Carnivore Coexistence Facilitated by Spatial and Dietary Partitioning and Fine-Scale Behavioural Avoidance in A Semi-Arid Ecosystem

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 16

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/359336958

Carnivore coexistence facilitated by spatial and dietary partitioning and fine-


scale behavioural avoidance in a semi-arid ecosystem

Article  in  Journal of Zoology · March 2022


DOI: 10.1111/jzo.12964

CITATION READS

1 143

5 authors, including:

Lana Müller Willem Briers-Louw


The Cape Leopard Trust Wildlife Conservation Alliance
6 PUBLICATIONS   56 CITATIONS    11 PUBLICATIONS   30 CITATIONS   

SEE PROFILE SEE PROFILE

Alison J Leslie
Stellenbosch University
63 PUBLICATIONS   1,004 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

The Gadoli and Manda Khal Biodiversity Conservation Project View project

Recovery of the Zambezi Delta Ecosystem View project

All content following this page was uploaded by Alison J Leslie on 19 April 2022.

The user has requested enhancement of the downloaded file.


Journal of Zoology. Print ISSN 0952-8369

Carnivore coexistence facilitated by spatial and dietary


partitioning and fine-scale behavioural avoidance in a
semi-arid ecosystem
Lana Müller1 , Willem Daniel Briers-Louw2 , Rajan Amin3 , Christiaan Stefanus Lochner1
& Alison Jane Leslie4
1
The Cape Leopard Trust, Cape Town, South Africa
2
Zambeze Delta Conservation, Marromeu, Mozambique
3
Conservation Programmes, Zoological Society of London, Regent’s Park, London, UK
4
Department of Conservation Ecology and Entomology, Stellenbosch University, Stellenbosch, Western Cape, South Africa

Keywords Abstract
activity patterns; camera trapping; Caracal
caracal; coexistence; diet; Panthera pardus; scat Sympatric carnivores compete for similar resources which may lead to dominant
analysis; spatial overlap. species influencing the ecology of subordinate ones. However, carnivores often
make use of coevolutionary strategies which enable them to minimise competition
Correspondence with dominant competitors and thus facilitate coexistence. We used camera trapping
Lana Müller, The Cape Leopard Trust, P.O. and scat analysis to investigate the potential competition between leopards (Pan-
Box 31139, Tokai, Cape Town, 7966, South thera pardus) and caracals (Caracal caracal) along spatial, temporal and dietary
Africa. axes to determine the mechanisms of their coexistence in the Cederberg, South
Email: lmuller1985@gmail.com Africa. Our results showed that both carnivores co-occurred at 39.73% of camera
trapping sites, but spatial overlap based on Pianka’s index was relatively low. We
Editor: Femke Broekhuis found a high overlap in daily activity patterns between these carnivore species both
Associate Editor: Meredith Palmer during winter and summer. Leopards and caracals exhibited fine-scale behavioural
avoidance of one another, with time-to-encounters between interspecific competitors
being significantly larger compared to intraspecific competitors. The two carnivores
Received 28 April 2021; revised 10 February had a relatively high diet overlap, although this was largely due to one prey spe-
2022; accepted 18 February 2022 cies, namely rock hyrax (Procavia capensis), which appears to be an important
shared prey item. However, we did find evidence of dietary niche segregation, as
doi:10.1111/jzo.12964 leopards consumed larger prey compared to caracals. Our study suggests that carni-
vore coexistence in the Cederberg is facilitated by a combination of population-
level partitioning in space-use and dietary habits, as well as fine-scale behavioural
avoidance at the individual level to reduce interference competition. This study pro-
vides insights into the competition and coexistence mechanisms between sympatric
carnivores and broadens our understanding of these ecological processes in carni-
vore guild systems.

partitioning along spatial, temporal and/or dietary axes to limit


Introduction the effects of interspecific competition (Durant, 1998; Karanth
Competition plays an important role within carnivore assem- & Sunquist, 1995; Périquet et al., 2015). Identifying and
blages by influencing the ability for carnivores to access lim- understanding the mechanisms facilitating carnivore coexistence
ited resources (exploitation competition) and through direct and is pertinent as these co-occurring species may provide essential
indirect antagonistic interactions between carnivores (interfer- ecosystem (e.g. mesopredator suppression; Ritchie & Johnson,
ence competition), thus shaping carnivore ecological niches 2009) and economic services (e.g. tourism and public health
(Linnell & Strand, 2000). More specifically, carnivore guilds benefits; Braczkowski et al., 2018; Sonawane et al., 2021),
are structured by the influence of dominant species on subordi- making it crucial to conserve these carnivore assemblages
nate predators through suppression of their population abun- (Estes et al., 2011; Rafiq et al., 2020).
dance and density (Caro & Stoner, 2003; Ritchie & Johnson, Interference competition has been described as a strong
2009), alteration of their spatial distribution (du Preez et al., determinant of carnivore spatial distribution, temporal patterns
2015) and modification of their feeding ecology (Elmhagen and behaviour (Berger & Gese, 2007; Broekhuis et al., 2013;
et al., 2010; Périquet et al., 2015). However, sympatric carni- Hayward & Slotow, 2009). Mesopredators are naturally held in
vores may coexist through coevolutionary strategies of resource their food web position by more dominant apex carnivores and

Journal of Zoology  (2022) – ª 2022 The Zoological Society of London. 1


Leopard and caracal coexistence L. Müller et al.

are buffered against population collapse by their ability to considerably different, as leopards occur at low densities
switch prey species (Prugh et al., 2009). However, removal of (0.25–2.3 leopards/100 km2) and display large home ranges
the apex predator from a system may result in the irruption of (74–910 km2, Martins, 2010), while caracals occur at high den-
these middle-ranking predators, known as the ‘mesopredator sities (23–47 caracals/100 km2) and have small home ranges
release hypothesis’ (Crooks & Soulé, 1999), whereby released (7–65 km2, Avenant & Nel, 1998; Norton & Lawson, 1985).
mesopredators can attain higher densities and/or broader distri- This spatial heterogeneity within the landscape may reduce
butions which may result in declines or extinctions of prey direct costly encounters between these felids. In terms of their
populations (Loehle & Eschenbach, 2012; Prugh et al., 2009). temporal ecology, both carnivores generally display nocturnal
For example, the extirpation of wolves (Canis lupus) in the behaviour with bimodal activity patterns, that is, two activity
American West released coyotes (Canis latrans) which resulted peaks, one after sunset and one before sunrise (Avenant &
in the killing of threatened and endangered species and the Nel, 1998; Martins & Harris, 2013).
decline of small mammal populations (Ripple et al., 2013). Quantifying carnivore dietary ecology is a fundamental com-
These indirect effects can cause wide-ranging cascading effects ponent in evaluating predator-prey dynamics and relationships
which can alter the structure and function of a whole ecosys- with sympatric carnivores (Vogel et al., 2019). Due to their
tem (Estes et al., 2011; Terborgh et al., 2001). Therefore, to dietary plasticity, leopard and caracal can have catholic diets,
ensure ecosystem structure and function are maintained, it is although they generally consume the most abundant, suitable
important to preserve interactions between apex predators and prey in a given area (Avenant & Nel, 2002; Hayward et al.,
mesopredators (Ritchie & Johnson, 2009). However, recent lit- 2006) and for leopards these species fall within a preferred
erature questions the ubiquity of mesopredator release in multi- weight range of 10–40 kg (Hayward et al., 2006). However,
predator communities, suggesting that impacts of apex preda- unlike in the rest of their range, leopards in the Western Cape
tors on mesopredators are not homogenous and thus largely are exposed to a low density, small-bodied prey base and as a
context-dependent (Haswell et al., 2017; Jachowski et al., result, this leopard population has a limited prey selection,
2020). showing a preference for smaller prey species such as small-
The threat of direct antagonistic encounters can create a to medium-sized ungulates and rock hyrax (Procavia capensis,
‘landscape of fear’ (Laundré et al., 2001), whereby subordinate Fröhlich et al., 2012; Martins et al., 2011; Norton et al.,
carnivores are limited to smaller or less suitable areas within a 1986). A recent comparative study from Anysberg Nature
landscape due to the presence of dominant carnivores (Swan- Reserve, a state-protected area in the Western Cape province
son et al., 2014). However, heterogeneity across the landscape of South Africa, showed that leopards consumed mostly wild
reduces the probability of encounters and therefore risk (Brown ungulates while caracals fed mostly on micromammals, with a
et al., 1999). This presents subordinate carnivores with oppor- dietary niche overlap of 53% (Drouilly, Nattrass et al., 2018).
tunities to evade dominant predators by utilising low-risk Although carnivore diets often reflect prey availability (Grif-
refuges (Durant, 1998), diurnal–nocturnal adaptations and vari- fiths, 1975; Karanth & Sunquist, 1995), prey selection may
ations in activity peaks (Hayward & Slotow, 2009) or modify- also depend on factors such as interspecific competition, with
ing their behaviour through habitat transitioning (du Preez larger carnivores potentially outcompeting smaller carnivores
et al., 2015) or vigilance in response to the degree of risk (Carbone et al., 1997; McVittie, 1979).
(Pangle & Holekamp, 2010; Switalski, 2003). Furthermore, Leopards in the Western Cape historically occurred sympatri-
where the purely spatial or temporal overlap is high, carnivores cally with a large carnivore assemblage including lions (Pan-
are also known to utilise fine-scale behavioural adaptations to thera leo) and spotted hyenas (Crocuta crocuta), although the
reduce the degree of encountering competitors which may latter competitors were extirpated around 300 years ago (Shor-
facilitate co-occurrence (Karanth et al., 2017). Such adaptations tridge, 1934; Skead, 1980). These leopards are genetically sim-
may involve fine-scale spatial–temporal avoidance (Harmsen ilar to the ‘bushveld’ leopard from the savanna region in
et al., 2009; Karanth et al., 2017; Lahkar et al., 2020; Périquet southern Africa, but their body size and mass differ signifi-
et al., 2021), reactive adjustment of subordinate carnivore cantly, with the former weighing between one-third to half that
behaviour to the nearest dominant carnivore (Broekhuis et al., of the latter, for example, mean mass of females: 21 kg
2013) or active tracking of dominant carnivores to reduce the (Cederberg) vs. 38 kg (Kruger National Park); and males: 35
risk of agnostic encounters for subordinate carnivores (López- vs. 58 kg (Bailey, 1993; Martins, 2010). Most carnivore guild
Bao et al., 2016). studies involving leopards have focused on competition with
Leopards (Panthera pardus) are large carnivores that are larger carnivores such as lions and spotted hyenas (e.g. du
considered apex predators, while caracals (Caracal caracal) Preez et al., 2015; Balme et al., 2017; Miller et al., 2018),
are medium-sized carnivores that are considered mesopredators. despite leopards co-occurring and competing with smaller guild
Both are highly adaptable, solitary felids that co-occur across members across their African range (Rafiq et al., 2020). As
much of sub-Saharan Africa (Caro & Stoner, 2003). Through- such, there is limited knowledge on how leopards interact and
out their range, these carnivores occupy different ecological coexist with smaller carnivores.
niches and are capable of variable spatial and temporal Caracals are a smaller carnivore species with females weigh-
responses to the presence of interspecific competitors (Avenant ing up to 12 kg and males up to 15 kg in the Western Cape
et al., 2016; Drouilly, Clark et al., 2018; Edwards et al., 2015; (Skinner & Chimimba, 2005). Carbone et al. (1999) showed
Rafiq et al., 2020). In the Western Cape province of South that carnivores weighing 21.5 kg or more, select vertebrate
Africa, their spatial distributions and population densities are prey that weighs >45% of their own body mass, whereas

