Походу капский лев жив!

P. l. melanochaita, known as the Cape Lion, became extinct in the wild around 1860. Results of mitochondrial DNA research do not support the status as a distinct subspecies. It seems probable that the Cape lion was only the southernmost population of the extant P. l. krugeri.

В настоящее время львы обитают лишь в Африке, к югу от Сахары, а также в виде нескольких сотен особей в Индии, в штате Гуджарат. Однако в доисторические, а также в исторические времена лев был гораздо более широкораспространённым животным. При попытке определения исторического ареала льва в целом и европейской части ареала в частности мы сталкиваемся с одной проблемой - кого считать львом современного типа (Panthera leo), а кого пещерным львом (Panthera (leo) spelaea), и что представляет собой пещерный лев - подвид или отдельный вид льва? Некоторые учёные (Сотникова и Никольский, 2006; Верещагин, 1971; Harington, 1969) выделяют пещерного льва в отдельный вид - Panthera spelaea, а его американского собрата, так называемого ужасного льва, или американского пещерного льва (Panthera (leo) atrox) его подвидом. Многие другие склонны считать пещерного льва всего лишь подвидом современного льва, то есть Panthera leo spelaea.
<lj-cut text="Дальше">В Европе львы (Panthera leo) известны со среднего плейстоцена по средние века. Здесь же в плейстоцене обитали пещерные львы (Panthera (leo) spelaea). Очень трудно разобраться, какие останки принадлежат льву, а какие к пещерному льву. Эти два вида (если это в действительности разные виды) очень схожи между собой. Их черепа практически одинаковы, за исключением некоторых незначительных различий, а также большего размера черепа пещерного льва. Однако все эти различия вполне могут быть внутривидовыми. К примеру, черепа разных подвидов волков (Canis lupus) отличаются друг от друга даже в большей степени, чем черепа льва и пещерного льва. Так что очень может быть, что пещерный лев является лишь подвидом льва современного типа. Таким образом картина с европейскими львами ещё более усугубляется. Европейский подвид льва, Panthera leo europaea, известен с плейстоцена Южной Европы. Как уже отмечалось выше, в плейстоцене в Европе обитал также пещерный лев (который встречался также много северней). Какие из останков европейских львов принадлежат к Panthera leo europaea, а какие к Panthera (leo) spelaea - вопрос довольно сложный.
Пещерный лев, будь он отдельным видом или подвидом современного льва известен довольно хорошо. Что же представлял собой европейский подвид современного льва - по большому счёту остаётся загадкой. Вполне возможно, что в Европе в эпоху позднего плейстоцена и голоцена обитали львы, того же подвида, что и в Азии - Panthera leo persica, то есть азиатские львы. Как бы то ни было, европейские львы были очень близки к современным азиатским.
В исторические времена львы современного типа были довольно широко распространены на Балканах, в Греции, Италии, Франции, Армении, Грузии и т.п. Последние львы Европы вымерли (а точнее были вытеснены человеком) в 10 веке на Кавказе. Сейчас трудно сказать, какому именно подвиду принадлежали львы, обитающие на Кавказе - к европейскому Panthera leo europaea или к азиатскому Panthera leo persica, тем более, что вполне возможно, что в данном случае речь идёт об одном и том же подвиде, или по крайне мере очень близкородственных между собой подвидах.
Генетически азиатские и европейские львы близки к ныне вымершим берберийским льва из Северной Африки, Panthera leo leo, одного из самых крупных подвидом льва. Азиатские львы не обладают столь пышной гривой, как берберийские, однако их грива, в отличие от современных африканских львов часто доходит до самого брюха, как и у их североафриканских родственников. Считается, что азиатские львы несколько уступают размерами своим африканским собратьям, однако мне это кажется сомнительным. Я видел довольно много как африканских, так и азиатских львом и все они были примерно одинакового размера. Мне кажется, что разница в размерах носит скорее индивидуальный, нежели межподвидовый характер. Добавлю также, что один из крупнейших львов, которых я видел в зоопарке принадлежал именно к азиатскому подвиду. Самый длинный азиатский лев, чей рекорд был зарегистрирован имел 292 см в длину, а самый высокий - 107 см в плече. Некто капитан Сми (Captain Smee) убил самца азиатского льва длиной в 268 см, который весил 222,3 кг (это вес крупного, хотя и не очень, африканского льва).
Следует отметить, что подвиды львов не так чётко очерчены, как например подвиды тигра (Panthera tigris). В настоящее время существует несколько подвидов львов в Африке, и один азиатский подвид. Азиатский подвид льва на сегодня является наиболее чётко очерченным подвидом льва. От своих африканских родичей он отличается менее пышной гривой (которая однако заходит дальше на брюхо), чуть более растянутым форматом и относительно немного укороченным черепом. Размеры азиатского льва в принципе сопоставимы с таковыми у его африканских сородичей.
Так что же собой представляют в действительности европейские львы? Для этого сначала следует установить, что собой представляют пещерные львы - вид или подвид. А если пещерный лев являет собой всё-таки отдельный от современного льва вид, то в таком случае надо как-нибудь отделить плейстоценовые останки европейского подвида льва современного типа от останков европейских пещерных львов.
Подводя итог, следует ещё раз отметить, что как бы то ни было, львы того же вида, что и современные были обычны по крайней мере для южных регионов Европы, включая практически весь Кавказ. Лев является очень пластичным видом, приспасабливающимся к самым разных средам обитания и некогда у него был один из самых обширных ареалов обитания среди хищных млекопитающих. А если лев и пещерный лев являются всё-таки одним и тем же видом, то в таком случае можно с уверенностью сказать, что данный вид обладал некогда самым обширным ареалом обитания среди всех хищных млекопитающих, так как включал в себя все континенты, кроме Австралии и Антарктиды.

Ну естественно мерзнут, но ведь это выход к выживанию!!
Больше нет примеров среди хищников, а конкретно среди кошачьих в такой организации в зимнее время!!

Разумеется будут мёрзнуть, так как толком не аклиматизированы. Тут нужно несколько поколений.

А действительно ли львы Южной Африки меньше по размерам и длинномордее остальных африканцев?

Отредактировано Revs (02 March 2010 21:41:41)

P. l. persica, known as the Asiatic Lion or South Asian, Persian, or Indian Lion, once was widespread from Turkey, across Southwest Asia, to Pakistan, India, and even to Bangladesh. However, large prides and daylight activity made them easier to poach than tigers or leopards; now around 300 exist in and near the Gir Forest of India.[23]
P. l. leo, known as the Barbary Lion, is extinct in the wild due to excessive hunting, although captive individuals may still exist. This was one of the largest of the lion subspecies, with reported lengths of 3–3.3 metres (10–10.8 ft) and weights of more than 200 kilograms (440 lb) for males. They ranged from Morocco to Egypt. The last wild Barbary lion was killed in Morocco in 1922.[24]
P. l. senegalensis, known as the West African Lion, is found in western Africa, from Senegal to Nigeria.
P. l. azandica, known as the Northeast Congo Lion, is found in the northeastern parts of the Congo.
P. l. nubica, known as the East African or Massai Lion, is found in east Africa, from Ethiopia and Kenya to Tanzania and Mozambique.
P. l. bleyenberghi, known as the Southwest African or Katanga Lion, is found in southwestern Africa, Namibia, Botswana, Angola, Katanga (Zaire), Zambia, and Zimbabwe.
P. l. krugeri, known as the Southeast African Lion or Transvaal Lion, is found in the Transvaal region of southeastern Africa, including Kruger National Park.
P. l. melanochaita, known as the Cape Lion, became extinct in the wild around 1860. Results of mitochondrial DNA research do not support the status as a distinct subspecies. It seems probable that the Cape lion was only the southernmost population of the extant P. l. krugeri.

http://urbawebsite.narod.ru/cats/textcats/lion.html

Азиатский лев по ряду признаков (невзрачная, жидкая грива либо ее совсем нет) отличается от своего африканского сородича. Исследования показали, что эти звери морфологически и генетически отличны от африканских.

Очень интересные слова. В частности, по поводу гривы. Вспоминаются безгривые (пещерные) львы с наскальных и пещерных рисунков и вот эти слова:

European lion (Panthera leo europaea or Panthera leo tartarica) could be an extinct subspecies of lion that inhabited southern Europe until historic times. This population is generally considered part of the Asiatic lion (Panthera leo persica), but others consider it a separate subspecies, the European lion (Panthera leo europaea). They also could possibly have been the last remnants of the Cave Lion (Panthera leo spelaea).

