First Camera Trap based Evidence of Grey Wolf Canis lupus in the Hanma National Nature Reserve, Inner Mongolia, China

Abstract

As an apex predator, the grey wolf (Canis lupus) is an ecologically important species. It is considered an ecologically important species due to its position as an apex predator. Grey wolves survive in a wide range of habitats including deserts, steppe, tundra, shrubs, coniferous and deciduous forests. Grey wolves have a cosmopolitan distribution, mostly found in the northern hemisphere. Due to historical and continued persecution; and reduced prey populations, its current range is restricted to remote areas. Thanks to conservation initiatives, grey wolves are beginning to reclaim parts of its historical distribution, currently listed as Least Concern by the IUCN. During a two year camera-trapping survey, we obtained the first photographic detection of grey wolves in Hanma National Nature Reserve, China. We deployed 113 camera traps spaced at least 1km apart, which ran for 27,607 trap nights. On October 11th 2017 at 16:40, a camera located in the coniferous forest detected two adult grey wolves. Thus, we report the first photographic detection of grey wolves in HNNR. We emphasize the need for more research to further determine the true distribution of grey wolves in China and suggest that wildlife managers can use the same conservation strategies applied in HNNR to other areas in order to assist grey wolf recovery.

Full text

Inuences of Predator Cues on the Incidence of
Ungulates, Mesopredators and Top Predators in the
Greater Khingan Mountains, Northeastern China
Hamenya Mpemba1, Fan Yang 1, Kirsty J. MacLeod2,3, Dusu Wen1, Yan Liu1,4 and
Guangshun Jiang1,*
1Feline Research Center of National Forestry and Grassland Administration, College
of Wildlife and Protected Area, Northeast Forestry University, 26 Hexing Road,
Harbin, Heilongjiang 150040, P.R. China
2Department of Ecosystem Science and Management, Pennsylvania National
University, Forest Resources Building, University Park, PA 16802, USA
3Department of Biology, Lund University, Sölvegatan 37, 223 62 Lund, Sweden
4General Station for Surveillance of Wildlife-Borne Infectious Diseases, State Forestry
and Grassland Administration, Shenyang, Liaoning Province, 110034, PR China
Hamenya Mpemba and Fan Yang have contributed equllay to this article.
Article Information
Received 13 February 2019
Revised 10 March 2019
Accepted 15 March 2019
Available online 07 March 2022
(early access)
Published 27 October 2022
Authors’ Contribution
GJ, HM and FY conceived and
designed the study. HM, FY, DW and
YL conducted eld work and collected
data. KM and HM analysed the data
and revised the manuscript. HM wrote
the manuscript.
Key words
Predator cues, Ungulates,
Mesopredators, Panthera tigris altaica,
Camera trap, Fear ecology.
Top predators can aect the behaviour of prey species via lethal (direct kill) or non-lethal eects (i.e.,
through predation risk). For example, prey species may move from areas perceived as risky to safer
spaces where predation risk is lower, which can have important consequences for investment in foraging,
movement, and mating, and for the behaviour and habitat use of other species, such as mesopredators.
These changes in prey and mesopredator behaviours are likely mediated by the presence of predator cues
in the environment. Here, we test how dierent predator cues (visual and odor) from familiar and novel
predators (brown bear and Amur tiger, respectively) inuence ungulate, mesopredator, and top predator
visitation rates to camera trap sites in a national nature reserve in China. The comparison of these predator
types is of particular interest in this region as Amur tigers may shortly be reintroduced here. We found
that visual but not odour cues signicantly aected ungulate visitation rates: ungulates showed reduced
visitation to sites with either a novel or familiar visual predator cue. When combined, mesopredators and
top predators also showed a small reduction in visitation rates to tiger cue sites compared to bear cue sites,
suggesting a possible novel predator eect. The generalisation and contextual importance of predator cues
for prey and mesopredators have been little studied. Understanding how species respond to novel cues
may help to determine extinction probabilities and overall plasticity in the face of change. This study is,
therefore, an important step forward in understanding predator cue responses at the community level. This
is also the rst study to test the ecological function of Amur tiger cues in the wild environment and may
serve as essential information in the rewilding process of captive Amur tiger plans.
INTRODUCTION
Top predators can aect the behaviour of prey species
via lethal (direct kill) or non-lethal eects (i.e.,
through predation risk) (Lima and Bedneko, 1999;
Tolon et al., 2009; Farnworth et al., 2016). Predation risk
* Corresponding author: jgshun@126.com
0030-9923/2023/0001-269 $ 9.00/0
Copyright 2023 by the authors. Licensee Zoological Society of
Pakistan.
This article is an open access article distributed under the terms
and conditions of the Creative Commons Attribution (CC BY)
license (https://creativecommons.org/licenses/by/4.0/).
their activity patterns when confronted with risk (Lima
and Bedneko, 1999; Tolon et al., 2009). Dierential use
of habitats due to variation in predation risk is known as
the landscape of fear phenomena (Laundré et al., 2001;
Hernández and Laundré, 2005; Brook et al., 2012). Prey
species such as ungulates tend to move from “risky” zones
(e.g., where predator density is higher, or refuge scarcer) to
safer spaces (e.g., reduced predator density, more refugia
available) to reduce their predation risk (Tolon et al. (2009).
Alterations in behavior can result in trade-os between
daily activities such as foraging, movement, and mating,
and safety behaviours, especially in high risk habitats.
For example, both ungulates and mesopredators adopt
“safety” behaviours like vigilance, reduced foraging time,
and shifting to safe habitats when under high predation
ABSTRACT
Pakistan J. Zool., vol. 55(1), pp 269-280, 2023 DOI: https://dx.doi.org/10.17582/journal.pjz/20190213140239