2 Journal of Zoology  (2022) – ª 2022 The Zoological Society of London.


L. Müller et al. Leopard and caracal coexistence

carnivores smaller than 21.5 kg select prey that weighs <45% was expected in the fynbos biome due to greater landscape
of their own body mass, with the latter selecting both verte- heterogeneity making it easier to coexist. Since the need for
brate and invertebrate prey. Thus, based on this energetic fine-scale interactions relies on competitors being relatively
model, leopards should occupy a distinct functional group close together, we expect fine-scale avoidance to be higher in
above caracals and maintain a competitive advantage over the summer when resources are limited and in the karoo biome
mesopredator, despite their smaller than usual body size and which has less heterogeneity. To test these predictions, we used
mass. Furthermore, leopards are known to affect the detection camera trapping for spatial, temporal and fine-scale interaction
probability of subordinate mesopredators (Ramesh et al., 2017) analyses and scat sampling for dietary analyses.
and interspecific killing of caracals by leopards has been docu-
mented (Martins, 2010). It is therefore likely that direct compe-
tition between these two felids may lead to caracals avoiding
Materials and methods
areas where leopards are present (Avenant et al., 2016).
Study area
The Cederberg Mountains of South Africa provides a unique
opportunity to investigate the mechanisms of co-occurrence The Cederberg (32°27’S; 19°25’E) is located 200 km north of
between these two carnivores where (1) other large, dominant Cape Town and extends from Pakhuis Pass to Grootrivier in
predators are absent, (2) leopards are physically smaller com- the Western Cape Province, South Africa, (Martins & Harris,
pared to the ‘bushveld‘ leopard and their body mass extends 2013). The study area covers approximately 1700 km2 of
either side of the 21.5 kg threshold of obligate carnivory (Car- rugged mountainous terrain, which includes CapeNature
bone et al., 1999) thus potentially reducing the ecological Reserves (Cederberg Wilderness Area and Matjiesrivier Nature
niche gap with caracals and (3) prey species are both small- Reserve), Bushmans Kloof Private Reserve, Cederberg Conser-
bodied and limited, which might result in the increased dietary vancy (an array of privately-owned farms where landowners
overlap. agreed not to kill leopards or caracals) and farmland (Fig. 1).
The aim of this study was to determine how leopards and Livestock farming was the dominant land use within the
caracals coexist within the Cederberg by investigating their Cederberg until the 1990s, but other agricultural practices such
spatial distributions, temporal patterns and feeding ecologies. as citrus, olive and rooibos tea farming have become more
We hypothesised that (1) caracals spatially avoid leopards, (2) common. Private reserves within the Cederberg area have rein-
caracals avoid peak activity of leopards where they overlap troduced previously extirpated herbivores such as gemsbok
spatially, (3) fine-scale behavioural avoidance of interspecifics (Oryx gazella), red hartebeest (Alcelaphus buselaphus caama),
will exceed that of conspecifics and (4) carnivores display diet- springbok (Antidorcas marsupialis) and Cape mountain zebra
ary segregation. We also investigated the effect of season and (Equus zebra zebra, Martins & Harris, 2013). The carnivore
biomes on the ecology of leopards and caracals. We expected community in the Cederberg consists of a variety of species
less spatial and temporal avoidance during the summer when such as bat-eared fox (Otocyon megalotis), Cape fox (Vulpes
food may be limited thus increasing the likelihood of inter- chama), small-spotted genet (Genetta genetta), large-spotted
species competition. Greater spatial and temporal avoidance genet (Genetta tigrina), striped polecat (Ictonyx striatus),

(a) (b)

Figure 1 Map of the camera trap sites (n = 73) and the presence of leopards, caracals or both carnivores at each site (a) and scat locations for
leopards and caracals (b) in the Cederberg study site. Inset: box indicates the study site (black square) in the Western Cape Province (grey
shade) of South Africa.

Journal of Zoology  (2022) – ª 2022 The Zoological Society of London. 3


Leopard and caracal coexistence L. Müller et al.

black-backed jackal (Canis mesomelas), aardwolf (Proteles 15–20 hairs from each scat sample (Wang & Macdonald,
cristata), African wildcat (Felis silvestris), caracal and leopard. 2009) to create hair cross-sections following the methods
The climate in the Cederberg is classified as semi-arid and described by Douglas (1989). Cross-sections were analysed
the area has two distinct seasons: a cool-wet winter between under a Leica DM 2000 light microscope at 20–40× magnifi-
April and September and hot-dry summer between October cation, and photographs and measurements of cross-sections
and March (Martins & Harris, 2013). Temperatures range from were made using the LAS Core V4.0 software (Leica
a minimum of –7°C in winter to a maximum of 47°C in sum- Microsystems Ltd, Wetzlar, Germany). Prey items were identi-
mer (Martins & Harris, 2013). There are two main biomes: fied by comparing hair cross-sections to reference collections
fynbos, consisting of both fynbos (restioids, ericoids and pro- from Rhodes University and published literature (Keogh, 1979,
teoids) and renosterveld vegetation (Renosterbos spp.), and the 1983, 1985). Cross-section analysis was aided by the macro-
succulent karoo comprising of small shrubs and succulents scopic appearance of hairs as well as the presence of other
(Martins et al., 2011). Mean annual rainfall is 669 mm in the remains such as bones, footpads, hooves, teeth and claws (Nor-
fynbos and 179 mm in the karoo (Martins & Harris, 2013). ton et al., 1986). All hairs were sorted to species-level, except
The rugged and mountainous topography is divided by valleys for rabbits and hares (Lepus capensis, Lepus saxatilis and
and ravines that are either densely wooded or open, with fyn- Pronolagus saundersiae) and species from the genus Otomys
bos having larger mountains and karoo having larger, flat open (Otomys irroratus and Otomys saundersiae) which were
areas. Altitudes range from 200 to 2026 m in the fynbos and grouped together due to similarities in hairs of individual spe-
258 to 1446 m in the karoo (Martins & Harris, 2013). cies (Mann et al., 2019). Isolated leopard and caracal hairs
within their own scats were assumed to be as a result of self-
grooming (Norton et al., 1986). Remains of birds, reptiles and
Camera trapping
invertebrates were identified but not classified into lower taxo-
Leopard and caracal presence data were collected from 73 cam- nomic groups.
era trap stations distributed across the study area continuously
from October 2017 to September 2018 (Fig. 1). The camera trap Data analysis
survey area was divided into grid cells of 50 km2 based on the
minimum home range size of female leopards in the Cederberg Spatial
(Martins, 2010). Each grid cell contained two camera trap sta-
tions with a mean intercamera station distance of 2.78 km We had insufficient data to conduct two-species occupancy mod-
(range = 0.5–6 km) to maximise the chance of capturing indi- elling as there were too few sites where only caracal was
viduals within the grid cells (Balme et al., 2009). Camera sites detected. So to investigate the site-specific association between
were placed on trails, jeep tracks or along natural features such leopards and caracals, we first calculated the relative abundance
as drainage lines where focal carnivores were likely to move index (RAI, O’Brien et al., 2003) at the camera trap station by
(Mann et al., 2015). Each camera trap station consisted of two dividing the independent number of captures by the total trap
white-flash camera traps (Cuddeback X-Change Color Model nights at any given station (Ramesh et al., 2012). A capture was
1279) which were placed about 2–3 m from the path. Cameras considered independent when there was at least a 30 min interval
were placed perpendicular to the path at ~40 cm above the between the individuals of the same species at the same site
ground to obtain a full lateral body image of a passing animal. (O’Brien et al., 2003). A Spearman’s rank correlation test (Zar,
All camera traps were set to operate for 24 h per day and to take 1999) was then conducted between RAI of leopard and caracal at
three photo-bursts every time the sensor was triggered, with a each station to assess the interpredator spatial association
delay of 1-s between successive triggers. Camera trap images (Ramesh et al., 2012). Each camera trap station was considered
were processed in Camera Base version 1.3 (Tobler, 2007) and spatially independent (Zhao et al., 2019) and we used the RAI for
manually identified to species level. each camera station to compute the spatial overlap index which is
used extensively to assess niche overlap between
1=2 species (Pianka,
1973): Oab ¼ ð∑nPia Pib Þ= ∑nP2ia ∑nP2ib , where Oab is spa-
Scat collection
tial overlap between species a and b; Pia and Pib are the propor-
Scat samples were collected opportunistically along footpaths tions of item i used by species a and b, respectively. Pianka’s
and trails between May 2017 and February 2018. Due to most overlap index values range from 0 (no spatial overlap) to 1 (com-
felid species also defecating for territorial markings, a small plete spatial overlap). The analysis was done in R using the pack-
amount of scat was left behind (Martins et al., 2011). Scats age spaa version 0.2.2 (Zhang, 2016). We used this index merely
were identified as felid scat in the field following Walker as a measure of niche overlap as opposed to automatically imply-
(1996). These scats were then differentiated between leopard ing competition from these values (Tsafack et al., 2021).
and caracal based on the size, shape and appearance of each
scat following methods described by Drouilly, Nattrass et al.
Temporal activity patterns
(2018), and later confirmed in the lab by the presence of hairs
from self-grooming (Mann et al., 2019). Carnivore scats of To estimate the temporal overlap between leopards and cara-
doubtful identity were excluded from the analysis. cals, we determined the diel activity patterns of the species’
All scat samples were washed in water and air-dried at 30– using the time of detection on the camera trap photographs.
40°C in preparation for hair analysis. We randomly sampled The activity time data were transformed to daylight, that is,

