Content uploaded by Sascha K Hooker
Author content
All content in this area was uploaded by Sascha K Hooker on Jun 26, 2014
Content may be subject to copyright.
New Zealand Journal of Zoology, 1995, Vol. 22: 291-304
0301-4223/2203-0291 $2.50/0' © The Royal Society of New Zealand 1995
291
Ranging behaviour of forest-dwelling ship rats, Rattus rattus,
and effects of poisoning with brodifacoum
SASCHA HOOKER
Department of Zoology
Pembroke College. Oxford University
Oxford. England*
'"Present address: Department of Biology, Dalhousie
University. Halifax. Nova Scotia. Canada B3H
4J1.
JOHN INNESt
Manaaki Whenua - Landcare Research
Private Bag 3052
Rotorua. New Zealand±
^Present address: Manaaki Whenua - Landcare
Research. Private Bag 3127, Hamilton. New
Zealand.
Abstract We radio-tracked five male and four fe-
male rats for 6 nights in primary forest at Rotoehu.
North Island. New Zealand. From trapping we esti-
mated rat density at the study site to be 6.2 rats/ha.
Radio-tracking revealed mean (+ SE) restricted poly-
gon home ranges to be three times greater in males
HA ± 0.29 ha) than females (0.3 + 0.04 ha). Male
ranges overlapped considerably, whereas those of fe-
males were largely exclusive. The ranges of males
encompassed several iemale ranges. Fourradio-col-
iared rats were retrapped and administered a lethal
dose of the anticoagulant poison brodifacoum. Dur-
ing the 3-5 nights after poisoning but before death,
we detected no significant change in home range area
or utilisation, arboreality. or movements. Further re-
search is required to determine if rats prey on other
fauna while fatally intoxicated or cause secondary
poisoning after being eaten by other predator spe-
cies.
Keywords ship rat: Rattus rattus; radio-tracking;
poisoning; brodifacoum; home range
'To whom all correspondence should be sent.
Received 22 June 1994; accepted 13 December 1994
INTRODUCTION
Two of the most important factors in the loss of is-
land fauna are habitat destruction (loss or fragmen-
tation) and the artificial introduction of exotic or
alien species (Moors 1985). One of
the
most destruc-
tive alien species is the ship rat or black rat, Rattus
rattus, the world-wide success of which stems from
its ability to disperse, its competitive superiority over
similar species in disturbed or secondary habitats,
and its ability to reproduce successfully in a wide
variety of habitats (Clark 1980). Endemic in the In-
dian subcontinent, the ship rat spread to Britain with
the Romans around 20 B.C. (Reumer 1986). Since
then, with humankind as its dispersal agent, it has
colonised six continents and thousands of islands,
including New Zealand (Atkinson 1985).
Today, ship rats are widespread and important
pests in New Zealand indigenous forests (Atkinson
1973.
1978; Innes 1990). Control operations have
covered areas ranging from <1 ha to reduce the risk
of rats boarding ships (Hickson et al. 1986) to
1400 ha to protect nesting kokako (Innes et al. in
press).
Behavioural data obtainable by radio-track-
ing (the amount of time the rats spend in trees, the
distances they move, their home range areas, and
social organisation) are essential prerequisites of
any
effective management strategy and may facilitate the
design of more efficient control operations.
In buildings and laboratory colonies, ship rats
occupy a group territory with a dominance-based
social hierarchy (Barnett 1967; Ewer 1971). How-
ever, little is known of the social organisation of
forest-dwelling ship rats, in New Zealand or else-
where. Nocturnal, arboreal, and alert, ship rats are
difficult to observe directly. Trapping, footprint-
tracking studies, and nest observations have indi-
cated that they are not colonial, but that individuals
or family groups are dispersed rather evenly through
available habitat (Daniel 1972; Innes & Skipworth
1983).
The first part of this study was an investiga-
tion of the home range size, social organisation, and
movement patterns of several male and female ship
rats,
using radio-tracking.
292
Our second aim was to investigate the effect of
the anti-coagulant brodifacoum on rat movements.
Rats that are fatally intoxicated with poison may still
prey on other fauna, or they may be eaten by preda-
tors which might themselves be poisoned as second-
ary kills. Many of the problems of commensal rat
population control are the result of neophobia
(Cowan 1977) and the social learning of diet
pref-
erence and aversion (Galef 1988). Control strategies
can minimise opportunities for social learning. Poi-
sons used for long-term control of
rat
populations are
therefore usually of the chronic type like brodi-
facoum, so that social transmission of bait aversion
is avoided. Brodifacoum is slow-acting, taking 3-7
days to cause death with a single dose.
By recapturing four of the radio-collared rats and
feeding them poison bait, we were able to track them
during the period after ingestion and before death
and to determine experimentally whether poisoning
altered the rats' movements.
New Zealand Journal of Zoology, 1995, Vol. 22
300
200
100
N
I
100
200
300 m
Fig.
1
Map of ship rat study site, Rotochu Forest, show-
ing central trapping grid and altitude contours (m a.s.l.).
Crosses in the central 68 x 80 m grid show cage-trap
positions.
METHODS
Study site
The 9 ha study site was located in tall (to 30 m) for-
est of tawa, Beilschmiedia tawa, and kohekohe,
Dysoxylum spectabile, within the current North Is-
land kokako research study area (37°58'S, 176°32'E)
in the Pongakawa Ecological Area of Rotoehu For-
est, North Island, New Zealand. (Plant nomenclature
follows Allan 1961 and Connor
&
Edgar 1987.) The
nearest weather station was formerly located at
Rotoehu Forest Headquarters, 5 km to the north.
Over the period
1941
-70,
mean annual rainfall was
1680 mm and mean temperatures ranged from 7.7°C
(July) to 17.8°C (February) (Leathwicket al. 1983).
To record the locations of rats, numbered pegs
marked with reflective tape were laid out in a 10 x
16.7 m grid system over the 300 x 300 m (9 ha)
study area.
Trapping and poisoning
Four trapping sessions were undertaken between 1
December 1991 and 21 January 1992. During the
periods 7-12 and 20-22 December, 24 rats were
live-trapped in 45 cage-traps set at the grid markers
in a rectangle of 9 x 5 stations covering 68 x 80 m
(0.54 ha) in the centre of the 9 ha study area, in or-
der to estimate population density and to attach ra-
dio-transmitters to particular rats (Fig. 1). The first
four adult males and four adult female rats to be cap-
tured were radio-collared. Subsequently, a further
adult male rat was captured and collared when one
of the original males (M5) could not initially be re-
located.
