ArticlePDF Available

Abstract and Figures

Predator-protected populations of threatened fauna are important for species conservation, although these animals can quickly become predator naïve and can lack appropriate antipredator behaviour to enable them to persist once released. Controlled predator exposure can improve predator recognition and encourage avoidance behaviour, but little is known about the escape responses or fleeing behaviour of prey species. We compared the escape behaviour of a small marsupial, the burrowing bettong, Bettongia lesueur, between two fenced populations: one that had been purposely exposed to feral cats, Felis catus, while the other had been maintained without exotic predators. To quantify escape behaviour, bettongs were trapped and released into a temporary runway and a threatening stimulus was introduced to encourage them to flee. Measures relating to reactivity (escape initiation), escape speed and flight path (protean characteristics: agility, path irregularity and straightness) were recorded from video footage. Cat-exposed bettongs were significantly heavier than those from the cat-naïve population. We found a significant effect of the interaction of treatment (‘cat-exposed’ or ‘cat-naïve’) and body mass on overall escape behaviour. These differences were attributed to increased reactivity and escape speed in cat-exposed bettongs, but not protean characteristics of their flight path. Cat-exposed bettongs fled at an intensity where body size affected their escape performance (larger animals performed longer bounds and achieved faster speeds), while this body size effect was not evident for cat-naïve animals. This result suggests the cat-naïve animals were not as motivated to flee. Introducing low levels of predation pressure can successfully promote the development of antipredator behaviour through selection and/or individual learning, including a heightened escape response. Controlled predator exposure may be able to address some types of prey naïvety and lead to increased survival outside predator-free sanctuaries.
Content may be subject to copyright.
Predator exposure enhances the escape behaviour of a small
marsupial, the burrowing bettong
Natasha E. Tay
, Patricia A. Fleming
, Natalie M. Warburton
Katherine E. Moseby
Harry Butler Institute, Murdoch University, Murdoch, Australia
Centre for Ecosystem Science, School of Biological, Earth and Environmental Sciences, University of New South Wales, Sydney, Australia
Arid Recovery Ltd., Roxby Downs, Australia
article info
Article history:
Received 28 May 2020
Initial acceptance 3 August 2020
Final acceptance 21 December 2020
MS number 20-00395R
antipredator behaviour
escape manoeuvres
prey naïvety
Predator-protected populations of threatened fauna are important for species conservation, although
these animals can quickly become predator naïve and can lack appropriate antipredator behaviour to
enable them to persist once released. Controlled predator exposure can improve predator recognition
and encourage avoidance behaviour, but little is known about the escape responses or eeing behaviour
of prey species. We compared the escape behaviour of a small marsupial, the burrowing bettong, Bet-
tongia lesueur, between two fenced populations: one that had been purposely exposed to feral cats, Felis
catus, while the other had been maintained without exotic predators. To quantify escape behaviour,
bettongs were trapped and released into a temporary runwayand a threatening stimulus was introduced
to encourage them to ee. Measures relating to reactivity (escape initiation), escape speed and ight path
(protean characteristics: agility, path irregularity and straightness) were recorded from video footage.
Cat-exposed bettongs were signicantly heavier than those from the cat-naïve population. We found a
signicant effect of the interaction of treatment (cat-exposedor cat-naïve) and body mass on overall
escape behaviour. These differences were attributed to increased reactivity and escape speed in cat-
exposed bettongs, but not protean characteristics of their ight path. Cat-exposed bettongs ed at an
intensity where body size affected their escape performance (larger animals performed longer bounds
and achieved faster speeds), while this body size effect was not evident for cat-naïve animals. This result
suggests the cat-naïve animals were not as motivated to ee. Introducing low levels of predation pres-
sure can successfully promote the development of antipredator behaviour through selection and/or in-
dividual learning, including a heightened escape response. Controlled predator exposure may be able to
address some types of prey naïvety and lead to increased survival outside predator-free sanctuaries.
©2021 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved.
Translocations are the intentional movement of organisms from
one site to another for a conservation benet(IUCN/SSC, 2013).
Predation is a key cause of mortality in many translocation pro-
grammes, and a lack of appropriate antipredator behaviour in
reintroduced animals has been well documented (e.g. Biggins,
Miller, Hanebury, &Powell, 2011;Moseby et al., 2011;Shier &
Owings, 2006). This is especially apparent when translocated ani-
mals are sourced from populations isolated from predators
(Aaltonen, Bryant, Hostetler, &Oli, 2009;Moseby, Carthey, &
Schroeder, 2015;Tetzlaff, Sperry, &DeGregorio, 2019;van Heezik,
Seddon, &Maloney, 1999). Therefore, enhancing appropriate
antipredator behaviours in naïve prey(animals that have lost the
ability to recognize a specic predator and its cues or have not
evolved with ecologically and morphologically similar predators) is
of great benet to wildlife managers.
There are two broad categories of antipredator responses of
prey: (1) avoidance behaviourdesigned to minimize detection by
predators, such as time spent vigilant or in hiding, and adjusting
activity times and habitat use; and (2) escape behaviourused
during an encounter where the predator has detected the prey and
initiated an attack (Lima &Dill, 1990;Sih, 1985). Aspects of escape
behaviour, such as ight initiation, speed and protean characteris-
tics (unpredictable paths that make targeting prey more difcult;
Jones, Jackson, &Ruxton, 2011;Richardson, Dickinson, Burman, &
Pike, 2018), can differ depending on the perceived risk (Edut &
Eilam, 2003). Comparisons of escape behaviour under different
*Corresponding author.
E-mail address: (N. E. Tay).
Contents lists available at ScienceDirect
Animal Behaviour
journal homepage:
0003-3472/©2021 The Association for the Study of Animal Behaviour. Published by Elsevier Ltd. All rights reserved.
Animal Behaviour 175 (2021) 45e56
levels of risk have been carried out using ight initiation distances
(FID, that is, how close the prey animal will let a threat approach
before they ee; Blumstein, Samia, &Cooper, 2016;Li, Belasen,
Palis, Bednekoff, &Foufopoulos, 2014;Worrell, Admiraal,
Bateman, &Fleming, 2017), but other aspects of how different
levels of predation can affect escape behaviour are less commonly
A novel approach to improving antipredator behaviour in pop-
ulations of predator-naïve animals has been to introduce low
densities of predators to them to gain in situ predator exposure so
that either learning or selection for predator-savvy individuals may
occur (Moseby, Blumstein, &Letnic, 2016). This has been trialled in
Australia using feral cats, Felis catus, and burrowing bettongs, Bet-
tongia lesueur (Moseby, Letnic, Blumstein, &West, 2019). Initial
results of in situ predator exposure have been promising. Cat-
exposed bettongs show increased predator avoidance behaviour:
they are more docile when trapped (a representation of hiding
behaviour) and approach feed trays more slowly and cautiously in
the presence of cat odour (Saxon-Mills, Moseby, Blumstein, &
Letnic, 2018;West, Letnic, Blumstein, &Moseby, 2018). Bettongs
also increase their FID in response to cat exposure (West, Letnic,
et al., 2018), showing the rst documented change in escape
behaviour. However, FID only looks at when the animal decides to
run, but not how it continues to react to the threat in a chase.
In this study we compared the escape behaviour of bettongs from
cat-exposed and cat-naïve fenced populations to identify any dif-
ferences in how they ee from a threat. The two populations lived in
similar environmental conditions but differed in density because of
the different time since establishment and the inuence of predation
on population size and demography. Our aim was to identify
whether individuals from the cat-exposed population had enhanced
escape behaviour that could allow them to survive better in the wild
after release. We did not aim to disentangle the relative contribu-
tions of selection and learning to these behavioural differences as we
did not test individual bettongs before and after cat exposure. We
chose a human-derived stimulus that incorporated a pursuit instead
of using indirect cues (i.e. predator odour such as urine or faeces) as
these do not necessarily indicate that a predator is still present and
that the prey has encountered or will encounter the predator, nor
would it elicit a eeing response, which is the focus of this study. We
predicted that cat-exposed bettongs would show an improvement in
escape behaviour manifested as more immediate reactions, faster
escape speeds and/or more protean behaviour following the pre-
sentation of the threatening stimulus and chase, compared to in-
dividuals from the population protected from cats.
Ethical Note
This project was conducted under permit from the South
Australian Department of Environment and Water (permit number:
Y26748) and with animal ethics approval from Murdoch University,
Western Australia (Project No. R2974/17). No anaesthesia proced-
ure was required while taking physical measurements of animals.
Study Species
Burrowing bettongs are a small, social marsupial weighing
approximately 0.8e2.0 kg (Short &Turner, 1993). They are
nocturnal and shelter in warrens or burrows during the day. There
is trivial sexual dimorphism with little size differentiation once
adult (Freegard, Calver, Richards, &Bradley, 2007). Females mature
sexually at approximately 7e8 months of age, producing one young
after 21 days gestation (Short &Turner, 1999). They can raise two
young per year (Burbidge &Short, 2008). Bettongs are primarily
bipedal (Stodart, 1966), switching from quadrupedal locomotion at
relatively slow speeds. Once widely distributed across two-thirds of
the Australian mainland (Burbidge &Short, 2008), bettongs were
lost from mainland Australia in the 1960s and persist naturally on
only four offshore islands (Short &Turner, 1993). As their body
mass lies within the critical weight rangeof marsupials (35
ge5.5 kg, Burbidge &McKenzie, 1989;Johnson &Isaac, 2009),
their decline is thought to have been caused primarily by predation
from exotic predators (Woinarski, Burbidge, &Harrison, 2014).
Reintroductions of bettongs into predator-free enclosures have
been successful; however, attempts to release these predator-naïve
animals outside these enclosures have failed due to predation by
feral cats, red foxes, Vulpes vulpes, and dingoes, Canis familiaris
(Bannister, Lynch, &Moseby, 2016;Moseby et al., 2011;Short,
Kinnear, &Robley, 2002).
Study Site
Arid Recovery (30
S, 136
E) is a 123 km
fenced reserve
located approximately 20 km north of Roxby Downs, South
Australia. The 1.8 m high wire-netting fence around the reserve
excludes introduced terrestrial mammalian predators (red foxes
and feral cats), as well as European rabbits, Oryctolagus cuniculus.
