You are on page 1of 11

Cooperative Hunting in Lions: The Role of the Individual

Author(s): P. E. Stander
Source: Behavioral Ecology and Sociobiology, Vol. 29, No. 6 (1992), pp. 445-454
Published by: Springer
Stable URL: http://www.jstor.org/stable/4600646
Accessed: 06/08/2010 12:53

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
http://www.jstor.org/action/showPublisher?publisherCode=springer.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.

Springer is collaborating with JSTOR to digitize, preserve and extend access to Behavioral Ecology and
Sociobiology.

http://www.jstor.org
Behav Ecol Sociobiol (1992) 29:445X54 Behavioral Ecology
and Sociobiology
? Springer-Verlag
1992

Cooperative hunting in lions: the role of the individual


P.E. Stander*
Etosha Ecological Institute, Ministryof Wildlife,Conservationand Tourism,P.O. Okaukuejo,Namibia
and Departmentof Zoology, Universityof Cambridge,UK

ReceivedMarch 27, 1991/AcceptedOctober 5, 1991

Summary.Individuallyidentified lions (Pantheraleo) fenceof young and the maintenanceof long termterrito-
wereobservedon the open, semi-aridplainsin Namibia. ries (Packeret al. 1990).
Data from 486 coordinatedgroup hunts were analysed Schaller's (1972) observations of 274 group hunts
to assesscooperationand individualvariationin hunting providedevidencethat not all lions presentin the group
tactics. Group hunts generally involved a formation participatedin grouphunts,lendingsupportto the theo-
whereby some lionesses ("wings") circled prey while retical concept of "cheating" (Maynard Smith 1982;
others ("centres") waited for prey to move towards Packer and Ruttan 1988). Packer and Ruttan (1988),
them. Those lionesses that occupied "wing" stalking using game theoreticalmodels, weighed the costs and
roles frequentlyinitiated an attack on the prey, while benefitsof cooperativehuntingand showedthat cooper-
lionesses in "centre" roles moved relativelysmall dis- ation was not advantageouswhenlions captureda single
tancesand most often capturedpreyin flightfrom other preyanimalper hunt,and that some individualsin larger
lionesses.Each lioness in a given priderepeatedlyoccu- group sizes may "cheat" (exploit the foragingactivities
pied the same position in a hunting formation. Hunts of conspecifics) during cooperative hunts. They also
where most lionesses present occupied their preferred showed that animalsshould be more cooperativewhen
positions had a high probabilityof success. Individual the hunting success of a solitary hunter is low. Scheel
huntingbehaviourwas not inflexible,however,but var- and Packer (1991) presented new data on communal
ied according to different group compositions and to hunting in the Serengeti,revealingthat cooperation is
variationsin the behaviourof other individualspresent. common duringhunts on largerand more difficultprey
The role of cooperativehuntingand its apparentadvan- and that cheating is more common during communal
tages within the semi-aridenvironmentof Etosha Na- huntswherethe successrate of solitariesis high.
tional Park,Namibia,are discussed. Lions in the semi-aridenvironmentof EtoshaNation-
al Park, Namibia, occur at low densities (2.8 lions
100 km-2; Stander1991), and live in fission-fusionso-
cial groups with no social hierarchyamong pride lion-
Introduction esses (Schaller1972;Packerand Pusey 1982, 1985),simi-
lar to those seen in other studies in Africa. Lionesses
Cooperativehunting aids social carnivoresin securing (2 yearsand older)do almost all of the huntingin mixed
prey in open habitats(Wilson 1975)and in three-dimen- groups(Schaller1972;Scheeland Packer1991)although
sional habitats(Norrisand Schilt 1988). Group hunting males may hunt on their own (Schaller1972). Previous
by terrestrialcarnivoressuch as the lion (Pantheraleo) work in Etosha N.P. has shown that lionesses rarely
is thought to reduceproblemsof locating and subduing scavenged, and hunted, on average, in groups of 3.6
large prey (Bertram 1979; Lamprecht 1981; Packer (Stander1992). Most hunts involved coordinatedstalk-
1986),and to improvekill efficiency(Schaller1972;Car- ing by groupsand aredefinedas cooperativehunts,anal-
aco and Wolf 1975). Packer (1986), however, suggests ogous to Scheeland Packer's(1991) "pursuers".Hunts
that group foragingdoes not exhibit conclusiveadvan- by solitarieswereinfrequent(1%) and had a low success
tages for lions and furthermore presents evidence rate (2.3%). Based on Packerand Ruttan's(1988)mod-
(Packeret al. 1990) that foraging success does not ex- els, one would thereforeexpect Etosha lions to be more
plain the groupingpatternsof female lions. They attri- cooperative,and less likely to "cheat" than elsewhere
bute groupingpatternsto other factors, such as the de- (Scheeland Packer1991). In this paper I examineonly
cooperativegroup hunts to considerthe value of group
* Presentaddress:P.O. Box 285, Omaruru,9000, Namibia hunting under semi-aridconditions, and to emphasize
446

individualspecializationas an aspect of lions' coopera- a.