Капский лев не вымер - это наиболее южная популяция южноафриканских львов (крюгери). Он не заметно крупнее. Восточноафриканские львы могут изредка достигать веса в 250 кг.
Любопытно, что капский лев внешне почти не отличим от берберийского льва, но генетически берберийский ближе всего к азиатским львам, а капский скорее близок к другим африканским львам чем к берберийскому и тем более азиатскому.
Тут дело в схожей среде обитания.
Таким образом смею предположить что львы Грузии выглядели примерно также как берберийские и капские львы, так как условия жизни похожи. Природа юга ЮАР довольно похожа на грузинскую.

И мне кажется что современные азиатские львы принадлежат разным подвидам. Азиатские львы из зоопарка в Иерусалиме отличаются от индийских львов в Гирском лесу, хотя по ним видно что они азиатские. Я думаю между ними такая же разница как между масайскими и скажем сенегальскими львами.
Если в Сенегале, Анголе и Танзании живут разные подвиды львов, то глупо было бы предположить, что в Греции, на Балканах, Грузии и Индии жил всего один подвид.
Там где в Грузии обитал лев природа как я уже говорил похожа больше на южноафриканскую, нежели на природу Гуджарата. И так как там холоднее, то скорее всего кавказские львы были крупнее индийских и внешне мало чем отличались от капских.

Я сам не знаю откуда они его взяли, в этой ветке есть его фотки. Он огромен, 220 кг мне кажется как минимум. Он больше африканских львов из двух других израильских зоопарков (Хайфского и Ришонского) и больше африканского льва из Тбилисского зоопарка. В Израиле в Сафари паруке есть целый прайд львов, они живут на обширной территории, но крупные самцы всегда дрыхнут и я толком не смог определить их размеры, но они крупные. Но в любом случае этот азиатский лев, ему сейчас около 9 лет, зовут Лайдер, крупнее львов из вышепречисленных зоопарков, по крайней мере крупнее трёх африканских львов которых я видел в зоопарках Израиля и Грузии. Так что выборка немаленькая. Я хочу лишь сказать что азиатские львы не меньше африканских. Всё в первую очередь зависит от конкретной особи а не от подвида. Хотя масайские львы как правило крупнее сенегальских.

Crazy Zoologist :

Аллозавр,
P. l. persica - азиатский лев
P. l. leo - берберийский лев
P. l. senegalensis - сенегальский лев - довольно некрупный лев.
P. l. azandica - его не всегда выделяют.
P. l. nubica - масайский лев (наиболее "типовой" из современных львов)
P. l. bleyenberghi - катангский, или ангольский лев - пышная и очень светлая грива, легко отличим от других подвидов.
P. l. krugeri - южноафриканский лев, возможно тот же подвид что и капский лев который числиться вымершим.

Берберийский подвид умер.А меня интересуют, которые ещё живут.
какие?

P. l. persica - азиатский лев
P. l. leo - берберийский лев
P. l. senegalensis - сенегальский лев
P. l. azandica -
P. l. nubica - масайский лев
P. l. bleyenberghi - катангский, или ангольский лев
P. l. krugeri - южноафриканский лев,

Какие размеры?