The reserve comprises six large paddocks (Fig. 1), of which cats,
foxes and rabbits have been removed from four predator-free
paddocks (totalling 60 km
). Bettongs were rst introduced to the
predator-free paddock surveyed in this experiment in 1999
(Moseby et al., 2011) and had a density of 64.29 bettongs per km
the time of this experiment. In 2014, a population of bettongs
(N¼353) was moved from the predator-free paddock into a
predator paddockwith a low density of feral cats intended to
replicate the lower range of feral cat densities reported in arid
Australia (Legge et al., 2017). Cat density within this paddock
ranged from 0.04 to 1.84 per km
over the 4 years since bettongs
were translocated, with an estimated maximum of 48 individuals
(Moseby et al. 2019, 2020). As necessary to allow the bettong
population to survive and recruit young, this feral cat density was
likely to be at a much lower predator density than at the time
bettongs became extinct on mainland Australia when rabbit prey
were abundant (Moseby et al., 2019). Bettong density in the larger
predator paddock was estimated at 4.62 animals per km
with 120
bettongs. Population growth rate in the predator-free paddock was
approximately twice that of the predator-paddock (Moseby et al.,
Trapping Procedure
Bettongs were trapped over 4 nights each from the predator
paddock and the predator-free paddock in May 2018 using Shefeld
wire cage traps baited with peanut butter and rolled oats. For each
paddock, bettongs were trapped for 2 nights at each of two loca-
tions per paddock, focused on different bettong warrens. Each trap
had a hessian sack placed over the back half to provide shelter and
protection for the animal. Traps were opened at approximately
1600 hours every day and rst checked at 1900 hours. They were
reset for a second round and checked again at 2200 hours. Traps
were closed after the second check. Only adult bettongs (>950 g)
were included in this study. Individuals unsuitable for the escape
behaviour trials (i.e. low body condition, females with pouch
young) were released immediately at their point of capture. We did
not trap any bycatch.
N. E. Tay et al. / Animal Behaviour 175 (2021) 45e5646
All trapped bettongs were given a Monel self-piercing ear tag
with a unique number for identication. Unique ear tags allowed us
to identify any founder individuals translocated to the predator
paddock in 2014. Bettongs were grouped into two age classes based
on their minimum estimated age (Appendix Table A1): young
adults (2 years old) and mature (>2 years old). Bettong longevity
at this study site is commonly 4e5 years, although a few individuals
have been known to live up to 7 years. We recorded three measures
of body size: body mass (±0.01 kg; HDB 5K5N, KERN &Sohn, Bal-
ingen, Germany), head length (±0.01 mm; digital Vernier callipers,
Kincrome, Australia) and pes length (hind foot; ±0.01 mm). Ani-
mals were restrained within their capture bag during handling,
exposing only each body segment to measure as required.
Escape Behaviour Trials
To assess escape behaviour and path characteristics, we used a
temporary runway to lm bettongs eeing. It was constructed on
cleared off-road vehicle tracks approximately 200 m away from
trap sites and comprised two 1.2 m high, 20 m long cloth walls,
placed 3 m apart and without a roof. The runway was open ended
to allow for animals to disperse after the trial with the opening
oriented towards their original point of capture. After being
measured, animals were held in their capture bags and placed in a
quiet environment for approximately 1 h prior to their escape trial.
All escape trials were conducted between 2000 and 0200 hours to
coincide with normal activity times of bettongs (Stodart, 1966) and
all escape behaviour trials occurred before moon rise. The runway
was illuminated by infrared oodlights and trials were lmed using
cameras (SJCAM SJ4000) with their infrared lter removed. Each
animal was lmed in HD (1280 x 720 pixels) at 30 fps from ve
camera angles mounted at animal height (Fig. 2). All videos were
synced using a ash of light at the beginning of each trial. One
observer (N.T.) inspected all the recorded footage and used BORIS
software (Friard and Gamba 2016) to assign time points for calcu-
lating escape variables.
To measure bound lengths and the exact path travelled, we
dusted each bettong's feet with uorescent powder to mark foot-
prints under ultraviolet (UV) light. After the animal had left the
runway, we recorded the coordinates of each footfall within the
runway using a handheld UV light to compare with results from the
video footage. Coordinates were recorded as the deviation (cm)
from the middle of the runway (X) and the distance along the
runway (Y;Fig. 2).
For each trial, the bettong was gently introduced to the runway
while still held in its capture bag. The bettong was placed at the
start line, oriented to face forwards with its feet positioned on the
ground, but gently held in place by the handler. The capture bag
was then lifted over the animal's head and the handler let go of the
animal allowing it to leave at will. To encourage the bettong to
escape, a second person (chaser) introduced a threatening stim-
ulus by loudly shaking and thumping a ball (120 mm in diameter)
containing bells against the ground approximately 300 mm behind
and to the side of the bettong. The apparatus was attached to the
end of a 1.5 m pole and the chaser followed behind the bettong as it
moved through the runway, matching its speed. For consistency,
one person performed the role of chaser for all animals. The chaser
and pole remained behind the bettong and the pole never touched
it. Animals could freely move away from the pole stimulus once
released from their bag. The stimulus was only in one direction
(from behind the animal), and once it exited the runway the animal
was free to return to its warren. Overall, animals were exposed to
the runway and stimulus for 7.5 ±0.40 s (maximum 13 s). Animals
were only tested once per night. If trapped again after their escape
behaviour trial on the same night, animals were checked over
quickly then released at their point of capture.
Our articial set-up allowed for a standardized environment for
comparison of escape behaviour at different locations within each
paddock. For this study, a human stimulus was sufcient to
encourage an escape response. Humans have been successfully
used in other studies to encourage eeing behaviour in animals,
including studies that measure FID (e.g. Li et al., 2014;Moore,
Cooper, Biewener, &Vasudevan, 2017;Wynn, Clemente, Nasir, &
Wilson, 2015).
Escape Behaviour Responses
Video footage and the path coordinates were used to score
escape behaviour for eight measures relating to the animal's reac-
tivity, escape speed and protean characteristics. To determine how
Predator paddock
paddock 0510
© Nathan Beerkens
Figure 1. The Arid Recovery Reserve showing the location of the predator paddock and predator-free paddock used in this experiment and a burrowing bettong, Bettongia lesueur,
with an ear tag. Photo: Nathan Beerkens.
N. E. Tay et al. / Animal Behaviour 175 (2021) 45e56 47
quickly an animal reacted to the novel experimental set-up, we
recorded three measures of reactivity (i.e. how the animal initiated
its escape): (1) latency to leave measured as the time between
when the animal's head was uncovered until its exit from the bag
(paused>1sored immediately <1 s); (2) gait used during its
initial movement out of the bag (quadrupedalor bipedal)to
represent its hesitancy to disperse; and (3) direction of rst
movement relative to the position of the pole (towards pole,
forward regardless of pole position,away from pole). We used
two measures of escape speed: (4) maximum bound length (m),
recorded during the trial as a proxy for locomotor ability (the dis-
tance between consecutive footprints identied by the UV uo-
rescent powder); and (5) average speed, calculated by dividing the
total path length (m; sum of all bound lengths) by the time (s) to
exit the 20 m long runway. We quantied protean behaviour in
three ways. (6) Agility was measured by how quickly each bettong
changed directions during its escape: the angular change between
two consecutive UV uorescent footprints was multiplied by its
corresponding bound speed (bound length divided by time differ-
ence between the two points). The maximum value recorded was
used as a proxy for agility as bettongs with higher values performed
sharper turns at faster speeds. (7) Path irregularity was calculated
as the coefcient of variation of the speed of change in direction
values. (8) Straightness was measured as the overall extent of de-
viation (d) along a path using the straightness index:
where Dis the straight-line distance between the start and end of
the path, and Lis the length of the path (Batschelet 1981).
Statistical Analyses
Analyses were conducted in R version 3.5.1 (R Core Team, 2018)
unless otherwise specied. A critical alpha of 0.05 was used to
dene signicance.
Data were transformed to t the assumption of normal distri-
bution of residuals where necessary. We used a BoxeCox trans-
formation for agility and path irregularity and the ordered quantile
technique for straightness (Peterson &Cavanaugh 2020).
As we trapped only ve founder bettongs from the predator
paddock, we did not statistically compare the escape behaviour of
founder animals against their progeny. Data points representing
founder individuals are, however, highlighted in all gures where
We assessed potential body size differences between pop-
ulations using analysis of variance (ANOVA) with treatment (cat-
exposed or cat-naïve), sex and age class as additive variables. We
analysed body mass, head length and pes length separately. Head
length violated Levene's test for homogeneity of variance between
treatment means (F
¼8.76, P¼0.006) but we retained this
measure in our analysis as its residuals approached a normal dis-
tribution on inspection of the model's QeQ plot and due to its
biological relevance. Collinearity between body size measures was
assessed separately for each treatment using Pearson correlation.
As the escape measures were not independent of each other, we
analysed all eight variables together in a matrix. Escape variables
were mean standardized then subjected to nonmetric multidi-
mensional scaling (nMDS) using Euclidean distances in the VEGAN
package (Oksanen et al., 2019) and differences in overall escape
behaviour were visualized in an ordination plot. We used multi-
variate permutational analysis of variance (PERMANOVA) to
determine whether overall escape behaviour differed by treatment
(cat-exposed or cat-naïve). Body mass and its interaction with
treatment were included in this PERMANOVA model to account for
any potential treatment-specic body size effects on overall escape
performance, and age class was included as an additive variable to
account for possible ontogenetic effects on behaviour. This was
followed by a similarity percentage analysis (SIMPER) conducted in
PAST v3.2 (Hammer, Harper, &Ryan, 2001) to calculate the per-
centage contribution of each escape variable to the dissimilarity
between treatments.
2000 cm
–150 cm
150 cm
0 cm
Handler releasing animal
Chaser with pole
Figure 2. The runway and camera set-up showing the horizontal XY coordinate system for measuring UV uorescent powder footprints within the runway.