tive huntingstrategy.
A G

Methods BF
Between May 1984 and July 1988 four prides were observed for
52 observation periods, each of which lasted from 1 (24 h) to
15 days. These prides inhabited the short grass plains of Etosha C E
N.P. (Le Roux et al. 1988). Most lions (83%; n=48) were perma- D
nently markedwith an individuallyrecognizablehot brand (Orford
et al. 1988). Brand marks, approximately 7 x 7 cm in size, were
placed either on a foreleg or a hind limb of each lion. With night-
viewing equipment these brandmarkswere visible up to roughly Directionof Initialapproach
50 m. One to three lionesses of each pride were radio-collared.
During standard immobilizations (Van Wyk and Berry 1986;
Stander and Morkel 1991) lions were marked, collared and aged
by the extent of tooth eruption, wear and discolouration (Smuts Scale 100 - 200 m
et al. 1978).Two unmarkedlionesseswere individuallyrecognizable b.
from natural scars. All lions were allocated into age classes de-
scribed by Schaller(1972) and Smuts et al. (1978). Radio-collared
lionesses were located using ground and aerial radio-telemetry.
When analysing hunting behaviour, caution was exercised in
defining a hunt (Mech 1970; Kruuk 1972; Schaller 1972; Mills
1990). Two or more lionesses were considered to be hunting "co-
operatively"when all lionesses stared at prey with an alert posture
and facial expression (Schaller 1972) and some or all individuals
subsequently stalked the prey. Some individuals stalked directly
towards the prey or remained stationary, while others took an ABC DFG
indirect route or even encircled prey. Only hunts that met these
criteria(n = 486) were consideredfor the following analysis. Hunts
by solitaries and those by groups where all individuals stalked
directly at prey (n=233) are similar to Scheel and Packer's(1991) I
Directionof initialapproach
"conformists", and were less successful (14%) than cooperative
hunts (27%), and the latter contributed to 68% of the observed Fig. 1 a, b. Schematicdiagramof seven typical stalking roles occu-
kills (Stander1992). pied by lionesses during cooperative hunts. a The final positions,
Since the lions hunted mostly on the plains, and almost exclu- in relation to the prey and the observer's position, occupied by
sively at night (Stander 1992), they were viewed from a vehicle lionesses in the different stalking roles. Lionesses then either
at distances of 20-100 m with the aid of image intensifyingnight- chargedthe prey or the prey fled. These positions correspondwith
glasses or a 75-W red-filteredspotlight and low-light binoculars, the typical stalking routes illustratedin Fig. 1b, and indicate "left
or with an infra-red sensitive video camera. The study area was wing" (A-B), "centre" (C-E) and "right wing" (F-G) stalk-cate-
subject to high tourist pressure which was advantageous as both gories. b The most typical stalking routes taken by lionesses in
lions and prey were habituated to the presence of vehicles. Care occupyingthe correspondingfinal stalking role positions (a)
was however taken not to disturb the animals or dazzle the prey
when the red-filteredlight was used.
Lions hunted 16 species, of which three (springbokAntidorcas
marsupialis,zebra (Equus burchelliand wildebeest Connochaetes final position relative to the observerand prey position (Fig. 1a).
taurinus)contributed83% of the prey animals killed (n= 156) and Stalking was measured from the start of the hunt until the first
89% of the biomass consumed (n = 6244 kg) (Stander 1992). The lioness chargedthe prey, or the prey fled. The most common stalk-
analysis of cooperative hunts was based on these three species, ing routes taken by lionesses which resulted in the occupation of
plus 28 hunts on gemsbok (Oryx gazella), giraffe (Giraffacamelo- the correspondingpositions are displayedin Fig. 1b. These stalking
pardalis)and ostrich (Struthiocamelus).Togetherthese are referred roles are relative to the position of the prey and the point from
to as "large and fleet-footedprey". which lionesses initially approached the prey, which is also the
observer's position. A lioness, irrespectiveof the hunting group
Definitionsand stalkingroles. During standardobservationsof co- size, could occupy any one of the seven stalking roles in a given
operative hunting, much attention was given to the individuals hunt and more than one lioness could occupy one stalking role.
in the hunting group, with emphasis on their positions and move- Without intending anthropomorphism,I have used three terms
ments in relation to other members of the group and to the prey. derived from common games like soccer and rugby (Anon. 1989)
A "group" was defined as individualsof the same pride who were to describe the positions of lionesses relative to one another and
within 200 m of each other (Packer et al. 1990). "Initiating an to the prey. The term "left wing" correspondsto positions A and
attack" occurred when the first lioness charged at prey during B, "right wing" to F and G, and "centre" to positions C, D
a hunt and "participation" occurredwhen a lioness joined in on and E in Fig. 1. "Stalk categories", hereafter, refer to "wings"
a chase initiated by another lioness. "Ambush" was defined as and "centres".
a crouchingposition from which a lioness could catch prey fleeing Distances stalked by lionesses were difficult to measureaccura-
from another lioness, and a "rush" was when a lioness initiated tely. Rough estimates of stalking distances were made, based on
an attack and captured prey after a chase of up to 150 m. Data measurementson the vehicle's odometer, during a subset (n= 278)
were recorded by tape recorder and/or video. Rough diagrams of the data. These estimates may not be accurate, but exact dis-
were also drawn during and shortly after hunts, indicating the tances stalked by lionesses in the differentroles were not important
movements of individuallions and of the prey. to the presentanalysis.The focus is ratheron the patternof stalking
Each individual lioness participatingin a hunt was classed in in relation to the prey and other lions. When a lion could not
one of seven possible stalking roles (A to G) depending on their individuallybe identified,the stalkingrole it occupiedwas recorded
447