4 Journal of Zoology  (2022) – ª 2022 The Zoological Society of London.


L. Müller et al. Leopard and caracal coexistence

sunrise and sunset using the dates and spatial locations of the will each have a value of 0.5 (Klare et al., 2011). Both FO
observations (Nouveliet et al., 2012). The analysis was per- and CFO account for rare prey species, and while CFO is the
formed in the overlap package (version 0.3.2) in the R soft- preferred qualitative method for dietary analysis, FO was
ware (Ridout & Linkie, 2009). The Δ4 overlap coefficient, included here for comparative purposes (Klare et al., 2011).
which is recommended for sample sizes >75, was calculated Dietary niche breadth, as a measure of diet specialisation for
for the summer and winter seasons and the two biomes leopards and caracals, was calculated using Levin’s index
(Meredith & Ridout, 2014; Ridout & Linkie, 2009). Values (Levins, 1968): B ¼ 1=∑P2i , where Pi represents the frequency
from the Δ4 overlap coefficient range from 0 (no overlap) to 1 of occurrence of the prey item in the scat. To allow for com-
(complete overlap). The estimated 95% confidence intervals parison, we standardised dietary niche breadth as a proportion
were obtained from 10 000 bootstrap samples. We followed using the equation: BA ¼ B  1=n  1 (Levins, 1968), where
Monterroso et al. (2014) and defined low overlap when Δ4 BA represents Levin’s standardised niche breadth and n the
was <0.5, moderate when Δ4 was between 0.5 and 0.75 and number of prey species in carnivore diets. The values range
high overlap when Δ4 was >0.75. from 0 (specialist diet) to 1 (generalist diet). To determine diet-
ary overlap between leopards and caracals we used Pianka’s
overlap index (Pianka, 1973).
Fine-scale behavioural interactions
Biomass calculation models provide the most accurate esti-
To assess fine-scale behavioural adaptations, we calculated mate of true carnivore diet and are the most relevant parame-
time-to-encounters, in decimal days, between consecutive ters in dietary analysis because of their application in ecology
‘caracal-caracal (CC)’, ‘leopard-leopard (LL)’, ‘caracal-leopard such as the impact of predation on prey populations (Klare
(CL)’ and ‘leopard-caracal (LC)’ capture events. We tested for et al., 2011). To calculate biomass consumed, we used an
interspecific time-to-encounter exceeding intraspecific time-to- asymptotic regression model developed by Chakrabarti et al.
encounter, as we expected caracals (subordinate carnivore) to (2016) for obligate carnivores: Y = (0.033–0.024exp–4.284X/Z),
display fine-scale temporal avoidance of leopards (dominant where Y is the mass of prey consumed per collectable scat, X
carnivore). We fitted a linear mixed-effects model with cross is prey body mass and Z is predator body mass. A mean body
factors ‘capture 1’ as the initial capture either caracal or leop- mass of 35 kg was used for leopards based on Martins (2010),
ard, and ‘capture 2’ the subsequent capture as either caracal or which was chosen for comparative purposes with other leopard
leopard in R statistical package lme4 (version 1.1-26, Bates populations in the Western Cape (e.g. Mann et al., 2019). For
et al., 2015). Camera location was added as a random-effect caracal and prey species, we obtained mean adult body mass
(Harmsen et al., 2009). The model ‘time-to-encounter’ from Skinner and Chimimba (2005), while for feral pigs (Sus
response variable was log10-transformed to approximate scrofa), we based this on recently published literature (Mann
normal distribution of the residuals and equal variances. We et al., 2019). For larger ungulate species (gemsbok), we multi-
calculated the differences between model predicted time-to- plied adult mass by 0.3 (Radloff & du Toit, 2004) due to the
encounters for (1) LC and CC events and (2) CL and LL assumption that leopards would typically prey on individuals
events, and the 95% confidence intervals of the differences closer to their optimal prey size (Hayward et al., 2006). Mam-
using a bootstrap. A significant difference was inferred by its malian prey were grouped into large (40–63 kg), medium (10–
95% confidence interval excluding zero. 40 kg), small (1–10 kg), very small (<1 kg) species and live-
stock categories, to compare the importance of each class to
carnivore diets. Non-mammalian prey items were excluded
Dietary
from biomass calculations due to difficulties in species identifi-
We followed Glen and Dickman (2006) in calculating the ade- cation. We used chi-square analysis to compare the frequency
quacy of the dietary sampling effort using a Brillouin index: of occurrence of mammalian prey size categories between car-
H b ¼ ðlnN !  ∑lnni !Þ=N, where Hb is the diversity of prey, N nivore diets. All statistical analyses were conducted in R ver-
is the total number of prey items in all samples and ni is the sion 3.6.3 (R Development Core Team, 2019).
number of individual prey items in the ith category (Brillouin,
1956). An Hb curve was generated by bootstrapping the sam-
ple 10 000 times with replacement to obtain a mean Hb, vary-
Results
ing the sample size from three to n, in increments of two.
Spatial
Additionally, we calculated a curve using the percentage incre-
mental change of Hb. Scat sampling effort for each carnivore We recorded 741 independent leopard captures at 69 (95.00%)
was considered adequate if the diversity curve reached an sites and 153 independent caracal captures at 33 (45.00%)
asymptote and the incremental change declined below 1% sites. Leopards were found exclusively at 54.79% of sites and
(Hass, 2009). caracals were found exclusively at only 5.48% of sites (Fig. 1).
Frequency of occurrence (FO) was calculated as the percent- Sites where both predators occurred totalled 39.73% and leop-
age of scats containing a particular prey item (Klare et al., ards used these sites more often than caracals with a ratio of
2011). A corrected frequency of occurrence (CFO), which 3:1 based on trap rate. Predator co-occurrence at the same site
accounts for multiple prey items per scat, was calculated by was more common in winter (30.00%) than in summer
designating a weighting value of 1 per scat, and each prey (18.57%), and the percentage of carnivore-exclusive sites was
item as a proportion of that scat, for example, two prey items marginally higher in summer (81.43%) than in winter

Journal of Zoology  (2022) – ª 2022 The Zoological Society of London. 5


Leopard and caracal coexistence L. Müller et al.