Table 1
Morphometrics for B. lesueur captured from two populations at Arid Recovery, South Australia
Mean cat-exposed
Mean cat-naïve
Mean difference
between treatments (%)
ANOVA results
Treatment Sex Age class
Body mass (kg) 1.37 ±0.12 1.18 ±0.13 0.19 (16.1) F
¼21.93, P<0.001 F
¼0.02, P¼0.888 F
¼2.48, P¼0.125
Head length (mm) 76.6 ±3.5 74.8 ±2.1 1.8 (2.4) F
¼3.34, P¼0.077 F
¼0.13, P¼0.720 F
¼0.02, P¼0.879
Pes length (mm) 104.1 ±3.0 103.2 ±2.1 0.9 (0.9) F
¼1.15, P¼0.292 F
¼2.45, P¼0.128 F
¼1.80, P¼0.189
Values are means ±1SD.
N. E. Tay et al. / Animal Behaviour 175 (2021) 45e5648
0.9 Cat-naïve Cat-exposed
Body mass (kg)
70 72.5 75 77.5 80 82.5
Body mass (kg)
Head length (mm)
70 72.5 75 77.5 80 82.5
Pes length (mm)
Head len
th (mm)
98 100.5 103 105.5 108 110.5
Body mass (kg)
Pes len
th (mm)
r16 = –0.03, P = 0.898
r15 = 0.58, P = 0.014
r15 = 0.30, P = 0.236
r16 = 0.41, P = 0.095
r15 = 0.50, P = 0.039
r16 = 0.07, P = 0.756
(a) (b)
(c) (d)
Figure 3. Treatment sample differences in (a) body mass, (b) the relationship between head length and body mass, (c) pes length and body mass and (d) head and pes length of cat-
exposed (N¼18, yellow triangles) and cat-naïve (N¼17, blue circles) Bettongia lesueur. Hollow symbols depict young adults (2 years) and lled symbols mature adults (>2 years).
Filled symbols with black borders show the original founders translocated to the predator paddock in 2014 (N¼5). For (a), bold horizontal lines represent the mean, the box
represents the standard deviation above and below the mean, and whiskers are the minimum and maximum values. Different letters denote a statistically signicant difference
between treatments. For (b-d), Pearson correlation and Pvalues for each treatment trendline are shown.
N. E. Tay et al. / Animal Behaviour 175 (2021) 45e56 49
As specic variables (i.e. speed) may be strongly associated with
body size, we also investigated body mass effects on individual
escape variables. We performed generalized linear models on each
of the eight escape variables separately using treatment (cat-
exposed or cat-naïve), body mass, their interaction and age class.
We used normal distribution with an identity link function for
continuous escape variables (1e8). For latency to leave and initial
gait we used binomial models with a logit link function, and for
direction of rst movement (0towards, 1forward, 2away) we
used an ordinal logistic regression.
Values are presented as means ±1SD throughout [95% con-
dence interval, CI, where applicable]. Effect sizes are presented as
) for analysis of variance models (Peters &Van Voorhis,
1940) and partial R
for generalized linear models (Edwards,
Muller, Wolnger, Qaqish, &Schabenberger, 2008).
Over 8 nights of trapping (total 160 trap nights), we trapped 18
bettongs from the predator paddock (cat-exposed) and 17 bet-
tongs from the predator-free paddock (cat-naïve). Five individuals
were recorded twice on successive nights; only data for their rst
escape behaviour trial were included in the analyses.
Three-quarters (76%) of cat-naïve animals trapped were 2
years of age compared with 50% of the cat-exposed bettongs. Of the
18 cat-exposed bettongs, ve individuals were the original released
stock translocated in 2014 (5 years old).
Body Size
Cat-exposed bettongs trapped during this experiment were on
average 16.1% heavier than those from the cat-naïve treatment
¼0.41 [0.15e0.61]; Table 1,Fig. 3a). There was less evidence of a
treatment effect on head length (
¼0.10 [0.00e0.32]) and no
evidence for pes length (
¼0.04 [0.00e0.23]; Table 1). We found
no sex or age class differences for the three body measures
(Table 1); however, only four bettongs tested from each treatment
were female.
The trends in body morphology recorded in our sample differed
between treatments. Head length was positively correlated with
body mass in the cat-naïve treatment (r
¼0.58, P¼0.014
[0.14e0.83]) but not in the cat-exposed treatment (r
P¼0.898 [-0.49e0.44]; Fig. 3b). While pes length increased with
body mass in both treatments (Fig. 3c), the correlation only
approached signicance for cat-exposed animals (r
P¼0.095 [-0.08e0.73]) and there was no signicant relationship
for cat-naïve animals (r
¼0.30, P¼0.236 [-0.21e0.68]). Correla-
tion between head and pes lengths was only evident within the cat-
naïve population (r
¼0.50, P¼0.039 [0.03e0.79]; Fig. 3d).
Escape Behaviour
Comparing the escape matrix between treatments, we found
some evidence of an overall difference in the presentation of escape
behaviour in cat-exposed versus cat-naïve bettongs (F
¼0.06). However, the distinction between treatments
was more apparent when considering the interaction of body mass
and treatment (F
¼3.24, P¼0.011,
¼0.09) even though body
mass alone had no inuence (F
¼0.59, P¼0.761,
¼0.02). We
found no evidence of age class affecting overall escape behaviour in
the model (F
¼0.75, P¼0.617,
¼0.02). The ve original
founder bettongs translocated from the predator-free paddock to
the predator paddock did not cluster together in the ordination, nor
was there a cluster by age class (Fig. 4).
SIMPER analysis revealed all eight variables contributed
somewhat equally to the overall dissimilarity in escape behaviour
between the two bettong populations (Tab le 2 ,Fig. 5). Latency to
–4 –2 0 2 4
NMDS Axis1
NMDS Axis 2
Faster s
eeds with lon
er bounds l Slower speeds with shorter bounds
More agile and irregular paths l Straigther paths
Stress = 0.199
Figure 4. Nonmetric multidimensional scaling ordination of escape behaviour in
Bettongia lesueur from trials from cat-exposed (N¼18, yellow triangles) and cat-naïve
(N¼17, blue circles) populations. Hollow symbols depict young adults (2 years) and
lled symbols mature adults (>2 years). Filled symbols with black borders show the
original founders translocated to the predator paddock in 2014 (N¼5). Overall escape
behaviour was derived from eight variables describing their reactivity (latency to leave,
initial gait, direction of rst movement), escape speed (maximum bound length,
average speed) and protean behaviour (agility, path irregularity and straightness). The
stress value of 0.199 represents a fairt of the projected 2D ordination compared to
the multidimensional data. NMDS axis 1 was correlated with escape speed (average
speed and maximum bound length), while NMDS axis 2 was correlated with protean
characteristics (straightness and path irregularity).
Table 2
Results of the similarity percentage analysis (SIMPER) comparing escape behaviour of cat-exposed (N¼18) and cat-naïve (N¼17) B. lesueur
Category Escape variable Average dissimilarity Contributing %
Reactivity Latency to leave 2.16 13.15
Initial gait 2.14 13.03
Direction of rst movement 2.04 12.46
Escape speed Maximum bound length 2.06 12.54
Average speed 2.12 12.96
Protean behaviour Agility 1.96 11.94
Path irregularity 1.94 11.84
Straightness 1.98 12.08
Escape behaviour was characterized by eight variables relating to reactivity, escape speed and protean characteristics.
N. E. Tay et al. / Animal Behaviour 175 (2021) 45e5650
0Cat-naïve Cat-exposed
Fled immediately Paused
0Cat-naïve Cat-exposed
Bipedal Quadrupedal
0Cat-naïve Cat-exposed
Cat-naïve Cat-exposed
Average speed (m/s)
ForwardAway Toward
0Cat-naïve Cat-exposed
Maximum bound length (m)
0Cat-naïve Cat-exposed
Agility index
Cat-naïve Cat-ex
Path irregularity
0Cat-naïve Cat-ex
(a) (b)
(c) (d)
(e) (f)
(g) (h)
Figure 5. Population summary of escape behaviour variables (a) latency to leave, (b) initial gait, (c) direction of rst movement, (d) maximum bound length, (e) average speed, (f)
agility, (g) path irregularity and (h) straightness. Escape behaviour was recorded from Bettongia lesueur from cat-exposed (N¼18, yellow) and cat-naïve (N¼17, blue) populations.
For box plots (d-h), bold horizontal lines represent the mean, the box represents the standard deviation above and below the mean, and whiskers are the minimum and maximum
N. E. Tay et al. / Animal Behaviour 175 (2021) 45e56 51
leave contributed 13.15% to treatment dissimilarity, with pausing
on release more evident in naïve bettongs (82%) than in cat-
exposed animals (50%; Fig. 5a). Initial gait contributed 13.03%
dissimilarity and more cat-exposed bettongs used the bipedal gait
when exiting the bag (67%) compared to 35% of cat-naïve animals
(Fig. 5b). For the direction of rst movement (SIMPER: 12.46%),
two-thirds (66.7%) of the cat-exposed bettongs escaped forward
with no regard to the position of the pole (Fig. 5c). By contrast,
three-quarters (76.47%) of the cat-naïve bettongs responded to the
position of the pole, moving towards (23.53%) or away from it
In terms of escape speed, maximum bound length for cat-
exposed bettongs averaged 1.49 ±0.30 m compared with
1.35 ±0.30 m for naïve bettongs (SIMPER: 12.54%; Fig. 5d). Average
speed was faster in cat-exposed bettongs (3.60 ±0.89 m/s)
compared to cat-naïve bettongs (3.07 ±0.74 m/s; SIMPER: 12.96%;
Fig. 5e).
The three protean measures were comparable between treat-
ments and together contributed 35.86% to the dissimilarity be-
tween cat-exposed and cat-naïve escape behaviour. Agility scores
were slightly higher in the cat-exposed bettongs compared to cat-
naïve bettongs (exposed: 3.86 ±1.85; naïve: 3.53 ±1.63), while
there was little difference in path irregularity (exposed:
0.97 ±0.27; naïve: 0.97 ±0.24) and straightness (exposed:
0.92 ±0.07; naïve: 0.91 ±0.08; Fig. 5f, g, h).