Each live-trapping session was preceded by 2
days of prebaiting to accustom the rats to using the
traps.
A carrot disk smeared with peanut butter was
used as bait. Traps were set during the afternoon and
cleared between 2400 and 0300 h NZST to reduce
the time rats spent in traps. Traps were also covered
with metal tunnels to further reduce stress on the
trapped rats. No trapped rats died during the study.
Captured rats were anaesthetised in a plastic bag
using cotton wool soaked in ether, then sexed,
weighed, and individually marked by toe-clipping.
Four adult female and five adult male rats were ra-
dio-collared.
Subsequently, between 3 and 8 January, two fe-
male and two male collared rats were retrapped and
fed brodifacoum poison ("Talon WB50", ICI
Tasman Ltd) in their cages between 2400 and
0100 h. They were then released and radio-tracked
(nights only) until they died.
At the end of the study (16-21 January), a kill-
trapping programme was carried out in order to es-
timate population density, using 45 Fenn Mk IV
traps and 52 wooden rat snap traps ("Ezeset
Supremes") laid out through the central 0.54 ha of
the study site. The traps were set during the after-
noon and cleared the following morning. This pro-
vided a minimum count of rats exposed to the
trapping site, from which an estimation of rat den-
sity was calculated. To check that all rats had been
Hooker & Innes—Ranging behaviour of ship rats
293
caught, baited tracking tunnels (King & Edgar 1977)
were placed throughout the grid and monitored dur-
ing the kill-trapping programme.
All traps and tracking tunnels were set on the
ground, since previous studies found that no rats
were entirely arboreal (Daniel 1972; Innes &
Skipworth 1983).
Radio-tracking techniques
While each rat was under light anaesthesia, an AVM
SS-1 transmitter (Biotrack U.K. Ltd, Wareham,
Dorset) was fitted around its neck with a non-release
nylon cable tie collar. Rats were released at the same
site after recovering full locomotor function. The
transmitter was generally detectable over 25-30 m,
but the range varied according to the rat's (and there-
fore the antenna's) orientation and to intervening
obstructions.
AVM receivers were used, together with a three-
element hand-held Yagi aerial. A radio-fix record-
ing the location of each rat to an accuracy of ±2 m
was taken every 10 min. The radio-fixes were taken
5-10 m away from the rat, at which distance no evi-
dence of disturbance was detected.
Four rats were radio-tracked each night from dusk
until dawn, for a minimum of
5
nights, except dur-
ing the poisoning study when each rat was tracked
until death (between 3 and 5 nights). Rats M2, M4,
F1,
and F2 were tracked between 15 and
21
Decem-
ber;
M1,
M3.
F3. and F4 were tracked on 22 and 26-
30 December and 3 January; M5 was tracked on 4
and 10-12 and
14
January;
M1,
M2,
F1,
and F2 were
poisoned and tracked between 4 and
11
January; and
the neighbouring non-poisoned females, F3 and F4,
were tracked on 9 and 12 January.
Data were analysed using the "Wildtrak" compu-
ter program on Macintosh computers (Todd 1992).
Arborealitx and nests
To obtain an estimation of arboreality, the approxi-
mate height of the rat above ground was recorded at
each fix. Heights were banded as 0-2,2-8, and >8 m.
Nest-sites were located to within 5 m
3
by radio-fix-
ing the position of the rat once it became stationary
at the end of the night.
Error estimation
We determined the accuracy of fixes in the field by
placing test transmitters in various unknown (to ob-
servers) positions and habitats throughout the study
area and taking fixes on these. The horizontal com-
ponent error averaged 2.5 m over 60 trials (standard
deviation 3.7 m; range 0-24.9 m).
The error check showed the vertical fix to be the
least accurate of
the
position data, with
10
of
60
fixes
placed in the wrong height band. Most inaccuracies
arose when the test transmitter was placed near the
2 m boundary between lower and middle bands.
Response of neighbours to poisoning
The removal of rats by poisoning was more likely
to effect changes in the ranges of adjacent females
rather than males, since previous work (Daniel 1972;
Innes & Skipworth 1983; Hickson et al. 1986) indi-
cated that female ranges were probably discrete,
whereas those of males were probably overlapping.
Therefore we radio-tracked two non-poisoned fe-
male rats for 2 nights during and immediately after
the poisoning study, to check for home range
changes following the poisoning of their neighbours.
Home range analysis
The home range is defined after Burt (1943) as the
area normally traversed by an individual during its
activities of food-gathering, mating, and caring for
young. Three non-statistical techniques (min. con-
vex polygons, restricted polygons, and grid cells)
were used to analyse various aspects of the home
range for each rat. These methods permit the use of
temporally autocorrelated fixes (fixes which are not
independent) when the number of fixes and the time
period over which they are taken are constant be-
tween ranges, or when the ranges are exhaustively
documented (Harris et al. 1990).
The time interval required for independence of
fixes was found for each rat by plotting values of
Schoener's Index for increasing time intervals
(Swihart & Slade 1985). This indicated that more
than an hour between successive fixes would be re-
quired for independence, and non-statistical meth-
ods of data analysis should be used.
These analyses allow for use of autocorrelated
data, but require that the data have been collected
over a sufficiently long time to obtain a representa-
tive sample. This can be checked visually by plot-
ting the percentage home range area against the
number of fixes (Voigt & Tinline 1980). Such analy-
ses (Fig. 2) showed that 80% of each range had been
described by about the 200th fix. All rats were there-
fore tracked for long enough to describe their home
range accurately.
Minimum convex polygons
The minimum convex polygon (MCP) method
(Mohr 1947) is still the most frequently used tech-
nique. The range boundary is constructed by drawing
294
New Zealand Journal of Zoology, 1995, Vol. 22
No.
fixes
Fig. 2 Mean % revealed home range area (restricted
polygon method) against number ofradio-lelemetry fixes
for male (A) and female (B) ship
rats.
Bars show standard
errors.
a line around the outermost locations recorded for
an animal. This limits its utility, since range size is
strongly influenced by peripheral fixes, and the
nominal range area may include large areas which
are never visited. The standard modification to over-
come this problem (White & Garrott 1990) is to
eliminate the "outliers" by calculating
95%
polygons
based on ordering criteria for the locations.
Restricted polygons
The restricted polygon (RP) method (Wolton 1985)
determines home range by joining up the outermost
of the set of fixes gathered for the animal, but with
the proviso that the maximum length of a line be-
tween two peripheral points is limited to the mean
distance between fixes and the arithmetic mean cen-
tre of activity. Some outlying fixes may not be
included into the polygon, but these fixes are treated
as temporary excursions.