(71.43%). While all sites used by caracals in the karoo were midnight (Fig. 2). In winter, leopards displayed a bimodal
also used by leopards, 8.89% of the sites in fynbos were activity pattern with peaks in activity at sunset and between
exclusively used by caracal. Despite caracals co-occurring with midnight and sunrise (Fig. 2). Caracals showed a trimodal
leopards at many camera trap sites, the spatial overlap based activity pattern with two main peaks close to midnight and
on Pianka’s index (which accounts for the proportion of use) sunrise and an additional slight peak around noon. Temporal
between leopard and caracal was relatively low (0.11) and overlap between leopard and caracal was high in summer
caracal space use was negatively associated with leopard space (0.77; 95% CI: 0.68–0.85) and winter (0.77; 95% CI: 0.68–
use (rs = –0.37, P < 0.01). 0.86). Similarly, temporal overlap was high between fynbos
(0.79; 95% CI: 0.71–0.87) and karoo (0.79; 95% CI: 0.70–
0.88; Fig. 2).
Temporal activity patterns
Overall, leopard and caracal were both crepuscular and noctur-
Fine-scale behavioural interactions
nal. In summer, leopards displayed a bimodal pattern of activ-
ity with peaks around sunset and sunrise, whereas caracals The number of intraspecific and interspecific photo-capture
displayed a unimodal pattern of activity with a peak around events were 139 (CC), 47 (CL), 46 (LC) and 724 (LL). Our

Figure 2 Overlap (grey area) in daily activity patterns between leopard and caracal between summer and winter, as well as between fynbos and
karoo biomes.

6 Journal of Zoology  (2022) – ª 2022 The Zoological Society of London.


L. Müller et al. Leopard and caracal coexistence

model revealed the time-to-encounters (in log10 scale) for leopards (83.09%) and caracals (70.83%). Our sampling
between LC captures (1.25) and CL capture (1.38) were sig- effort was adequate for both species, with diversity curves
nificantly longer than those between consecutive same-species reaching an asymptote and the incremental change decreasing
captures, CC (0.85) and LL (1.11). The predicted difference to <1% at ≥43 and ≥37 samples, respectively (Fig. 3).
between interspecific and intraspecific time-to-encounter events Based on FO, the most frequently consumed prey items for
were (1) LC and CC (0.40, 95% CI: 0.18–0.63) and (2) CL leopards were rock hyrax (49.04%), klipspringer (Oreotragus
and LL (0.27, 95% CI: 0.08–0.47). As expected, the LC–CC oreotragus, 29.64%) and common duiker (Sylvicapra grimmia,
difference was larger than CL–LL. Time-to-encounters varied 6.82%, Table 1). For caracals, rock hyrax (39.58%), Namaqua
by season and biome, with greater LC and CC differences rock mouse (Micaelamys namaquensis, 23.96%) and Karoo or
compared to CL and LL in summer and in the karoo biome bush rats (Otomys spp., 22.92%) were most frequently con-
(Appendix S1). sumed. Mean prey weight was higher for leopards
(12.38  3.57 [SE] kg) compared to caracals (4.46  3.86
[SE] kg), based on the five top contributing species. Based on
Dietary
biomass consumed, rock hyrax and klipspringer remained the
A total of 547 prey items were recorded in leopard scats two most important species for leopards, although grey rhebok
(n = 469), representing 27 different prey species, while 133 (Pelea capreolus) contributed the next highest proportion to
prey items were recorded in caracal scats (n = 96), representing total biomass consumed (7.72%). For caracals, rock hyrax
20 different prey species (Appendix S2). The mean number of (50.27%), rabbits and hares (Lepus spp. and Pronolagus saun-
prey items per scat was 1.17 for leopards and 1.39 for cara- dersiae, 10.23%) and Namaqua rock mouse (8.27%) repre-
cals, while the majority of scats contained single prey items sented the highest contributions to overall biomass consumed.

Figure 3 Brillouin diversity curves (black triangles) and percentage (incremental) change curves (grey dots) for leopard and caracal scat samples.
Mean values were obtained by resampling with replacement 10 000 times.

Journal of Zoology  (2022) – ª 2022 The Zoological Society of London. 7


Leopard and caracal coexistence L. Müller et al.

Table 1 Frequency of occurrence per scat (FO, %), corrected frequency of occurrence (CFO, %), mean prey body mass (kg), corrected biomass
consumed (kg) and relative biomass consumed (%) of prey species in leopard and caracal scat in the Cederberg, South Africa

Mean Corrected Relative


CFO mass biomass biomass
FO (%) (%) (kg) (kg) (%)
Prey species Leopard Caracal Leopard Caracal Leopard Caracal Leopard Caracal
Aardwolf Proteles cristatus 0.21 0.00 0.21 0.00 9.50 0.88 0.00 0.24 0.00
African wild cat Felis silvestris 0.43 0.00 0.43 0.00 4.30 1.28 0.00 0.34 0.00
Cape fox Vulpes chama 0.21 1.04 0.21 0.52 2.65 0.52 0.15 0.14 0.67
Cape grey mongoose Galerella pulverulenta 0.00 1.04 0.00 0.35 0.80 0.00 0.06 0.00 0.27
Cape grysbok Raphicerus melanotis 4.05 1.04 3.59 1.04 10.25 15.24 0.41 4.11 1.91
Caracal Caracal caracal 0.21 0.00 0.21 0.00 8.70 0.85 0.00 0.23 0.00
Chacma baboon Papio ursinus ursinus 2.13 1.04 2.03 1.04 23.60 10.51 0.43 2.84 1.96
Common duiker Sylvicapra grimmia 6.82 4.17 5.86 3.65 16.10 28.41 1.48 7.67 6.83
Common mole-rat Cryptomys hottentotus 0.00 2.08 0.00 1.56 0.00 0.00 0.17 0.00 0.80
Domestic pig Sus scrofa 0.21 0.00 0.21 0.00 130.00 1.15 0.00 0.31 0.00
Elephant shrew Elephantulus edwardii 0.00 2.08 0.00 1.30 0.00 0.00 0.14 0.00 0.64
Forest shrew Myosorex varius 0.00 1.04 0.00 0.52 0.00 0.00 0.05 0.00 0.24
Gemsbok Oryx gazella 0.85 0.00 0.75 0.00 63.00a 4.04 0.00 1.09 0.00
Goat Capra hircus 1.49 3.13 1.39 3.13 50.00 7.49 1.28 2.02 5.87
Grey climbing mouse Dendromus melanotis 0.00 1.04 0.00 0.35 0.00 0.00 0.02 0.00 0.16
Grey rhebuck Pelea capensis 6.18 0.00 5.65 0.00 20.00 28.60 0.00 7.72 0.00
Hairy-footed gerbil Gerbillurus paeba 0.43 2.08 0.32 0.69 0.03 0.42 0.07 0.11 0.32
Klipspringer Oreotragus oreotragus 29.64 1.04 23.53 0.52 11.90 104.94 0.21 28.32 0.96
Hares & rabbits Lepus/Pronolagus spp. 2.77 9.38 2.56 7.81 2.35 5.98 1.67 1.61 9.56
Marsh mongoose Atilax paludinosus 0.21 0.00 0.11 0.00 2.80 0.27 0.00 0.07 0.00
Namaqua rock mouse Micaelamys 0.64 23.96 0.64 18.40 0.05 0.86 1.91 0.23 8.79
namaquensis
Karoo/Bush rat Otomys spp. 1.07 22.92 0.96 14.84 0.13 1.32 1.67 0.36 7.66
Porcupine Hystrix africaeaustralis 2.99 0.00 2.20 0.00 12.15 9.89 0.00 2.67 0.00
Rock hyrax Procavia capensis 49.04 39.58 43.00 33.33 3.66 120.18 10.56 32.43 48.57
Sheep Ovis aries 2.77 2.08 2.56 2.08 40.00 13.78 0.85 3.72 3.92
Small-spotted genet Genetta genetta 0.21 0.00 0.21 0.00 1.90 0.46 0.00 0.12 0.00
Springbok Antidorcas marsupialis 1.71 0.00 1.71 0.00 31.50 9.09 0.00 2.45 0.00
Steenbok Raphicerus campestris 0.85 0.00 0.85 0.00 11.10 3.72 0.00 1.00 0.00
Striped field mouse Rhabdomys pumilio 0.21 0.00 0.11 0.00 0.04 0.14 0.00 0.04 0.00
Striped polecat Ictonyx striatus 0.21 0.00 0.21 0.00 0.68 0.35 0.00 0.09 0.00
Yellow mongoose Cynictic pencillata 0.21 1.04 0.11 1.04 0.83 0.18 0.18 0.05 0.83
Bird 0.85 7.29 0.39 3.30 – – – – –
Invertebrates 0.00 7.29 0.00 2.86 – – – – –
Squamates 0.00 4.17 0.00 1.65 – – – – –

Boldface indicates the most frequently consumed (>5%) prey items for both carnivores.
a
Is the juvenile mass for larger ungulate species.

Prey size categories differed significantly between the two species, mostly owing to the substantial contribution of rock
species (χ2 = 86.36, d.f. = 3, P < 0.05). Leopards consumed hyrax in both carnivore diets.
larger prey, with medium-sized mammals (10–40 kg) making
up 56.78% of their diet, while caracals consumed relatively
smaller prey with small mammals (1–10 kg) making up
Discussion
58.83% of their diet (Fig. 4). Livestock contributed to both Effectively conserving carnivores in their natural environment
leopard and caracal diets, but the total biomass consumed was requires an understanding of how these carnivores coexist
<10% for both carnivores. without competitively excluding one another. Sympatric carni-
Niche dietary breadth was 3.97 for leopard and 6.59 for vores in this study did not segregate along the temporal axis.
caracal and standardisation revealed that dietary breadth was However, we found these carnivores displayed some degree of
low for leopard (0.11) and caracal (0.29), representing spe- spatial and dietary niche partitioning as well as fine-scale beha-
cialised diets for both carnivores. A relatively high degree of vioural avoidance which likely facilitates carnivore coexistence
dietary overlap (0.66) was recorded between the two carnivore in this semi-arid landscape.