J. Zool., Lond. (2005) 267, 309–322 C _ 2005 The Zoological Society of London Printed in the United Kingdom doi:10.1017/S0952836905007508
Prey preferences of the lion (Panthera leo)
Matt W. Hayward* and Graham I. H. Kerley
Terrestrial Ecology Research Unit, Department of Zoology, Nelson Mandela Metropolitan University, P.O. Box 77000, Port Elizabeth, 6031,
Eastern Cape, South Africa
(Accepted 7 March 2005)
Abstract
Lions Panthera leo are generally thought to prey on medium to large ungulates. Knowledge of which species are
actually preferred and which are avoided is lacking, however, as is an understanding of why such preference or
avoidance may arise. An analysis of 32 studies over 48 different spatial locations or temporal periods throughout
the distribution of the lion shows that it preferentially preys upon species within a weight range of 190–550 kg.
The most preferred weight of lion prey is 350 kg. The mean mass of significantly preferred prey species is 290 kg
and of all preferred species is 201 kg. Gemsbok, buffalo, wildebeest, giraffe and zebra are significantly preferred.
Species outside the preferred weight range are generally avoided. Species within the preferred weight range that
are not significantly preferred (such as roan, sable and eland) generally have features that reduce predation either
morphologically (e.g. sable horns), ecologically (e.g. roan and sable occurring at low density), or behaviourally
(e.g. the large herd size and increased vigilance of eland). Warthog are below the preferred weight range yet are
taken in accordance with their availability and this is probably due to their sympatry with lion, their relatively slow
evasion speed and their lower level of vigilance. Plots of prey preference against prey body mass follows a bell
curve with a right skew that, we argue, is caused by collective hunting by lions of larger-bodied prey. Our methods
can be used on all large predators and are likely to be useful in assessing competition in sympatric communities of
predators, cooperative hunting and predicting predator diets. This will allow us to move beyond descriptive dietary
studies to improve our predictive understanding of the mechanisms underlying predator–prey interactions.
Key words: buffalo, gemsbok, giraffe, Jacobs’ index, optimal foraging, predation preference, preferred prey weight
range, zebra, wildebeest, warthog, lion, Panthera leo
INTRODUCTION
The strategies of a predator are forged by natural selection
to maximize nutrient intake while being tempered by
a wide range of ecological constraints, such as prey
density and habitat, that differ throughout its geographical
distribution (Sunquist & Sunquist, 1997). As long as a
predator can increase its survival chances or reproductive
success by hunting more efficiently, natural selection will
favour efficient, optimally foraging predators (J. R. Krebs,
1978).
For lions Panthera leo L. this means preying upon a
broad range of medium- and large-sized mammals (Hanby
& Bygott, 1991). Yet a review of the literature reveals
they commonly take species as small as warthog (e.g.
Savuti, Botswana; Viljoen, 1993) and as large as buffalo
(e.g. Kafue National Park, Zambia; Mitchell, Shenton &
Uys, 1965). Lions also take unusual prey items such as
* All correspondence to: M. W. Hayward.
E-mail: hayers111@aol.com
seals (Bridgeford, 1985), rhinoceroses(W.M. Elliot, 1987;
Brain, Forge & Erb, 1999; Matipano, 2004) and elephant
(Ruggiero, 1991a).
A summary such as this reveals very little about the
ecology of the lion, other than its catholic tastes. These
dominant prey species may be the only ones present or
may be the most abundant. Alternatively, these data may
reflect innate preferences by lions. If two species are
equally abundant, why is one preferred? To determine
how and why a predator selects its prey, cognisance of
prey availability is imperative. If a predator kills a species
more frequently than expected based on its availability
then we can consider it a preferred prey species, while if
the predator takes proportionally fewer prey than expected
based on availability then it is an avoided prey species.
This definition of preference is potentially misleading,
as a kill signifies not just prey availability but also
its vulnerability (Schaller, 1972). The capture of prey
by a predator involves behaviours relating to searching,
stalking, attacking and subduing, and the susceptibility
of a prey species to a particularly predator relates to the
matching of these events (J. P. Elliott, McTaggart Cowan&
310 M. W. HAYWARD AND G. I. H. KERLEY
Holling, 1977). Whether a predator initiates this series of
behaviours depends on the difference between the energy
expenditure required during the hunt and that gained
in its ingestion (J. P. Elliott et al., 1977). This energy
differential must also be weighted against the potential
risks associated with capturing it.
Factors affecting prey preferences
Factors affecting the vulnerability of prey to a certain
predator are its abundance, size, temporal and spatial distribution,
defences and anti-predatory tactics (Sunquist &
Sunquist, 1997). Environmental factors that influence
hunting success in lions include grass/shrub cover, time
of day, moon presence and terrain (van Orsdol, 1984;
Stander & Albon, 1993; Funston, Mills & Biggs, 2001).
Features of the predator that affect hunting success include
hunting-group size and composition, and hunting
method used (Sunquist & Sunquist, 1997). Climatic conditions
are also influential (Mills, Biggs & Whyte, 1995).
Unfortunately the majority of these are difficult to measure
(Sunquist & Sunquist, 1997), such that Radloff & du Toit
(2004) suggest that the complexity of the interactions
between these factors, compounded by opportunistic
predatory behaviour, will confound attempts to explain
predator–prey relationships in terms other than body mass.
Encounter rate
At each site, there is a given amount of energy available
as food for carnivores and higher herbivore density means
more food for predators (Sunquist & Sunquist, 1997).
High encounter rate reflects similar habitat use, activity
patterns and temporal and spatial distribution patterns
(Sunquist & Sunquist, 1997). Lions have a very diverse
habitat tolerance ranging from semi-desert to dense woodland
(Stuart & Stuart, 2000). They hunt mostly at night or
in cooler daytime periods although they are highly opportunistic
(Schaller, 1972). Lions are also able to travel long
distances in search of prey (Eloff, 1984) and regularly hunt
in groups (Schaller, 1972). This suggests that lions have
a relatively high encounter rate with their prey. We can
therefore expect that lions will encounter the full range of
potential prey species occurring in their habitat, and a reanalysis
of Scheel’s (1993, table 1) data suggest that there
is significant linear relationship between herd availability
and encounter rate (r2 =0.978, n=10, P<0.001).
Body mass of prey
Although large felids can subsist on small, abundant prey,
they are morphologically specialized to take prey their
own size or larger, and readily do so (Sunquist&Sunquist,
1997). Larger vertebrate predators tend to forage optimally
on larger-bodied prey (see review by Pyke, Pulliam &
Charnov, 1977) as energy maximizers (Griffiths, 1975),
and so are expected to take the largest prey that could
safely be killed (Sunquist & Sunquist, 1997). Prey
digestibility and nutrient quality are unlikely to be important
as large predators possess constant and efficient
digestive systems (Ewer, 1973), so profitability for carnivores
is correlated with prey size (Lendrem, 1986).
Studies from the Serengeti confirm this for lions as their
preference for wildebeest and zebra during the migration,
and warthog and buffalo during prey scarcity suggests
they are risk-sensitive foragers that maximize food intake
(Scheel & Packer, 1995) such that their modal prey
size is 150 kg (Packer, 1986). Differences in food habits
of a species between sites reflects prey availability and
vulnerability (Sunquist & Sunquist, 1997).
Conversely, larger-bodied prey species are also thought
to avoid predation through their larger size (Schaller 1972;
J. P. Elliott et al., 1977; Ruggiero 1991a; Sinclair, 1995b),
which would reduce the size of preferred prey. Smaller
prey depend less on their ability to evade predators at the
subduing stage, and instead aim to evade predators during
search, stalk and attack phases (J. P. Elliott et al., 1977).
Herd size
Larger groups convey safety from predation by increased
detection of the predator through increased group vigilance
(Hamilton, 1971) or by the confusion effect (Estes,
1976), suggesting predators should prefer solitary individuals
over large herds, as occurs in Kruger National Park
(Funston, Mills&Biggs, 2001). Conversely, predators can
more easily locate large groups, either visually or through
their smell (Krause & Godin, 1995), and approach closer
to them (Walther, 1969), suggesting herding species will
be preferentially preyed upon. Small groups of kob detected
lions hunting at night significantly more than larger
groups of topi (van Orsdol, 1984), despite topi being the
most vigilant prey species (Schaller, 1972).
Habitat use of prey
Habitat features are also important drivers of prey selection.