Grid cells
The overlaying of grid cells (Siniff & Tester 1965)
is useful as a representation of spatial utilisation and
static interaction between individuals, but is less
useful for calculating home range area. The area over
which an animal has moved is dissected using a grid
of cells, or blocks. The number of animal fixes is
tabulated for each of these cells, and the sum of the
areas of cells containing locations is taken as the
estimate of home range area. The main problem with
this method is the disjointedness that arises with low
sampling intensity—an animal may pass through
many grid cells but never actually be recorded there.
A
technique to overcome this is to add influence cells
around the cells containing fixes and then to remove
(trim) those that do not border a certain proportion
of cells containing fixes. This fills in holes but does
not add to the outer boundaries of the range. A grid
cell resolution of
5
m, at least twice the accuracy of
the data, was used for this study.
In this study, the RP method was preferred for
quantitative home range estimation, but the MCP
method was also included for comparison with other
studies (as recommended by Harris et al. 1990). Grid
cells were used in analysis of home range utilisation.
Impact of poisoning on range
To investigate the impact of poisoning, we compared
the ranging behaviour of the four poisoned rats over
3 nights before poisoning with their behaviour for 3
nights afterwards. In this way the numbers of fixes
taken in the two periods were similar. The same com-
parison was also made for four non-poisoned rats,
which acted as a non-treatment group for compari-
son with poisoned rats. For this comparison, track-
ing dates "before" (Time 1) versus "after" (Time 2)
were 15-17 December versus 18-21 December for
M4;
22, 27-28 December versus 29, 30 December,
and 3 January for M3; 26-28 December versus 30
December and 3, 9 January for F3; and 22, 26, 27
December versus 28, 30 December and
3
January for
F4.
Change in home range overlap
The spatial overlap of two home ranges is termed
"static interaction" (Dunn & Gipson 1977). We in-
vestigated the home range overlap of the same rats
before and after poisoning. The grid cell method was
used to calculate the percentage of shared cells be-
tween the two ranges, and this provided an indication
Hooker & Innes—Ranging behaviour
of
ship rats
295
as
to
whether home range
was
significantly differ-
ent after poisoning.
The
home range sizes
(RP
method) were compared using paired /-tests
for
rats
before and after poisoning, and, for rats not poisoned,
at Time
1
and
Time
2.
Rat movement
The distance travelled per night was calculated as the
total distance between successive fixes.
The
mean
distance travelled per night was compared before and
after poisoning and,
for
rats not poisoned,
at
Time
1
and Time
2
using paired /-tests.
Home range utilisation
Grid cell plots (5 m. trimmed influences)
of
fixes
for
the four poisoned rats were compared before
and
after poisoning for changes in patterns
of
home range
RESULTS
Population density
Initial cage-trapping
Twelve adult male
and 12
adult female rats were
caught.
All
were
of the
frugivorus (agouti; white-
bellied) morph. Means
(±SD) of rat
weights were
155.9
± 20.9 g for
males (range 121-190 g),
and
141.4
±
22.9
g for
females (range 108-187 g).
Cri-
teria
of
maturity were perforate vagina
for
females
and scrotal testes
for
males.
Final kill-trapping
Twenty rats—eight adult females,
10
adult males,
and
two
juveniles—were kill-trapped after
the two
Table
1
Home range length, width, and average diameter
'.Av. D.) (m) (defined in text) for ship rats, Rotoehu
Forest, calculated from minimum convex polygon home
r
anges based on radio-tracking data taken between 15
December 1991 and3January 1992 for all rats except M5
i4-14 January).
Len- Av. Len- Av.
Males ath Width D. Females jith Width D.
Ml
M2
M3
M4
M5
Mean 194
SE 28
'97
!79
143
150
300
73
129
104
104
95
101
9
135
154
123
127
197
i47
13
Fl
F2
F3
F4
109
1
II
109
82
103
7
91
70
54
64
70
7
100
90
81
73
86
6
adult males and two adult females died from poison-
ing. The assumption that
all
rats resident on the grid
were caught
is
likely
to
have been true, since
all ra-
dio-collared rats except one male (which could
not
be located) were caught
in the
kill traps,
and no rat
tracks were detected in 45 baited tracking tunnels on
the grid
at
the end
of
trapping, although tracks were
recorded regularly beforehand.
To calculate ship
rat
population density
it is
nec-
essary
to
know
or
estimate
the
area covered
by the
counts.
The
area
of
exposure
to
trapping will differ
for males
and
females because males have bigger
home ranges
and so are
more likely
to
encounter
traps.
These areas
can be
calculated
by
adding
a
border
of
one-half
of
the mean home range "diam-
eter"
to the
trapping grid, representing
the
average
distance outside the grid included within the ranges
of
the
trapped animals (Dice 1938). Home range
"diameter'" assumes circularity
of
ranges,
so is in-
applicable
for
many
of
the ranges observed. We
es-
timated
it
by
a
parameter
Av.D..
the average
of
range
length and width. Range length
is
the longest possi-
ble straight line inside the range,
and
range width
is
the length
of
the
line
at
right angles to this and meas-
ured
at its
midpoint. Using MCP range dimensions,
Table 2 Home range areas (ha) for ship rats in Rotoehu
Forest before and after poisoning (RP = restricted polygon,
MCP = minimum convex polygon; M = male. F = female).
Pre-poison ranges were derived from radio-tracking data
taken from 15 December 1991 to 3 January 1992 for all
rats except M5 (4-14 January), and post-poison ranges
from data taken during 4-10 January.
Rat RP
MCP MCP
(100%) (95%)
No.
fixes
Pre-poison Ml
M2
M3
M4
M5
Mean (SE)
Fl
F2
F3
F4
Mean (SE)
Post-poison Ml
M2
Fl
F2
0.665
0.766
0.562
1.319
2.133
1.08
(0.29)
0.283
0.411
0.289
0.218
0.30
(0.04)
0.905
0.626
0.109
0.368
1.155
1.186
1.066
1.654
2.564
1.52
(0.28)
0.685
0.547
0.397
0.352
0.49
(0.07)
1.137
0.894
0.187
0.496
1.007
0.597
0.636
1.36
2.178
1.15
(0.29)
0.292
0.366
0.259
0.197
0.28
(0.03)
1.01
0.62
0.126
0.396
243
302
342
287
233
281
(20)
338
291
216
281
281
(25)
273
171
162
173
296
New Zealand Journal of Zoology, 1995, Vol. 22
Av.D.
for males was calculated as 147 m and for
females as 86 m (Table 1). The trap-exposed area
(trapping grid area + Av.D./2) was thus calculated
to be 2.6 ha for adult females and 4.9 ha for adult
males.