8 Journal of Zoology  (2022) – ª 2022 The Zoological Society of London.


L. Müller et al. Leopard and caracal coexistence

Figure 4 Biomass consumed (%) for each of the prey weight classes found in the leopard and caracal diet in the Cederberg.

karoo which is likely due to the landscape heterogeneity which


Spatial
provides more refugia (e.g. vegetation cover) for caracals, to
As expected, our study revealed some level of spatial segrega- which they may be positively associated (Verschueren et al.,
tion between leopards and caracals. In the Cederberg, the like- 2021). This emphasises the importance of conserving natural
lihood of encountering a leopard may be relatively low for a habitat heterogeneity in human-dominated landscapes and sup-
caracal as leopards maintain large home ranges and occur at porting land-use types that are conducive to carnivore conser-
low densities (Martins, 2010), in contrast to the comparatively vation.
higher densities and smaller ranges of caracals in the region
(Avenant & Nel, 1998). Resource availability is a strong pre-
Temporal activity patterns
dictor of carnivore spatial distribution (Rich et al., 2017; Ver-
schueren et al., 2021) and spatial overlap between is usually Carnivores exhibited a high degree of temporal overlap in our
higher when prey densities are lower to maximise resource study with no differences between seasons or biomes, which is
acquisition (Karanth et al., 2017). Thus, given the spatial seg- likely due to relatively low carnivore densities minimising the
regation found here, it is likely that prey densities are suffi- likelihood of competitive or agnostic interactions (Hayward &
cient to support current carnivore densities and reduce Hayward, 2007). Romero-Muñoz et al. (2010) also suggested
competition between carnivores. However, contrary to our pre- that low carnivore densities may reduce the probability of
diction, spatial avoidance was highest in summer. One hypoth- competitor encounters, and resource partitioning may not even
esis is that caracals can spatially avoid leopards more in be required in such cases. As a result, subordinate species do
summer due to density peaks in rodents (Avenant, 1993), not necessarily need population-level modifications of their
which contribute substantially to their diet, resulting in less activity patterns to facilitate carnivore coexistence (Lahkar
direct competition with leopards. However, we acknowledge et al., 2020). However, subtle variations in activity peaks did
that spatial analyses were not based on robust two-species exist in our study. Caracal activity peaked around midnight,
occupancy analyses. while leopard activity peaked around dawn and dusk, these dif-
Spatial heterogeneity in the landscape may also play a key ferences were more pronounced during summer. Although it is
role in facilitating carnivore coexistence through the availabil- clear that carnivore temporal overlap is high, given that tempo-
ity of different habitats and other physical features, as well as ral activity patterns have evolved to satisfy a variety of selec-
uneven resource distribution (Chesson, 2000; Stein et al., tion forces that differ amongst species to minimise interactions
2014) which can provide vital refuges for subordinate species with competitors, while increasing chances of hunting success
to reduce encounters with dominant carnivores (Davies et al., (Hayward & Slotow, 2009; Karanth et al., 2017), it is plausible
2021; Durant, 1998). Our results indicated that spatial avoid- that caracals could alter their peak activity to optimise this
ance was slightly higher in the fynbos biome compared to the trade-off.

Journal of Zoology  (2022) – ª 2022 The Zoological Society of London. 9


Leopard and caracal coexistence L. Müller et al.

Fine-scale behavioural interactions constituting mostly mammals, but the importance of prey
items, like rodents and antelope, varies widely between studies
Resource partitioning is not always distributed along clear
(Avenant et al., 2016). Thus, these findings highlight the
space and time axes, but could also be ‘hidden’ in finer-scale
importance of area-specific studies to improve the understand-
behavioural interactions between carnivores (Karanth et al.,
ing of carnivore feeding behaviour at a finer scale (Briers-
2017; Lahkar et al., 2020). Karanth et al. (2017) found that
Louw & Leslie, 2020).
fine-scale spatio–temporal interactions facilitated coexistence
Despite the smaller gap between carnivore body sizes in the
among tigers (Panthera tigris), leopards and dholes (Cuon
Western Cape, leopards still consumed significantly larger-sized
alpinus) in India. Similarly, Swanson et al. (2016) found that
prey groups compared to caracals, with leopards feeding pre-
cheetahs avoided dominant lions on a moment-to-moment basis
dominantly on medium-sized prey (10–40 kg) and caracals on
and suggested that fine-scale avoidance strategies may prove
small-sized prey (1–10 kg), which coincides with other studies
less costly for subordinate predators compared with long-term
(Avenant & Nel, 1998; Drouilly, Nattrass et al., 2018; Hay-
avoidance of suitable habitats. Furthermore, Broekhuis et al.
ward et al., 2006). Although carnivores displayed selective pre-
(2013) stated that carnivore responses to predation risk may be
dation of different sized prey groups which may facilitate
reactive rather than predictive, with subordinate predators
carnivore coexistence (Karanth & Sunquist, 1995), we found a
adjusting their behaviour to the nearest position of dominant
high dietary overlap between the two sympatric carnivores,
predators. Our study found that LC and CC differences were
which was attributed to the contribution of rock hyrax. The
greater than CL and LL differences, consistent with our
continued utilisation of rock hyrax by leopards in the Western
hypothesis and this literature. The fine-scale avoidance found
Cape is certainly interesting and its importance has been docu-
in our study (i.e. LC and CC > CL and LL) was higher in
mented for over 35 years (Norton et al., 1986). This could be
summer, as expected, which is likely when carnivores are
related to density-dependent processes as the rock hyrax is the
attracted to limited resource hotspots. Fine-scale avoidance was
most abundant prey species within the study site (Drouilly,
also higher in the karoo biome which supports less landscape
Nattrass et al., 2018b; Martins & Harris, 2013; Palmer & Fair-
heterogeneity compared to fynbos, and therefore caracals likely
all, 1988). High prey density has also been highlighted as a
have to respond reactively to reduce predation risk by leopards
potential mediator of carnivore coexistence (Marneweck et al.,
in this biome. This suggests that fine-scale avoidance responses
2019). Therefore, the high abundance of rock hyrax in the
play an important role in shaping carnivore assemblages both
Cederberg could be facilitating coexistence between leopards
seasonally and across habitats and we recommend the inclusion
and caracals. In conjunction with previous studies (Fröhlich
of these analyses in future studies.
et al., 2012; Jansen et al., 2019; Mann et al., 2019; Martins
et al., 2011; Norton et al., 1986), our findings suggest that this
prey species contributes to a fundamental part of both leopard
Dietary
and caracal diet in the Western Cape and is therefore a crucial
Both carnivores displayed specialised diets, which could be species to manage and conserve.
explained by the lack of large-bodied prey and limited prey A contentious issue in the farmland surrounding the Ceder-
diversity in the study area (Martins et al., 2011). Previous berg is the depredation of livestock by leopard and caracal,
studies suggest that leopards have generalised diets (e.g. Hay- and the subsequent retaliatory killings of these carnivores by
ward & Kerley, 2008), however, the leopard diet may be more landowners. Lethal control of leopards (and likely caracals)
specialised than previously thought (Balme et al., 2007; Pitman appears to have decreased over time due to long-term conser-
et al., 2013). Interestingly, the leopard diet in the Cederberg vation efforts (Martins, 2010), but human-carnivore conflict
was more specialised than those found in the Boland and Little still occurs today. It has been suggested that livestock preda-
Karoo regions of the Western Cape, with the Boland being a tion can be predicted from the biomass of wild prey, with live-
more mesic region dominated by fynbos and the Little Karoo stock predation rates increasing significantly when prey
having similar climate, rainfall and topography to the Ceder- biomass decreases below 540 kg/km2 (Khorozyan et al., 2015).
berg (Mann et al., 2019). This variation in feeding patterns Another study found that carnivores consumed less livestock
across sites concurs with a previous comparative study (Norton than expected due to the high biomass of preferred prey (Wang
et al., 1986). Balme et al. (2020) found that leopard popula- & Macdonald, 2009). Findings from this and other studies
tions may consist of a heterogeneous mix of specialist and (Martins et al., 2011; Norton et al., 1986) show that livestock
generalist individuals with the primary driver of dietary spe- may provide an alternative food resource for leopard and cara-
cialisation being ecological opportunity which is influenced by cal, but the contribution remains less than 10% of their overall
phenotypic constraints (e.g. sex of individuals or experience) diets. Therefore, it is likely that current wild prey populations
and environmental conditions (e.g. prey diversity), as predicted may sufficiently support leopard and caracal populations that
by the optimal foraging theory (Stephens & Krebs, 1986). coexist in the Cederberg landscape, reducing the need for car-
Therefore, we hypothesise that variations in the types and nivores to find alternative food sources.
amount of prey available to leopards as well as individual Interestingly, in a range-wide study of leopards (which
traits may be influencing the degree of dietary specialisation excluded leopards from the Cape), the mean body mass of pre-
between leopards in these different regions. Our study found ferred prey was ~43% of their body mass (Hayward et al.,
that the caracal diet was specialised and consisted mostly of 2006), whereas in our study, the mean prey body mass of leop-
rodents. Caracals typically have a broad dietary range ard selected prey was ~35% of their body mass. This suggests