a) No. 4;n= 124 b) No. 27; n =84 c) No. 34; n=281 d) No. 39; n=266
40 ~~~~~~~~~~40 120 150

100
30

20
A B CD EFG
[]
AB CD E FG
60A BOCD EFG AB CD EF G

10H
e)No.40; n =83 f) 20No. 46; n= 108 - -. 1~~~~~~~~0
g) No. 49; n =40
E.0 10

0 30300 ~~~~~~~~0 0

0'~~~~~~~~~3

A BOCDE F G A BOCDE F G A BOCDE F G


Stalking roles
Fig. 2a-g. Stalking roles occupied by seven lionesses of the Okon- occupying stalking roles different from random (G test; df= 42;
deka pride, irrespective of the hunting group size or individuals P<0.001)
present in the group; n=398. Lionesses showed heterogeneity in

anonymously. As with most behavioural studies, data could not Table 1. Differences between the stalking roles occupied by seven
be normalized and non-parametric statistics (Siegel 1956) were lionesses of the Okondekapride
used. All P values are two-tailed.
Identification 4 27 34 39 40 46 49
number of
Results lionesses

By means of direct observation, I recorded 486 coopera- 4 - 0.11 0.64 0.12 0.24 0.44 0.49
NS * NS ** ***
tive hunts (123 hunts were recorded on video tape) in 27 - - 0.72 0.05 0.31 0.51 0.57
which the stalking roles occupied by individual lions in *** NS ** **
four prides were noted. The majority of these hunts were 34 - - - 0.72 0.52 0.53 0.68
on the plains, and the average prey group size was 17 *** *** *** ***

animals (SE=4.06; n=320; range 1-200). All lionesses 39 - - - - 0.37 0.57 0.62
*** *** ***
present in a hunting group participated in every hunt.
40 - - - - - 0.24 0.25
From rough estimates, lionesses that occupied "wing" * NS
stalking roles stalked an average distance of 320 m 46 - - - - - - 0.11
(SD=256; n=271) whereas "centres" stalked on aver- NS
age 30 m (SD= 31; n=278). Nevertheless, all were con- 49 - - - - _ - _
sidered participants as supporting data will demonstrate
(see below). Comparisonsbetween individualsare based on Kolmogorov-Smir-
nov two-sampletests. P values are two-tailed (NS = not significant;
* <0.05; ** <0.01; *** <0.001)

Okondekapride (Fig. 2)

The stalking roles played by seven adult lionesses of and 39 were "centres" while Nos. 40, 46 and 49 were
the Okondeka pride were observed during 398 hunts. viewed as "right wings".
These hunts represented all observations, including dif- To test whether lionesses in these three stalking cate-
ferent sub-groups each of which may have consisted of gories behaved differently from each other, the stalking
a variety of individuals, and occasionally the entire roles of each lioness were compared (Table 1). "Left
pride. The seven lionesses showed distinct individual wing" No. 34 occupied mostly stalking roles A and B,
preferences for particular positions in the hunting con- showing a significant partiality for these stalking roles
figuration (Fig. 2), and behaved differently from each when compared to all other lionesses. "Centres" (Nos.
other. Lioness No. 34 was a "left wing", Nos. 4, 27 4, 27 and 39) occupied similar stalking roles, reflected
448

a) No.45;n=17 b) No. 47;n=14 adult lionesses and the cubs behaved differently from
6 6 each other (Kolmogorov-Smirnov test; P < 0.05).
5 5