As cat-exposed bettongs were signicantly heavier than cat-
naïve bettongs and we detected different effects of body mass on
overall escape behaviour between treatments, we also investigated
whether body mass amplied any treatment differences in per-
formance for each escape measure (Table 3). We found signicant
interactions of treatment and body mass for initial gait (partial
¼0.15 [0.15e2.86]), maximum bound length (partial R
[0.03e0.55]) and average speed (partial R
¼0.26 [0.00e0.08];
Table 3). Age class did not affect any individual escape variable
(Table 3). Of the bettongs that started their escape with a bipedal
gait, cat-exposed bettongs were signicantly heavier than cat-naïve
bettongs (mean ±SD: exposed: 1.40 ±0.11 kg, N¼18; naïve: 1.10
±0.07 kg, N¼17). There were signicant positive correlations with
body mass for both maximum bound length (r
¼0.60, P¼0.009
[0.18e0.83]) and average speed (r
¼0.56, P¼0.015 [0.13e0.82])
in the cat-exposed population but not for the cat-naïve population
(Fig. 6).
We compared the behavioural responses of cat-exposed and
cat-naïve bettongs to a threatening stimulus and found that cat-
exposed bettongs had more immediate reactions, faster escape
speeds and/or more protean behaviour. Our results were
confounded by differences in body size between the two pop-
ulations, where heavier cat-exposed bettongs improved their
Table 3
Results of GLM models testing the effects of treatment, body mass, their interaction and age class on individual escape behaviour measures recorded from cat-exposed (N¼18)
and cat-naïve (N¼17) B. lesueur
Escape variable Treatment Body mass Interaction Age class
Test statistic PTest statistic PTest statistic PTest statistic P
Latency to leave Z¼0.89 0.372 Z¼0.35 0.726 Z¼-0.76 0.445 Z¼1.24 0.214
Initial gait Z¼-2.05 0.041 Z¼-1.37 0.170 Z¼2.18 0.029 Z¼0.00 0.999
Direction of rst movement t¼-1.39 0.164 t¼-1.35 0.177 t¼1.53 0.127 t¼-0.93 0.352
Maximum bound length t¼2.84 0.008 t¼2.81 0.009 t¼-2.91 0.007 t¼-0.98 0.334
Average speed t¼3.05 0.005 t¼2.72 0.011 t¼-3.21 0.003 t¼-0.50 0.623
Agility t¼1.47 0.152 t¼0.40 0.695 t¼-1.57 0.127 t¼-1.42 0.166
Path irregularity t¼0.42 0.680 t¼0.32 0.753 t¼-0.38 0.705 t¼-0.93 0.359
Straightness t¼-0.19 0.851 t¼0.79 0.438 t¼0.20 0.841 t¼0.21 0.832
Signicant Pvalues are shown in bold.
0.8 1 1.2 1.4 1.6
Maximum bound length (m)
1 1.2 1.4 1.6
Average speed (m/s)
r16 = 0.60, P = 0.009
r15 = –0.25, P = 0.326
r16 = 0.56, P = 0.015
r15 = –0.41, P = 0.098
(a) (b)
mass (k
Figure 6. Different treatment effects of body mass on escape speed variables: (a) maximum bound length and (b) average speed recorded during escape behaviour trials of Bettongia
lesueur from cat-exposed (N¼18, yellow triangles) and cat-naïve (N¼17, blue circles) populations. Hollow symbols depict young adults (2 years) and lled symbols mature adults
(>2 years). Filled symbols with black borders show the original founders translocated to the predator paddock in 2014 (N¼5). Pearson correlation and Pvalues for each treatment
trendline are shown.
N. E. Tay et al. / Animal Behaviour 175 (2021) 45e5652
escape behaviour (taking larger bounds and escaping at faster
speeds) while we found no similar trend with body mass for the
cat-naïve population. We also found no evidence of age affecting
escape behaviour. Overall, we have shown that exposing bettongs
to exotic predators for a period of 4 years was sufcient to
contribute to differences in their escape behaviour during eeing.
The observed differences could be due to selection and/or in-
dividual learning, but we did not measure escape behaviour before
cat exposure in individual bettongs to allow us to distinguish the
potential mechanism. However, no obvious trend or difference in
escape behaviour between founder bettongs and progeny within
the cat-exposed population could be detected (see Figs. 4 and 6).
There was also no obvious effect of age class on behaviour for the
bettongs tested.
Cat-exposed bettongs were larger than individuals from the cat-
naïve population. The difference may reect greater vulnerability
for smaller bettongs and thus selection for bigger animals in the
predator paddock. Larger body size may confer an escape advan-
tage for bettongs when released into the wild as they are difcult
for predators to handle (Grifths 1980) and can pose a greater risk
of injury to the predator (Fleming et al., 2020). Although our sample
size was relatively small, the result is consistent with larger cat-
exposed bettongs recorded by Moseby, Letnic, Blumstein, and
West (2018) for the same populations at Arid Recovery. Although
difcult to test, there is likely to be an interaction between learning
by individuals and direct selection on body size from predation.
Rapid responses play a central role in successfully evading a
sudden predator attack (Weihs &Webb, 1984;Whitford,
Freymiller, Higham, &Clark, 2019). In our study, cat-naïve bet-
tongs showed delayed escape initiation, more commonly pausing
on release, with some animals appearing to inspect the equip-
ment before deciding to ee. This behaviour of approaching an
unknown stimulus is not limited to the bettongs in our study; for
example, when presented with a cat model, 75% of naïve Aegean
wall lizards, Podarcis erhardii, approached the model before
eeing compared to only 35% of cat-exposed lizards (Li et al.,
2014). The act of inspecting a potential predator can provide an
animal with information on the nature of the threat (Dugatkin &
Godin, 1992), but pausing also increases the probability of being
captured and killed (Dugatkin &Godin, 1992;FitzGibbon, 1994).
Cats typically rely on stealth and approach their prey as close as
possible before pouncing (Turner &Meister, 1988); it would
therefore be benecial for prey to ee immediately when they
detect a potential threat, instead of pausing to inspect it. Cat-
exposed bettongs showed less hesitancy to leave their capture
bag and maximized the distance between them and the threat
by eeing forward without investing additional time in
inspecting the novel stimulus. Predator exposure has similarly
been shown to improve prey reaction times in desert kangaroo
rats. Dipodomys deserti (Freymiller, Whitford, Higham, &Clark,
We recorded the initial gait of bettongs as they exited their
capture bag, either bipedal hopping or quadrupedal bounding, as a
measure of hesitancy. Bettongia spp. transition to hopping at rela-
tively slow speeds (Webster &Dawson, 2003) and only use a
quadrupedal gait during slow activities such as foraging or wary
inspection behaviour (Stodart, 1966;Saxon-Mills et al., 2018).
Similarly, bipedalism is also consistently associated with faster
speeds and more protean manoeuvres during escape in sympatric
desert rodents (e.g. Djawdan &Garland, 1988;Taraborelli,
Corbalan, &Giannoni, 2003;Moore et al., 2017). More cat-naïve
bettongs (65%) used a quadrupedal gait upon release, slowly
inspecting the novel stimulus and runway environment. By
contrast, 67% of cat-exposed bettongs used a bipedal gait imme-
diately from exiting the bag.
Reactivity is likely to be inuenced by social environment,
which in this experiment could be reected in differences in pop-
ulation density between the paddocks. Higher bettong numbers
within the predator-free paddock could reect more intraspecic
competition for resources compared to the less dense predator
paddock, potentially contributing to their lower body mass. While
bettongs maintain social groupings and share warrens regardless of
the paddock in which they are trapped (Moseby, Blumstein, Letnic,
&West, 2018;Sander, Short, &Turner, 1997), density could affect
social interactions. More isolated individuals may naturally
perceive a greater predation risk than larger groups. While not
specically identied in bettongs, this group effect can be seen in
other social macropods where individuals decrease antipredator
vigilance when scanning for predators is shared among conspe-
cics (Favreau, Goldizen, &Pays, 2010).
Escape Speed
The difference in body mass alone could have caused behav-
ioural differences between the treatments, regardless of predator
exposure. Larger animals might be expected to perform longer
bounds and thus be capable of reaching faster speeds (Biewener &
Baudinette, 1995;Christiansen, 2002). Cat-exposed bettongs were
signicantly larger than those from the naïve population, but only
the cat-exposed bettongs ed at an intensity where body size
affected their escape performance; there was no body size effect for
the cat-naïve population. These results suggest that cat-naïve ani-
mals were not as motivated to ee as the cat-exposed bettongs.
Faster animals are at an advantage when attempting to outrun a
predator during a pursuit (Moore &Biewener, 2015). Average speed
obtained from our trials was approximately 3.3 m/s, which is
slower than Bettongia penicillata on treadmills under laboratory
settings (6.2 m/s; Webster &Dawson, 2003), but similar to free-
running Bettongia gaimardi at top speed (3.4 m/s; Bennett, 1987).
However, we only report an average speed over the 20 m long
runway; these values do not reect their maximum speeds.
Protean Behaviour
In a pursuit scenario, prey must balance speed and manoeuvring,
depending on the manoeuvrability of themselves and the predator
(Clemente &Wilson, 2015;Wynn et al., 2015). If the prey can simply
outrun its pursuer, it is less likely to invest energy in employing
protean behaviours. While we expected cat-exposed animals to
exhibit more protean characteristics in their escape path, our results
suggest that these characteristics of escape may not be as exible as
reactivity (e.g. Freymilleret al., 2017)or escape speed (e.g. Webster &
Dawson, 2003). While repeated experimental exposure to a stim-
ulus has been shown to increase FID in cat-exposed prey (Li et al.,
2014;West, Letnic, et al., 2018), other aspects of escape behaviour
may be less plastic and therefore less responsive to change. For
example, altering the unpredictability of their escape path may
require more complex cognitive control (i.e. Eilam, 2005) or require
N. E. Tay et al. / Animal Behaviour 175 (2021) 45e56 53
different morphologies (i.e. Moore et al., 2017), which are both likely
to require longer evolutionary time to perfect.