10 Journal of Zoology  (2022) – ª 2022 The Zoological Society of London.


L. Müller et al. Leopard and caracal coexistence

that leopard body size may influence prey size selection,


although the availability of medium- and large-bodied prey
Acknowledgements
which make up their diets may also play an important role. We thank Ewan Brennan, Hannes de Kok, Barbara Seele,
Meanwhile, for caracals, the absence of a dominant apex preda- Ismail Wambi and Grant Baker, for assisting with the camera
tor may influence their diet selection with caracals potentially trap survey. A special thanks also to Rika du Plessis from
taking down larger-bodied prey typically associated with larger CapeNature and all the Cederberg landowners for their support
carnivore diets (Palmer & Fairall, 1988). However, rodents are and for giving us access to their properties. We thank Ford
still reported as their main prey group across most studies both Wildlife Foundation for providing a vehicle to the research
with and without more dominant competitors (e.g. Avenant & team. We thank Mikaela Liltved and the Department of Con-
Nel, 2002; Jansen et al., 2019; Melville et al., 2004). servation Ecology & Entomology at Stellenbosch University
Throughout Africa, leopards and caracals naturally co-occur for assisting with the caracal scat analysis, as well as Margaret
within systems that support intact carnivore guilds, including and Graham Avery from the Iziko South African Museum for
large-bodied and abundant prey as well as larger-bodied leop- osteological reference collections. We acknowledge the Cape
ards (e.g. Melville et al., 2004). In these systems, one would Leopard Trust for providing the resources to enable this
expect a broader ecological niche gap to exist based on the fol- research. We are also very grateful to our donors, Abax Foun-
lowing factors. Firstly, the body size difference between these dation, Bushmans Kloofs Wilderness Reserve, Lomas Wildlife
two carnivores is greater compared to our scenario, which will Protection Trust and the Rolf Stephan Nussbaum Foundation.
result in different energetic requirements (Carbone et al., 1999) Without their support this work would not have been possible.
and species may respond differently along space, time and diet Finally, we want to thank Professor Dan Parker for reviewing
axes. For example, a study from Welgevonden Game Reserve (a this paper.
fenced system with an intact carnivore guild) found that sym-
patric carnivore activity overlap with dominant lions decreased
as the body mass of sympatric carnivores decreased (Greco References
et al., 2021). Secondly, the availability of larger-bodied prey
Avenant, N. L. (1993). The caracal, Felis caracal caracal
may reduce the need for leopards to consume smaller-bodied
Schreber, 1776, as predator in the West Coast National Park.
prey, that is naturally consumed by caracals, according to ener-
getic expenditure versus energetic intake rates (Carbone et al., MSc thesis, University of Stellenbosch, South Africa.
2007), resulting in lower levels of direct competition. For exam- Avenant, N. L., Drouilly, M., Power, R. J., Thorn, M., Martins,
ple, in the Kgalagadi Transfrontier Park (an intact, semi-arid sys- Q., Neils, A., du Plessis, J., & San, D. L. E. (2016). A
tem), the leopard diet consisted largely of springbok and other conservation assessment of Caracal caracal. In M. F. Child, L.
medium-to-large-sized mammals (Mills, 1984), while the caracal Roxburgh, D. L. E. San, D. Raimondo, & H. T. Davies-
diet consisted mostly of rodents (Melville et al., 2004). Based Mostert (Eds.), The Red List of Mammals of South Africa,
on our results, there was evidence that leopards and caracals Swaziland and Lesotho (pp. 1–13). South African National
operated in different ecological niches, although the niche gap Biodiversity Institute and Endangered Wildlife Trust.
was not exceptionally broad and resource overlap did occur. Avenant, N. L., & Nel, J. A. J. (1998). Home-range use,
This could be explained by a lack of large prey available to activity, and density of caracal in relation to prey density.
leopards, hyper abundance of rock hyrax providing plenty of African Journal of Ecology, 36, 347–359.
food for both carnivores, low densities of the apex carnivore and Avenant, N. L., & Nel, J. A. J. (2002). Among habitat variation
both carnivores occurring relatively close to the 21.5 kg weight in prey availability and use by caracal Felis caracal.
threshold between small and large carnivores (Carbone et al., Mammalian Biology, 67, 18–33.
1999). Ultimately, despite a narrower ecological gap, caracals Bailey, T. N. (1993). The African leopard: ecology and behavior
are still able to coexist alongside leopards in the Cederberg land- of a solitary felid. Columbia University Press.
scape. However, quantitative evidence comparing ecological Balme, G. A., Hunter, L. T. B., & Slotow, R. (2007). Feeding
niche gaps among carnivore species is lacking and studies inves- habitat selection by hunting leopards Panthera pardus in a
tigating this could provide important insights into carnivore
woodland savanna: Prey catchability versus abundance.
competition across various systems. Today, intact large carnivore
Animal Behavior, 74, 589–598.
guilds only cover 8% of their former range due to human popu-
Balme, G. A., Hunter, L. T. B., & Slotow, R. (2009). Evaluating
lation growth and agricultural development (Wolf & Ripple,
methods for counting cryptic carnivores. The Journal of
2017). Therefore, systems with incomplete carnivore guilds (like
the Cederberg) are widespread across the continent and carni- Wildlife Management, 73, 433–441.
vores have adapted to coexist within these landscapes. Our find- Balme, G. A., le Roex, N., Rogan, M. S., & Hunter, L. T. B.
ings emphasise the importance of studying competition and (2020). Ecological opportunity drives individual dietary
coexistence mechanisms between sympatric carnivores across specialization in leopards. Journal of Animal Ecology, 89,
their ranges to better understand the ecological processes that 589–600.
govern carnivore community structure and function, particularly Balme, G. A., Pitman, R. T., Robinson, H. S., Miller, J. R. B.,
since carnivore populations are threatened by habitat loss and Funston, P. J., & Hunter, L. T. B. (2017). Leopard distribution
fragmentation, prey loss and direct poaching (Rayan & Linkie, and abundance is unaffected by interference competition with
2016). lions. Behavioral Ecology, 28, 1348–1358.

Journal of Zoology  (2022) – ª 2022 The Zoological Society of London. 11


Leopard and caracal coexistence L. Müller et al.

