Content uploaded by Luke T. B. Hunter
Author content
All content in this area was uploaded by Luke T. B. Hunter on Sep 11, 2015
Content may be subject to copyright.
Ecological Journal, Volume 6, 2004: Balme & Hunter
1
External Research
Mortality in a protected Leopard population, Phinda Private Game Reserve, South Africa:
A population in decline?
Guy Balme1 & Luke Hunter2
1Phinda Private Game Reserve
2Wildlife Conservation Society: lhunter@wcs.org (author for correspondence)
Abstract
We investigated the causes and rates of mortality in a protected Leopard population in the Phinda Private Game
Reserve, South Africa. Data from 16 radio-tagged Leopards and their cubs were used to determine the causes of
mortality and annual mortality rates for various age and sex classes in the population. Intra-specific strife accounted for
the greatest number of deaths followed by human-related mortality. Males died mainly as a result of human activity
whereas females died from natural causes. The mortality rate for males was significantly higher than for females, and
the annual mortality rate for the population was higher than any previously recorded in Leopards. Rapid turnover of
adult males due to human persecution may have reduced recruitment into the population because social instability
prevented females from raising cubs. If the present rates of mortality and recruitment are maintained, Phinda may
represent a population sink for Leopards with poor conservation and tourism prospects.
Introduction
Mammalian carnivores are vulnerable to local
extinction in fragmented landscapes due mainly to their
low densities, large ranges and inevitable conflict with
humans (Noss et al., 1996; Woodroffe & Ginsberg,
1998). Outside protected areas, accidental or
intentional killing by people drives local extinction of
large carnivores or reduces their numbers, and it is a
matter of fact that many human-dominated landscapes
will always be unsuitable for some large carnivores. As
such, large, inviolate protected areas remain critical to
the conservation of top carnivores.
However, there is increasing evidence that even
protected areas may not effectively protect large
carnivores. Using data from 22 intensive studies of
carnivores inside protected areas, Woodroffe and
Ginsberg (1998) demonstrated that 74% of known-
cause deaths were directly caused by people. They
showed that high levels of persecution of carnivores
along the border regions of protected areas were
sufficient to create population sinks. The impact of
such sinks was greatest in small reserves with high
perimeter:area ratios where the wide-ranging behaviour
of carnivores leads them to cross reserve boundaries
frequently and suffer high rates of human-caused
mortality. Irrespective of population size, those
carnivores that suffered most from human-caused
mortality along reserve boundaries were the species
most likely to disappear from reserves (Woodroffe &
Ginsberg, 1998).
In 2002, we initiated the Munyawana Leopard Project
to assess the impact of human-caused deaths on a
protected Leopard Panthera pardus population in the
Phinda Private Game Reserve, Kwa-Zulu Natal, South
Africa. Leopards are fully protected in Phinda and in
the Mkhuze Game Reserve on its western boundary
(Figure 1). However the land to the south and east of
Phinda comprises a mosaic of pastoral Zulu
communities, livestock farms and private game farms
where Leopards are often killed as perceived or real
problem animals, or by commercial trophy hunting
operations. Despite electrified game-fencing along
most borders, Leopards move freely between adjacent
properties and, because Phinda is long and narrow, few
individual Leopards have their entire home range
within the boundaries of the reserve (Balme & Hunter,
unpub. data). The result is that most individual
Leopards considered protected on Phinda are actually
exposed to high levels of hunting due to frequent
movements off the reserve.
In this paper, we present the results of the first 29
months of the study, addressing three main questions:
1. What are the causes of mortality to Leopards in the
Phinda population?
2. What is the annual mortality rate of the Phinda
Leopard population for different age and sex classes?