Limitations of This Study
We acknowledge that our study cannot properly mimic the ef-
fect of presence of a predator in the wild and that our test scenario
was articial. However, any approach by a human (e.g. West, Letnic,
et al., 2018) is not the same as a potential predator in terms of speed
or tenacity of pursuit. Similarly, chasing potential prey with cat
mounts or dead cats (e.g. Moseby et al., 2012) or the use of predator
scents to suggest their presence (e.g. Saxon-Mills et al., 2018) also
makes many assumptions about the relevance of the cues and the
naturalness of the animal's escape behaviour. Low sample size is
inherent in studies on threatened species, but the differences be-
tween experimental treatments recorded in this study reached
statistical signicance and showed moderate effect sizes, the
strongest being the signicant positive correlations with body mass
for both maximum bound length and average speed in the cat-
exposed population (but no correlations for the cat-naïve
Traditional approaches to instilling antipredator behaviour in
naïve animals involve classical conditioning by repeatedly pre-
senting a cue (commonly a model, scent or sound) associated with
the predator and pairing it with a fearful stimulus until the animal
forms a negative association (Grifn, Blumstein, &Evans, 2000).
However, these approaches only address the rst component of
predator naïvety (teaching an animal to recognize a predator) and
there is no mechanism to instil appropriate escape responses other
than avoidance. In situations where the predator is already in close
proximity and may initiate an attack, employing appropriate and
efcient escape behaviour to evade capture and therefore death or
signicant injury is essential. Investigating changes in escape
behaviour, in addition to measuring traditional avoidance behav-
iours, can provide wildlife managers with information on the ef-
cacy of antipredator programmes and how animals are responding
to new predators beyond just recognition.
Historically, the extinction of many small-to-medium-sized
species on mainland Australia coincided with spread of exotic
predators supplemented by high densities of European rabbits (see
Short, 1998). Effective control of both rabbits and exotic predators
outside fenced areas is nowadays possible due to combinations of
targeted trapping, biocontrol and baiting programmes. These more
favourable conditions, coupled with prerelease in situ exposure to a
low density of predators, may present an opportunity for native
Australian species, like the burrowing bettong, to re-establish
outside fenced enclosures. As bettongs in this study responded to
a general threat stimulus, the heightened antipredator behaviour in
cat-exposed bettongs may be applicable for escaping predators in
general, enabling these animals to persist in the wild among a
range of predation threats.
Our results suggest that cohabitating with predators appears to
have altered the bettong's perception of risk, either through direct
selection for individuals that have survived previous interactions
with cats and/or through learned behaviour by individuals that
have escaped a previous predation attempt or from observing the
predation of congeners. There is likely to be an interaction between
individual learning and selection with both leading to a population
level change in antipredator behaviour. This exposure to predators
appears to have led to a heightened sensitivity to any potential
threats they may encounter in the future. As a result, cat-exposed
bettongs responded more readily to the presentation and pursuit
by a threatening stimulus than their naïve counterparts. Our
research has shown that exposing previously naïve populations to
low levels of predation leads to changes in escape behaviour that
may improve survival in the presence of exotic predators. These
results add to a growing body of work (Blumstein, Letnic, &Moseby,
2019;Moseby, Letnic, et al., 2018;Ross, Letnic, Blumstein, &
Moseby, 2019;Saxon-Mills et al., 2018;West, Blumstein, Letnic, &
Moseby, 2018) that suggests in situ predator exposure can lead to
tangible changes in antipredator traits. Such changes are impera-
tive if we wish to ensure a future for these species outside predator-
free sanctuaries where they can coexist with introduced predators
in the wild.
This project was funded by the Holsworth Wildlife Research
Endowment eEquity Trustees Charitable Foundation &the
Ecological Society of Australia. This research was conducted at Arid
Recovery, a conservation research initiative supported by BHP Bil-
liton, The South Australian Department for Environment, The
University of Adelaide and Bush Heritage Australia. Thank you to
the volunteers and staff at Arid Recovery who assisted with eld
work. Thanks to K. Tuft for logistical support and M. Letnic, D.
Blumstein and R. West for initial work to establish the cat-exposed
bettong population.
Aaltonen, K., Bryant, A. A., Hostetler, J. A., & Oli, M. K. (2009). Reintroducing en-
dangered Vancouver Island marmots: Survival and cause-specic mortality
rates of captive-born versus wild-born individuals. Biological Conservation, 142,
Bannister, H. L., Lynch, C. E., & Moseby, K. E. (2016). Predator swamping and sup-
plementary feeding do not improve reintroduction success for a threatened
Australian mammal, Bettongia lesueur.Australian Mammalogy, 38,177e187.
Batschelet, E. (1981). Circular statistics in biology. New York: NY: Academic Press.
Bennett, M. B. (1987). Fast locomotion of some kangaroos. Journal of Zoology, 212,
Biewener, A., & Baudinette, R. (1995). In vivo muscle force and elastic energy storage
during steady-speed hopping of tammar wallabies (Macropus eugenii). Journal
of Experimental Biology, 198, 1829e18 41.
Biggins, D. E., Miller, B. J., Hanebury, L. R., & Powell, R. A. (2011). Mortalityof Siberian
polecats and black-footed ferrets released onto prairie dog colonies. Journal of
Mammalogy, 92,721e731.
Blumstein, D. T., Letnic, M., & Moseby, K. E. (2019). In situ predator conditioning of
naive prey prior to reintroduction. Philosophical Transactions of the Royal Society
B, 374, 20180058.
Blumstein, D. T., Samia, D. S., & Cooper, W. E. (2016). Escape behavior: Dynamic
decisions and a growing consensus. Current Opinion in Behavioral Sciences, 12,
Burbidge, A. A., & McKenzie, N. L. (1989). Patterns in the modern decline of western
Australia's vertebrate fauna: Causes and conservation implications. Biological
Conservation, 50,143e198. 006-3207(89)90009-8
Burbidge, A. A., & Short, J. C. (2008). Burrowing bettong, Bettongia lesueur. In S. Van
Dyck, & R. Strahan (Eds.), The mammals of Australia (3rd ed., pp. 288e290).
Sydney, Australia: Reed New Holland.
Christiansen, P. (2002). Locomotion in terrestrial mammals: The inuence of body
mass, limb length and bone proportions on speed. Zoological Journal of the
Linnean Society, 136, 685e714.
Clemente, C. J., & Wilson, R. S. (2015). Balancing biomechanical constraints: Optimal
escape speeds when there is a trade-off between speed and maneuverability.
Integrative and Comparative Biology, 55,1142e1154.
Djawdan, M, & Garland, T (1988). Maximal running speeds of bipedal and
quadrupedal rodents. Journal of Mammalogy, 69(4), 765e772.
Dugatkin, L. A., & Godin, J.-G. J. (1992). Prey approaching predators: A cost-benet
perspective. Annales Zoologici Fennici, 29, 233e252.
Edut, S., & Eilam, D. (2003). Rodents in open space adjust their behavioral response
to the different risk levels during barn-owl attack. BMC Ecology, 3,10.
Edwards, L. J., Muller, K. E., Wolnger, R. D., Qaqish, B. F., & Schabenberger, O.
(2008). An R
statistic for xed effects in the linear mixed model. Statistics in
Medicine, 27,6137e6157.
N. E. Tay et al. / Animal Behaviour 175 (2021) 45e5654
Eilam, D. (2005). Die hard: A blend of freezing and eeing as a dynamic defense-
dimplications for the control of defensive behavior. Neuroscience &Biobehav-
ioral Reviews, 29,1181e1191 .
Favreau, F.-R., Goldizen, A. W., & Pays, O. (2010). Interactions among social moni-
toring, anti-predator vigilance and group size in eastern grey kangaroos. Pro-
ceedings of the Royal Society B, 277, 2089e2095.
FitzGibbon, C. D. (1994). The costs and benets of predator inspection behaviour in
Thomson's gazelles. Behavioral Ecology and Sociobiology, 34,139e14 8.
Fleming, P. A., Crawford, H. M., Auckland, C. H., & Calver, M. C. (2020). Body size and
bite force of stray and feral catsdAre bigger or older cats taking the largest or
more difcult-to-handle prey? Animals, 10,707.
Freegard, C. J., Calver, M. C., Richards, J. D., & Bradley, J. (2007). Age estimation of
pouch young and growth of the burrowing bettong, Bettongia lesueur,(Mar-
supialia: Potoroidae) on Heirisson Prong, Shark Bay, western Australia. Austra-
lian Mammalogy, 29,157e167.
Freymiller, G. A., Whitford, M. D., Higham, T. E., & Clark, R. W. (2017). Recent in-
teractions with snakes enhance escape performance of desert kangaroo rats
(Rodentia: Heteromyidae) during simulated attacks. Biological Journal of the
Linnean Society, 122,651e660.
Friard, O., & Gamba, M. (2016). BORIS: a free, versatile open-source event-logging
software for video/audio coding and live observations. Methods in Ecology and
Evolution, 7, 1325e1330.
Grifn, A. S., Blumstein, D. T., & Evans, C. (2000). Training captive-bred or trans-
located animals to avoid predators. Conservation Biology, 14,1317e1326. https://
Grifths, D. (1980). Foraging costs and relative prey size. American Naturalist, 116,
743 e752.
Hammer, Ø., Harper, D. A., & Ryan, P. D. (20 01). PAST: paleontological statistics
software package for education and data analysis.
van Heezik, Y., Seddon, P. J., & Maloney, R. F. (1999). Helping reintroduced houbara
bustards avoid predation: Effective anti-predator training and the predictive
value of pre-release behaviour. Animal Conservation, 2,155
IUCN/SSC. (2013). Guidelines for reintroductions and other conservation trans-
locations. Gland, Switzerland: IUCN Species Survival Commission.
Johnson, C. N., & Isaac, J. L. (2009). Body mass and extinction risk in Australian
marsupials: The 'critical weight range' revisited. Austral Ecology, 34,35e40. 42-9993.2008.01878.x
Jones, K. A., Jackson, A. L., & Ruxton, G. D. (2011). Prey jitters; protean behaviour in
grouped prey. Behavioral Ecology, 22,831e836.
Legge, S., Murphy, B., McGregor, H., Woinarski, J., Augusteyn, J., Ballard, G., et al.