All felids use physical environmental features extensively
to get as close as possible to prey while hunting
(Sunquist & Sunquist, 1997). Grass to 0.4 m height was
required for a successful daytime hunt (J. P. Elliott et al.,
1977) and hunting success increased with grass up to
0.8 m (van Orsdol, 1984). Yet it is difficult to classify and
compare habitat constraints amongst sites (Sunquist &
Sunquist, 1997).
Localized features, driven by climate-induced habitat
change, affect the susceptibility of prey to a predator.
Drought is known to increase the susceptibility of buffalo
to lion predation (Prins & Iason, 1989), while wet periods
decrease this susceptibility for buffalo, as well as for
waterbuck and kudu (Smuts, 1978). Increased water
availability led to increased lion predation on roan antelope
(Harrington et al., 1999) probably through altered
habitat conditions (Mills & Funston, 2003). Similarly,
rainfall variation affects zebra and wildebeest predation
rates in the Serengeti (Scheel & Packer, 1995) and both
Lion prey preferences 311
species declined in high rainfall years in the Kruger due to
increased lion and hyena predation (Smuts, 1978; Mills &
Shenk, 1992).
Anti-predator behaviour
Aspecies’ anti-predatory strategy can involve crypsis (e.g.
kudu), speed of evasion (e.g. tsessebe, topi, hartebeest),
mobbing predators (e.g. springbok, Thomson’s gazelle) or
active defence (e.g. warthog, zebra stallions, gemsbok and
buffalo herd bulls) (Estes, 1991). Analysing the degree
of crypsis a species uses in avoiding predation requires
knowledge of the visual acuity of the predator relative to
the pelage colour and pattern of the prey. Similarly, colours
and patterns providing crypsis in an open grassland are
unlikely to do so in denser vegetation, so habitat features
also need to be accounted for.
The minor amount of published work on prey evasion
speed suggests that lion have an initial acceleration
advantage but this is lost within the first few seconds, after
which Thomson’s gazelle rapidly outpace them, followed
by zebra with wildebeest only fractionally faster than
lion (J. P. Elliott et al., 1977). Buffalo also have a slight
speed advantage over lion (Prins & Iason, 1989), while
warthog are relatively slow compared to the ungulates
(Estes, 1999).
Migratory or nomadic prey species are largely unaffected
by top-down, predatory influences (Schaller, 1972;
Kruuk, 1972; Mills, 1990), however resident herbivores
at low density can be regulated by predation (Fryxell,
Greever & Sinclair, 1988). Yet variation in movement
patterns within sites, such as the sedentary and nomadic
wildebeest populations in the Kalahari (Mills & Retief,
1984), precludes this from being investigated more
broadly.
The result of anti-predatory strategies from the
predator’s perspective varies from an unsuccessful hunt
to bodily harm and even death. Elephant and rhinoceros
actively defend themselves and their young (Makacha &
Schaller, 1969; Ruggiero, 1991a; Berger & Cunningham,
1994). Buffalo are known to mob predators and lions
have been killed by them (Mangani, 1962; Beyers,
1964; Mitchell et al., 1965; Makacha & Schaller, 1969).
Giraffe, sable (Wilson, 1981) and gemsbuck (Eloff, 1964)
have also been observed killing lions and a thorough
literature search is likely to reveal more such incidents.
Any injury that incapacitates a predator has serious consequences
on its survival chances (Sunquist & Sunquist,
1997). These species, and hippopotamus, therefore constitute
the greatest category of threat to predators based on
their size, strength and aggressive nature.
Aims and summary
In 1978, Smuts stated that ‘predation needs to be analysed
for each particular situation and its influence may vary not
only spatially, but also temporally within a certain area’.
Since then, numerous studies have done just that and
enough data has accumulated to investigate the issue of
lion prey preferences from a species, rather than population
perspective. In this study therefore, the aim was to
use kill and prey availability data collected from various
studies conducted throughout the distribution of the lion
to determine which, if any, prey species the lion prefers
and those it avoids. Further, why particular species are
preferred or avoided is discussed. Our results will allow
us to predict the diet of the lion at other sites of natural
or reintroduced populations, and thereby facilitate the expansion
of predator dietary ecology beyond the descriptive
to the predictive. These same techniques can be used
on all predators, to predict their diet at unstudied
sites and to investigate other features such as predator
coexistence.
By necessity, our analyses are generalized and descriptive
because of the nature of the data we are restricted in
using. Consequently, it was not possible to use varying
ecological conditions or migratory/sedentary behaviour
of prey because this information is rarely stated in the
studies used, there are migratory or sedentary populations
within individual species (e.g. wildebeest; Estes, 1976;
Mills & Retief, 1984), or not a large enough sample
size was available. Numerous studies have explored
factors that make certain species susceptible to predation
(e.g. J. P. Elliott et al., 1977; Stander & Albon, 1993;
Mills & Shenk, 1992; Funston, Mills & Biggs, 2001).
Hence, rather than reanalyse this work, our study attempts
to explain the preference and avoidance of prey by
lions and relate this to general characteristics of those
species.
METHODS
Aliterature survey revealed 32 different studies describing
the diet of the lion which included some measure of prey
availability (either actual or relative) (Table 1). Several
of these studies were conducted over a long-term and
these allowed temporally separated prey preferences to be
calculated as prey abundance changed over time (Table 1).
Others provided detailed information on prey and
availability of lions in different study regions (Table 1).
Such partitioning has been used previously in the
study of carnivore ecology (see Creel & Creel, 2002).
Consequently, 48 assessments of prey preference were
calculated from sites throughout the distributional range
of the lion.We do not believe that autocorrelation exists by
using data from the same area at different prey abundance,
as one of the fundamental rules of whether a species is
captured and killed is the probability of coming in contact
with the predator and this varies with prey density.
Numerous studies provided excellent descriptive information
on lion diet but insufficient or no information on
prey availability (Mitchell et al., 1965; Eloff, 1973; 1984;
Smuts, 1979; McBride, 1984; Stander, 1992; Scheel,
1993; Scheel & Packer, 1995; Funston, Mills, Biggs et al.,
1998; Cronje, Reilly & MacFadyen, 2002). Unless
other sources could be found that provided prey
312 M. W. HAYWARD AND G. I. H. KERLEY
Table 1. Sites and sources of prey preference data used in this study
Country Site Years/Period Source
Botswana Savuti, Chobe National Park Wet season Viljoen, 1993
Dry season Viljoen, 1993
Central African Manovo-Gounda-St. Floris 1982–84 Ruggiero, 1991b
Republic National Park
Congo Albert National Park 1950s Bourliere & Verschuren, 1960
India Gir Forest, Gujarat 1970s Joslin, 1984
1986–89 Chellam & Johnsingh, 1993
Kenya Nairobi National Park 1966 Foster & McLaughlin, 1968
1967 Foster & McLaughlin, 1968
1968–72 Rudnai, 1974
Masai-Mara Game Reserve 1973–74 Saba, 1979
Namibia Etosha National Park 1975–78 Berry, 1981
Unstated Stander & Albon, 1993
South Africa Hluhluwe-Umfolozi 1965–68 Steele, 1970
Umfolozi Whateley & Brooks, 1985
Hluhluwe Whateley & Brooks, 1985
1990s Maddock et al., 1996
Kalahari Gemsbok National Park 1971–82 Eloff, 1984
1974–82 Mills, 1984
Kruger National Park 1956–65 North Pienaar, 1969
1956–65 South Pienaar, 1969
1956–65 Central Pienaar, 1969
1969 Bryden, 1976
1970 Bryden, 1976
1971 Bryden, 1976
Early 1990s Mills & Biggs, 1993
1990s Harrington et al., 1999
Madjuma Game Reserve 1998 Power, 2002
1999 Power, 2002
Phinda Resource Reserve 1992–96 Hunter, 1998
Timbavati Game Reserve 1964 Hirst, 1969
1965 Hirst, 1969
1966 Hirst, 1969
1967 Hirst, 1969
Tanzania Lake Manyara 1967–68 Makacha & Schaller, 1969
Ngorongoro 1970–72 J. P. Elliott & McTaggart Cowan, 1978
Selous 1974 Rodgers, 1974
1993–2002 Creel & Creel, 2002
Serengeti 1950s Wright, 1960
1965–66 Kruuk & Turner, 1967
1966–69 Masai Pride Schaller, 1972
1966–69 All Schaller, 1972
Uganda Queen Elizabeth Ishashi Pride Van Orsdol, 1982, 1984
(Ruwenzori) National Park Mweya Pride Van Orsdol, 1982, 1984
Zambia Kafue National Park 1962–64 Mitchell et al., 1965
Zimbabwe Mana Pools National Park 1968–69 Dunham, 1992
1981–84 Dunham, 1992
1985–89 Dunham, 1992
availability at the appropriate time, these studies could
not be used in this analysis (see Table 2).
The data collected in these studies was derived from
both incidental observations and continuous follows.
Continuous follows are widely regarded as the superior
method of ascertaining the diet of a predator (Bertram,
1979; Mills, 1992; although see also Radloff & du Toit,
2004). Incidental observations are biased toward larger
prey; however, this bias against smaller items is generally
alleviated by the undercounting of small prey species in