Bates, D., Mächler, M., Bolker, B. M., & Walker, S. C. (2015). Drouilly, M., Nattrass, N., & O’Riain, M. J. (2018). Dietary niche
Fitting linear mixed –effects models using lme4. Journal of relationships among predators on farmland and a protected area.
Statistical Software, 67, 1–48. The Journal of Wildlife Management, 82, 507–518.
Berger, K. M., & Gese, E. M. (2007). Does interference Du Preez, B., Hart, T., Loveridge, A. J., & Macdonald, D. W.
competition with wolves limit the distribution and abundance (2015). Impact of risk on animal behaviour and habitat
of coyotes? Journal of Animal Ecology, 76, 1075–1085. transition probabilities. Animal Behavior, 100, 22–37.
Braczkowski, A. R., O’Bryan, C. J., Stringer, M. J., Watson, J. Durant, S. M. (1998). Competition refuges and coexistence: An
E. M., Possingham, H. P., & Beyer, H. L. (2018). Leopards example from Serengeti carnivores. Journal of Animal
provide public health benefits in Mumbai, India. Frontiers in Ecology, 67, 370–386.
Ecology and the Environment, 16, 176–182. Edwards, S., Grange, A. C., & Wiesel, I. (2015). Spatiotemporal
Briers-Louw, W. D., & Leslie, A. J. (2020). Dietary partitioning resource partitioning of water sources by African carnivores
of three large carnivores in Majete Wildlife Reserve, Malawi. on Namibian commercial farmlands. Journal of Zoology, 297,
African Journal of Ecology, 58, 371–382. 22–31.
Brillouin, L. (1956). Science and information theory. Academic Elmhagen, B., Ludwig, G., Rushton, S. P., Helle, P., & Lindén,
Press. H. (2010). Top predators, mesopredators and their prey:
Broekhuis, F., Cozzi, G., Valeix, M., McNutt, J. W., & Interference ecosystems along bioclimatic productivity
Macdonald, D. W. (2013). Risk avoidance in sympatric large gradients. Journal of Animal Ecology, 79, 785–794.
carnivores: Reactive or predictive? Journal of Animal Ecology, Estes, J. A., Terborgh, J., Brashares, J. S., Power, M. E., Berger,
82, 1098–1105. J., Bond, W. J., Carpenter, S. R., Essington, T. E., Holt, R.
Brown, J. H., Laundre, J. W., & Gurung, M. (1999). The D., Jackson, J. B. C., Marquis, R. J., Oksanen, L., Oksanen,
ecology of fear: Optimal foraging, game theory, and trophic T., Paine, R. T., Pikitch, E. K., Ripple, W. J., Sandin, S. A.,
interactions. Journal of Mammalogy, 80, 385–399. Scheffer, M., Schoener, T. W., . . . Wardle, D. A. (2011).
Carbone, C., du Toit, J. T., & Gordon, I. J. (1997). Feeding Trophic downgrading of planet Earth. Science, 333, 301–306.
success in African wild dogs: Does kleptoparasitism by Fröhlich, M., Berger, A., Kramer-Schadt, S., Heckmann, I., &
spotted hyenas influence hunting group size? Journal of Martins, Q. (2012). Complementing GPS cluster analysis with
Animal Ecology, 66, 318–326. activity data for studies of leopard (Panthera pardus) diet.
Carbone, C., Mace, G. M., Roberts, S. C., & Macdonald, D. W. South African Journal of Wildlife Research, 42, 104–110.
(1999). Energetic constraints on the diet of terrestrial Glen, A. S., & Dickman, C. R. (2006). Diet of the spotted-tailed
carnivores. Nature, 402, 286–288. quoll (Dasyurus maculatus) in eastern Australia: Effects of
Carbone, C., Teacher, A., & Rowcliffe, J. M. (2007). The costs season, sex and size. Journal of Zoology, 269, 241–248.
of carnivory. PLoS One, 5, e22. Greco, I., Chizzola, M., Meloro, C., Swanepoel, L., Tamagnini,
Caro, T. M., & Stoner, C. J. (2003). The potential for D., & Dalerum, F. (2021). Similarities between lions and
interspecific competition among African carnivores. Biological sympatric carnivores in diel activity, size and morphology.
Conservation, 110, 67–75. Hystrix Italian Journal of Mammalogy, 32, 1–8.
Chakrabarti, S., Jhala, Y. V., Dutta, S., Qureshi, Q., Kadivar, R. Griffiths, D. (1975). Prey availability and the food of predators.
F., & Rana, V. J. (2016). Adding constraints to predation Ecology, 56, 1209–1214.
through allometric relation of scats to consumption. Journal of Harmsen, B. J., Foster, R. J., Silver, S. C., Ostro, L. E. T., &
Animal Ecology, 85, 660–670. Doncaster, C. P. (2009). Spatial and temporal interactions of
Chesson, P. (2000). General theory of competitive coexistence in sympatric jaguars (Panthera onca) and pumas (Puma
spatially-varying environments. Theoretical Population concolor) in a neotropical forest. Journal of Mammalogy, 90,
Biology, 58, 211–237. 612–620.
Crooks, K. R., & Soulé, M. E. (1999). Mesopredator release and Hass, C. C. (2009). Competition and coexistence in sympatric
avifaunal extinctions in a fragmented system. Nature, 563, bobcats and pumas. Journal of Zoology, 278, 174–180.
563–566. Haswell, P. M., Kusak, J., & Hayward, M. W. (2017). Large
Davies, A. B., Tambling, C. J., Marneweck, D. G., Ranc, N., carnivore impacts are context-dependent. Food Webs, 12, 3–13.
Druce, D. J., Cromsigt, P. G. M., le Roux, E., & Asner, G. P. Hayward, M. W., & Hayward, G. J. (2007). Activity patterns of
(2021). Spatial heterogeneity facilitates carnivores coexistence. reintroduced lion Panthera leo and spotted hyaena Crocuta
Ecological Society of America, 102, e03319. pauta in the Addo Elephant National Park, South Africa.
Douglas, R. M. (1989). A new method of cross-sectioning hair African Journal of Ecology, 45, 135–141.
of larger mammals. South African Journal of Wildlife Hayward, M. W., Henschel, P., O’Brien, J., Hofmeyr, M.,
Research, 9, 73–76. Balme, G., & Kerley, G. I. H. (2006). Prey preferences of the
Drouilly, M., Clark, A., & O’Riain, M. J. (2018). Multi-species leopard (Panthera pardus). Journal of Zoology, 270, 298–313.
occupancy modelling of mammal and ground bird Hayward, M. W., & Kerley, G. I. H. (2008). Prey preferences
communities in rangeland in the Karoo: A case for dryland and dietary overlap amongst Africa’s large predators. South
systems globally. Biological Conservation, 224, 16–25. African Journal of Wildlife Research, 38, 93–108.

12 Journal of Zoology  (2022) – ª 2022 The Zoological Society of London.


L. Müller et al. Leopard and caracal coexistence

Hayward, M. W., & Slotow, R. (2009). Temporal partitioning of López-Bao, J. V., Mattisson, J., Persson, J., Aronsson, M., &
activity in large African carnivores: Tests of multiple Andrén, H. (2016). Tracking neighbours promotes the
hypotheses. South African Journal of Wildlife Research, 39, coexistence of large carnivores. Scientific Reports, 6, 23198.
109–125. Mann, G. K. H., O’Riain, M. J., & Parker, D. M. (2015). The
Jachowski, D. S., Butler, A., Eng, R. Y. Y., Gigliotti, L., Harris, road less travelled: Assessing species-specific detection
S., & Williams, A. (2020). Identifying mesopredator release in probabilities with camera traps in an arid biodiversity hotspot.
multi-predator systems: A review of evidence from North Biodiversity and Conservation, 24, 531–545.
America. Mammal Review, 50, 367–381. Mann, G. K. H., Wilkinson, A., Hayward, J., Drouilly, M.,
Jansen, C., Leslie, A. J., Cristescu, B., Teichman, K. J., & O’Riain, M. J., & Parker, D. M. (2019). The effects of aridity
Martins, Q. (2019). Determining the diet of an African on land use, biodiversity and dietary breadth in leopards.
mesocarnivore, the caracal: Scat or GPS cluster analysis? Mammalian Biology, 98, 43–51.
Wildlife Biology, 1, 1–8. Marneweck, C., Marneweck, D. G., van Schalkwyk, O. L.,
Karanth, K. U., Srivathsa, A., Vasudev, D., Puri, M., Beverley, G., Davies-Mostert, H. T., & Parker, D. M. (2019).
Parameshwaran, R., & Kumar, N. S. (2017). Spatio-temporal Spatial partitioning by a subordinate carnivore is mediated by
interactions facilitate large carnivore sympatry across a conspecific overlap. Oecologia, 191, 531–540.
resource gradient. Proceedings of the Royal Society B: Martins, Q. E. (2010). The ecology of the leopard Panthera
Biological Sciences, 284, 20161860. pardus in the Cederberg Mountains. PhD thesis, University of
Karanth, K. U., & Sunquist, M. E. (1995). Prey selection by Bristol, UK.
tiger, leopard and dhole in tropical forests. Journal of Animal Martins, Q., & Harris, S. (2013). Movement, activity and
Ecology, 64, 439–450. hunting behaviour of leopards in the Cederberg mountains,
Keogh, H. J. (1979). An atlas of hair from southern African South Africa. African Journal of Ecology, 51, 571–579.
mammal species with reference to its taxonomic and Martins, Q., Horsnell, W. G. C., Titus, W., Rautenbach, T., &
ecological significance. PhD thesis, University of Pretoria, Harris, S. (2011). Diet determination of the Cape Mountain
South Africa. leopards using global positioning system location clusters and
Keogh, H. J. (1983). A photographic reference system of the scat analysis. Journal of Zoology, 283, 81–87.
microstructure of the hair of southern African bovids. South McVittie, R. (1979). Changes in the social behaviour of south
African Journal of Wildlife Research, 13, 89–131. west African cheetah. Madoqua, 11, 171–184.
Keogh, H. J. (1985). A photographic reference system based on Melville, H. I. A. S., Bothma, J. P., & Mills, M. G. L. (2004).
the cuticular scale patterns and groove of the hair of 44 Prey selection by caracal in the Kgalagadi Transfrontier Park.
species of southern African Cricetidae and Muridae. South South African Journal of Wildlife Research, 34, 67–75.
African Journal of Wildlife Research, 15, 109–159. Meridith, M., & Ridout, M. (2014). Overview of the overlap
Khorozyan, I., Ghoddousi, A., Soofi, M., & Waltert, M. package. https://cran.cs.wwu.edu/web/packages/overlap/
(2015). Big cats kill more livestock when wild prey reaches vignettes/overlap.pdf
a minimum threshold. Biological Conservation, 192, Miller, J. R. B., Pitman, R. T., Mann, G. K. H., Fuller, A. K., &
268–275. Balme, G. A. (2018). Lions and leopards coexist without
Klare, U., Kamler, J. F., & Macdonald, D. W. (2011). A spatial, temporal or demographic effects of interspecific
comparison and critique of different scat-analysis methods for competition. Journal of Animal Ecology, 87, 1709–1726.
determining carnivore diet. Mammal Review, 41, 294–312. Mills, M. G. L. (1984). Prey selection and feeding habits of the
Lahkar, D., Ahmed, M. F., Begum, R. H., Das, S. K., & large carnivores in the southern Kalahari. Koedoe, 27, 281–
Harihar, A. (2020). Inferring patterns of sympatry among large 294.
carnivores in Manas National Park – a prey-rich habitat Monterroso, P., Alves, P. C., & Ferreras, P. (2014). Plasticity in
influenced by anthropogenic disturbances. Animal circadian activity patterns of mesocarnivores in Southwestern
Conservation, 24, 1–13. Europe: Implications for species coexistence. Behavioral
Laundré, J. W., Hernández, L., & Altendorf, K. B. (2001). Ecology and Sociobiology, 68, 1403–1417.
Wolves, elk, and bison: reestablishing the ‘landscape of fear’ Norton, P. M., & Lawson, A. B. (1985). Radio tracking of
in Yellowstone National Park, U.S.A. Canadian Journal of leopards and caracals in the Stellenbosch area, Cape Province.
Zoology, 79, 1401–1409. South African Journal of Wildlife Research, 15, 17–24.
Levins, B. (1968). Evolution in changing environments. Norton, P. M., Lawson, A. B., Henley, S. R., & Avery, G.
Princeton University Press. (1986). Prey of leopard in four mountain areas of the south-
Linnell, J. D. C., & Strand, O. (2000). Interference interactions, western Cape Province. South African Journal of Wildlife
co-existence and conservation of mammalian carnivore. Research, 16, 47–52.
Diversity and Distributions, 6, 169–176. Nouveliet, P., Rasmussen, G. S. A., Macdonald, D. W., &
Loehle, C., & Eschenbach, W. (2012). Historical bird and Courchamp, F. (2012). Noisy clocks and silent sunrises:
terrestrial mammal extinction rates and causes. Diversity and Measurement methods of daily activity pattern. Journal of
Distributions, 18, 84–91. Zoology, 286, 179–184.