2 7 0
risk (Creel et al., 2005; Li et al., 2011; Zheng et al., 2013;
Kuijper et al., 2014). Thus, top predators can indirectly
control the behavior and habitat use of mesopredators, as
well as prey (Palacios et al., 2016).
Changes in prey and mesopredator behaviours
are likely mediated by the presence of predator cues in
the environment (Apfelbach et al., 2005; Caro, 2005).
According to Creel et al. (2008) presence of cues indicating
the availability of predators within the habitat can be
sucient enough for prey to assess the predation risk within
the area. Top predator odour cues like feces, urine, and fur,
are used by most mammalian prey species to detect the
presence of danger during foraging; cues, therefore, help
to recognize the extent of risk (Garvey et al., 2016, 2017).
The presence of predator odours can trigger prey species
to respond by altering their behavior (Apfelbach et al.,
2005; Caro, 2005), such as decreasing visitation to specic
areas (Nersesian et al., 2012) and increased protection of
young (Schulte et al., 2013). In the presence of wolves,
elk change behaviours like vigilance levels and movement
(Creel et al., 2005, 2008). The accuracy of risk perception
based on predator cues may increase once dierent cues
are combined, or used in dierent contexts; for example
lizards and newts use predator cues in a habitat-dependent
manner, relying more on visual cues in an open area,
and odour cues in a dense vegetative habitat (Mathis and
Vincent, 2000). Visual and odour cues have the potential
to convey dierent information (Smith and Belk, 2001),
and may also dier in intensity and longevity (Brown and
Cowan, 2000; Kats and Dill, 2016), as well as detectability
by dierent species (e.g., some may have a greater ability
to detect visual cues than others based on dierences in
visual acuity (Gonzálvez and Rodríguez-Gironés, 2013).
A number of studies have shown the eects of
predator cues on ungulate and/or mesopredator behavior
(e.g., Hughes et al., 2010; Cremona et al., 2014; Kuijper et
al., 2014; Wikenros et al., 2015; Natt et al., 2017; Suárez-
Tangil and Rodríguez, 2017; Wikenros et al., 2017;
Goullaud et al., 2018). However, less has been done to
identify specic cues used to assess the presence of risk at
the community level (Winnie, 2012; Beschta and Ripple,
2013; Kuijper et al., 2014). For example, although prey
get information about the proximity of predators via some
cue channels, for the most part in the type of visual and
odour cues (Kelley and Magurran, 2003), most studies
have considered only one cue. The relative importance of
cue type for dierent species (of both predator and prey)
is therefore relatively understudied. Further, the study on
the eects of predator cues on prey species behavior vastly
outweighs that on the eects on mesopredator species,
though this is likely to be equally important in structuring
communities, and some clear eects have been shown. For
example, Palacios et al. (2016) showed that experimental
modication of apex predator cues (e.g., predator
odour and visual cues) reduced the distance swum, the
area used and even foraging behaviours carried out by
mesopredators. Mesopredators also respond especially
quickly to predator species that have recently consumed
conspecics (Cremona et al., 2014).
The Siberia Tiger Park in Heilongjiang, China has
bred a reasonable Amur tiger number and is planning a
program to train and reintroduce them in the wild (Wang et
al., 2018). Thus, comparison of these predator types is of
particular interest in this region, as for how native prey and
mesopredator species respond to these novel cues gives
us insight into how predator-prey dynamics are likely to
unfold in this region going forward. Currently, we are in
a period of anthropogenic change; species are shifting
ranges, and animals may be encountering novel predators
that are moving into their range (Chen et al., 2011; Van
Dyck, 2012; Wong and Candolin, 2015). Knowing how
species respond to novel cues might help to determine
extinction probabilities and overall plasticity in the face of
change. To our current understanding, this is the rst study
that involves captive Amur tiger ecological inuence on
prey site visitation rate in the wild, which as we have noted
is likely to be important in the future as this species regains
some of its prevalence in the region.
Here, we test how dierent predator cues (visual and
odour) from familiar and novel predators (brown bear vs.
Amur tiger) inuence ungulate, mesopredator, and top
predator visitation rates to camera trap sites in Hanma
National Nature Reserve, China. Previous work has
demonstrated reductions in visitation rate to sites where
predator cues are prevalent in both ungulates (Kuijper et
al., 2014; Wikenros et al., 2015; Venter et al., 2017), and
mesopredators (Palacios et al., 2016; Wikenros et al., 2017;
Haswell et al., 2018; Sivy et al., 2018). We, therefore,
predict that overall, ungulate and mesopredator visitation
rates should be reduced at camera trap sites where predator
cues are present, versus sites where control cues are
present. We also predict that cue type might dierentially
inuence prey and mesopredator visitation rates to camera
trap sites. Dierent prey responds dierently towards
dierent predator cues, but also the intensity of predator
cues determines the intensity of response (Gonzálvez and
Rodríguez-Gironés, 2013). We predict that odour cues
might induce a weaker response in terms of visitation rate
because they are more aected by wind and are likely to be
less long-lived than visual cues (Brown and Cowan, 2000;
Kats and Dill, 2016). We also predict dierences in prey and
mesopredator response to predator cues based on predator
type. Correct identication and response to predator cues
are important for prey species tness and often relies on
experience (Carthey and Blumstein, 2017; Saxon-Mills
H. Mpemba et al.