(2017). Enumerating a continental-scale threat: How many feral cats are in
Australia? Biological Conservation, 206, 293e303.
Li, B., Belasen, A., Palis, P., Bednekoff, P., & Foufopoulos, J. (2014). Effects of feral
cats on the evolution of anti-predator behaviours in island reptiles: Insights
from an ancient introduction. Proceedings of the Royal Society B, 281, 20140339.
Lima, S. L., & Dill, L. M. (1990). Behavioral decisions made under the risk of
predationea review and prospectus. Canadian Journal of Zoology, 68,619e640.
Moore, T. Y., & Biewener, A. A. (2015). Outrun or outmaneuver: Predatoreprey in-
teractions as a model system for integrating biomechanical studies in a broader
ecological and evolutionary context. Integrative and Comparative Biology, 55,
118 8e11 97.
Moore, T. Y., Cooper, K. L., Biewener, A. A., & Vasudevan, R. (2017). Unpredictability
of escape trajectory explains predator evasion ability and microhabitat prefer-
ence of desert rodents. Nature Communications, 8, 440.
Moseby, K. E., Blumstein, D. T., & Letnic, M. (2016). Harnessing natural selection to
tackle the problem of prey naivete. Evolutionary Applications, 9, 334e343.
Moseby, K., Carthey, A., & Schroeder, T. (2015). The inuence of predators and prey
naivety on reintroduction success: Current and future directions. In
D. Armstrong, M. Hayward, D. Moro, & P. Seddon (Eds.), Advances in reintro-
duction biology of Australian and New Zealand fauna (pp. 29e32). Melbourne,
Australia: CSIRO.
Moseby, K. E., Blumstein, D. T., Letnic, M., & West, R. (2018). Choice or opportunity:
Are post-release social groupings inuenced by familiarity or reintroduction
protocols? Oryx, 54,215e221.
Moseby, K. E., Letnic, M., Blumstein, D. T., & West, R. (2018). Designer prey: Can
controlled predation accelerate selection for anti-predator traits in naïve pop-
ulations? Biological Conservation, 217,213e221.
Moseby, K. E., Letnic, M., Blumstein, D. T., & West, R. (2019). Understanding predator
densities for successful co-existence of alien predators and threatened prey.
Austral Ecology, 44, 409e419.
Moseby, K. E., Cameron, A., & Crisp, H. A. (2012). Can predator avoidance training
improve reintroduction outcomes for the greater bilby in arid Australia? Animal
Behaviour, 83(4), 1011e1021. https://10.1016/j.anbehav.2012.01.023
Moseby, K. E., McGregor, H., & Read, J. L. (2020). Effectiveness of the Felixer
grooming trap for the control of feral cats: A eld trial in arid South Australia.
Wildlife Research.
Moseby, K. E., Read, J. L., Paton, D. C., Copley, P., Hill, B. M., & Crisp, H. A. (2011).
Predation determines the outcome of 10 reintroduction attempts in arid South
Australia. Biological Conservation, 144, 2863e2872.
Oksanen, J., Blanchet, F. G., Friendly, M., Kindt, R., Legendre, P., McGlinn, D., et al.
(2019). vegan: Community Ecology Package. R package version 2.5-5.https://cran.
Peters, C. C., & Van Voorhis, W. R. (1940). Statistical procedures and their mathe-
matical bases. New York, NY: McGraw-Hill.
Peterson, R. A., & Cavanaugh, J. E. (2020). Ordered quantile normalization: a sem-
iparametric transformation built for the cross-validation era. Journal of Applied
Statistics, 47(13-15), 2312e2327.
R Core Team. (2018). R: A Language and Environment for Statistical Computing.
Vienna, Austria: R Foundation for Statistical Computing.
Richardson, G., Dickinson, P., Burman, O. H. P., & Pike, T. W. (2018). Unpredictable
movement as an anti-predator strategy. Proceedings of the Royal Society B, 285,
Ross, A. K., Letnic, M., Blumstein, D. T., & Moseby, K. E. (2019). Reversing the effects
of evolutionary prey naivet
e through controlled predator exposure. Journal of
Applied Ecology, 56,1761e176 9.
Sander, U., Short, J., & Turner, B. (1997). Social organisation and warren use of the
burrowing bettong, Bettongia lesueur (Macropodoidea: Potoroidae). Wildlife
Research, 24,143e157.
Saxon-Mills, E. C., Moseby, K., Blumstein, D. T., & Letnic, M. (2018). Prey naïvet
e and
the anti-predator responses of a vulnerable marsupial prey to known and novel
predators. Behavioral Ecology and Sociobiology, 72,151.
Shier, D. M., & Owings, D. H. (2006). Effects of predator training on behavior and
post-release survival of captive prairie dogs (Cynomys ludovicianus). Biological
Conservation, 132,126e135.
Short, J. (1998). The extinction of rat-kangaroos (Marsupialia: Potoroidae) in New
South Wales, Australia. Biological Conservation, 86, 365e377.
Short, J., Kinnear, J., & Robley, A. (2002). Surplus killing by introduced predators in
Australiadevidence for ineffective anti-predator adaptations in native prey
species? Biological Conservation, 103, 283e301.
Short, J., & Turner, B. (1993). The distribution and abundance of the burrowing
bettong (Marsupialia: Macropoidea). Wildlife Research, 20, 525e533.
Short, J., & Turner, B. (1999). Ecology of burrowing bettongs, Bettongia lesueur
(Marsupialia : Potoroidae), on Dorre and Bernier islands, western Australia.
Wildlife Research, 26,651e669.
Sih, A. (1985). Evolution, predator avoidance, and unsuccessful predation. American
Naturalist, 125,153e157.
Stodart, E. (1966). Observations on the behaviour of the marsupial Bettongia lesueri
(Quoy &Gaimard) in an enclosure. Wildlife Research, 11,91e99.
Taraborelli, P., Corbalan, V., & Giannoni, S. (2003). Locomotion and escape modes in
rodents of the Monte desert (Argentina). Ethology, 109,475e485. https:// 03.00884.x
Tetzlaff, S. J., Sperry, J. H., & DeGregorio, B. A. (2019). Effects of antipredator training,
environmental enrichment, and soft release on wildlife translocations: A review
and meta-analysis. Biological Conservation, 236, 324e331.
Turner, D. C., & Meister, O. (1988). Hunting behaviour of the domestic cat. In
D. C. Turner, & P. Bateson (Eds.), Domestic cat: the biology of its behaviour (pp.
111e121). Cambridge, U.K.: Cambridge University Press.
Weihs, D., & Webb, P. W. (1984). Optimal avoidance and evasion tactics in predator-
prey interactions. Journal of Theoretical Biology, 106,189e206.
10.1016/0 022-5193(84)90019-5
West, R., Letnic, M., Blumstein, D. T., & Moseby, K. E. (2018). Predator exposure
improves anti-predator responses in a threatened mammal. Journal of Applied
Ecology, 55,147e156.
West, R. S., Blumstein, D. T., Letnic, M., & Moseby, K. E. (2018). Searching for an
effective pre-release screening tool for translocations: Can trap temperament
predict behaviour and survival in the wild? Biodiversity &Conservation, 28,
Whitford, M. D., Freymiller, G. A., Higham, T. E., & Clark, R. W. (2019). Determinants
of predation success: How to survive an attack from a rattlesnake. Functional
Ecology, 33, 1099e1109.
Woinarski, J. C., Burbidge, A. A., & Harrison, P. L. (2014). The action plan for Australian
mammals 2012. Melbourne, Australia: CSIRO.
Worrell, T., Admiraal, R., Bateman, P., & Fleming, P. (2017). Are tourism and con-
servation compatible for island tamespecies? Animal Conservation, 20,
155 e163.
Wynn, M. L., Clemente, C., Nasir, A. F., & Wilson, R. S. (2015). Running faster causes
disaster: Trade-offs between speed, manoeuvrability and motor control when
running around corners in northern quolls (Dasyurus hallucatus). Journal of
Experimental Biology, 218, 433e439. 82
Webster, K. N., & Dawson, T. J. (20 03). Locomotion energetics and gait character-
istics of a rat-kangaroo, Bettongia penicillata, have some kangaroo-like features.
Journal of Comparative Physiology B, 173,549e557.
N. E. Tay et al. / Animal Behaviour 175 (2021) 45e56 55
Table A1
Details of the 35 Bettongia lesueur trapped and tested during this study
Treatment Bettong ID Sex Year of rst capture Estimated age (years) Age class
Cat-exposed 2861 M 2014 5 Mature
Cat-exposed 3067 F 2014 5 Mature
Cat-exposed 3093 M 2014 5 Mature
Cat-exposed 3172 M 2014 5 Mature
Cat-exposed 3286 M 2014 5 Mature
Cat-exposed 4165 M 2016 3 Mature
Cat-exposed 4107 M 2016 3 Mature
Cat-exposed 370 M 2016 3 Mature
Cat-exposed 4167 F 2016 3 Mature
Cat-exposed 4274 F 2017 2 Young
Cat-exposed 3894 M 2018 1 to 2 Young
Cat-exposed 4879 F 2018 1 to 2 Young
Cat-exposed 4811 M 2018 1 to 2 Young
Cat-exposed 4823 M 2018 1 to 2 Young
Cat-exposed 4808 M 2018 1 to 2 Young
Cat-exposed 4816 M 2018 1 to 2 Young
Cat-exposed 4878 M 2018 1 to 2 Young
Cat-exposed 5251 M 2018 1 to 2 Young
Cat-naïve 2751 M 2016 3 Mature
Cat-naïve 3314 F 2016 3 Mature
Cat-naïve 3472 M 2016 3 Mature
Cat-naïve 3759 M 2016 3 Mature
Cat-naïve 4841 M 2018 1 to 2 Young
Cat-naïve 4873 M 2018 1 to 2 Young
Cat-naïve 4913 M 2018 1 to 2 Young
Cat-naïve 4966 M 2018 1 to 2 Young
Cat-naïve 5249 M 2018 1 to 2 Young
Cat-naïve 5273 M 2018 1 to 2 Young
Cat-naïve 5274 F 2018 1 to 2 Young
Cat-naive 5275 M 2018 1 to 2 Young
Cat-naïve 4980 F 2018 1 Young
Cat-naïve 4988 M 2018 1 Young
Cat-naïve 5223 M 2018 1 Young
Cat-naïve 5247 F 2018 1 Young
Cat-naïve 5250 M 2018 1 Young
M: male; F: female. Age is based on sexual development on capture (i.e. pouch condition and testes development).