Rat density was therefore calculated as 10/2.6 =
3.8 adult females/ha and 12/4.9 = 2.4 adult males/
ha, giving a total rat density of 6.2 adults/ha. This
density estimate is roughly confirmed by the home
range data obtained. Mean female home range size
was 0.3 ha (RP method, Table 2). If female ranges
were contiguous and non-overlapping (the pattern
found for the radio-tracked females), this would be
equivalent to a density of 3.3 females/ha.
Rats released from cage traps before 0200 h
showed no or little damage to their noses from rub-
bing on cage bars, although some released after this
time did. No rats died in cages, although death of
trapped rats was problematic for Daniel (1972),
Innes (1977), and Hickson et al. (1986). We suggest
that clearing traps during the night on which they are
set rather than during the following day greatly re-
duces stress to the rats, both by detaining them for
less time and by releasing them into darkness rather
than daylight.
Home ranges
RP home ranges for male and female rats are shown
in Fig. 3A, B. The female ranges were discrete,
whereas the male ranges overlapped each other and
partially overlapped the ranges of several females.
The trapping data corroborated
this:
no other females
were trapped >10 m within the range of
a
radio-col-
lared female.
All tracked rats spent the entire night moving
around their ranges. They left the nest at or just be-
fore dusk (1930-2000 h) and returned to the same
nest or a different one at daybreak (0500-0600 h).
Occasional observations of rats showed them to be
very active, running along supplejack vines (Ripo-
gonum scandens) or up and down trees as well as
moving laterally around their home range. The rats
covered most of their range each night, although
apparently not in any systematic way.
Arboreality and nests
Overall, 26% of
all
fixes were recorded in the 0-2 m
height band, 56% in the 2-8 m band, and 18% in the
+8 m height band (but see Error
estimation
in Meth-
ods).
There were no significant differences between
the foraging heights of males and females nor be-
tween poisoned and non-poisoned rats. Repeated
measures ANOVA (split-plot analysis) with
300
m
200-
•
100-
o-
\V^v
!
/
i
;--'
l
".'>M2
CT
;.\
:
-.^M3
u
B
0 100 200 300 m
100
200
300 m
Fig. 3 Home ranges (restricted polygon method) of
(A)
male and
(B)
female
ship rats at
Rotoehu Forest, shown
on
300 x 300
m
grid.
sex/poison as the whole-plot factor and position as
the sub-plot factor gave F
](t
=
0.03, P = 0.873 and
F,.
6
= 0.73, P
=
0.425, alternately.
Unusually high numbers of ground fixes were
observed for M5 and F4, both of which had >40%
of fixes between 0-2 m
(X
2
=
123.01 forM5, against
expected values based on average of other males, P
< 0.001 (2 d.f.); and X
2
=
66.7 for F4, against ex-
pected based on average of other females, P
<
0.001
(2 d.f.)). M5 utilised two centres of activity 230 m
apart and travelled on the ground between the two.
F4,
on the other hand, had young unweaned rats in
the nest, but we do not know why this required her
to spend more time on the ground.
Day nests were always up trees, but from the
Hooker & Innes—Ranging behaviour of ship rats
297
ground we could not locate exact nest
sites.
Only one
nest (for M5) was recovered; this was a loosely
woven structure in a small rimu tree.
Two females (F2 and F4) each used the same nest
tree during the pre-poisoning phase of
the
study (15-
21 December and 22 December -
4
January, respec-
tively),
but were later recorded using a different tree.
F4 had young in the nest at this time, and F2 prob-
ably also did, since she later appeared to come into
oestrus. The other rats each used 3-5 different nest
trees during the study. We detected no regular pat-
tern of
use
during the period studied; some rats used
one tree for 3-4 consecutive days then moved to a
new one, whereas others changed after only
1
day.
Rats were never observed to share the same nest tree.
Rat activity and behaviour
Rats were active from dawn until dusk, regardless
of moon phase or weather, although they sometimes
stopped moving during heavy rain. Mean speed of
movement over the ground (all fixes) showed that
males were significantly less active after midnight
than before (mean speed of males =1.59 (±SD 0.17)
m/min before midnight and 1.3 (±0.29) m/min after
midnight: two-sample Mest—before > after, T
=
1.94, P
—
0.05, n
=
5).
A similar though non-signifi-
cant trend was found for females.
We often saw and heard rats moving, though usu-
ally between fixes. They were actually in view dur-
ing only 3% of fixes. Rats were often noisy, rustling
the vegetation or making squeaking vocalisations.
Several untagged rats were known to be present in
the study area (15 of the initial 24 cage-trapped rats
were not collared), and were sometimes seen near
tagged rats.
Although they had individual home ranges, rats
were occasionally located together. Up to two col-
lared males were located in the vicinity of
a
collared
female, but there was no evidence of long-term as-
sociations between individuals. On one occasion a
male and a female were tracked (and observed) to-
gether for 10-20 min. In another incident, a male
near a female was observed apparently chasing an-
other male away. The commonest observation was
of one rat following another.
One female (F4) had
three
juveniles present in her
home range, which was the smallest of
all
the female
ranges. The three juveniles were seen travelling with
their mother several times before their cage-capture
over 3-5 January, when they weighed 22, 24, and
27 g. The latter was 10 cm in body length and 11 cm
in tail length. On 4 January, two of the juveniles were
cage-trapped together, and the third juvenile kept
running up to the cage and the observer. Two rats,
each weighing 42 g, were kill-trapped on 16 and 19
January and were probably these juveniles. Before
this,
the adult female, who had apparently just re-
jected the juveniles, had at least two adult males
around her, possibly indicating oestrus. No females
were pregnant or lactating when necropsied at the
end of the study.
Poisoning
Poison baits were weighed before and after con-
sumption by rats. The rats each consumed on aver-
age 10.7 g of bait (range 7.3-13.9 g), which would
contain 0.54 mg of brodifacoum, a lethal dose.
Each rat was regarded as dead at the last detected
movement. Fl died on the fourth night after poison-
ing, F2 and M2 died on the fifth night, and Ml died
on the sixth night. Three of the rats died in trees; two
(M1 and M2) in nest trees already known, one (Fl)
in an epiphyte clump likely to have been a nest (this
later fell to the ground). The other (F2) died in a hole
in the ground below her nest tree.
Home ranges before and after poisoning
There was no significant difference between the
home range areas before and after poisoning (paired
Mest, t
3
= 0.60, P
=
0.59), nor was there a signifi-
cant difference between the home range areas of non-
poisoned rats at Time 1 and Time 2 (paired Mest,
r, =
-0.71,
P
=
0.53). There was also no significant
difference in the percentage of shared cells for the
poisoned rats compared with the non-poisoned rats
(two-sample unpaired Mest, t
=
1.26, P
=
0.25).