2 7 1
et al., 2018). Due to this, prey species may show a very
minimal or no response to cues from a novel predator or
one that has not existed in a system for many generations
(Saxon-Mills et al., 2018). Amur tigers have been absent
from this region for many decades with no reported
sightings in our study area in recent history (Turvey et al.,
2017). We predict that prey and mesopredator response to
familiar predator cues (brown bear visual odour cues) will
be greater (a more signicant drop in visitation rate) than
the response to novel predator cues (tiger visual and odour
cues).
MATERIALS AND METHODS
Study site
This study was conducted between July 2016 and
June 2017 in the Greater Khingan Mountains of Hanma
National Nature reserve (51°20′02″‒51°49′48″N,
122°23′34″-122°52′46″E), close to the small city of
Genhe, Inner Mongolia, North East China (Fig. 1). In
total, Hanma comprises an area of 1073.48 km2. Habitats,
including vegetation and species characteristics, are
described elsewhere (Zhai-Penghui, 2015; Guo et al.,
2017). Predators and mesopredators present in this system
include the Eurasian lynx (Lynx lynx), wolf (Canis lupus),
wolverine (Gulo gulo), and brown bear (Ursus arctos).
Naturally occurring ungulate species include roe deer
(Capreolus pygargus), moose (Alces alces), musk deer
(Moschus sifanicus), reindeer (Rangifer tarandus), and red
deer (Cervus elaphus) (Zhai-Penghui, 2015).
Experiment design
Design
First, we established treatment plots or grids, with
one square plot for each treatment type: the visual cue
experiment comprised three plots (Aa- tiger visual, Ab-
bear visual, Ac- control), and the odour cue experiment
comprised two plots (Ba- tiger, Bb- control; Fig. 1). Plot
grids were determined randomly using the shnet method
in ArcGIS 9.3 (Krivoruchko, 2011), which laid grid cells
of specied areas within the study area. Here, we set up
camera traps, at least 1km from each other. At rst, the
odour group grids were designed to be the same as the
visual cue experiment (i.e., with one plot for the treatment
and another plot for the control), but given unavoidable
dierences in environmental conditions across the
designated grid sites (i.e., mountains and swamp areas).
We, therefore, redesigned the layout for this experiment so
that each grid contained both treatment and control (i.e.,
Ba and Bb, respectively; Fig. 1) to avoid confounding
eects of habitat type. We installed a total of 113 camera
traps (LTL ACORN scouting camera; Ltl 5210 5MP and
Ltl 5210A 12MP, and Nighthawk Bestguarder SG-990V
infrared sensor trigger self-timer digital camera) in all over
the Hanma national nature reserve. In our designed study
grids there were a total of 86 cameras. We additionally
set up 27 no treatment cameras to establish baseline
visitation rates with no manipulation present; these were
placed randomly within and outside the grids to the areas
with high chances to capture mammals, i.e., nearby water
points, areas with numerous animal ways and droppings.
Cameras were set to a video mode to record for 20s at each
5s interval when triggered by an animal passing across.
The treatments and controls used were: photos of Amur
tigers Panthera tigris tigris (Kitchener et al., 2017) and
brown bear Ursus arctos (Blanco et al., 2011) as visual
cues and a blank model (vegetation) as their control.
Lastly, we used Amur tiger feces as odour cue and soil as
its control (Fig. 2).
Visual cue plots were specied as grids of 3.2 km ×
3.2 km; the odour cue experiment plots were specied as
grids of 1.6 km × 1.6 km. The latter was smaller due to the
topography of this area having lower accessibility. Each
visual cue plot had 18 camera traps, and the smaller odour
cue plots had 8 camera traps each. In total we established
seven plots; two for treatments (i.e., Fig. 1A, C) and one
as control (Fig. 1C) for the visual cue experiment. The
two other plots for odour experiment were divided each
into treatment and its control due to the reasons stated
above; making it four as previously stated. We chose to
have separate plots for each treatment type as opposed to
dispersing predator cue and control sites randomly due to
the small size of the reserve which might have resulted in
treatments inuencing the visitation rates of nearby camera
trap sites. The separation of treatment types ensured that
this was not the case. We are condent that there was no
eect of grid placement on visitation rates, again due to
the small size of the reserve, and because all grids were
similar in habitat and vegetation density.
Study 1: Visual cue experiment
In tiger visual grid (Fig. 1Aa), tiger visual cues (a
life-size photo of a standing tiger printed on a canvas;
Figure 2 following studies by Li et al. (2011) and Fischer
et al. (2017)) were placed at pre-designated plot points
by tying the print to two trees located adjacent to each
other. Then, cameras were tied about 50cm height from
the ground to one of the trees to record incidence of
visitation and ungulate/mesopredator behaviours. The bear
visual grid (Fig. 1Ac) was set up in the same way, with
brown bear visual cues (life-size photo of a standing bear
printed on a canvas). A control grid (Fig. 1Ab), contained
canvases printed with a photo of vegetation/habitat-like
environment. Each visual cue grid contained 18 camera
trap plots (18 x tiger, 18 x bear, 18 x controls).
Fear Eects of Predator Cues on Prey 271

2 7 2
Fig. 1. Experimental sites and camera trap distribution in Hanma National Nature Reserve. Black boxes indicate reserve areas
selected for experimental treatment group sites as determined by the shnet method in ArcGIS 9.3. Then A- visual cue experiment
(Aa- tiger, Ab- bear, Ac- control), and the odour cue experiment comprises of (Ba- tiger, Bb- control). Tiger visual (Aa) includes a
life-size photo of a standing tiger printed on a canvas which is tied to two trees that are side by side, with a camera tied to one of
the trees; the same size and settings were for the bear visual (Ac). For the control visual (Ac), a canvas with the same size as tiger
and bear visual cues was printed with vegetation colour to mimic the habitat and then tied side by side to two trees with a camera
tied to one of the trees to record visitation incidences. For the tiger odour cue experiment (Ba, b), a plastic bottle with tiger feces
(i.e., Ba groups) was tied to a tree, and then a camera is tied to the opposite tree to record the visitation rate; the same was done to
odour cue control (Bb) groups, but here the bottle was lled with soil. Finally, there were cameras at no treatment sites indicated
as white round; spots with neither treatment nor control. Red circles indicate individual camera trap sites each of which includes
one motion-triggered camera.
H. Mpemba et al.

2 7 3
Fear Eects of Predator Cues on Prey 273
Fig. 2. Experiment, and control visual and odour cues. From top left to bottom right is a tiger, bear, and control visual cues; and
tiger odour cue, respectively. Tiger odour control cue looks similar to tiger control cue with the dierences in their content (tiger
feces vs. soil).
Study 2: Odour cue experiment
The Amur tiger feces from Harbin Siberian Tiger Park
were collected by hand and refrigerated (maximum of 6
days in storage) before being transferred to 150 ml plastic
bottles, pierced with small holes in order to let the tiger
feces odour escape. Plastic bottles (150 ml) with the same
volume of soil were used as controls. In both treatment
and control plots, the plastic bottle containing either soil
or feces was tied at a tree bark (Fig. 2D) at the height that
is equivalent to the camera trap height (located at opposite
tree) so that the camera could record animal incidences
of visitation at the station and behaviours expressed. Each
odour cue experiment grid contained 8 camera trap sites
(i.e., 8 x tiger feces, and 8 x controls at Figure 1Ba; and 8
x tiger feces, and 8 x controls at Figure 1Bb).
Data collection
After one year of recording, we collected the memory
cards (SD) from the cameras and brought them back to
the laboratory for analysis. Using focal animal sampling
and all-occurrence methodology (Altmann, 1974; Lehner,
1992; Martin et al., 1993; Margulis, 2016) we extracted
the required data (namely – incidences in which an
ungulate or mesopredator triggered the camera to record).
Triggered recordings that were within 30 min of each other
and contained the same animal were counted as one visit.
Ungulate species detected were roe deer, moose, musk
deer, red deer, and wild boar. Mesopredators were dened
as predators or population of predators that are classied
to be in the mid-size of the available predators within the
given ecosystem (Groom et al., 2006; Prugh et al., 2009;
Wallach et al., 2015); under this denition, we detected
sable Martes zibellina and weasels Mustela nivalis in this
system. Top predators (apex predators, alpha predators
or mega-predators) were dened as large predators with
no natural predators to feed on them within the food
chain (Groom et al., 2006; Prugh et al., 2009; Sukhdeo,
2012; Wallach et al., 2015); we detected bear, lynx, and