N. E. Tay et al. / Animal Behaviour 175 (2021) 45e5656
... Mirza & Chivers 2003;de Oliveira Mesquita & Young 2007;West et al. 2018) or allow prey to learn to respond to predators either by observing peers (Arai et al. 2007;Alfieri & Dugatkin 2009;Griesser & Suzuki 2017) or by associating predator presence with other alarm cues (Berejikian et al. 1999;Vilhunen 2006). This type of training is commonly called antipredator training, and is also referred to as predatoravoidance training (Moseby et al. 2012;Edwards et al. 2020), predator exposure training (West et al. 2018), predator recognition training (Ross et al. 2019;Tay et al. 2021), and predator awareness training (Rowell et al. 2020). ...
Captive‐rearing is commonly used to support efforts to restore extirpated or diminished wild populations. Many rearing programs include antipredator training in an effort to reduce post‐release mortality due to predation, but the value such training adds to restoration efforts is not fully understood. Analysis of antipredator training impacts often focus on changes in prey following training and not the actual effects when prey are exposed to predators. Studies also commonly assess prey traits instead of fitness. However, changes in prey following training may not translate to desired increase in fitness upon release for a self‐sustaining population. To consider these relationships, we identified studies that analyzed changes in organisms following both training and lethal exposure to predators. Notably, we identified only 12 studies that analyzed outcomes in both training and lethal exposure settings and found the majority of measurements focused on traits. Although meta‐analysis indicated that antipredator training changed prey traits and increased prey survival, relationships between changes in traits and fitness could not be assessed given the rarity of studies that included both outcomes. While these results support the general use of antipredator training, they also suggest that trait changes cannot be reliably used as proxies for fitness impacts. Future work should seek to more clearly measure impacts on fitness and determine when changes in traits can truly be used as a surrogate for training efficacy. These issues reflect recent concerns regarding research on non‐consumptive effects and highlight potential avenues for future collaborative research on fear. This article is protected by copyright. All rights reserved.
... This could have significant implications for biodiversity conservation with research demonstrating the capacity of threatened species to coexist in the presence of feral cats, but not foxes Miritis et al. 2020;Rendall et al. 2021a). The capacity of these threatened species to persist is associated with actively managed low feral cat densities , altered behavioural traits including increased wariness due to feral cat exposure Tay et al. 2021), or associated with habitat complexity (Miritis et al. 2020). Our results suggest that the regulation and control of red foxes, for the purpose of species conservation, may be an appropriate management method even when control of feral cats is not possible or is infeasible-either due to legislative or technical limitations (Johnston et al. 2020). ...
Full-text available
Invasive species control has resulted in unintended consequences where the control of one species causes increased impacts of a co-occurring species. Red foxes (Vulpes vulpes) and feral cats (Felis catus) co-occur throughout Australia-with control and regulation of fox's commonplace, a greater understanding of how feral cats will respond in the absence of foxes is needed. In this study we use feral cat diet to assess potential prey switching after red fox control. Feral cat stomach contents were sampled between 1983 and 1994 on Phillip Island, Victoria, Australia when both feral cats and red foxes were abundant; and again from 2016 to 2019 when foxes were eradicated. A total of 277 feral cats were sampled between 1983-1994 and 415 from 2016 to 2019. Although we did not track feral cat diet throughout the red fox decline, diet comparisons between the two time periods suggest no substantial shift in feral cat diet with only slight increases in invertebrates and black rats observed. Invasive prey (European rabbits, black rats, and house mice) still formed the majority of feral cat diet. We further reinforce the role of seasonality, surrounding land use and sex as factors influencing cat diet. Our results suggest red fox control in the presence of cats may still achieve conservation benefits with feral cats maintaining a comparable diet, dominated by invasive species, despite fox control-though this may be context dependent. Given the abundance and frequency of invasive prey species within cat diet, we further support the idea that invasive prey control could be a viable indirect method to control the impacts of feral cats.
... Despite awareness that behavioural traits such as boldness or shyness can influence conservation outcomes, measuring such traits is rarely incorporated into monitoring and management [16,43]. Anti-predator responses have only recently been identified as a potential barrier to the success of conservation projects [13,[37][38][39], and while an array of academic literature exists that details various methods for measuring these behaviours [15,38,39,[44][45][46][47][48], accessing the methodologies, comparing them for rigor and identifying the most appropriate measure is labour intensive. Stakeholders, such as conservation and population managers, are likely to be seeking this information, but also likely to be limited by the time and resources necessary to find it. ...
Full-text available
Background Mammals, globally, are facing population declines. Strategies increasingly employed to recover threatened mammal populations include protecting populations inside predator-free havens, and translocating animals from one site to another, or from a captive breeding program. These approaches can expose predator-naïve animals to predators they have never encountered and as a result, many conservation projects have failed due to the predation of individuals that lacked appropriate anti-predator responses. Hence robust ways to measure anti-predator responses are urgently needed to help identify naïve populations at risk, to select appropriate animals for translocation, and to monitor managed populations for trait change. Here, we outline a protocol for a systematic review that collates existing behavioural assays developed for the purpose of quantifying anti-predator responses, and identifies assay types and predator cues that provoke the greatest behavioural responses. Methods We will retrieve articles from academic bibliographic databases and grey literature sources (such as government and conservation management reports), using a Boolean search string. Each article will be screened for the satisfaction of eligibility criteria determined using the PICO (Population—Intervention—Comparator—Outcome) framework, to yield the final article pool. Using metadata extracted from each article, we will map all known behavioural assays for quantifying anti-predator responses in mammals and will then examine the context in which each assay has been implemented (e.g. species tested, predator cue characteristics). Finally, with mixed effects modelling, we will determine which of these assays and predator cue types elicit the greatest behavioural responses (standardised difference in response between treatment and control groups). The final review will highlight the most robust methodology, will reveal promising techniques on which to focus future assay development, and will collate relevant information for conservation managers.
Full-text available
As carnivorans rely heavily on their head and jaws for prey capture and handling, skull morphology and bite force can therefore reflect their ability to take larger or more difficult-to-handle prey. For 568 feral and stray cats (Felis catus), we recorded their demographics (sex and age), source location (feral or stray) and morphological measures (body mass, body condition); we estimated potential bite force from skull measurements for n = 268 of these cats, and quantified diet composition from stomach contents for n = 358. We compared skull measurements to estimate their bite force and determine how it varied with sex, age, body mass, body condition. Body mass had the strongest influence of bite force. In our sample, males were 36.2% heavier and had 20.0% greater estimated bite force (206.2 ± 44.7 Newtons, n = 168) than females (171.9 ± 29.3 Newtons, n = 120). However, cat age was the strongest predictor of the size of prey that they had taken, with older cats taking larger prey. The predictive power of this relationship was poor though (r2 < 0.038, p < 0.003), because even small cats ate large prey and some of the largest cats ate small prey, such as invertebrates. Cats are opportunistic, generalist carnivores taking a broad range of prey. Their ability to handle larger prey increases as the cats grow, increasing their jaw strength, and improving their hunting skills, but even the smallest cats in our sample had tackled and consumed large and potentially ‘dangerous’ prey that would likely have put up a defence.
Full-text available
Normalization transformations have recently experienced a resurgence in popularity in the era of machine learning, particularly in data preprocessing. However, the classical methods that can be adapted to cross-validation are not always effective. We introduce Ordered Quantile (ORQ) normalization, a one-to-one transformation that is designed to consistently and effectively transform a vector of arbitrary distribution into a vector that follows a normal (Gaussian) distribution. In the absence of ties, ORQ normalization is guaranteed to produce normally distributed transformed data. Once trained, an ORQ transformation can be readily and effectively applied to new data. We compare the effectiveness of the ORQ technique with other popular normalization methods in a simulation study where the true data generating distributions are known. We find that ORQ normalization is the only method that works consistently and effectively, regardless of the underlying distribution. We also explore the use of repeated cross-validation to identify the best normalizing transformation when the true underlying distribution is unknown. We apply our technique and other normalization methods via the bestNormalize R package on a car pricing data set. We built bestNormalize to evaluate the normalization efficacy of many candidate transformations; the package is freely available via the Comprehensive R Archive Network.
Full-text available
Inappropriate anti‐predator responses (naiveté) towards introduced predators is a key factor contributing to the extinction and endangerment of prey species world‐wide and the failure of wildlife reintroductions. Here, we test the idea that success of reintroduction can be improved by exposing a predator naïve prey species to introduced predators under controlled conditions (in situ predation) prior to reintroduction, such that prey adopt increased wary behaviours to aid in survival. We exposed a population of a naïve marsupial, the greater bilby (Macrotis lagotis), to a controlled number of introduced predators (feral cats, Felis catus) for 2 years within a large fenced paddock and then compared the pre‐release behaviour and post‐translocation survival of predator‐exposed and predator‐naïve bilbies over 40 days. Behavioural assays conducted in a small pen prior to reintroduction suggested that predator‐exposed bilbies were warier as they spent less time moving and more time in cover than predator naïve bilbies. After translocation, predator‐exposed bilbies were more likely to survive to 40 days and were less likely to be preyed upon by cats than predator‐free bilbies. Synthesis and applications. Naiveté towards predators is a major problem thwarting successful reintroductions world‐wide. Our study demonstrates that exposure to predators under controlled conditions can increase survival of reintroduced prey and is a promising approach to overcome the problem of naiveté towards introduced predators and the global problem of prey naiveté. Future conservation of naïve prey species may depend on such training methods prior to releasing into areas where predators are present.