There was also no significant difference between
mean distances travelled per night before and after
poisoning (paired Mest, Ho: no difference before and
after poisoning, f
3
= -1.26, P
=
0.30), nor was there
a significant difference between mean distances trav-
elled at Time
1
and Time 2 for the non-poisoned rats
(paired Mest, r, = 0.57, P
=
0.61) (Fig. 4). Thus, rat
movement did not appear to be restricted during the
nights before death.
These statistical tests are not very powerful be-
cause of the small sample sizes, but they would de-
tect large differences.
Home range utilisation
Visual inspection (Fig. 5) shows that there were few
differences in range utilisation for rats before and
after poisoning, nor were there significant differ-
ences in the foraging heights of poisoned and non-
poisoned rats (repeated measures ANOVA, P =
0.425). The differences in location of darker shaded
298 New Zealand Journal
of
Zoology, 1995, Vol.
2
11
E
1200
O)
1000 "
a 800
-
g
600-
(Z
CO
to
400 •
200
'
0
J
±JL
a Normal
• Poisoned
M1 M2
F1
F2
B
•^
800
Z
700 -
CD
=5
600 -
2
500 -
g
400 "
ra
300
-
to
•5
200 •
ro
100
•
CD
5 0
• Time
1
•
Time 2
M3
M4
F3
F4
Fig.
4 Mean distances moved per night by (A) poisoned
ship rats during 3 nights before and after poisoning, and
(B) non-poisoned rats in two 3-night periods to enable
comparison to those in (A). Bars are standard errors; /) = 4
in all instances.
areas represented change
in
nest sites used. These
changes from before
to
after poisoning followed
no
set pattern.
Inspection
of
grid cell plots (Fig.
6) for the two
neighbouring non-poisoned females
for the 5
days
of tracking before other rats were poisoned,
and the
2 days
of
tracking after, reveals little difference
in
home range use. There
are
several fixes which
ap-
pear
to
represent excursions from
the
home range,
although only one
of
these
is
towards
the
newly
va-
cated ranges.
DISCUSSION
Population density
The estimated density
of
6.2 ship rats/ha measured
at Rotoehu
in
January 1992
is
high compared with
densities measured
in
other
New
Zealand forests.
Previous estimates obtained
by
trapping
and
track-
ing were 2.0-2.5 rats/ha
on
Stewart Island
in
early
spring (Hickson
et
al. 1986) and 0.7-9.4 rats/ha over
8 years
in the
Orongorongo Valley (Daniel 1978).
The Rotoehu estimate was taken late in the breeding
season
at a
time
of
year when density
is
expected
to
be high, although rats do not produce major popula-
tion peaks regularly each year (Daniel 1978; Innes
1990).
These densities are much lower than some meas-
ured
in
subtropical zones, such
as the
12-15 rats/ha
in Cyprus macchie scrub (Watson 1951),
2-22
rats/
ha
in
uncultivated Hawaiian scrubland (Tomich
1970),
and up to 64
rats/ha
in
Hawaiian kiawe
for-
est (Tamarin & Malecha 1971). Clark (1980) found
0.4-19 rats/ha
in a
range
of
scrub
and
forest habi-
tats on the Galapagos Islands, where the density and
biomass
of
rats
was correlated with an index
of
veg-
etation biomass.
Home ranges
The observation of overlapping home ranges
of
some
male
but not
female ship rats
at
Rotoehu
is
consist-
ent with other studies
in New
Zealand forests
and
elsewhere (Granjon & Cheylan 1989). Daniel (1972)
first showed (but did not state) this, although his data
were from
a
small number
of
cage-trapping records
taken over many months,
and
may
not
have given
a
precise picture
of
the relationships between adjacent
ranges. Innes
&
Skipworth (1983)
and
Hickson
et
al.
(1986) combined cage-trapping
and
tracking
to
generate many more location records
in a
shorter
time;
both found largely exclusive female ranges,
but
obtained too few data on males to comment on them
in detail.
At
Rotoehu,
the
ranges
of
some radio-
tracked males overlapped considerably, and the pic-
ture may be even more complicated by the presence
of other males that were
not
radio-tracked.
It may
be misleading, however,
to
infer
the
combined
use
of common ground from overlapping MCP ranges.
For example,
the
range
of Ml is
mostly contained
inside that
of
M5
(Fig.
3A),
but in
fact M5 mostly
used two separated parts
of
his
range which avoided
Ml.
Although females" MCP ranges
at
Rotoehu were
about
the
same size (0.5
±
SE 0.07
ha, n
=
4) as the
monthly ranges measured
by
Hickson
et al.
(1986;
0.56 + 0.07 ha, n
=
8), their typical range shapes
dif-
fered.
On
Stewart Island, ranges were long and thin
(132
± 16
m
x
43
±
4 m), perhaps because they were
aligned
to
rivers
and the
coastline (Hickson
et al.
1986,
fig.
5).
At
Rotoehu, ranges in the more homo-
geneous forest were more nearly circular (103 +
7
m
x70±8m).
Radio-tracking quickly generates copious
amounts
of
location data, but
is
also very laborious.
Hooker & Innes—Ranging behaviour of ship rats
299
which restricts the possible sample size if adjacent
animals are followed simultaneously. Hence, this
study reports on rather few individuals. Cage-trap-
ping alone generates few location data, but trapping
and baited tracking together are much more effec-
tive (Innes & Skipworth
1983;
Hickson et al. 1986).
However, only radio-tracking permits monitoring of
the simultaneous movements of
neighbours,
detailed
study of range use and overlap between ranges which
is not dependent on trap or tunnel spacing, analysis
of temporal movement patterns in three dimensions
within a home range, and the location of nests and
verification of mortality and its possible causes.
The arboreal location of nest sites found in this
study supports previous observations that epiphytes
and tree-hollows are preferred sites, but that spar-
row-like nests (such as that recovered for M5) are
built if other sites are not available (Innes 1990).
Rotoehu ship rats were mostly arboreal, with 73%
of fixes above 2 m, but were nevertheless recorded
on the ground fairly regularly. The frequent (but not
prolonged) use of the forest floor by rats to move
from tree to tree provides intermittent opportunities
for feral cats (for which ship rats are important
prey—Fitzgerald & Karl 1979; Karl & Best 1982)
and other predators to hunt them without climbing.