2 7 4
H. Mpemba et al.
wolverine as top predators in this system.
For the visual cue experiment, data were collected
from a total of 83 deployed cameras. Some cameras
stopped functioning soon after deployment (within 30
days) and were not included in data analysis (n = 3 tiger
photo cameras, n = 2 bear photo cameras, n = 6 control
cameras) resulting in a sample of 69 cameras. In this
sample, the mean length of camera activity was 280.0
days ± 11.1, and the total number of mammals captured on
camera was 997.
For the odour cue experiment, data were collected
from a total of 61 deployed cameras. Again, we excluded
cameras that were active for fewer than 30 days (n = 1
control cameras, no predator odour cameras excluded),
resulting in a sample of 57 cameras. In this sample, the
mean length of camera activity was 251.6 days ± 14.5, and
the total number of mammals captured on camera was 837.
We also collected data from 27 no treatment cameras,
again excluding cameras that were active for fewer than
30 days (n = 3 cameras), resulting in a total of 24 cameras.
Statistical analysis
For analysis of both experiments, the number of a)
ungulates, b) mesopredators, and c) top predators was set
as the dependent variable in three separate generalized
linear models (GLMs), with treatment, and number of
days of camera activity as independent variables. For the
photo cue experiment, treatment was a four-level factor:
no treatment, tiger cue, bear cue, and control cue. For the
odour cue experiment, treatment was a three-level factor:
no treatment, tiger odour cue, control cue. Given the
skew towards 0s and low numbers in mammal incidence
data in both experiments, we used a negative binomial
error structure to account for overdispersion, with a
log link. All models were constructed using the glm.nb
function (MASS package; Venables and Ripley, 2002) in
the statistical software, R version 3.2.3 (R Core Team,
2015). We compared models (the full model, single-term
models, a null model) using AIC, and report the best model
(lowest AIC value). Signicance values in models were
determined using likelihood ratio tests. Where treatment
was signicant, we used post-hoc tests (Tukey tests) to
determine whether some treatment types had stronger
eects than others, as predicted.
For analysis of the photo experiment, we initially
constructed models containing either all four treatments,
or the two predator cue treatments grouped (i.e., a predator
cue treatment level), and tested which better t the data
using AIC comparison (Akaike, 2011). If the model
containing the four-level treatment variable (including
both predator cue types) was within 2AICc of the model
containing the three-level treatment variables (predator
cue types grouped), we maintained the separate cue types,
as we were interested in potential dierences in responses
to dierent predators.
In models analysing the odour cue experiment, we
additionally included cue ID as an independent variable
to account for potentially varying eects of dierent
individual cues, as two dierent tiger odour cues and two
dierent control cues were used. In all three models, cue
ID was not a signicant predictor variable, so we do not
report results for this here.
Fig. 3. Ungulate visitation rates (individuals/day) at
camera trap sites under four dierent treatments: no cue,
predator photo cues (amur tiger/brown bear), and a control
cue. Treatment signicantly predicted ungulate visitation
rate (X2
3
= 18.53, P <0.001; Table I). Visitation to sites with
the tiger cue and the bear cue diered signicantly from
visitation to sites where no cue was present (tiger photo
Z = -4.51, P < 0.0001; bear photo Z = -2.97, P = 0.015,
respectively). Visitation to no cue and control cue sites did
not dier (Z = -1.65, P = 0.35). Visitation responses to tiger
and bear cues were not statistically dierent (Z = -1.46, P
= 0.46).
RESULTS
Photo cue experiment
Treatment signicantly predicted ungulate visitation
to camera trap sites (X2
3
= 18.53, P <0.001; Table I; Fig. 3).
Post-hoc tests revealed that visitation to sites with the tiger
cue and the bear cue was signicantly lower than visitation
to sites where no cue was present (tiger photo Z = -4.51, P
< 0.0001; bear photo Z = -2.97, P = 0.015, respectively).
Visitation to no cue and control cue sites did not dier,
as expected (Z = -1.65, P = 0.35). Interestingly, visitation
responses to tiger and bear cues were not statistically
dierent (Z = -1.46, P = 0.46). The length of camera
activity (days) was positively correlated with ungulate
visitation (X2
1
= 4.38, P <0.05).

2 7 5
Fear Eects of Predator Cues on Prey 275
Table I.- Results from best model (GLM) of ungulate
visitation to camera traps.
Estimate S.E. Z P
Intercept 1.82 0.44 4.10
Treatment 0.0003
No cue 0.00
Tiger photo cue -1.61 0.36 -4.51
Bear photo cue -1.01 0.34 -2.97
Control cue -0.66 0.40 -1.65
Days of camera activity 0.003 0.001 2.18 0.004
Table II.- Results from best model (GLM) of all
predator (mesopredator + top predator) visitation to
camera traps.
Estimate S.E. Z P
Intercept 0.68 0.31 2.22
Treatment 0.06
No cue 0.00
Tiger photo cue -0.83 0.59 -1.41
Bear photo cue 0.88 0.52 -1.70
Control cue -0.39 0.66 -0.59
Fig. 4. Predator (mesopredators and top predators
combined) visitation rates (individuals/day) at camera trap
sites under four dierent treatments: no cue, predator photo
cues (amur tiger/brown bear), and a control cue. Combined
mesopredators and top predators showed a small eect of
treatment on visitation rates (X2
3
= 7.43, P = 0.06; Table II):
post-hoc testing showed that the only signicant dierence
between cue types was that visitation to bear cue sites was
slightly higher than visitation to tiger cue sites (Z = -2.62,
P < 0.05).
Treatment did not signicantly predict mesopredator
visitation (X2
1
= 1.14, P = 0.77); the null model best
described this data. Treatment also did not signicantly
predict top predator visitation (X2
1
= 2.81, P = 0.24);
the null model best described this data. Combining
mesopredators and top predators showed a small eect of
treatment on visitation rates (X2
3
= 7.43, P = 0.06; Fig. 4):
post-hoc testing showed that the only signicant dierence
between cue types was that visitation to bear cue sites was
slightly higher than visitation to tiger cue sites (Z = -2.62,
P < 0.05).
Odour cue experiment
Odour cue treatment did not signicantly predict
ungulate visitation (X2
1
= 0.71, P = 0.70); mesopredator
visitation (X2
1
= 1.14, P = 0.77); or, top predator visitation
(X2
1
= 2.81, P = 0.24) to camera trap sites. The null model
best described these data in all three cases. The null
model also best described visitation rates of all predators
combined; treatment was not signicant (X2
2
= 2.12, P =
0.35).
DISCUSSION
In this study, we tested how visual and odour cues
from familiar and novel predators (brown bear vs. Amur
tiger) inuenced ungulate, mesopredator, and top predator
visitation rates to camera trap sites in Hanma National
Nature Reserve. We show that visual cues predicted
ungulate visitation rates and ungulates responded equally
to the novel and familiar predators. Odour cue, on the
other hand, did not inuence ungulate visitation rates.
Mesopredator visitation was not inuenced by either
cue type, contrary to predictions, though mesopredators
and top predators combined showed a weakly stronger
negative response to tiger versus bear cues, suggesting a
small novel predator eect on visitation rates. Predator
avoidance has vital biological and transformative
outcomes on communities and trophic cascades (Peacor
and Werner, 2001; Gonzálvez and Rodríguez-Gironés,
2013); nevertheless, the generalisation and contextual
importance of predator cues have also been little studied.
This study is, therefore, an essential and exciting step
forward in understanding predator cue responses at
the community level, which may have implications for
conservation in this region. There has been a study by
Wang et al. (2018) on the captive Amur tiger responses
towards its natural prey cues; here we test the response
of prey towards the Amur tiger, providing insight into its
ecological functioning in the wild environment to which it
may soon be reintroduced.
Photo cue experiment
There has been work by various authors on studying
the inuence of predation risk to ungulates, but less has