Journal of Zoology  (2022) – ª 2022 The Zoological Society of London. 13


Leopard and caracal coexistence L. Müller et al.

O’Brien, T., Kinnaird, M., & Wibisono, H. (2003). Crouching Ridout, M., & Linkie, M. (2009). Estimating overlap of daily
tigers, hidden prey: Pauta tiger and prey populations in a activity patterns from camera trap data. Journal of
tropical forest landscape. Animal Conservation, 6, 131–139. Agricultural, Biological, and Environmental Statistics, 14,
Palmer, R., & Fairall, N. (1988). Caracal and African wild cat 322–337.
diet in the Karoo National Park and the implications thereof Ripple, W. J., Wirsing, A. J., Wilmers, C. C., & Letnic, M.
for hyrax. South African Journal of Wildlife Research, 18, 30– (2013). Widespread mesopredator effects after wolf
34. extirpation. Biological Conservation, 160, 70–79.
Pangle, W. M., & Holekamp, K. E. (2010). Functions of Ritchie, E. G., & Johnson, C. N. (2009). Predator interactions,
vigilance behaviour in a social carnivore, the spotted hyaena, mesopredator release and biodiversity conservation. Ecology
Crocuta pauta. Animal Behavior, 80, 257–267. Letters, 12, 982–998.
Périquet, S., Fritz, H., Revilla, E., Macdonald, D. W., Loveridge, Romero-Muñoz, A., Maffei, L., Cuéllar, E., & Noss, A. J.
A. J., Mtare, G., & Valeix, M. (2021). Dynamic interactions (2010). Temporal separation between jaguar and puma in the
between apex predators reveal contrasting seasonal attraction dry forests of southern Bolivia. Journal of Tropical Ecology,
patterns. Oecologia, 195, 51–63. 26, 303–311.
Périquet, S., Valeix, M., Calypole, J., Drouet-Hoguet, N., Shortridge, G. C. (1934). The mammals of South West Africa.
Salnicki, J., Mudimba, S., Revilla, E., & Fritz, H. (2015). Heinemann.
Spotted hyaenas switch their foraging strategy as a response Skead, C. J. (1980). Historical mammal incidence in the Cape
to changes in intraguild interactions with lions. Journal of Province: The Western and Northern Cape, Vol. 1. Cape
Zoology, 297, 245–254. Department of Nature & Environmental Conservation.
Pianka, E. R. (1973). The structure of lizard communities. Skinner, J. D., & Chimimba, C. T. (2005). The mammals of the
Annual Review of Ecology and Systematics, 4, 53–74. southern African subregion. Cambridge University Press.
Pitman, R. T., Kilian, P. J., Ramsay, P. M., & Swanepoel, L. H. Sonawane, C., Yirga, G., & Carter, N. H. (2021). Public health
(2013). Foraging and habitat specialization by female leopards and economic benefits of spotted hyenas Crocuta Crocuta in a
(Panthera pardus) in the Waterberg Mountains of South peri-urban system. Journal of Applied Ecology, 58, 2892–2902.
Africa. South African Journal of Wildlife Research, 43, 167– Stein, A., Gerstner, K., & Kreft, H. (2014). Environmental
176. heterogeneity as a universal driver of species richness across
Prugh, L. R., Stoner, C. J., Epps, C. W., Bean, W. T., Ripple, taxa, biomes and spatial scales. Ecology Letters, 17, 866–880.
W. J., Laliberte, A. S., & Brashares, J. S. (2009). The rise of Stephens, D. W., & Krebs, J. R. (1986). Foraging theory.
the mesopredator. BioScience, 59, 779–791. Princeton University Press.
R Development Core Team (2019). R: A language and Swanson, A., Arnold, T., Kosmala, M., Forester, J., & Packer,
environment for statistical computing. R Foundation for C. (2016). In the absence of a “landscape of fear”: How lions,
Statistical Computing. hyenas, and cheetahs coexist. Ecology and Evolution, 6, 1–12.
Radloff, F. G. T., & du Toit, J. T. (2004). Large predators and Swanson, A., Caro, T., Davies-Mostert, H., Mills, M. G. L.,
their prey in a southern African savanna: A predator’s size Macdonald, D. W., Borner, M., Masenga, E., & Packer, C.
determines its prey size range. Journal of Animal Ecology, 73, (2014). Cheetahs and wild dogs show contrasting patterns of
410–423. suppression by lions. Journal of Animal Ecology, 83, 1418–
Rafiq, K., Hayward, M. W., Wilson, A. M., Meloro, C., Jordan, 1427.
N. R., Wich, S. A., McNutt, J. W., & Golabek, K. A. (2020). Switalski, T. A. (2003). Coyote foraging ecology and vigilance
Spatial and temporal overlaps between leopards (Panthera in response to gray wolf reintroduction in Yellowstone
pardus) and their competitors in the African large predator National Park. Canadian Journal of Zoology, 81, 985–993.
guild. Journal of Zoology, 311, 246–259. Terborgh, J., Lopez, L., Nuñez, P., Rao, M., Shahabuddin, G.,
Ramesh, T., Kalle, R., & Downs, C. T. (2017). Staying safe Orihuela, G., Riveros, M., Ascanio, R., Adler, G. H., Lambert,
from top predators: Patterns of co-occurance and inter-predator T. D., & Balbas, L. (2001). Ecological meltdown in predator-
interactions. Behavioral Ecology and Sociobiology, 71, 41. free forest fragments. Science, 294, 1923–1926.
Ramesh, T., Kalle, R., Sankar, K., & Qureshi, Q. (2012). Tobler, M. W. (2007). Camera Base, version 1.3. Botanical
Spatiotemporal partitioning among large carnivores in relation Research Institute of Texas. http://www.atrium-biodiversity.org/
to major prey species in Western Ghats. Journal of Zoology, tools/camerabase/
267, 269–275. Tsafack, N., Wang, X., Xie, Y., & Fattorini, S. (2021). Niche
Rayan, M. D., & Linkie, M. (2016). Managing conservation overlap and species co-occurrence patterns in carabid
flagship species in competition: tiger, leopard and dhole in communities of the northern Chinese steppes. ZooKeys, 1044,
Malaysia. Biological Conservation, 204, 360–366. 929–949.
Rich, L. N., Miller, D. A. W., Robinson, H. S., McNutt, J. W., Verschueren, S., Briers-Louw, W. D., Monterroso, P., & Marker,
& Kelly, M. J. (2017). Carnivore distributions in Botswana. L. (2021). Local-scale variation in land use practice supports a
Are shaped by resource availability and intraguild species. diverse carnivore guild on Namibian multiple-use rangeland.
Journal of Zoology, 303, 90–98. Rangeland Ecology & Management, 79, 64–76.

14 Journal of Zoology  (2022) – ª 2022 The Zoological Society of London.


L. Müller et al. Leopard and caracal coexistence

Vogel, J. T., Somers, M. J., & Venter, J. A. (2019). Niche Zhao, G., Yang, H., Xie, B., Gong, Y., Ge, J., & Feng, L.
overlap and dietary resource partitioning in an Africa large (2019). Spatio-temporal coexistence of sympatric
carnivore guild. Journal of Zoology, 309, 212–223. mesocarnivores with a single apex carnivore in a fine-scale
Walker, C. (1996). Signs of the wild: a field guide to the spoor landscape. Global Ecology and Conservation, 21, e00897.
& signs of the mammals of Southern Africa. Struik Publishers.
Wang, S. W., & Macdonald, D. W. (2009). Feeding habits and
Supporting Information
niche partitioning in a predator guild composed of tigers,
leopards and dholes in a temperate ecosystem in central Additional Supporting Information may be found in the online
Bhutan. Journal of Zoology, 277, 275–283. version of this article:
Wolf, C., & Ripple, W. J. (2017). Range contractions of the Appendix S1. Model results for predicted time-to-encounters
world’s large carnivores. Royal Society Open Science, 4, for (1) leopard-caracal (LC) and caracal-caracal (CC) events
170052. and caracal-leopard (CL) and leopard-leopard (LL) events.
Zar, J. H. (1999). Biostatistical analysis. Prentice Hall. Appendix S2. Details of leopard and caracal scats collected in
Zhang, J. (2016). Spaa: Species association analysis. https:// the Cederberg study area between May 2017 and February
cran.r-project.org/web/packages/pa/pa.pdf 2018.

Journal of Zoology  (2022) – ª 2022 The Zoological Society of London. 15

View publication stats

You might also like