Social organisation
Observations of wild ship rats with young are rare
and therefore worth reporting. Three juvenile rats
(<30 g) were seen and eventually cage-trapped with
an adult female, presumably their mother. Rats of
such light weight are rarely trapped in New Zealand,
despite extensive kill-trapping, suggesting that these
young rats were undertaking some of their first trips
around their mother's range. In one sample of 307
ship rats snap-trapped at Tiritea in the northern
Tararua Range, the lightest rat trapped was 35.5 g
(J. Innes unpubl. data). This is not because the rats
are too light to set the traps off, because rat traps
regularly catch mice (Mus musculus) which weigh
20 g or less. Ship rats in captivity wean young at
21-28 days old (Cowan 1981). In one of the young
rats,
the tail was longer than the body, supporting the
observation by Ewer (1971) that tail length outstrips
body length by the time that the young leave the nest.
No adult male was seen accompanying the female
with young. Other anecdotes of wild ship rats with
young in a nest (van Riper 1974; Innes 1977) also
mention one adult only. In captivity, female ship rats
drive the male away shortly before giving birth, and
raise the litter alone (Cowan 1981). Ewer (1971)
studied a commensal wild colony and observed that
females abandoned their young quickly after wean-
ing; sometimes the young's first foray away from the
nest was without the parent female.
Rats were often seen together, usually one follow-
ing another. Some interaction between males and
females was observed. M2 and Fl moved together
for 10-20 min in the branches and fed together on
the ground. A female, perhaps in oestrus, was
followed by two collared males. One male left upon
the approach of the other, but returned later.
The quality of data provided by radio-tracking
allows inferences to be made about the study ani-
mal's social organisation. Home range configura-
tions over a 4-week period in the breeding season
were different for males and females. Females'
ranges were small and non-overlapping, whereas
males'
ranges were larger and some overlapped con-
siderably. Each male's range touched the ranges of
several females.
Compared with commensal ship rats, feral ship
rats at Rotoehu had very different home ranges and,
by implication, a different social system. Rats at
Rotoehu did not live in colonies but were dispersed
rather evenly through the study area. Data suggested
that there were no long-term associations between
males and females, and that females alone raised
young. Ewer (1971) recorded behavioural obser-
vations of ship rats in a commensal environment (the
laboratory roof and an adjacent courtyard). Ewer's
study groups each had a single dominant male and
linear male hierarchy, with one or more high-rank-
ing females depending on the group size. Females
were responsible for most of the routine defence of
the group's territory, and top-ranking females rarely
attacked each other. Territories were marked by rub-
bing cheeks and ventral surface on branches, and the
boundaries were defended less often the further they
were from the nest.
Ship rats can apparently form either group or ex-
clusive territories, depending on the clumped or scat-
tered supply of food and shelter and the costs of
defending these, as can Mus musculus, another cos-
mopolitan commensal rodent pest (Fitzgerald et al.
1981).
Ship rat social organisation can be explained as
food-determined female dispersion, which in turn
determines male dispersion. This type of social sys-
tem implying male promiscuity and female territo-
riality is not uncommon in small mammals (Gipps
1985;
Ostfeld 1985). However, few data have been
collected for free-living rat species, most studies
having been carried out on commensal rats. Multi-
male polygyny has been documented in "clans"
New Zealand Journal of Zoology, 1995, Vol.
m
00-
B
r
/
\
Qlnf
1 1
H 2
1 3
1
<*
• 5*
\
\
100 200
B
m
m
200-
J
T
5
\
i>
© V
\
• Inf
m
'
• 2
P 3
S
4
• 5*
—-
\
200
m
300
200
m
300
Fig. 5 Grid cell plots (5 m, trimmed influences) of home ranges of
the
four poisoned ship rats before (A) and after (B)
poisoning. Areas of progressively darker shading represent more intensive use.
(subgroups within a colony) of brown rats, R.
norvegicus, on agricultural land (Fenn 1989).
Poisoning
The mean lethal dose (LD
50
, the amount of poison
needed to kill 50% of
the
population) of brodifacoum
for wild ship rats is 0.69 mg/kg (Dubock &
Kaukeinen 1978). The mean weight of
rats
poisoned
at Rotoehu was 170 g; the average rat therefore
needed to ingest 0.117 mg brodifacoum to receive a
mean lethal dose. Average consumption for the poi-
soned rats was 4.6 times the LD
50
dose, and even the
rat which ate only 7.3 g of bait consumed over three
times the LD
50
dose.
The rats poisoned with "Talon" died 3-6 days af-
ter ingestion. Hickson et al. (1986) estimated 2-3
days to death, although their measurement was based
on the time to which footprints were no longer ob-
served in tracking tunnels.
Further research is needed to determine if fatally
intoxicated ship rats are still a predation threat to
other fauna during the few days before their deaths.
Cox & Smith (1992) showed that intake of food and
water declined rapidly in anticoagulant-treated,
caged Rattus norvegicus. They also observed a pre-
lethal reversal of light-dark activity pattern, but no
evidence of this was found for ship rats at Rotoehu,
which maintained normal nocturnal movement and
showed no nest change between dawn and dusk.
Implications for control of ship rats in New
Zealand
Direct observations of ship rats at Rotoehu con-
firmed that they were superbly agile in trees, that
they ranged widely throughout their home ranges
each night, and that they were not colonial but evenly
dispersed through the forest study area. These facts
indicate the potential threat of this species to
avifauna, since every bird nest is likely to be in the
range of at least two rats, and the rats are likely to
pass by each nest sooner or later. Clearly, to target
all individuals, the control effort must be spread
Hooker & Innes—Ranging behaviour of ship rats
Malel
301
m
300
throughout the entire area for which protection is
sought.
Traps,
poison baits or stations, and population
monitoring devices such as tracking tunnels can all
be set effectively on the ground, since the rats often
travelled there despite spending most time up trees.
They moved extensively (400-900 m) inside their
home ranges each night; range lengths averaged
c. 100 m for females and twice that for males, and
rats were active in all weathers. Rats are highly likely
to find traps or bait stations anywhere inside their
ranges, especially if they are lured. Bait stations on
a grid at 100 m intervals would expose most rats to
poison, and at 50 m intervals would expose all rats,
wherever rat density is similar to that at Rotoehu.
Ship rats have been eradicated by ground-poison-
ing from five New Zealand islands, the biggest of
them being Somes I. at 32 ha (Veitch & Bell 1990),
and were significantly reduced during a
5
ha ground-
poisoning operation on Stewart I. (Hickson et al.
1986).
Since 1990 several large-scale aerial
poisoning programmes (mostly using "1080"—so-
dium monofluoroacetate) have been targeted at ship
rats in forest on the mainland to protect North Island
kokako for the duration of the nesting season (Innes
et
al.
in
press).