2 7 6
H. Mpemba et al.
been done to identify the cue used by ungulates to assess
presence of risk (Winnie, 2012; Beschta and Ripple, 2013;
Kuijper et al., 2014), and equally little to understand
whether mesopredators use the same cues within a
community. Our results conrm that predation risk as
assessed using predator visual cues (both Amur tiger
and brown bear photos) signicantly inuences ungulate
visitation rate at predator cues sites. There was a dierence
in visitation rate of ungulates (roe deer, moose and musk
deer) to the camera sites with tiger and bear visual cues
compared to camera sites at the control sites as well as
at no treatment sites: more ungulates visited no treatment
and control sites as opposed to sites where there were
predator visual cues. This result is in line with a study
on tammar wallabies (Macropus eugenii), showing that
they respond to visual predator cues; tammar wallabies
increased vigilance behaviours and reduced feeding rate
after confronting fox visual cues (Blumstein et al., 2000).
Our results are contrary to a study by Venter et al. (2017),
which reported that large grazing herbivores (such as
zebra, red hartebeest, and eland) do not appear to rely on
visual cues while on foraging activities. How ungulate
species respond to visual predator cues may have a size-
dependent component. For example, small prey species
such as impala, warthog, waterbuck, and kudu minimize
their chances of encountering predators by evading the
use of same space with all predators, while larger ungulate
species like wildebeest, girae, and zebra only avoided
sharing space with lion and leopards (Thaker et al.,
2011). Our ungulate group contained moose, roe deer, and
musk deer, with roe deer a majority. Moose are relatively
large and comparable to zebra, red hartebeest, and eland,
whereas roe deer and musk deer are much smaller. This may
explain the dierence in our results from those of Venter et
al. (2017), indicating that visual cues inhibit visitation rate
to the predator visual cue camera sites. Unfortunately, our
small sample size of moose mean that we cannot further
interpret our results based on ungulate size classes, but this
would be a fruitful area of further study in this system.
We show that responses of ungulates to Amur tiger
and brown bear cues in terms of site visitation are not
statistically dierent, though tiger cues are relatively novel
given their long absence in this region, contrary to our
prediction that ungulates should show a stronger response
to the cue with which they were more familiar (Wiles et
al., 2003; Carthey and Blumstein, 2017; Saxon-Mills et
al., 2018). This indicates a degree of generalization of
predator visual cues that do not appear to rely strongly on
previous experience. This is similar to Dunlop-Hayden
and Rehage (2011) results which indicated no dierences
in prey reaction towards native vs. non-native predators.
Alternatively, ungulates may have been responding to
novelty per se, which in itself can elicit fear responses
(Sneddon et al., 2003). An appropriate way to test this
further might be to repeat the experiment using an
additional novel cue such as a photograph of a train or other
novel object, to determine whether the equal response to
tiger cues as bear cues is because tiger cues are inherently
novel, or because they are recognised as a predatory threat.
It is worth to note that there are no any literature which
show existence of tiger in the study area, even local people
suggest that tiger has never lived in this ecosystem.
Mesopredator release theory suggests that an
increase in apex predators also suppresses mesopredators
in the area (Haswell et al., 2018). Mesopredators are
likely to also use predator cues, for example, Switalski
(2003) suggested that when wolves are within the area
of study, coyotes use a visual cue to detect their predator
(wolves) availability. Additionally, Palacios et al. (2016)
proposed that that presence of any kind of apex predator
cue (coral whether visual, chemical or even combined
(coral trout Plectropomus leopardus) limits the distance
its mesopredator prey (dottyback Pseudochromis fuscus)
would swim or engage in other activities like foraging.
Interestingly our study contradicts the above studies
because both Amur tiger and brown bear cues showed no
signicant inuence on visitation rates of mesopredators
alone (though in combination with top predators,
mesopredators showed slightly lower visitation to tiger
cue sites). This may be because brown bear are few in our
study area compared to the incidence of top predators in the
coral trout and wolf systems, hence lessening the chance
that bears interact directly with mesopredators, implying
that mesopredators might not use bear visual cues (but
may still respond to a novel cue, such as the tiger).
Odour cue experiment
Wikenros et al. (2015) studied the response of red deer
and roe deer to Eurasian lynx olfactory cues, concluding
that both red deer and roe deer decreased their visitation
duration at the treatment sites. Also, Noell (2013) states
that ungulates have a strong ability to detect the smell,
but their response to predator cues vary (Apfelbach et al.,
2005). We predicted that odour cue would predict ungulate
visitation rates signicantly in the wild, but this was not
the case in our study. Nevertheless, our data support other
previous ndings that indicate that predator cues do not
aect deer visitation rate to the chemical cue environment
(Kimball et al., 2009; Elmeros et al., 2011). We believe
that wetness and wind have a role in determining the
eectiveness and longevity of odour cues. The odour cue
experimental sites were located in the mountain that faces
a large swamp area that contains water in summer and
autumn. This might have inuenced our result too due to