aerial counts. Furthermore, even continuous studies find
only a small proportion of kills are small species (Schaller,
1972; Packer, Scheel & Pusey, 1990; Hanby, Bygott &
Packer, 1995; Scheel & Packer, 1995). Consequently both
types of data were used.
Many indices have been used to describe selectivity;
however, none is generally superior to the rest and none
is without bias and increasing error at small proportions
(Chesson, 1978; Strauss, 1979). As a result, many researchers
have overstated the accuracy of their preference
results (Norbury & Sanson, 1992). This is particularly the
case with the most commonly used techniques, such as
the forage ratio and Ivlev’s electivity index (Ivlev, 1961).
These indices and their variations suffer from several
problems including non-linearity, bias to rare food
items, increasing confidence intervals with increasing
Lion prey preferences 313
Table 2. Assumptions made in determining prey availability for studies where it is not implicitly stated. Page and table numbers relate to
the source of the information
Study and section Assumptions made or source of availability data
Bourliere & Verschuren (1960) Data taken from Bourliere (1963)
Chellam & Johnsingh (1993) Wild prey availability data taken from Khan et al. (1996) and livestock taken from original paper.
Body mass of chital and sambar taken from Nowak & Paradiso (1983)
Dunham (1992) Availability data comes from Durham (1994). Elephant and hippopotamus were excluded because
they were scavenged
Mills (1984) Availability data comes from Mills & Retief (1984)
Eloff (1984) Availability data comes from Mills (1990) – dunes
Mills & Biggs (1993) Giraffe and hippopotamus were assumed to be in equal abundance from Mills & Biggs (1993
Fig. 2) and buffalo abundance came from Donkin (2000). Similarly kudu and waterbuck were
assumed to be of equal abundance. This is irrelevant as no individuals of these species were
preyed upon by lions
Mitchell et al. (1965) Availability data taken from Dowsett (1966)
Rudnai (1974) Availability data taken from 1967 game census (Foster & McLaughlin, 1968)
Saba (1979) Availability data taken from 1976 in Ottichilo et al. (2000) and wildebeest and zebra numbers
from Sinclair (1995a) based on presence in the Mara for 3–4 months
Schaller (1972) Resident prey of the 115 km2 range of the Masai pride is given in table 34 (p. 430), however
there is no mention of migratory species that were preyed upon. The total migratory population
of wildebeest (400,000), zebra (132,500) and Thomson’s gazelle (165,000) were multiplied by
the proportion of time they spent around Seronera (p. 211) to give their availability in theMasai
pride range
Steele (1970) Availability data taken from Brooks & Macdonald (1983)
Van Ordsol (1984) Prey availability for large species taken from Van Ordsol (1982) and density per km for duiker,
hippopotamus, elephant and reedbuck were calculated from Petrides & Swank (1965)
Wright (1960) Availability data comes from Grzimek & Grzimek (1960)
heterogeneity, being unbound or undefined, and lacking
symmetry between selected and rejected values (Jacobs,
1974). Confidence intervals also become excessive for
proportions below c. 10% (Strauss, 1979). There are
methods that minimize these biases (Krebs, 1989) and
we have chosen Jacobs’ index:
D = r − p
r + p − 2rp
where r is the proportion of the total kills at a site made
up by a species and p is the proportional availability of the
prey species (Jacobs, 1974). The resulting value ranges
from + 1 to −1, where + 1 indicates maximum preference
and −1 maximum avoidance (Jacobs, 1974). The mean
Jacobs’ index for each prey species across studies was
calculated (±1 SE wherever mean is shown) and these
values were tested for significant preference or avoidance
using t-tests against a mean of 0 if they conformed to
the assumptions of normality (Kolmogorov–Smirnov test)
(e.g. Palomares et al., 2000; Hayward, de Tores & Banks,
2005). Where transformation could not satisfy these
assumptions, the sign test (Zar, 1996) was used, although
the biological relevance of non-significant results where
several −1 (maximum avoidance) values were coupled
with a fractionally positive one must be questioned. The
value of this kind of analysis is that it is not biased by the
results from one particular area; it is not overly influenced
by the available community of prey species because for a
species to be significantly preferred or avoided it must be
so in diverse communities throughout its range; and that
it takes into account of varying hunting group sizes and
sex ratios by being collected in populations that hunt as
fission–fusion groups.
Several assumptions were made when calculating the
availability of prey from various studies and these are
listed in Table 2. Sources of error arise in preference indices
via inaccurate availability or choice data. Such errors
may occur in various studies here where availability data
is not taken for the exact time period during which
the prey data were collected. However, such errors are
minimal given the frequency with which studies presented
actual availability or coincided with other studies
that did.
The resulting mean Jacobs’ index values were analysed
using multiple regression of 3 variables that were initially
considered to influence lion predation after several
other variables were found to be partially correlated
(see Results). Significant relationships were plotted using
distance weighted least-squares and linear-regression fits
of transformed data. Lions are generally thought to eat
prey of medium body size and consequently 3/4 . mean
adult female body mass of prey species was used (Table 3)
to take account of calves and sub-adults eaten. This value
was used by Schaller (1972) to ascertain food intake in
lions and was confirmed by Scheel (1993) who found 25%
of kills in the Serengeti were calves or juveniles. Weights
were taken from Stuart & Stuart (2000).
Social organization of prey species was taken to be an
indicator of the ability of the prey to detect predators.
This was a categorical variable with 1 relating to solitary
individuals, 2 relating to species that exist in pairs, 3
relating to small family grouping species, 4 to small herds
(10–50) and 5 relating to large herds (>50; Table 3).
This is a simplification as solitary males exist alongside
larger breeding herds and smaller bachelor herds of several
species. Yet mean herd size of each sex is not provided
314 M. W. HAYWARD AND G. I. H. KERLEY
Table 3. Jacobs’ index values, number of studies recording the species as a potential prey item (n), mean percentage availability of each
species (out of a total of 4,581,156 species records), mean percentage that each species comprised of the total kills recorded (22,684),
body mass (3/4 of mean female body mass) and categories of herd size, habitat density and injury threat to predators used in modelling.
Specifics of each category are described in the text and their details were derived from Stuart & Stuart (2000) and Estes (1991). Symbols
+ refers to significantly preferred and −, significantly avoided
Jacobs’ index Availability (%) Prey (%) Body Herd
Species (±1 SE) n (±1 SE) (±1 SE) mass (kg) size Habitat Threat
Aardvark Orycteropus afer 0.03 (0) 1 0.24 ± 0 0.25 ±0 40 1 2 0
Baboon Papio cynocephalus −0.89 (0.11) 5 4.22 ± 1.89 1.25 ± 0.56 12 5 2 1
Black rhinoceros Diceros bicornis −1 (0) 2 1.19 ± 0.84 0 ± 0 800 1 3 2
Blue duiker Cephalophus monticola −1 (0) 1 0.01 ±0 0±0 3 1 3 0
Buffalo Syncerus caffer 0.32 (0.10)+ 30 13.29 ± 2.43 20.65 ± 3.77 432 5 2 2
Bushbuck Tragelaphus scriptus −0.53 (−0.20)− 9 1.14 ± 0.31 1.94 ± 0.52 23 1 3 0
Bushpig Potamochoerus larvatus 0.11 (0.26) 2 1.75 ± 1.24 1.91 ± 1.35 46 3 3 1
Chital deer Axis axis −0.81 (0.19) 2 42.14 ± 29.80 21.48 ± 15.19 30 5 1.5 0
Common duiker Sylvicapra grimmia −0.83 (0.07)− 7 1.84 ± 0.65 0.22 ± 0.08 16 1 2.5 0
Eland Tragelaphus oryx 0.18 (0.14) 24 1.23 ± 0.25 2.33 ± 0.48 345 5 2 1
Elephant Loxodonta africana −0.87 (0.12) 4 3.23 ± 1.14 1.01 ± 0.36 1600 3 2 2
Gemsbok Oryx gazella 0.70 (0.06)+ 4 14.80 ± 7.40 23.29 ± 11.65 158 4 1 2
Giraffe Giraffa camelopardalis 0.24 (0.10)+ 24 3.71 ± 0.76 6.32 ± 1.29 550 3 2 2
Grant’s gazelle Gazella granti −0.56 (0.22)− 8 7.08 ± 2.68 1.31 ± 0.49 38 4 1 0
Hartebeest Alcephalus buselaphus 0.02 (0.13) 17 9.55 ± 2.19 9.66 ± 2.22 95 4 1.5 0.5
Hippopotamus Hippopotamus amphibius −0.45 (0.33) 4 6.12 ± 3.06 5.75 ± 2.88 750 3 1.5 2
Impala Aepyceros melampus −0.73 (0.05)− 34 39.52 ± 6.78 9.94 ± 1.70 30 4 2 0
Klipspringer Oreotragus oreotragus −0.96 (0.01)− 4 0.89 ± 0.44 0.02 ± 0.01 10 2.5 3 0
Kob Kobus kob −0.31 (0.26) 4 43.07 ± 21.53 27.89 ± 13.95 47 4 2 0
Kudu Tragelaphus strepsiceros 0.13 (0.12) 25 3.20 ± 0.64 7.47 ± 1.49 135 3 2 0
Nilgai Boselaphus tragocamelus −1 (0) 1 0.31 ± 0 0 120 3 2 0
Nyala Tragelaphus angasi −0.32 (0.14) 7 19.91 ± 7.52 15.95 ± 6.03 47 3 2 0
Oribi Ourebia ourebi −0.72 (0.28) 2 5.53 ± 3.91 0.90 ± 0.64 14 2 1 0
Ostrich Struthio camelus −0.55 (0.17)− 11 1.23 ± 0.37 1.25 ± 0.38 70 3 1.5 0
Porcupine Hystrix africaeaustralis 0.58 (0) 1 0.47 ± 0 2.0 ± 0 10 1 2 1.5
Puku Kobus vardoni −.058 (0) 1 0.26 ± 0 1.0 47 4 1 0
Red duiker Cephalophus natalensis −1 (0) 1 1.78 ±0 0 10 1 3 0
Reedbuck Redunca sp. −0.57 (0.15)− 17 1.60 ± 0.39 0.61 ± 0.15 30 3 3 0