3. Does Phinda effectively protect Leopards?
Ecological Journal, Volume 6, 2004: Balme & Hunter
2
Figure 1: The study site showing land-use types in the region (Inset:
arrowhead indicates region shown). Leopards are legally protected
in areas shown in white but the degree of protection actually offered
them varies widely; Leopards are best protected in Phinda and
Mkhuze Game Reserve. Leopards are persecuted on game farms
though harvest rates depend on individual landowners. The
MunYaWana area is now contiguous with Phinda and although
Leopards are occasionally hunted there, this is likely to diminish or
cease entirely. Cattle areas are indicated where livestock is present
though these areas also have significant game populations; these
areas are hostile to Leopards. Leopards appear to be killed rarely on
communal land though these areas have insufficient habitat and prey
for permanent occupation.
Methods:
Detail of the study site and general methodology can be
found in Hunter (1998) and Hunter et al. (2003).
Between April 2002 and October 2003, we captured 16
Leopards and fitted them with VHF radio-collars. We
attempted to locate every radio-collared animal at least
once daily and recorded their location to the nearest
50m using a hand-held GPS receiver. From April 2002
to August 2004, we logged a total of 6527 transmitter-
days following Leopards.
We aged Leopards by the wear and eruption of teeth
(Stander, 1997), and classified them into three age
classes to investigate age-related mortality: juveniles
under one year old, sub-adults one to three years old,
and adults over three years old. All sub-adults were
independent from the mother but showed little or no
territorial behaviour, while all adults displayed frequent
territorial behaviour such as territorial vocalising and
scent-marking (Balme & Hunter, unpub. data). All
juveniles were dependent on the mother.
We established causes of mortality by direct
observation or by post-mortem and evidence collected
at the site such as tracks in the surrounding substrate.
We usually found dead Leopards within 24 hours of
their death and the cause was rarely ambiguous. We
considered three Leopards which disappeared outside
Phinda to have been killed by people. Although this is
speculative, in all cases the animal disappeared and its
radio-collar abruptly stopped transmitting within 24
hours of moving into an area known to be hostile to
Leopards. We were unable to search for missing
Leopards on these properties to confirm their fate.
However, we searched for radio-collar signals from
their boundaries or from the air, a technique which had
always yielded a signal in similar attempts made while
the Leopard was living. Further, radio-collars did not
fail during the study and had they, Leopards with faulty
collars would almost certainly have been re-sighted
given that all our collared animals during life were seen
consistently by game drives at Phinda (Balme &
Hunter, in prep.). The highly probable explanation in
these cases is that the radio-collars were destroyed by
people after the animal was illegally killed.
We calculated annual mortality rates (AMRs) after
Ferraras et al. (1992) using the formula:
number of deaths
number of radio-transmitter days x 365
We calculated the AMR for the population, and
separately for males, females, all adults and all sub-
adults using age/sex-class specific data (i.e. number of
deaths and radio-transmitter days). We did not have
sufficient observations to calculate the juvenile
mortality rate. We compared differences between age
and sex classes using non-parametric statistics (Siegel,
1956).
Results:
Causes of mortality
Eight radio-collared Leopards died during the study,
for which the cause of death was certain or probable in
seven cases (Table 1). Intra-specific conflict and
anthropogenic deaths were equally important causes of
mortality for adults and sub-adults combined, both
claiming three individuals. Natural causes (excluding
other Leopards) were responsible for one and possibly
two additional deaths.
Ecological Journal, Volume 6, 2004: Balme & Hunter
3
Human Intra-specific Inter-specific UnknownAge & Sex
AMR DEA AMR DEA AMR DEA AMR DEA
Adult male 0.497 20.248 10000
Adult female 0000000.160 1
Sub-adult male 0.484 1000.484 100
Sub-adult female 000.361 20000
Total 0.168 30.168 30.056 10.056 1
Table 1: Cause of death (DEA) and cause-specific annual mortality rates (AMR) for different age and sex categories.