2 7 7
Fear Eects of Predator Cues on Prey 277
the wet wind that blows towards the side of the camera trap
experiment site. The current study acts as a foundational
base for another ecologist to carry out further experiments
on the ecological functioning role of captive Amur tiger
odour in the wild.
Our result is in line with numerous other studies
which reported that there is a weaker response of
mammalian mesopredators (Garvey et al., 2016) towards
the risk posed by their top predator. Tiger odour cue did
not predict mesopredator visitation to the camera trap
sites. Sih et al. (2010) suggested that prey species may
need to have preceding experience toward the predator
chemical cue before an experiment can be done. We agree
with this idea; and we think that mesopredators of our
study area might have evolved a lack of fear of tigers due
to unfamiliarity (Suraci et al., 2017). This is because as
discussed in the visual cue experiment that there is no clear
record evidence that tiger ever existed in the area recently;
but some archives indicate that there are signs for the tiger
to have existed in the area in the distant past (Turvey et
al., 2017). The same reasons (wetness and windward side)
discussed in the above paragraph for ungulates applies here
too. It would be informative to repeat the experiment using
the familiar predator odour cue (brown bear), as the lack of
response by both ungulates and mesopredators toward the
tiger odour cue may have been due to lack of familiarity,
rather than a lack of response generally to chemical cues.
Unfortunately, we did not have access to suitable brown
bear cues for this study.
Visual vs odour cues inuence on visitation rate
Normally visual cue indicates the predator is present
and prey/mesopredator will be in immediate danger if
they come face to face with the cue, this means they have
to minimize unnecessary movements or opting to ee
from the area. However, feces may mean that there is a
tiger in the area but investigating the cue could provide
information on the sex of the donor, what it has eaten,
etc. Tigers are territorial and the feces were fresh, it could
mean that the donor is not likely to be back to this spot
for a while and therefore the area is safe in the immediate
future. Hence less fear, similar to our results. Apfelbach
et al. (2005) indicate that body odour is a better test for
antipredator responses than feces or urine, and studies of
antipredator responses are moving towards using body
odour if possible rather than feces as it is a less confusing
cue of immediate predator presence.
CONCLUSION
To our understanding, this is the rst study to test
the ecological function of Amur tiger cues in China
and possibly worldwide in the wild environment. We
demonstrate that captive Amur tiger visual and odour cues
do not predict mesopredators and top predator visitation
rates signicantly, but both captive Amur tiger and brown
bear visual cues predict the visitation rate of ungulates
signicantly to predator treatment sites. Our results
have laid down foundations on the understanding of the
ecological functioning of captive Amur tiger visual and
odour cues. Therefore our study together with that of Wang
et al. (2018) may save as key in the rewilding process of
captive Amur tiger plans.
ACKNOWLEDGMENTS
This work was supported by funds from the
Fundamental Research Funds for the Central Universities
(2572017PZ14), the fund from the National Key
Programme of Research and Development, Ministry of
Science and Technology (2016YFC0503200), and National
Nature Science Foundation of China (NSFC 31872241;
31572285). We thank Bao Heng and Zhai Penghu for their
support during eldwork. Thanks to the sta at Hanma
National Nature Reserve for their eld work support,
the management of Heilongjiang Amur Tiger Park who
provided us with the fresh Amur tiger feces. This article is
part of the principal author Ph. D. work.
Statement of conict of interest
The authors declare that they have no conict of
interests regarding the publication of this article.
REFERENCES
Altmann, J., 1974. Observational study of behavior:
Sampling methods. Behaviour, 49: 227-266.
https://doi.org/10.1163/156853974X00534
Apfelbach, R., Blanchard, C.D., Blanchard, R.J.,
Hayes, R.A. and McGregor, I.S., 2005. The eects
of predator odors in mammalian prey species: A
review of eld and laboratory studies. Neurosci.
Biobehav. Rev., 29: 1123-1144. https://doi.
org/10.1016/j.neubiorev.2005.05.005
Beschta, R.L. and Ripple, W.J., 2013. Are wolves
saving yellowstone’s aspen? A landscape-level
test of a behaviorally mediated trophic cascade:
Comment. Ecology, 94: 1420-1425. https://doi.
org/10.1890/11-0063.1
Blanco, J.C., Ballesteros, F., García-Serrano, A.,
Herrero, J., Nores, C. and Palomero, G., 2011.
Behaviour of brown bears killing wild ungulates
in the cantabrian mountains, southwestern europe.
Eur. J. Wildl. Res., 57: 669-673. https://doi.

2 7 8
H. Mpemba et al.
org/10.1007/s10344-010-0464-z
Blumstein, D.T., Daniel, J.C., Grin, A.S. and Evans,
C.S., 2000. Insular tammar wallabies (Macropus
eugenii) respond to visual but not acoustic cues
from predators. Behav. Ecol., 11: 528-535. https://
doi.org/10.1093/beheco/11.5.528
Brook, L.A., Johnson, C.N. and Ritchie, E.G., 2012.
Eects of predator control on behaviour of an apex
predator and indirect consequences for mesopredator
suppression. J. appl. Ecol., 49: 1278-1286. https://
doi.org/10.1111/j.1365-2664.2012.02207.x
Brown, G.E. and Cowan, J., 2000. Foraging trade-
os and predator inspection in an ostariophysan
sh: Switching from chemical to visual
cues. Behaviour, 137: 181-195. https://doi.
org/10.1163/156853900502015
Caro, T., 2005. Antipredator defenses in birds and
mammals. University of Chicago Press, Chicago,
Illinois, United States.
Carthey, A.J.R. and Blumstein, D.T., 2017. Predicting
predator recognition in a changing world. Trends
Ecol. Evol., 33: 106-115. https://doi.org/10.1016/j.
tree.2017.10.009
Chen, I.C., Hill, J.K., Ohlemüller, R., Roy, D.B. and
Thomas, C.D., 2011. Rapid range shifts of species
associated with high levels of climate warming.
Science, 333: 1024-1026. https://doi.org/10.1126/
science.1206432
Creel, S., Winnie, J., Maxwell, B., Hamlin, K. and
Creel, M., 2005. Elk alter habitat selection as an
antipredator response to wolves. Ecology, 86:
3387-3397. https://doi.org/10.1890/05-0032
Creel, S., Winnie, J.A., Christianson, D. and Liley,
S., 2008. Time and space in general models of
antipredator response: Tests with wolves and
elk. Anim. Behav., 76: 1139-1146. https://doi.
org/10.1016/j.anbehav.2008.07.006
Cremona, T., Crowther, M.S. and Webb, J.K., 2014.
Variation of prey responses to cues from a
mesopredator and an apex predator. Austral. Ecol.,
39: 749-754. https://doi.org/10.1111/aec.12138
Dunlop-Hayden, K. and Rehage, J.S., 2011.
Antipredator behavior and cue recognition by
multiple everglades prey to a novel cichlid
predator. Behaviour, 148: 795-823. https://doi.
org/10.1163/000579511X577256
Elmeros, M., Winbladh, J.K., Andersen, P.N., Madsen,
A.B. and Christensen, J.T., 2011. Eectiveness of
odour repellents on red deer (Cervus elaphus) and
roe deer (Capreolus capreolus): A eld test. Eur. J.
Wildl. Res., 57: 1223-1226. https://doi.org/10.1007/
s10344-011-0517-y
Farnworth, B., Innes, J. and Waas, J.R., 2016.
Converting predation cues into conservation tools:
The eect of light on mouse foraging behaviour.
PLoS One, 11: e0145432. https://doi.org/10.1371/
journal.pone.0145432
Fischer, S., Oberhummer, E., Cunha-Saraiva, F., Gerber,
N. and Taborsky, B., 2017. Smell or vision? The
use of dierent sensory modalities in predator
discrimination. Behav. Ecol. Sociobiol., 71: 143.
https://doi.org/10.1007/s00265-017-2371-8
Garvey, P.M., Glen, A.S., Clout, M.N., Wyse, S.V.,
Nichols, M. and Pech, R.P., 2017. Exploiting
interspecic olfactory communication to monitor
predators. Ecol. Appl., 27: 389-402. https://doi.
org/10.1002/eap.1483
Garvey, P.M., Glen, A.S. and Pech, R.P., 2016. Dominant
predator odour triggers caution and eavesdropping
behaviour in a mammalian mesopredator.
Behav. Ecol. Sociobiol., 70: 481-492. https://doi.
org/10.1007/s00265-016-2063-9
Gonzálvez, F.G. and Rodríguez-Gironés, M.A., 2013.
Seeing is believing: Information content and
behavioural response to visual and chemical cues.
Proc. R. Soc. London B: Biol. Sci., 280: 886. https://
doi.org/10.1098/rspb.2013.0886
Goullaud, E.L., de Zwaan, D.R. and Martin, K., 2018.
Predation risk-induced adjustments in provisioning
behavior for horned lark (Eremophila alpestris) in
british columbia. Wilson J. Ornithol., 130: 180-
190. https://doi.org/10.1676/16-150.1
Groom, M.J., Mee, G.K. and Carroll, C.R., 2006.
Principles of conservation biology. Sinauer
Associates, Sunderland.
Guo, K., Liu, H., Bao, H., Hu, J., Wang, S., Zhang, W.,
Zhao, Y. and Jiang, G., 2017. Habitat selection
and their interspecic interactions for mammal
assemblage in the Greater Khingan Mountains,
northeastern China. Wildl. Biol., 2017: 00261.
https://doi.org/10.2981/wlb.00261
Haswell, P.M., Jones, K.A., Kusak, J. and Hayward,
M.W., 2018. Fear, foraging and olfaction: How
mesopredators avoid costly interactions with apex
predators. Oecologia, 187: 573–583. https://doi.
org/10.1007/s00442-018-4133-3
Hernández, L. and Laundré, J.W., 2005. Foraging in the
landscape of fear and its implications for habitat
use and diet quality of elk cervus elaphus and bison
bison bison. Wildl. Biol., 11 : 215-220. https://doi.
org/10.2981/0909-6396(2005)11[215:FITLOF]2.0
.CO;2
Hughes, N.K., Price, C.J. and Banks, P.B., 2010.
Predators are attracted to the olfactory signals of