Roan Hippotragus equinus 0.15 (0.17) 16 0.65 ± 0.19 1.53 ± 0.46 220 3.5 2 0.5
Sable Hippotragus niger −0.05 (0.20) 12 0.94 ± 0.27 2.18 ± 0.63 180 4 2 0.5
Sambar deer Cervus unicolor −0.16 (0.84) 2 1.72 ± 1.22 7.39 ± 5.23 200 4 2 0
Sharpe’s grysbok Raphicerus sharpei −0.96 (0.03)− 4 1.16 ± 0.67 0.04 ± 0.02 7 1 2.5 0
Springbok Antidorcas marsupialis −0.59 (0.32) 4 50.09 ± 25.05 29.38 ± 14.69 26 5 1 0
Steenbok Raphicerus campestris −0.86 (0.08)− 4 4.34 ± 1.94 0.07 ± 0.03 8 1.5 1.5 0
Thomson’s gazelle Gazella thomsoni −0.62 (0.17)− 9 19.21 ± 6.79 10.33 ± 3.65 15 5 1 0
Topi/tsessebe Damaliscus lunatus 0.01 (0.13) 13 5.43 ± 1.51 5.56 ± 1.54 90 3 2 0
Vervet monkey Cercopithecus aethiops −1 (0) 4 0.96 ± 0.48 0 ± 0 3.5 4 2 0
Warthog Phacochoerus africanus 0.11 (0.09) 39 4.06 ± 0.65 8.07 ± 1.29 45 3 2 0
Waterbuck Kobus ellipsiprymnus 0.18 (0.12) 30 2.85 ± 0.52 7.98 ± 1.46 188 3.5 2 0.5
White rhinoceros Ceratotherium simum −1 (0) 2 1.94 ± 1.37 0 ± 0 1400 2 1.5 2
Wildebeest Connochaetes taurinus 0.27 (0.08)+ 38 17.14 ± 2.78 25.99 ± 4.22 135 5 1 0
Zebra Equus burchellii 0.16 (0.07)+ 40 11.24 ± 1.78 15.19 ± 2.40 175 3 2 1
in the majority of studies analysed here and nor are rates
of predation for each sex or age class. Hence we were
limited to the information provided. None the less, it is
still meaningful and similar categorization has been used
previously (e.g. Funston, Mills & Biggs, 2001).
A predator must encounter a prey species in order to
kill it and habitat type may affect predation rates. Also
the density of vegetation may affect the detectability of
prey. Although inherently difficult to classify (Sunquist &
Sunquist, 1997), a categorical variable of habitat density
was used, with 1 referring to open grasslands, 2 referring
to savanna, and 3 to densely vegetated areas. Obviously
a species may overlap these habitat types and in this case
an average of habitat use was applied (Table 3). Again
by necessity, this is a simplification; however, it serves to
highlight general trends.
Finally, the anti-predatory strategy a species uses will
affect its chances of becoming prey. Unfortunately, the
lack of studies comparing crypsis or evasion speed of
prey species (J. P. Elliott et al., 1977; Prins & Iason, 1989
excepted) meant the threat of injury to a hunter was all that
could be analysed. The categories of threat used were 0
(no threat), 1 (minor threat or active defence of young), 2
(severe threat; known deaths attributed to predators caused
by this species) (Table 3). Information for each of these
categories comes from Estes (1991) and Stuart & Stuart
Lion prey preferences 315
(2000), and Nowak & Paradiso (1983) for Indian prey
species. Categorization as used here is fairly rudimentary;
however, it serves to illustrate general trends adequately.
The analysis of this data was not separated by the sexes
or the group size of the hunting lions as no study looked
solely at the prey of solitary hunters or groups or either
sex. We aimed to investigate lion diet from a population
perspective whereby a population at each site contains
both males and females hunting as individuals or groups,
thereby taking account of hunting preferences of both
individuals and groups of males and females. We expect
our results to reflect the mean hunting group size at a site
(e.g. 2 in the Serengeti; Schaller, 1972) and the variation
in hunting group size (1–11; Packer et al., 1990) of both
males and females. The data available never indicated
the group size or sex of each kill and we leave such
detailed analyses to individual study sites where data has
been collected with this in mind (e.g. Radloff & du Toit,
2004). Similarly, as several of the studies used looked at
seasonal variation (e.g. Viljoen, 1993) and others included
varying climatic conditions (e.g. Hirst, 1969; Bryden,
1976; Dunham, 1992), we expect the mean Jacobs’ index
value for a species to reflect the variable susceptibility of
prey in drought or above average rainfall.
Generalized linear and additive models were not used
because of the requirement that there be 10 samples for
each predictor variable (Burnham & Anderson, 2002). As
Jacobs’ index for individual prey species was the dependent
variable, our analysis is limited to prey species that
also have their availability recorded.
RESULTS
Jacobs’ index scores for 22 684 kills of 42 species recorded
as prey in the literature are shown in Fig. 1 and
Table 3. Zebra are recorded as lion prey in 40 studies, with
warthog (39), blue wildebeest (38), impala (34), buffalo
(30) and waterbuck (30) also frequently taken (Table 3).
Springbok account for 50.1% of available prey at the four
sites where it was recorded, and kob account for 43.1% in
the four floodplain sites it was recorded (Table 3). Impala
(39.5% of prey available at 34 sites), Thomson’s gazelle
(19.2% at 9 sites), blue wildebeest (17.1% at 38 sites),
gemsbok (14.8% at 4 sites), buffalo (13.3% at 30 sites)
and zebra (11.2% at 40 sites)were the other most abundant
prey species (Table 3).
Springbok were the most commonly killed prey (29.4%
at the 4 sites it occurred), closely followed by kob (27.9%
at 4 sites). Both of these species are taken proportionally
less than they are available (Table 3). Blue wildebeest
(26.0% at 38 sites), gemsbok (23.3% at 4 sites) and buffalo
(20.7% at 30 sites) were taken proportionally more than
often than they were available (Table 3). A plot of the
percentage that each species made up of the diet of lion
and its body mass shows lions predominately eat prey
ranging from 40 to 251 kg with a peak at 115 kg (Fig. 2).
The mean body mass of the eight most frequently killed
lion prey is 160±73 kg.
Table 4. Regression statistics for the multiple regression model
Jacobs’ index = −1.04 + 0.09 (herd size) + 0.23 (log10(body
mass)−0.03 (log10(availability). Standard error of estimate=
0.481; r 2 = 0.174. Analysis of variance F3,38 = 2.67, P = 0.061.
All variables satisfied multiple normal distribution and had constant
variances (normality test P = 0.292; constant variance test P =
0.532; Power= 0.994).Body mass (italicized) predicted the Jacobs’
index value at α = 0.05
Variable Coefficient SE t Probability
Constant −1.039 0.255 −4.688 <0.001
Log10 (availability) −0.031 0.124 −0.252 0.803
Log10 (body mass) 0.228 0.112 2.033 0.049
Herd size 0.092 0.068 1.351 0.185
Gemsbok (t=11.05, d.f.=3, P=0.008), buffalo
(t=3.25, d.f.=28, P=0.003) blue wildebeest (t=3.61,
d.f.=37, P<0.001), giraffe (t=2.31, d.f.=23,
P=0.030) and zebra (t=2.45, d.f.=39, P=0.019) are
significantly preferred (Fig. 1). Eland, kudu, warthog,
waterbuck, roan, topi, tsessebe, hartebeest and sable
are all preyed upon in accordance with their abundance
(Fig. 1). A greater sample size through time from India’s
remnant Gir Forest lion population is required to ascertain
whether sambar, chital or nilgai are preferred or avoided.
Conversely, klipspringer (t=–65.54, d.f.=3, P<
0.001), Sharpe’s grysbok (t=–37.22, d.f.=3, P<0.001),
elephant (Sign test Z =2.47; n=8; P=0.013), steenbok
(t=–10.51, d.f.=3, P=0.002), common duiker (t=
−12.00, d.f.=7, P<0.001), impala (t=–13.77,
d.f.=33, P<0.001), reedbuck (t=–3.89, d.f.=16,
P=0.001), Grant’s gazelle (t=–2.57, d.f.=7, P=
0.037), ostrich (t=–3.23, d.f.=10, P=0.009), bushbuck
(t=–2.93, d.f.=10, P=0.012) and Thomson’s gazelle
(t=–3.70, d.f.=9, P=0.006) are all significantly
avoided by lion. Rhinoceros (black and white), baboon
and vervet monkey are also avoided (Fig. 1); however,
these differences were not significant according to the
sign test (P>0.05). It is likely that greater sample sizes
would result in kob, nyala, hippopotamus, springbok and
oribi also being significantly avoided (Fig. 1).
A strong positive correlation was found between
body mass and predator injury threat category (r=0.65;
P<0.05), no doubt as a result of larger prey species being
able to inflict more damage on lions than smaller species.
Irrespectively, there was no relationship between predator
injury threat and Jacobs’ index values (r2 =0.057;
P=0.211). Similarly, there was a negative correlation
between herd size class and habitat density category (r=
−0.50; P<0.05) indicating that species that live in big
herds prefer open habitat. Nonetheless, there was no
relationship between Jacobs’ index values and habitat
density category (r2 =0.071; P=0.163).
Subsequently, a multiple regression model of relative
availability, body mass and herd size category revealed
body mass significantly contributed to the prediction of
the Jacobs’ index dependent variable (Table 4). A distance
weighted least squares regression plot of mean Jacobs’
index values against body mass (3/4 mass of adult female)
316 M. W. HAYWARD AND G. I. H. KERLEY
-1 -0.5 0 0.5 1
Rhinoceros (Black)
Rhinoceros (White)
Klipspringer
Sharpe's grysbuck
Baboon
Elephant
Steenbock
Common duiker
Chital
Impala
Oribi
Springbock
Reedbuck
Grant's gazelle
Ostrich
Hippopotamus
Bushbuck
Thomson's gazelle
Nyala
Kob
Sambar
Sable
Topi/tsessebe
Hartebeest
Warthog
Bushpig
Kudu
Roan
Zebra
Waterbuck
Eland
Giraffe
Blue wildebeest
Buffalo
Gemsbock
Jacobs' index (1 SE)
Fig. 1. Lion Panthera leo dietary preferences based on Jacobs’ index (mean±1 SE) of 48 lion populations at differing prey densities.
Black bars, species taken significantly more frequently than expected based on their availability (preferred); grey bars, species taken in
