ecology of Rattus norvegicus on Whale Island,
Bay of Plenty, New Zealand. MSc thesis,
University of Auckland, Auckland, New
Zealand. 60 pp.
Bremner, A.G.; Butcher, C.F.; Patterson, G.B. 1984.
The density of indigenous invertebrates on three
islands in Breaksea Sound, Fiordland, in
relation to the distribution of introduced
mammals. Journal of the Royal Society of New
Zealand 14: 379-386.
Brown, K.P.; Moller, H.; Innes J.; Alterio, N. 1996.
Calibration of tunnel tracking rates to estimate
relative abundance of ship rats (Rattus rattus)
and mice (Mus musculus) in a New Zealand
forest. New Zealand Journal of Ecology 20:
271-275.
Clout, M.N. 1980. Ship rats (Rattus rattus L.) in a
Pinus radiata plantation. New Zealand Journal
of Ecology 3: 141-145.
Connor, H. E., Edgar, E. 1987. Name changes in the
indigenous New Zealand flora, 1960-1986 and
Nomina Nova IV, 1983-1986. New Zealand
Journal of Botany 25: 115-170.
Cunningham, D.M.; Moors, P.J. 1993. Guide to the
identification and collection of New Zealand
rodents. 2nd edition. Department of
Conservation, Wellington, N.Z. 16 pp.
Dowding, J.E.; Murphy, E.C. 1994. Ecology of ship
rats (Rattus rattus) in a kauri (Agathis australis)
forest in Northland, New Zealand. New Zealand
Journal of Ecology 18: 19-27.
Efford, M.G.; Karl. B.J.; Moller, H. 1988.
Population ecology of Mus musculus on Mana
Island, New Zealand. Journal of Zoology
(London) 216: 539-563.
Esler, A.E. 1974 (unpublished). Browns Island.
Botany Division, Department of Science and
Industrial Research, Auckland, N.Z. 12 pp.
Hickson, R.E.; Moller, H.; Garrick, A.S. 1986.
Poisoning rats on Stewart Island. New Zealand
Journal of Ecology 9: 111-121.
McCallum, J. 1986. Evidence of predation by kiore
upon lizards from the Mokohinau Islands. New
Zealand Journal of Ecology 9: 83-88.
Miller, C.J.; Miller, T.K. 1995. Population dynamics
and diet of rodents on Rangitoto Island, New
Zealand, including the effect of a 1080 poison
operation. New Zealand Journal of Ecology 19:
19-27.
Moors, P.J. 1983. Predation by mustelids and
rodents on eggs and chicks of native and
introduced birds in Kowhai Bush, New Zealand.
Ibis 125: 137-154.
Nelson, L.; Clark, F.W. 1973. Correction for sprung
traps in catch/effort calculations of trapping
results. Journal of Mammalogy 54: 295-298.
SAS Institute 1990. SAS/STAT User’s guide, version
6. SAS Institute, Cary, North Carolina, U.S.A.
1028 pp.
Stubbs, A. 1996 (unpublished). Management and
recommendation for the ecological
rehabilitation of Motuihe Island. MSc thesis,
51
WEIHONG, VEITCH and CRAIG: EFFICIENCY OF RODENT TRAPPING
University of Auckland, Auckland, N.Z.
199 pp.
Taylor, R.H.; Thomas, B.W. 1989. Eradication of
Norway rats (Rattus norvegicus) from Hawea
Island, Fiordland, using brodifacoum. New
Zealand Journal of Ecology 12: 23-32.
Taylor, R.H.; Thomas, B.W. 1993. Rats eradicated
from rugged Breaksea Island (170ha),
Fiordland, New Zealand. Biological
Conservation 65: 191-198.
Veitch, C.R. 1995. Habitat repair: a necessary
prerequisite to translocation of threatened birds.
In: Serena, M. (Editor), Reintroduction biology
of Australian and New Zealand fauna, pp.
97-104. Surrey Beatty & Sons, Chipping
Norton, U.K.
Veitch, C.R.; Bell, B.D. 1990. Eradication of
introduced mammals from the islands of New
Zealand. In: Towns, D.R.; Daugherty, C.H.;
Atkinson, I.A.E. (Editors), Ecological
Restoration of New Zealand Islands. pp. 137-
143. Conservation Sciences Publication No. 2,
Department of Conservation, Wellington, N.Z.
Whitaker, A. H. 1978. The effects of rodents on
reptiles and amphibians. In: Dingwall, P. R.;
Atkinson, I. A. E.; Hay, C. (Editors), The
ecology and control of rodents in New Zealand
nature reserves, pp. 75-86. Department of
Lands and Survey Information Series 4,
Wellington, N.Z.