2 7 9
Fear Eects of Predator Cues on Prey 279
prey. PLoS One, 5: e13114. https://doi.org/10.1371/
journal.pone.0013114
Kats, L.B. and Dill, L.M., 2016. The scent of death:
Chemosensory assessment of predation risk by
prey animals. Ecoscience, 5: 361-394. https://doi.
org/10.1080/11956860.1998.11682468
Kelley, J.L. and Magurran, A.E., 2003. Learned predator
recognition and antipredator responses in shes.
Fish Fish., 4: 216-226. https://doi.org/10.1046/
j.1467-2979.2003.00126.x
Kimball, B.A., Taylor, J., Perry, K.R. and Capelli, C.,
2009. Deer responses to repellent stimuli. J. chem.
Ecol., 35: 1461-1470. https://doi.org/10.1007/
s10886-009-9721-6
Kitchener, A.C., Breitenmoser-Würsten, C., Eizirik, E.,
Gentry, A., Werdelin, L., Wilting, A., Yamaguchi,
N. and Johnson, W.E., 2017. A revised taxonomy of
the felidae: The nal report of the cat classication
task force of the iucn cat specialist group. Cat
News, NZP Sta Publications.
Krivoruchko, K., 2011. Spatial statistical data analysis
for gis users. Esri Press, Redlands.
Kuijper, D.P., Verwijmeren, M., Churski, M., Zbyryt, A.,
Schmidt, K., Jędrzejewska, B. and Smit, C., 2014.
What cues do ungulates use to assess predation risk
in dense temperate forests? PLoS One, 9: e84607.
https://doi.org/10.1371/journal.pone.0084607
Laundré, J.W., Hernández, L. and Altendorf, K.B.,
2001. Wolves, elk, and bison: Reestablishing the
landscape of fear in Yellowstone National Park,
USA. Can. J. Zool., 79: 1401-1409. https://doi.
org/10.1139/z01-094
Lehner, P.N., 1992. Sampling methods in behavior
research. Poult. Sci., 71: 643-649. https://doi.
org/10.3382/ps.0710643
Li, C., Yang, X., Ding, Y., Zhang, L., Fang, H., Tang,
S. and Jiang, Z., 2011. Do pere david’s deer lose
memories of their ancestral predators? PLoS
One, 6: e23623. https://doi.org/10.1371/journal.
pone.0023623
Lima, S.L. and Bedneko, P.A., 1999. Temporal
variation in danger drives antipredator behavior: The
predation risk allocation hypothesis. Am. Natural.,
153: 649-659. https://doi.org/10.1086/303202
Margulis, S., 2016. Sampling animal behavior. Animal
Behavior Society Workshop, July 30, 2016.
Martin, P., Bateson, P.P.G. and Bateson, P., 1993.
Measuring behaviour: An introductory guide.
Cambridge University Press. https://doi.
org/10.1017/CBO9781139168342
Mathis, A. and Vincent, F., 2000. Dierential use of
visual and chemical cues in predator recognition
and threat-sensitive predator-avoidance responses
by larval newts (Notophthalmus viridescens). Can.
J. Zool., 78: 1646-1652. https://doi.org/10.1139/
cjz-78-9-1646
Natt, M., Lönnstedt, O.M. and McCormick, M.I., 2017.
Coral reef sh predator maintains olfactory acuity in
degraded coral habitats. PLoS One, 12: e0179300.
https://doi.org/10.1371/journal.pone.0179300
Nersesian, C.L., Banks, P.B. and McArthur, C., 2012.
Behavioural responses to indirect and direct
predator cues by a mammalian herbivore, the
common brushtail possum. Behav. Ecol. Sociobiol.,
66: 47-55. https://doi.org/10.1007/s00265-011-
1250-y
Noell, S., 2013. Eects of brown bear (Ursus arctos)
odour on the patch choice and behaviour of
dierent ungulate species. Dept. of Wildlife, Fish
and Environmental Studies, SLU, Umeå. Available
at: https://stud.epsilon.slu.se/6003/ (Accessed on
August 31, 2019).
Palacios, M., Warren, D.T. and McCormick, M.I., 2016.
Sensory cues of a top-predator indirectly control
a reef sh mesopredator. Oikos, 125: 201-209.
https://doi.org/10.1111/oik.02116
Peacor, S.D. and Werner, E.E., 2001. The contribution
of trait-mediated indirect eects to the net eects
of a predator. Proc. natl. Acad. Sci., 98: 3904-3908.
https://doi.org/10.1073/pnas.071061998
Prugh, L.R., Stoner, C.J., Epps, C.W., Bean, W.T.,
Ripple, W.J., Laliberte, A.S. and Brashares, J.S.,
2009. The rise of the mesopredator. Bioscience, 59:
779-791. https://doi.org/10.1525/bio.2009.59.9.9
Saxon-Mills, E.C., Moseby, K., Blumstein, D.T. and
Letnic, M., 2018. Prey naïveté and the anti-predator
responses of a vulnerable marsupial prey to known
and novel predators. Behav. Ecol. Sociobiol., 72:
151. https://doi.org/10.1007/s00265-018-2568-5
Schulte, L.M., Schulte, R. and Lötters, S., 2013.
Avoiding predation: The importance of chemical
and visual cues in poison frog reproductive
behaviour. In: Chemical signals in vertebrates.
Springer, pp. 309-321. https://doi.org/10.1007/978-
1-4614-5927-9_25
Sih, A., Bolnick, D.I., Luttbeg, B., Orrock, J.L., Peacor,
S.D., Pintor, L.M., Preisser, E., Rehage, J.S.
and Vonesh, J.R., 2010. Predator–prey naïveté,
antipredator behavior, and the ecology of predator
invasions. Oikos, 119: 610-621. https://doi.
org/10.1111/j.1600-0706.2009.18039.x
Sivy, K.J., Pozzanghera, C.B., Colson, K.E., Mumma,
M.A. and Prugh, L.R., 2018. Apex predators and
the facilitation of resource partitioning among