accordance with their relative abundance; unfilled bars, species taken significantly less frequently than expected based on their availability
(avoided). Only species recorded in lion diet more than once are included.
shows the preferred weight range of lion prey is between
190 kg and 550 kg (Fig. 3). Instead of a clear peak in
preference the distribution is skewed to the right (Fig. 3).
Exclusion of the megaherbivores (>600 kg) led to a
significant linear relationship between Jacobs’ index and
body mass (Fig. 4). This relationship still existed for
prey species with body masses ranging from 30 kg to
600 kg (i.e. excluding many significantly avoided species)
(r2 =0.583; n=19; P<0.001). While herd size was not
considered by the multiple regression to be useful in
predicting the Jacobs’ index of a species, there is still
a significant linear relationship between the two variables
(Fig. 5).
The most preferred weight of lion prey species appears
to be approximately 350 kg (Fig. 3). The mean body mass
of the significantly preferred prey specieswas 290±84 kg
and for all preferred species was 201±44 kg. There is
no relationship between prey availability and lion prey
preference (Table 4).
DISCUSSION
Lions prefer large prey species within a weight range of
190 to 550 kg (Fig. 3) irrespective of their availability
(Table 4), yet they predominately take prey substantially
smaller than this (Fig. 2) reflecting their opportunistic
hunting behaviour. In the Serengeti, they preferred
prey ranging from 170 to 250 kg (Sinclair, Mduma &
Brashares, 2003), which is lower but comparable to that
Lion prey preferences 317
0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8 3.0 3.2
Log10(body mass)
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
Log10(proportion prey + 1)
Fig. 2. Distance weighted least-squares relationship between mean
percentage of lion Panthera leo diet (log10 + 1) and prey body mass
(0.75 . adult female body mass) (r =0.19, P =0.28).
0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8 3.0 3.2
Log(mass)
-1.0
-0.8
-0.6
-0.4
-0.2
0.0
0.2
0.4
Jacobs’ index
Fig. 3. Distance weighted least squares relationship between lion
Panthera leo prey preference (mean Jacobs’ index) and the log10 of
prey body mass (r =0.401, P =0.015).
found here for lions throughout their range. While the
modal prey size of felids is usually less than their body
weight (Packer, 1986), lions are expected to take prey
>45% of their body mass as they are larger than Carbone
et al.’s (1999) 21.5 kg threshold, but the preferred weight
range of lions is greater than 100% of their body mass.
This weight range encompasses adults of larger
antelope, buffalo and the young of massive (or mega-)
herbivores, such as hippopotamus, elephant and white and
black rhinoceros. Within this range they prefer species
that weigh 350 kg (Fig. 4) which is much larger than
the largest recorded weight of lion (260 kg according to
Berry, 1983 in Estes, 1991: 369). This is far heavier than
the 100–200 kg previously hypothesized (Bertram, 1979).
The preferred weight range calculated here is likely to
be an underestimate given that males of most species are
preferentially taken (Schaller, 1972; Rudnai, 1974; Mills,
1984; Prins & Iason, 1989), and we have conservatively
0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8
Log10(body mass)
-1.0
-0.8
-0.6
-0.4
-0.2
0.0
0.2
0.4
0.6
Jacobs’ index
Fig. 4. Linear relationship between lion Panthera leo prey preference
(mean Jacobs’ index) and the log10 of prey body mass
excluding elephant, rhinoceros and hippopotamus (r 2 =0.823,
P <0.001, y =1.57 + 0.75x).
0.2 0.3 0.4 0.5 0.6 0.7 0.8
Log10(herd size category)
-1.0
-0.8
-0.6
-0.4
-0.2
0.0
0.2
0.4
Jacobs’ index
Fig. 5. Linear relationship between lion Panthera leo prey preference
(mean Jacobs’ index) and prey herd size (r 2 =0.186,
P <0.017, y= −1.15 + 1.39x).
used Schaller’s (1972) figure of three-quarters of the
female body mass in our analyses.
We had hypothesized that lion prey preferences would
follow a normal distribution when plotted against prey
body mass based on optimal foraging theory (Pulliam,
1974).We expected some species to be too small for lions
to obtain enough energy from hunting to be sustainable
(Rosenzweig, 1966; J. P. Elliott et al., 1977; Earle, 1987)
and others to be too large to easily and safely be taken
(J. P. Elliott et al., 1977). Generalized capture success
curves reflect this distribution with peak efficiency
between inefficient small and excessively large prey sizes
(D. S. Wilson, 1975). This is also supported by long-term
data from the Serengeti where carnivores are inefficient at
catching prey outside their preferred size range (Sinclair
et al., 2003). In essence, our data support two predictions
of optimal foraging theory (J. R. Krebs, 1978: 29); namely
that lions prefer more profitable prey (Fig. 4) and largely
318 M. W. HAYWARD AND G. I. H. KERLEY
ignore unprofitable prey which are outside the optimal
set (Fig. 3) regardless of how common they are. That
they do take such prey reflects their opportunistic hunting
behaviour (Schaller, 1972).
The body mass distribution of lion prey species shows a
skew to the right which we suggest stems from the group
hunting strategy of lionswhich enables exceptionally large
prey items to be taken (see review by Gittleman, 1989).
Other predators that hunt as groups, such as African
wild dog Lycaon pictus, spotted hyena Crocuta crocuta,
Indian dhole Cyon alpinus, bush dogs Speothos venaticus,
and grey wolf Canis lupus, are also likely to exhibit
a skewed preference toward larger bodied prey in plots
similar to Fig. 4, as group hunting predators are able to
kill larger prey relative to their own size (Kruuk, 1975).
This prediction was partially met by Radloff & du Toit
(2004)who found a significant linear relationship between
predator and prey body mass in the Mala Mala Game
Reserve. There, group hunting species, such as lions and
wild dogs, killed the broadest range of prey, from the
largest to the smallest, with a predator:prey body mass
ratio ranging from 1:1.0 for lionesses to 1:2.1 for lions
and 1:1.2 for wild dogs (Radloff & du Toit, 2004). We
have essentially shown this occurs for lion throughout
their range (Fig. 2), but it is clearly not their preferred
option, which is to kill increasingly large prey (Fig. 4).
We also hypothesize that individually hunting predators,
such as caracal Caracal caracal, leopard Panthera pardus,
tiger P. tigris, jaguar P. onca,mountain lion Felis concolor,
Ethiopian wolf Canis simensis and red fox Vulpes vulpes
would have optimal preference for prey species with body
masses closer to the mean of their own species and with
less of a skew toward large bodied prey. For solitary hunting
cheetah and leopard at Mala Mala, the predator:prey
ratio was less than 1:1 (Radloff & du Toit, 2004).
While body mass in predators is governed by the frequency
distribution of the size of available prey (McNab,
1971), diverse communities of prey require predators
to have evolved competition avoidance or minimization
mechanisms. These may be morphological, through the
segregation of prey by body mass; ecological, through
habitat segregation; or behavioural, via the use of refuges
(e.g. Durant, 1998). The linear relationship between
Jacobs’ index and prey body mass with exclusion of
the megaherbivores (Fig. 4) suggests that lions increasingly
prefer larger prey species, irrespective of other ecological
or behavioural features of the prey. For example,
there seems to be no preference for water dependent prey
species given that buffalo, giraffe and wildebeest arewater
dependent and preferred and gemsbok is water independent
and also preferred (Table 5). Avoided species can
also be water dependent (impala and kob) or independent
(springbok) (Table 5). The nominally water independent
larger species (roan, sable and eland) are all taken in
accordance with their availability (Table 5). This could not
be analysed separately because the exact nature of water
dependency is not known for most species and variation
exists within species (e.g. there are water dependent and
independent populations of wildebeest in the Kalahari;
Mills & Retief, 1984).
Table 5. Comparison between species preferred, avoided and killed
in accordance with their abundance separated by their water
dependency (if recorded in Estes 1991)
Water dependent Water independent
Preferred prey Buffalo Gemsbok
Giraffe
Warthog
Wildebeest
Zebra
Taken in accordance Bushpig Eland
with availability Hippopotamus Hartebeest
Kob Kudu
Nyala Oribi
Topi/tsessebe Roan
Waterbuck Sable
Springbok
Avoided prey Black rhinoceros Ostrich
Bushbuck Steenbok
Grant’s gazelle Thomson’s gazelle
Impala
Klipspringer
Reedbuck sp.
White rhinoceros
The preference of lions for specific prey species is also
interesting. Blue wildebeest, with males weighing 250 kg
(Stuart & Stuart, 2000), are within the lion’s preferred
body mass range, occur in large herds, occur in habitat
used by lion, and are unlikely to cause injury to hunting
lions. Their maximum speed is very similar to that of lions
and much slower than other potential prey species, and
their ability to detect predators is much lower than zebra