Most future large-scale operations are
likely to be aerial, for reasons of cost, and it is
important to estimate an appropriate sowing rate.
Current aerial operations with 1080 are targeted pri-
marily at brushtail possums, Trichosurus vulpecula,
another widespread introduced forest pest, usually
at a minimum sowing rate of
8
kg of baits/ha
(1
bait/
7.5 m
2
). Such a sowing rate delivers 1470 and 400
baits to average Rotoehu male and female rat ranges,
respectively. Since each bait contains many times the
required lethal dose for a ship rat, bait density for
operations concerned with rats rather than possums
could be hugely reduced and still be adequate to kill
all rats.
Most rats died in their nests after poisoning, sug-
gesting that few dead rats will be found in the open
after a Talon poisoning operation. However there is
New Zealand Journal of Zoology, 1995, Vol. 22
B
o
fXXXXXXXjjXoJI
"J Inf
*| 1
I 2
Si 3
• 4
• 5*
\
(
100
B
200-
m
200
I
\
m
i
H 2
@ 3
1 4
• 5-
t
100
100
m
200
Fig. 6 Grid cell plots (5 m, trimmed influences) of home ranges of Female 3 and Female 4 (neighbours of poisoned
rats) for six tracking-days before the other rats were poisoned (A) and two tracking-days after (B). Areas of darker
shading represent more intensive use.
still a risk of non-target poisoning of native avian
predators such as the ruru (Strigidae, Ninox novae-
seelandiae), or introduced mammal predators such
as cats and stoats, from the normally ranging intoxi-
cated rats. This is especially so if
the
rats leave blood
trails or react sluggishly to touch as observed by Cox
& Smith (1992) with Norway rats.
Ship rats fed brodifacoum poison and radio-
tracked at Rotoehu died after 4-6 days, suggesting
that the interval between pulses of brodifacoum
should be 7 days. This time to death is less than the
average of 10 days (range 5-14) reported for ship
rats in a laboratory (Redfern & Gill 1976). The lack
of
a
colony-based hierarchical social system among
forest-dwelling ship rats may reduce the time needed
to poison them relative to commensal rats, because
dominant individuals in colonies keep subordinates
away from feeding stations until the dominant rats
die (Timm & Salmon 1988). In the Philippines, West
et al. (1975) found significantly greater bait con-
sumption by R. rattus from several small bait con-
tainers compared with one large one, and
hypothesised that this was due to reduced interac-
tion between rats.
ACKNOWLEDGMENTS
This study is based on fieldwork carried out during the
Oxford University undergraduate expedition "Kokako
'91".
Many thanks to expedition members Vickie Heaney,
Pete Buston, Neil Davies, Richard Bradbury, and Zoe
Billinghurst; all sponsors for the financial aid enabling
the expedition to go ahead; New Zealand Department of
Conservation and the Foundation for Research, Science
and Technology (CO 9495) for funding the participation
of John Innes; Tom Tew (Department of Zoology,
Oxford) for his help and advice throughout the project;
Ian Todd (Department of Zoology, Oxford) for providing
Hooker & Innes—Ranging behaviour of ship rats
303
his computer package "Wildtrak" for the analysis of
radio-tracking data; Dale Williams for field assistance;
Mike Fitzgerald, John McLennan, Murray Efford, and
Cleveland Duval (Manaaki Whenua - Landcare Research)
for advice and manuscript review; Carolyn King
(Biological Sciences Department, Waikato University)
for guiding the initial contact between the authors and
improving the manuscript, and an anonymous referee
for helpful manuscript review.
REFERENCES
Allan. H. H. 1961: Flora of New Zealand, Vol 1. Wel-
lington. Government Printer. 1085 p.
Atkinson. I. A. E. 1973: Spread of the ship rat (Rattus r.
raUus
L.) in New Zealand. Journal of
the
Royal
Society of New Zealand
3:
457^-72.
Atkinson, I. A. E. 1978: Evidence for effects of rodents
on the vertebrate wildlife of New Zealand is-
lands.
In: Dingwall, P. R.; Atkinson, I. A. E.;
Hay, C. ed. The ecology and control of rodents in
New Zealand nature reserves. Department of
Lands and Survey information series
4:
7—30.
Atkinson. I. A. E. 1985: The spread of commensal spe-
cies of Rattus to oceanic islands and their effects
on island avifaunas. In: Moors, P. J. ed. Conser-
vation of island birds. International Council for
Bird Preservation technical publication 3: 35-
81.
Barnett, S. A. 1967: Rats. Scientific American 216: 78-
85.
Burt, W. H. 1943: Territoriality and home range concepts
as applied to mammals. Journal of mammalogy
24: 346-352.
Clark, D. B. 1980: Population ecology of Rattus rattus
across a desert-montane forest gradient in the
Galapagos Islands. Ecology
61:
1422-1433.
Connor, H. E.; Edgar, R. L. 1987: Name changes in the
indigenous New Zealand flora 1960-1986 and
Nomina Nova IV, 1983-1986. New Zealand jour-
nal of botany
25:
115-170.
Cowan. P. E. 1977: Neophobia and neophilia: new object
and new place reactions of three Rattus species.
Journal of comparative & physical psychology
91:63-7].
Cowan, P. E. 1981: Early growth and development of
roof
rats,
Rattus rattus L. Mammalia 45
(2):
239-
250.
Cox. P.; Smith, R. H. 1992: Rodenticide ecotoxicology:
pre-lethal effects of anticoagulants on rat behav-
iour.
In:
Borrecco, J. E; Marsh, R. E. ed.
Proceed-
ings of the 15th Vertebrate Pest Conference.
Davis.
University of California. Pp. 165-170.
Daniel, M. J. 1972: Bionomics of the ship rat (Rattus r.
rattus) in a New Zealand indigenous forest. New
Zealand journal of science
15:
313-341.
Daniel, M. J. 1978: Population ecology of ship and Nor-
way rats in New Zealand. In: Dingwall, P. R.;
Atkinson, I. A. E.; Hay, C. ed. The ecology and
control of rodents in New Zealand nature re-
serves. Department of Lands and Surx'ev infor-
mation series
4:
145-154.
Dice,
L. R. 1938: Some census methods for mammals.
Journal of wildlife management
2:
119-130.
Dubock, A. C; Kaukeinen, D. E. 1978: Brodifacoum
(Talon™ rodenticide), a novel concept. In: Pro-
ceedings of 8th Vertebrate Pest Conference.
Sacramento, California. Pp. 127-137.