2 8 0
H. Mpemba et al.
mesopredators. Oikos, 127: 607-621. https://doi.
org/10.1111/oik.04647
Smith, M.E. and Belk, M.C., 2001. Risk assessment
in western mosquitosh (Gambusia anis): Do
multiple cues have additive eects? Behav. Ecol.
Sociobiol., 51: 101-107. https://doi.org/10.1007/
s002650100415
Sneddon, L.U., Braithwaite, V.A. and Gentle, M.J.,
2003. Novel object test: Examining nociception
and fear in the rainbow trout. J. Pain, 4: 431-440.
https://doi.org/10.1067/S1526-5900(03)00717-X
Suárez-Tangil, B.D. and Rodríguez, A., 2017. Detection
of iberian terrestrial mammals employing olfactory,
visual and auditory attractants. Eur. J. Wildl. Res.,
63: 93. https://doi.org/10.1007/s10344-017-1150-1
Sukhdeo, M.V., 2012. Where are the parasites in food
webs? Parasites Vectors, 5: 239. https://doi.
org/10.1186/1756-3305-5-239
Suraci, J.P., Roberts, D.J., Clinchy, M. and Zanette,
L.Y., 2017. Fearlessness towards extirpated large
carnivores may exacerbate the impacts of naïve
mesocarnivores. Behav. Ecol., 28: 439-447. https://
doi.org/10.1093/beheco/arw178
Switalski, T.A., 2003. Coyote foraging ecology and
vigilance in response to gray wolf reintroduction in
yellowstone national park. Can. J. Zool., 81: 985-
993. https://doi.org/10.1139/z03-080
Thaker, M., Vanak, A.T., Owen, C.R., Ogden, M.B.,
Niemann, S.M. and Slotow, R., 2011. Minimizing
predation risk in a landscape of multiple predators:
Eects on the spatial distribution of african
ungulates. Ecology, 92: 398-407. https://doi.
org/10.1890/10-0126.1
Tolon, V., Dray, S., Loison, A., Zeileis, A., Fischer, C.
and Baubet, E., 2009. Responding to spatial and
temporal variations in predation risk: Space use of a
game species in a changing landscape of fear. Can.
J. Zool., 87: 1129-1137. https://doi.org/10.1139/
Z09-101
Turvey, S.T., Crees, J.J., Li, Z., Bielby, J. and Yuan, J.,
2017. Long-term archives reveal shifting extinction
selectivity in china’s postglacial mammal fauna.
Proc. R. Soc. B, 284: 20171979. https://doi.
org/10.1098/rspb.2017.1979
Van Dyck, H., 2012. Changing organisms in rapidly
changing anthropogenic landscapes: The
signicance of the umwelt-concept and functional
habitat for animal conservation. Evolut. Applic.,
5: 144-153. https://doi.org/10.1111/j.1752-
4571.2011.00230.x
Venter, J.A., Prins, H.H., Mashanova, A. and Slotow, R.,
2017. Ungulates rely less on visual cues, but more
on adapting movement behaviour, when searching
for forage. PeerJ, 5: e3178. https://doi.org/10.7717/
peerj.3178
Wallach, A.D., Izhaki, I., Toms, J.D., Ripple, W.J. and
Shanas, U., 2015. What is an apex predator? Oikos,
124: 1453-1461. https://doi.org/10.1111/oik.01977
Wang, Q., Liu, D., Holyoak, M., Jia, T., Yang, S., Liu, X.,
Kong, X. and Jiang, G., 2018. Innate preference for
native prey and personality implications in captive
amur tigers. Appl. Anim. Behav. Sci., 210: 95-102.
https://doi.org/10.1016/j.applanim.2018.10.006
Wikenros, C., Jarnemo, A., Frisén, M., Kuijper, D.P.
and Schmidt, K., 2017. Mesopredator behavioral
response to olfactory signals of an apex predator.
J. Ethol., 35: 161-168. https://doi.org/10.1007/
s10164-016-0504-6
Wikenros, C., Kuijper, D.P., Behnke, R. and Schmidt, K.,
2015. Behavioural responses of ungulates to indirect
cues of an ambush predator. Behaviour, 152: 1019-
1040. https://doi.org/10.1163/1568539X-00003266
Wiles, G.J., Bart, J., Beck, Jr. R.E. and Aguon, C.F.,
2003. Impacts of the brown tree snake: Patterns
of decline and species persistence in guam’s
avifauna. Conserv. Biol., 17: 1350-1360. https://
doi.org/10.1046/j.1523-1739.2003.01526.x
Winnie, J.A., 2012. Predation risk, elk, and aspen: Tests
of a behaviorally mediated trophic cascade in the
greater yellowstone ecosystem. Ecology, 93: 2600-
2614. https://doi.org/10.1890/11-1990.1
Wong, B. and Candolin, U., 2015. Behavioral responses
to changing environments. Behav. Ecol., 26: 665-
673. https://doi.org/10.1093/beheco/aru183
Zhai-Penghui, L.Y., Zhengshan, L., 2015. The current
situation and evaluation of the biodiversity of
hanma national nature reserve. China Acad. J.
Electron. House, (Original article in Chinese
published 2015).
Zheng, W., Beauchamp, G., Jiang, X., Li, Z. and
Yang, Q., 2013. Determinants of vigilance in
a reintroduced population of pere david’s deer.
Curr. Zool., 59: 265-270. https://doi.org/10.1093/
czoolo/59.2.265