and gazelles (J. P. Elliott et al., 1977). Consequently, they
are one of the most preferred prey species. The degree
of preference is limited by the synchronous breeding
exhibited by wildebeest that minimizes predation on
calves (Estes, 1976), although they are belowthe preferred
weight range, and by the migratory nature of various populations
(see Estes, 1976; Mills & Retief, 1984). Mills &
Shenk (1992) concluded that the sedentary behaviour of
wildebeest in Kruger meant lion predation impacted on
their population more severely than on the semi-migratory
zebra.
Buffalo are amuch riskier species to hunt (e.g.Mangani,
1962; Mitchell et al., 1965; Packer, 1986) and have a
slightly faster maximum speed than lions (Prins & Iason,
1989); however, the returns in energy investment make
the risks rewarding, such that some lions hunt buffalo
almost exclusively (Makacha & Schaller, 1969; Funston,
Mills, Biggs et al., 1998). Large buffalo herds are easy
to detect through their noise and smell, and this may
increase the preference of lion for buffalo. Lion in the
Serengeti rely on buffalo while the migratory species
are absent (Scheel & Packer, 1995) and buffalo seem to
be much more susceptible to predation during droughts
(Mills et al., 1995). Studies where availability data were
not included suggest that buffalo are also preferentially
preyed upon (e.g.McBride, 1984; Scheel&Packer, 1995).
Lion prey preferences 319
The inclusion of this data is likely to positively increase
the Jacobs’ index value of this species.
Gemsbok occur in arid, prey-impoverished environments,
such as the Kalahari and Etosha, where they are
independent of water. With so little other medium-sized
to large prey available, gemsbok are highly preferred as
prey by lion. Despite their formidable horns, gemsbok are
within the preferred weight range, occur in herds and in a
habitat where lions are relatively abundant.
There have been suggestions that the stripes of zebra
evolved to reduce the risk of predation (see review by
Ruxton, 2002). Kingdon (1984) discounted this hypothesis
because he contended that zebra were killed in proportion
to their availability. Our data support this and
zebra should be preferentially preyed upon by lion based
on their size, grouping strategy and habitat choice, while
tempered slightly by their active predator defence, leading
to a 28.6% hunt failure, and their relatively high detection
distance (J. P. Elliott et al., 1977). Lions have 6 s when
attacking zebra before they start being outrun as the zebra
reaches top speed (J. P. Elliott et al., 1977). In Etosha
and the Kruger, lions preferentially prey upon zebra foals
(Stander, 1992; Mills & Shenk, 1992). Once wildebeest
and zebra were encountered there was no difference in
the likelihood of either becoming prey. However, lions
encountered wildebeest more frequently than zebra during
continuous follows (0.71 wildebeest encounters per zebra
encounter compared to 0.51 wildebeest per zebra in the
population; Mills & Shenk, 1992; Tables 1 & 4). This
greater than expected encounter rate, coupled with a better
ability to determine the outcome of hunts (measured
by proportion of encounters that did not lead to hunts;
Mills & Shenk, 1992), may further assist in explaining
why wildebeest are more preferred than zebra despite
differences in body mass predicting otherwise.
Giraffe are at the upper end of the preferred weight
range and are preyed upon more frequently than expected
based on their availability. Between 50% and 75% of
calves are preyed upon in their first few months (Estes,
1999). The height of giraffe, and the associated increased
predator detection capability, and the threat of injury to
predators from its hooves, which are actively used in
defending adults and offspring, would be expected to
minimize the preference. This is not the case.
Klipspringer, grysbok, steenbok, common duiker,
impala, reedbuck, Grant’s gazelle, ostrich, bushbuck and
Thomson’s gazelle are all less than half of the preferred
body mass of lion prey and are significantly avoided. This
is not to say that these species are unimportant items in the
diet of lion as many of them are commonly taken (Table 3,
Fig. 2).Rather they are killed less frequently than expected
based on their availability. Whether this is the result of
active avoidance (not attacked when encountered) by lion
or by a reduced encounter rate is unknown, although reanalysis
of data presented in Funston, Mills & Biggs
(2001) suggests the former. Encounters of small prey
(steenbok, duiker, hare and porcupine) led to hunts on
68.7% of occasions (46 hunts of 67 encounters from Table
2; Funston, Mills & Biggs, 2001). For impala, only 61.3%
of 312 encounters led to hunts (Funston, Mills & Biggs,
2001). Medium-sized prey (wildebeest, zebra, kudu and
waterbuck) were encountered on 420 occasions and
hunted on 76.2% of them,while buffalo,were encountered
104 times and hunted on 84.6% of occasions (Funston,
Mills & Biggs, 2001). Separate analysis of wildebeest and
zebra found that 61.2% of 98 encounters led to hunts of
wildebeest while 77.9% of 140 zebra encounters led to
hunts (Mills & Shenk, 1992). This reanalysis illustrates
that larger prey species are preferentially hunted over
smaller species. Furthermore, lions invest more energy
in capturing larger prey (31.6 min per stalk of wildebeest
and zebra) compared to small prey (6.9 min per stalk of
gazelles) (J. P. Elliott et al., 1977).
Yet within the preferred weight range, there exists
species that are not preferred. Optimal foraging theory
suggests that a predator should be able to distinguish
between items of differing profitability and prefer the most
profitable types (J. R. Krebs, 1978). For roan and sable this
lack of preference may be because of their low abundance
throughout their range (Stuart & Stuart, 2000), reducing
their encounters with lions to the point that searching
becomes too energetically costly (Sunquist & Sunquist,
1997), although during altered environmental conditions
they can be substantially taken (Harrington et al., 1999).
The large body mass of eland coupled with their active
defence against predators, their vigilance (through large
herd size) and their weaponry (Estes, 1991), probably
reduce their likelihood of predation, such that they are
killed in accordance with their abundance. Waterbuck
weigh between 250 and 270 kg (195 kg adjusted) which
puts them within the preferred prey weight range and their
requirement for grassland habitat near water (Stuart &
Stuart, 2000) means they are likely to be encountered by
lion frequently. Yet they are only preyed upon by lion in
accordance with their abundance, possibly arising from
taste aversion stemming from its lack of scent glands
that results in a greasy, musky smelling coat (Estes,
1999).
Warthog are well below the preferred weight range of
lion prey and so would be expected to be avoided, yet
they are taken in accordance with their abundance. It is
unlikely that this is an artefact of censusing or counting
kills as the underestimation of the population size of small
species is likely to be counteracted by the undercounting
of the carcasses of these small species which are almost
totally consumed. More likely is that warthogs are common
in habitats where lions are, and are slower and have
less endurance than most savanna ungulates (Estes, 1999).
Confounding this is the use of burrows by warthogs,
although this is by no means a secure refuge (Estes, 1999).
Analyses such as this can only be achieved by using the
results of numerous detailed studies from different parts
of the lions distribution and in different time periods. This
highlights the importance of replicated dietary studies for
other predators which take into account prey availability.
While South Africa and Tanzania lead the way in such
studies on lions (Table 1), it is crucial to increase the
number of studies from other parts of Africa.
The methods used here can be used on all other
large predators to determine their preferred weight range
320 M. W. HAYWARD AND G. I. H. KERLEY
of prey, as well as determining if other characteristics
influence predation patterns. Eventually, the degree of
competition in a community of predators, as occurs in
Africa, Australia, North America, Asia or South America,
can be investigated. Comparison between such predators
may also shed light on the evolution of cooperative
hunting. Additionally, wildlife managers can use this data
to predictwhat lions will eatwhen reintroduced by solving
Jacobs’ index with the value calculated here and the
game count data of the site in question, thereby allowing
them to plan rather than simply respond. Consequently,
the information presented here will allow us to move
beyond descriptive dietary studies to improving our
predictive understanding of the mechanisms underlying
predator/prey interactions.
Acknowledgements
MWH was funded by a University of Port Elizabeth Postdoctoral
fellowship while this manuscript was prepared.
Ben Russell obtained and sent references unobtainable in
Africa to MWH from Australia. Gina Dawson assisted
with field work. Budget Rent-A-Car, C. R. Kennedy Pty
Ltd, Eveready and Continental Tyres provided support for
this project. This manuscript was improved by reviews by
Rob Slotow and an anonymous reviewer.