Sex & Age AMR RD IND DEA
Adult male 0.745 1469 4 3
Adult female 0.160 2280 5 1
Sub-adult male 0.968 754 3 2
Sub-adult female 0.361 2024 5 2
Total male 0.821 2223 7 5
Total female 0.254 4304 9 3
Total adult 0.389 3749 9 4
Total sub-adult 0.526 2778 8 4
Total 0.447 6527 16* 8
Table 2: Annual mortality rates (AMR), number of radio-transmitter days (RD), number of monitored individuals (IND) and number of deaths
(DEA) for different age/sex classes of leopards (*Total shows number of individuals radio-tracked during study. Classes combined may exceed this
number because some individuals were tracked while belonging to different age classes.)
Intra-specific clashes killed one adult male and two
sub-adult females. M1 was a resident territorial male
when killed by the adult male, M5. The sub-adult
female F10 was killed by an uncollared adult female
leopard when almost three years old and displaying the
first signs of territorial behaviour (‘sawing’ and scent-
marking). The sub-adult female F15 was 20 months of
age when killed by a male Leopard, the sub-adult M14.
Additionally, although we have not included juvenile
deaths in the estimation of mortality rates, three
juveniles were killed by the adult male M13 that had
recently become established and was not their sire.
Human-related deaths killed three males. Two adult
males M5 and M13 were established territorial males at
the time of their disappearance on properties adjacent
to Phinda. The sub-adult male, M7 was last located on
private property near the town of Hluhluwe,
approximately 11km south of Phinda’s southern
boundary, when his signal disappeared.
One Leopard was killed by another carnivore. The sub-
adult male, M4 died from septicemia arising from
severe bite wounds on the neck, shoulders and
hindquarters. We found evidence at the site of a
prolonged fight between the Leopard and an adult
Spotted Hyaena Crocuta crocuta which was consistent
with the bite marks. We do not know the cause of death
of the adult female F2 whose carcass we examined
approximately two weeks after death when
decomposition was advanced.
Mortality rates
The average annual mortality rate (AMR) of the
population between April 2002 and August 2004 was
0.447 (Table 2). The mortality rate for males was
significantly higher than for females (p = 0.004, df =
1). Sub-adult males (0.968) had the greatest annual
mortality of any cohort, followed by adult males
(0.745). Adult females had the lowest mortality rate
(0.160) with only one death recorded for the study
period.
Discussion:
Causes of mortality
Leopards at Phinda were killed chiefly by other
Leopards. Two deaths (M1 and F10) were the result of
territorial clashes between same-sex pairs. Leopards
are known to defend their territories from same-sex
intruders sometimes leading to fatalities (Le Roux &
Skinner, 1989) though the proportion of deaths here
caused by other Leopards is higher than reported in
other detailed studies (Bailey, 1993; Stander et al.,
1997). The killing of sub-adult female F15 by a sub-
adult male (M14) is unusual, particularly as these
animals had met on previous occasions. F15 was not
reproductively mature and M14 typically rebuffed
curious approaches from her in past encounters we
observed, but his aggression was largely demonstrative
and never escalated to physical contact. A similar
instance of a male Leopard killing a younger female
was documented at Londolozi Private Game Reserve,
South Africa (Hes, 1991). Males might regard young,
reproductively unavailable females as competitors for
food resources and attack them as such, though it is
unclear why it happened later rather than sooner in this
case.
All radio-tagged Leopards that died due to human-
related causes were males. Males are more desirable to
trophy hunters due to their larger size, and males also
utilise larger home ranges and cover greater daily
distances than females, increasing their chances of
moving off the reserve into areas where they can be
hunted (Mizutani & Jewel, 1998; Hunter & Balme,
unpubl. data). Importantly, as far as we know, males
Ecological Journal, Volume 6, 2004: Balme & Hunter
4
killed outside Phinda were not shot legally by
international hunters with CITES permits. Nonetheless,
all three deaths occurred during the legal trophy
hunting season between April to November. This may
be due to Leopards being mistaken for another legally
hunted species but more likely they are killed
intentionally by South Africans. Leopards are
persecuted intensely by various landowners in the
region and there is little chance of prosecution for
illegal killing. We do not know whether the risks for
Leopards are elevated during the legal hunting season.