Dunn, J. E.; Gipson, P. S. 1977: Analysis of radio-telem-
etry data in studies of home range. Biometrics 33:
85-101.
Ewer, R. F. 1971: The biology and behaviour of a free-
living population of black rats (Rattus rattus).
Animal behaviour monographs
4:
127-174.
Fenn, M. G. P. 1989: Ecology of the Norway rat (Rattus
norvegicus) on lowland farms in England. Un-
published D. Phil, thesis, Oxford University.
Fitzgerald, B. M.; Karl, B. J. 1979: Foods of feral house
cats (Felis catus L.) in forest of the Orongorongo
Valley, Wellington. New Zealand journal of
zo-
ology
6:
107-126.
Fitzgerald, B. M.; Karl, B. J.; Moller, H. 1981: Spatial
organisation and ecology of a sparse population
of house mice (Mus musculus) in a New Zealand
forest. Journal of animal ecology
50:
489-518.
Galef,
B. G. Jr 1988: Social learning of food preference
by rats
In:
Zentall, T. R.;
Galef,
B. G. Jr
ed.
Social
learning: psychological and physiological perspec-
tives.
Lawrence Erlbaum Assoc.
Gipps, J. H. W. 1985: Spacing behaviour and male repro-
ductive ecology in voles of the genus Clethrio-
nomys. Annales zoologici fennici
22:
343—351.
Granjon, L.; Cheylan, G. 1989: The fate of black rats
(Rattus rattus L.) introduced on an island, as
revealed by radio tracking. Comptes rendus de
I'Academie des Sciences 309: 571-575.
Harris, S.; Cresswell, W. J.; Forde. P. G.; Trewhella, W.
J.; Woolard, T.; Wray, S. 1990: Home range analy-
sis using radio-tracking data—a review of prob-
lems and techniques particularly as applied to the
study of mammals. Mammal review
20:
97-123.
Hickson, R. E.; Moller, H.; Garrick, A. S. 1986: Poison-
ing rats on Stewart Island. New Zealand journal
of ecology 9:
111-121.
Innes,
J. G. 1977: Biology and ecology of the ship rat
Rattus rattus rattus (L.) in Manawatu (N.Z.) for-
ests.
Unpublished M.Sc. thesis, Massey Univer-
sity. Palmerston North, New Zealand.
304
New Zealand Journal of Zoology, 1995, Vol. 22
Innes,
J. G. 1990: Ship rat. In: King, C. M. ed. The
handbook of New Zealand mammals. Auckland,
Oxford University Press. Pp. 99-1 13.
Innes,
J. G.; Skipworth, J. P. 1983: Home ranges of ship
rats in a small New Zealand forest as revealed by
trapping and tracking. New Zealand journal of
zoology
10:
99-110.
Innes,
J. G.; Warburton, B.; Williams, D.; Speed, H.;
Bradfield, P. in press: Large-scale poisoning of
ship rats (Rattus rattus) in indigenous forests of
the North Island, New Zealand. New Zealand
journal of ecology.
Karl,
B.
J.; Best, H. A. 1982: Feral cats on Stewart Island;
their foods, and their effects on kakapo. New
Zealand journal of zoology
9:
287-294.
King, C. M.; Edgar, R. L. 1977: Techniques for trapping
and tracking stoats (Mustela erminea); a review,
and a new system. New Zealand journal of zool-
ogy
4:
193-212.
Leathwick, J. R.; Hay, J. R.; Fitzgerald, A. E. 1983: The
influence of browsing by introduced mammals
on the decline of the North Island kokako. New
Zealand journal of ecology
6:
55-70.
Mohr, C. O. 1947: Table of equivalent populations of
North American small mammals. American mid-
land naturalist
37:
223-249.
Moors, P. J. 1985: Conservation of island birds. ICBP
technical publication 3: 271 p.
Ostfeld, R. S. 1985: Limiting resources and territoriality
in microtine rodents. American naturalist
126:
1-
15.
Redfern, R.; Gill, J. E. 1976: Laboratory evaluation of
WBA
81 19
as a rodenticide for use against warfa-
rin-resistant and non-resistant rats and
mice.
Jour-
nal of hygiene Cambridge
77:
419-^26.
Reumer, J. W. F. 1986: Note on the spread of the black
rat, Rattus rattus. Mammalia
50:
118-119.
Siniff,
D. B.; Tester, J. R. 1965: Computer analysis of
animal movement data obtained by telemetry.
Bioscience 15: 104-108.
Swihart, R. K.; Slade, N. A. 1985: Testing for independ-
ence of observations in animal movements. Ecol-
ogy
66:
1176-1184.
Tamarin, R. H.; Malecha, S. R. 1971: The population
biology of Hawaiian rodents: demographic pa-
rameters. Ecology
52:
383-394.
Timm, R. M; Salmon, T. P. 1988: Behavior.
In:
Prakash,
I. ed. Rodent pest management. Florida, CRC
Press.
Pp. 225-236.
Todd, I. A. 1992: Wildtrak: non-parametric home range
analysis for Macintosh computers. Department of
Zoology, University of Oxford.
Tomich, P. Q. 1970: Movement patterns of field rodents
in Hawaii. Pacific science 24: 195-234.
van Riper, C. Ill 1974: Unusual uses of ricebird's nests in
Hawaii. Elepaio 35 (5): 5.
Veitch, C. R.; Bell, B. D. 1990: Eradication of introduced
animals from the islands of New Zealand. In:
Towns, D. R.; Daugherty, C. H.; Atkinson, I. A.
E. ed. Ecological restoration of New Zealand Is-
lands.
Conservation sciences publication 2. Wel-
lington, Department of Conservation. Pp.
137-146.
Voigt, D. R.; Tinline, R. R. 1980: Strategies for analysing
radio-tracking
data.
In:
Amlaner, C. J.; Macdonald,
D.
W. ed. A handbook of biotelemetry and radio-
tracking. Oxford, Pergamon Press.
Watson, J. S. 1951: The rat problem in Cyprus. Colonial
research publications 9. London, Colonial Of-
fice.
66 p.
West, R. R.; Fall, M. W.; Libay, J. L. 1975: Reducing
interactions among rats to improve bait accept-
ance.
Philippine agriculture
59:
31-36.
White, G. C; Garrott, R. A. 1990: Analysis of wildlife
radio-tracking data. London, California Academic
Press Inc.
Wolton, R. J. 1985: The ranging and nesting behaviour of
wood mice, Apodemus svlvaticus (Rodentia:
Muridae), as revealed by radio-tracking. Journal
of zoology London (A) 206: 203-224.