Full-text available
1. The selection pressures that arise from capturing prey and avoiding predators are some of the strongest biotic forces shaping animal form and function. Examining how performance (i.e., athletic ability) affects the outcomes of encounters be- tween free-ranging predators and prey is essential for understanding the determinants of predation success rates and broad scale predator–prey dynamics, but quantifying these encounters in natural situations is logistically challenging. 2. The goal of our study was to examine how various metrics of predator/prey performance determine predation success by studying natural predator–prey interactions in the field with minimal manipulation of the study subjects. 3. We used high-speed video recordings of free-ranging sidewinder rattlesnakes (predator, Crotalus cerastes) and desert kangaroo rats (prey, Dipodomys deserti) to study how performance at various stages of their encounters alters the outcome of their interactions. 4. We found that predation success depends on (a) whether the rattlesnake struck accurately, (b) if the rattlesnake strike was accurate, the reaction time and escape manoeuvres of the kangaroo rat, and (c) if the kangaroo rat was bitten, the ability of the kangaroo rat to use defensive manoeuvres to avoid subjugation by the snake. 5. The results of our study suggest that the role of performance in predator–prey interactions is complex, and the determinants of predation success are altered by both predator and prey at multiple stages of an interaction in ways that may not be apparent in many experimental contexts.
Full-text available
Individuals often respond to threatening situations in consistently different ways and these differences may predict later translocation success. Thus, the ability to easily identify these differences prior to translocation may assist in improving conservation outcomes. We asked whether burrowing bettongs (Bettongia lesueur), a marsupial species that has undergone significant decline since the introduction of exotic predators to Australia, responded in consistently different ways to capture in traps, and if so, whether this was related to anti-predator behaviour, ranging behaviour and survival following translocation. Behavioural responses of 40 bettongs were measured and included response to removal from traps (trap docility), latency to leave a trap or bag and escape behaviour upon release. We used flight initiation distance to measure escape behaviour, and distance moved from diurnal refuges during nocturnal foraging to measure ranging behaviour. Survival was measured through radiotracking after release. Behaviours scored during removal from a trap were consistent and repeatable, and formed a behavioural syndrome with anti-predator and ranging behaviour. Less docile bettongs foraged closer to refuges and had longer flight initiation distances. Less docile bettongs were also more likely to survive after release, although the sample size of mortalities was small. Our results suggest that behaviours scored during trapping could be a useful metric for pre-release screening in translocation programs to enhance the chances of individual survival post-release.
ContextFeral cats pose a significant threat to wildlife in Australia and internationally. Controlling feral cats can be problematic because of their tendency to hunt live prey rather than be attracted to food-based lures. The Felixer grooming trap was developed as a targeted and automated poisoning device that sprays poison onto the fur of a passing cat, relying on compulsive grooming for ingestion. AimsWe conducted a field trial to test the effectiveness of Felixers in the control of feral cats in northern South Australia where feral cats were present within a 2600-ha predator-proof fenced paddock. Methods Twenty Felixers were set to fire across vehicle tracks and dune crossings for 6 weeks. Cat activity was recorded using track counts and grids of remote camera traps set within the Felixer Paddock and an adjacent 3700-ha Control Paddock where feral cats were not controlled. Radio-collars were placed on six cats and spatial mark–resight models were used to estimate population density before and after Felixer deployment. Key resultsNone of the 1024 non-target objects (bettongs, bilbies, birds, lizards, humans, vehicles) that passed a Felixer during the trial was fired on, confirming high target specificity. Thirty-three Felixer firings were recorded over the 6-week trial, all being triggered by feral cats. The only two radio-collared cats that triggered Felixers during the trial, died. Two other radio-collared cats appeared to avoid Felixer traps possibly as a reaction to previous catching and handling rendering them neophobic. None of the 22 individually distinguishable cats targeted by Felixers was subsequently observed on cameras, suggesting death after firing. Felixer data, activity and density estimates consistently indicated that nearly two-thirds of the cat population was killed by the Felixers during the 6-week trial. Conclusions Results suggest that Felixers are an effective, target-specific method of controlling feral cats, at least in areas in which immigration is prevented. The firing rate of Felixers did not decline significantly over time, suggesting that a longer trial would have resulted in a higher number of kills. ImplicationsFuture studies should aim to determine the trade-off between Felixer density and the efficacy relative to reinvasion.
Many translocations and introductions to recover threatened populations fail because predators kill prey soon after release; a problem exacerbated for predator-naive prey. While pre-release training has been shown to work in some situations, it is time consuming and relies on using inferred predator cues and treating small groups. We review a relatively new and very promising management tool: in situ , pre-release predator conditioning. Here, the goal is to allow prey in large enclosures to live with low densities of predators to accelerate selection for antipredator traits (in an evolutionary sense) or provide prey essential experience with predators that they will later encounter. We review the published results of a large-scale, controlled experiment where we have permitted burrowing bettongs ( Bettongia lesueur ) and greater bilblies ( Macrotis lagotis ) to live with low densities of feral cats ( Felis catus ), a species implicated in their widespread decline and localized extinction. We found that both species could persist with cats, suggesting that future work should define coexistence thresholds—which will require knowledge of prey behaviour as well as the structure of the ecological community. Compared to control populations, predator-naive prey exposed to cats has a suite of morphological and behavioural responses that seemingly have increased their antipredator abilities. Results suggest that predator-conditioned bilbies survive better when released into a large enclosure with an established cat population; future work will determine whether this increased survival extends to the wild. This article is part of the theme issue ‘Linking behaviour to dynamics of populations and communities: application of novel approaches in behavioural ecology to conservation’.
Wildlife translocations can have conservation value but results have been mixed regarding animal behavior and survival post-release. Practitioners have adopted antipredator training, environmental enrichment, and soft release as pre-release conditioning tactics to encourage adaptive behavior and improve post-release survival, but their utility has not been broadly quantified. We performed a formal literature review and conducted meta-analysis on 108 effects from 41 studies experimentally testing how these tactics affected survival, movement, or site fidelity compared to unconditioned animals. We further investigated how each conditioning tactic, animal source (wild-to-wild translocated or captive-released), age, and taxonomic group (birds, fish, mammals, and reptiles) influenced outcomes. Relative to unconditioned animals, conditioned individuals were 1.5 times more likely to survive, had reduced movement, and were three times more likely to show site fidelity. Each of the three conditioning tactics resulted in improved survival. Juveniles released from captivity derived the greatest survival benefit from conditioning. Across taxa, conditioning most benefitted survival of fish. Conditioning also had positive effects on survival of mammals and reptiles, albeit with less certainty than for fish. Estimates comparing survival of conditioned to unconditioned birds were much more variable, suggesting avian translocation programs using conditioning generally need improvement. Soft release consistently reduced movement and increased site fidelity; this was an especially viable technique for adult wild-to-wild translocated animals. We provide quantitative evidence that behavioral conditioning can aid wildlife translocations, and we encourage continued experiments to further elucidate how refined tactics could advance conservation efforts using translocation as a management tool.
The Action Plan for Australian Mammals 2012 is the first review to assess the conservation status of all Australian mammals. It complements The Action Plan for Australian Birds 2010 (Garnett et al. 2011, CSIRO Publishing), and although the number of Australian mammal taxa is marginally fewer than for birds, the proportion of endemic, extinct and threatened mammal taxa is far greater. These authoritative reviews represent an important foundation for understanding the current status, fate and future of the nature of Australia. This book considers all species and subspecies of Australian mammals, including those of external territories and territorial seas. For all the mammal taxa (about 300 species and subspecies) considered Extinct, Threatened, Near Threatened or Data Deficient, the size and trend of their population is presented along with information on geographic range and trend, and relevant biological and ecological data. The book also presents the current conservation status of each taxon under Australian legislation, what additional information is needed for managers, and the required management actions. Recovery plans, where they exist, are evaluated. The voluntary participation of more than 200 mammal experts has ensured that the conservation status and information are as accurate as possible, and allowed considerable unpublished data to be included. All accounts include maps based on the latest data from Australian state and territory agencies, from published scientific literature and other sources. The Action Plan concludes that 29 Australian mammal species have become extinct and 63 species are threatened and require urgent conservation action. However, it also shows that, where guided by sound knowledge, management capability and resourcing, and longer-term commitment, there have been some notable conservation success stories, and the conservation status of some species has greatly improved over the past few decades. The Action Plan for Australian Mammals 2012 makes a major contribution to the conservation of a wonderful legacy that is a significant part of Australia’s heritage. For such a legacy to endure, our society must be more aware of and empathetic with our distinctively Australian environment, and particularly its marvellous mammal fauna; relevant information must be readily accessible; environmental policy and law must be based on sound evidence; those with responsibility for environmental management must be aware of what priority actions they should take; the urgency for action (and consequences of inaction) must be clear; and the opportunity for hope and success must be recognised. It is in this spirit that this account is offered. Winner of a 2015 Whitley Awards Certificate of Commendation for Zoological Resource.
The high failure rate of threatened species translocations has prompted many managers to fence areas to protect wildlife from introduced predators. However, conservation fencing is expensive, restrictive and exacerbates prey naïveté reducing the chance of future co‐existence between native prey and introduced predators. Here, we ask whether two globally threatened mammal species protected in fenced reserves, with a history of predation‐driven decline and reintroduction failure, could co‐exist with introduced predators. We defined co‐existence as population persistence for at least 3 years and successful recruitment. We manipulated the density of feral cats within a large fenced paddock and measured the impact on abundance and reproduction of 353 reintroduced burrowing bettongs and 47 greater bilbies over 3 years. We increased cat densities from 0.038 to 0.46 per square km and both threatened species survived, reproduced and increased their population size. However, a previous reintroduction trial of 66 bettongs into the same paddock found one red fox (Vulpes vulpes), at a density of 0.027 per square km, drove the bettong population extinct within 12 months. Our results show that different predator species vary in their impact and that despite a history of reintroduction failure, threatened mammal species can co‐exist with low densities of feral cats. There may be a threshold density below which it is possible to maintain unfenced populations of reintroduced marsupials. Understanding the numerical relationships between population densities of introduced predators and threatened species is urgently needed if these species are to be re‐established at landscape scales. Such knowledge will enable a priori assessment of the risk of reintroduction failure thereby increasing the likelihood of reintroduction success and reducing the financial and ethical cost of failed translocations.