It is likely that opportunistic killing of Leopards occurs
year-round though increased numbers of local hunters
seeking other species in the area during the legal
season might result in more Leopards killed then.
Bailey (1993) reported that starvation (mainly of sub-
adults) accounted for the greatest proportion of
Leopard deaths in Kruger National Park (KNP). He
suggested that sub-adults were more likely to starve
due to a related set of factors that included a lack of
hunting experience, loss of condition due to increased
parasitic infestations, competition for resources with
other predators, and seasonal changes in prey
abundance and availability of cover. We found no
evidence of starvation contributing to Leopard deaths
at Phinda. One individual, M7 was emaciated at
capture but this was due to serious injuries, probably
incurred from a conspecific. His condition improved
dramatically post-capture and he made a full recovery.
He was clearly foraging successfully for 12 months
following capture until February 2004 when he was
killed outside Phinda.
The only adult female to die during the study probably
succumbed to natural causes. At the time of her death,
she was due to give birth and we made no effort to
approach her, assuming she was localised with
newborn cubs. By the time we decided to investigate,
autolysis of the carcass was too advanced to determine
a cause but there was no evidence to suggest her death
was related to human activity. We found no snares and
the site was not close to a boundary where the risk of
snaring at Phinda is greatest. She may have died due to
complications arising from birth. This is considered
unusual in felids (Apps & Du Toit, 2000) but is
occasionally recorded: for example, an otherwise
healthy Lioness Panthera leo in Pilanesberg National
Park died from secondary septicemia due to dystocia
(G. van Dyk, Pers. comm.).
Mortality rates and the future of the Phinda population
Our results agree with other detailed studies of African
Leopards in demonstrating sex dependent mortality.
Similar to Bailey’s (1993) study of Leopards in Kruger
National Park, males at Phinda suffered higher
mortality rates than females. However, there are some
significant contrasts. The AMR of the Phinda Leopard
population is high, with effectively half the population
dying each year, many of them directly or indirectly
due to human factors. This is considerably higher than
mortality rates recorded for the protected population in
the KNP, South Africa where all deaths were natural
(see Table 2: KNP adult AMR = 0.185, subadult AMR
= 0. 320; Bailey, 1993). In particular, the AMR of
Phinda males is close to three times that of Kruger
males (AMR = 0.250; Bailey, 1993) though differences
between the methods used to calculate mortality rates
may have given rise to some of the discrepancy.
High levels of mortality among adult males at Phinda
may have had an additive effect on mortality in the
population by lowering the reproductive success of
females. Although male Leopards provide no parental
care to cubs, the presence of the sire allows mothers to
raise cubs with a reduced risk of infanticide by foreign
males (Hunter et al., in press). There are few reliable
observations of infanticide in leopards (see: Ilani, 1986;
1990; Scott & Scott, 2003) but new males entering the
population are likely to kill existing cubs. We saw this
once during the study period, following the illegal
killing of M5. The resulting vacancy was rapidly filled
by the male M13 who killed the 4-month old cubs of
F12 (which M5 probably sired). Although we observed
infanticide only on this occasion, there was limited
evidence of successful reproduction in general. During
the 29-month study, we observed consorting pairs on
18 occasions involving seven adult females, multiple
males and 305 actual matings, yet only seven cubs in
three litters were born. Only two cubs are still alive at
the time of writing, one of them still dependent on its
mother.
That few cubs were produced during the study may be
a further consequence of high turnover among males.
In Lions, high levels of infanticide further impact
reproductive output by reducing the rate at which
females conceive (Packer & Pusey, 1983). Lionesses
display a period of reduced fertility immediately
following the take-over of a pride by a new male
coalition. This presumably allows females to assess the
fitness of new males and postpone conception until the
males are established and the threat of another take-
over is reduced. Rapid turnover of male Leopards at
Phinda might be driving female Leopards into a
reproductive dead-end in which cubs are killed at high
rates and subsequent conception is delayed. In an
isolated Leopard population in the Judean Desert,
Israel, infanticide was the chief reason that not a single
individual was recruited into the adult population
during a five-year period (Ilani, 1986; 1990).
At Phinda, human persecution of males outside the
reserve is a possible ultimate cause. Whitman et al.
(2004) demonstrated that excessive trophy hunting of
male Lions under a certain age reduced the chance of
population persistence because Lionesses failed to raise
cubs to independence due to the constant cycle of
infanticide and delayed conception. We do not yet have
sufficient data to draw the same conclusion here;
Ecological Journal, Volume 6, 2004: Balme & Hunter
5
however there are coarse patterns that warrant concern.
Only one individual born into the population reached
independence and became established on Phinda during
the study period and, at the time of writing, she is still
too young to reproduce. In contrast, eleven individuals
died, four of which were breeding residents. For a
population to remain stable, birth and mortality rates
need to be equivalent, and immigration and emigration
rates need to be comparable (Eltringham, 1979). At
Phinda, deaths far exceeded the number of Leopards
being born in the population so high levels of
immigration will need to occur for densities to remain
stable. We do not yet have a clear picture of
immigration and emigration at Phinda to assess this.
However, even if immigration is balancing mortality, it
may be depressing reproduction if it encourages
constant incursions by new males that kill cubs.
It is possible that the study period represents a
temporary period of unusually elevated social flux in
Leopards that will eventually stabilise. We do not yet
know whether male territorial tenure is reduced across
the population, or that the low success rates of female
reproduction observed here are permanent. Continued
monitoring (the project will proceed until at least the
end of 2006) will investigate whether the population
attains greater stability in the future.
In the meantime, we need to consider the possibility
that Phinda Leopards are in a state of ongoing
instability. Geographically narrow, and located
between the (presumably) well-protected population in
Mkhuze Game Reserve to the west and the private and
communal lands to the south and east where Leopards
are persecuted, Phinda might represent a population
sink for Leopards, despite their fully protected status
on the reserve. There is little reason to think that levels
of killing of Leopards adjacent to Phinda will decline.
Indeed, with the recent (October 2004) CITES approval
granted to South Africa to double its quota of Leopard
trophies to 150, legal off-take will increase in the area.
Assuming that levels of illegal killing remain the same,
the patterns described here are likely to persist or
possibly worsen.
Provided a well-protected source population (Mkhuze)
persists, it is unlikely that Leopards will disappear from
Phinda. However, at the very least, ongoing social
perturbation will affect the tourism potential of the
species on the reserve. Regular replacement of known,
viewable animals with shy immigrants and a lack of
cubs are two immediate consequences already evident.
To address these problems and the potentially more
serious conservation implications for the species will
require a coordinated approach from all stakeholders
with an interest in the Leopard. The formation of the
Munyawana Leopard Conservancy (Hunter et al.,
2003) represents a beginning toward this end.
However, to date there is little evidence that this has
fostered greater tolerance for the Leopard in the region
outside protected areas where the species is already
valued.
Acknowledgments:
We are grateful to CC Africa for permission to conduct
the research on Phinda and for providing ongoing
financial support and accommodation for the research
team. Our gratitude extends to all at Phinda for their
assistance. Special thanks to Kevin Pretorius for his
unfailing support without which the research could not
proceed; and to Jaco Mattheus whose assistance in
capturing leopards has been invaluable. Our thanks also
to all the rangers and trackers for their sensitivity and
patience in their interactions with Leopards, and for
their terrific help in locating animals for us to capture.
Thanks to all of our volunteer field assistants whose
support has been critical. We are particularly grateful
to Tim Driman, Sue and Jim MacIntosh, Christian
Sperka, Scott Salzman and the Environmental Action
Group, Furman University (South Carolina) for their
generous donations for the purchase of equipment.
Finally, special mention must go to Victoria Mitchell
for her fund raising initiatives and ongoing efforts to
support the project.
References:
APPS, P. & DU TOIT, R. 2000. Creatures of Habit: Understanding
African Animal Behaviour. Struik, Cape Town.
BAILEY, T.N. 1993. The African leopard: Ecology and Behaviour of
a Solitary Felid. Columbia University Press, New York.
BALME, G.A. & HUNTER, L.T.B. in prep. Radio-collars on safari;
quantifying the benefits for ecotourism of ‘invasive’
research on leopards Panthera pardus in KwaZulu-Natal,
South Africa.
ELTRINGHAM, S.K. 1979. The Ecology and Conservation of Large
African Mammals. Macmillan, London.
FERRERAS, P., ALDAMA, J.J., BELTRAN, J.F. & DELIBES, M.
1992. Rates and causes of ortality in a fragmented
population of Iberian lynx Felis pardina Temminck,
1824. Biological Conservation 1992: 197-202.
HES L. 1991. The Leopards of Londolozi. Struik Winchester, Cape
Town.
HUNTER, L.T.B. 1998. The behavioural ecology of reintroduced
lions and cheetahs in the Phinda Private Game Reserve,
Kwa-Zulu-Natal, South Africa. Unpublished PhD thesis,
University of Pretoria.
HUNTER, L.T.B., BALME G.A., WALKER, C., PRETORIUS, K.
& ROSENBERG, K. 2003. The landscape ecology of
leopards in northern KwaZulu-Natal, South Africa: a
preliminary project report. Ecological Journal 5: 24-30.
HUNTER, L. & BALME, G. 2004. Trash or in trouble? The
leopard’s predicament. Africa Geographic 2(1): 40-47.
HUNTER, L., RAY, J. & HENSCHEL, P. in press. Leopard
Panthera pardus. In: Kingdon, J. & Butynski, T. (eds.).
The Mammals of Africa, Vol. 4 Carnivora etc. Academic
Press.
ILANI, G. 1986. Preliminary observations on the ecology of the
leopard (Panthera pardus jarvisi) in he Judean Desert. In:
Miller, S.D. & Everett, D.D. (eds.). Cats of the World:
Biology, Conservation and Management. National
Wildlife Federation, Washington D.C. p.199.
ILANI, G. 1990. Leopard Panthera pardus in Israel. Cat News 12: 4-
5.
Ecological Journal, Volume 6, 2004: Balme & Hunter
6
LE ROUX, P.G. & SKINNER, J.D. 1989. A note on the ecology of
the leopard (Panthera pardus Linnaeus) in the Londolozi
Game Reserve, South Africa. African Journal of Ecology
27: 167-171.
MIZUTANI, F. & JEWEL, P.A. 1998. Home-range and movements
of leopards (Panthera pardus) on a livestock ranch in
Kenya. Journal of Zoology (London) 244. 269-286.
NOSS, R.F., QUIGLEY, H.B., HORNOCKER, M.G., MERILL, T.
& PAQUET, P.C. 1996. Conservation biology and
carnivore conservation in the Rocky Mountains.
Conservation. Biology 10: 949-963.
PACKER, C. & PUSEY, A.E. 1983. Adaptations of female lions to
infanticide of incoming males. American Naturalist 121:
716-728.
SCOTT, J. & SCOTT, A. 2003. Big Cat Diary: Leopard. Harper
Collins, London.
SIEGEL, S. 1956. Non-parametric Statistics for the Behavioural
Sciences. McGraw-Hill, New York.
STANDER, P.E. 1997 Field age determination of leopards by tooth
wear. African Journal of Ecology 35: 156-161
STANDER, P.E., HADEN, P.J., KAQECE & GHAU. 1997. The
ecology of asociality in Namibian leopards. Journal of
Zoology (London) 242: 343-364.
WHITMAN, K., STARFIELD, A.M., QUADLING, H.S. &
PACKER, C. 2004. Sustainable trophy hunting of African
lions. Nature 428 (6979): 175-178.
WOODROFFE, R. & GINSBERG, J.R. 1998. Edge effects and
extinction of populations inside protected areas. Science
280: 2126-2128.