A4 4 Ombikapride (Fig. 5)
Fg
0
3 3
Observations on individual positioning of the six lion-
02 2 esses in this pride were limited (Fig. 5), nonetheless, these
E observations do show distinction between lionesses. The
"left wing" position was filled by No. 37, with Nos.
7, 17 and 36 frequenting "centre" positions. No. 33
ABC D E FG A B.CD EF G
tended to favour the "right wing", while No. 13 filled
Stalking roles roles both to the immediate left and right of "centre".
Fig. 3 a, b. Stalkingroles occupied by two lionesses of the Okaukue-
jo pride; n=20
Variationswithin the Okondekapride

by the low Dmaxvalues, and they varied significantly Stalking positions taken up by lionesses of the Okondeka
from the "left wing" and "right wings". Among the pride are presented in terms of different group sizes and/
"right wings" (Nos. 40, 46, 49), stalking roles were the or individuals present in the hunting group. During 145
same except for a statistical difference between Nos. 40 of the observed hunts, only two lionesses (Nos. 34 and
and 46. Variations in stalking roles filled by individual 39) were present. They were each identifiable by brand
lionesses at different group sizes are discussed below. marks, and moreover, one of them had a black radio
collar. When hunting in this group composition, as in
all group configurations for this pride (Fig. 2c, d), lio-
Okaukuejopride (Fig. 3) ness No. 34 occupied "left wing" and No. 39 "centre"
roles (Fig. 6) (D=0.648; P<0.001).
This pride consisted of five lionesses, but due to the When comparing the distribution of roles occupied
difficulty of identifying some individuals, the stalking by the two lionesses (Table 2) it appears that these indi-
roles of only two lionesses were recorded (Fig. 3). Lio- viduals were compensating for variations in each others'
ness No. 45, a "wing ", occupied stalking roles on both stalking roles. When No. 34 played roles on the left
the left and right sides, while No. 47 frequented the side, No. 39 would occupy roles in the centre and to
centre roles (D= 0.47; P<0.05). the right of centre, while on the few occasions that No.
34 stalked on the right side, No. 39 would compensate
by occupying stalking roles to the left of centre. From
Rietfontein pride (Fig. 4) an observer's point of view, it appeared that the two
lionesses attempted to keep the prey on the imaginary
Two adult lionesses (Nos. 611 and 612) and four large straight line between themselves. The advantage of this
female cubs, which formed the nucleus of this pride, behaviour is discussed below.
were all observed hunting in different roles (Fig. 4). Lioness No. 40, a "wing", occupied both left and
None of the large cubs were individually recognizable right hand stalking roles (Fig. 2e). During 89 observa-
and the stalking roles they occupied were combined and tions the stalking roles of No. 40 were compared to
compared with those of the two adult lionesses. No. hunts when the "left wing" No. 34 was present, or not
611 formed the "right wing", No. 612 the "left wing" present, in the hunting group (Fig. 7). When No. 34 was
while the large cubs played centre roles (Fig. 4). Both present occupying the "left wing", No. 40 hunted on

a) No. 611 n =34 b) No. 612; n =29 c) Cubs; n= 123


20 20 40

C 30

0
10 10 20

n 0~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~1
E~ ~~~~~E
0 ~~~~~~~~~~~~~~~0 0
A B C D E F G A B C D E F G A B C D E F G
Stalking roles
Fig. 4 a-c. Stalking roles occupied by two adult lionesses and four large cubs of the Rietfontein pride; n =39. Both lionesses and the
cubs occupied stalking roles distinctlydifferentfrom each other (G= 99.9; df= 12; P <0.001)
449

a) No. 7;n =13 b) No.13;n=9 c) No. 17; n=16


6 4 6

5 3
4

2 2~~~~~~~~~~~~

A B C D E FG A B C D E F G A B C D E F G

d)tNalki;ng1ONo.r37nl14 e) No.36n13 f)
5 ~~~~~~~~~6 8

U,
05~~~~~~~~~~~~~
4
O- 0ofteObkprd;n2
H ~~~~4H
6

0
3 I
~~~~~~~~~~~~~~~~4
0
.0
~~~~~~~~~~~~~2Fl
nIl ~~~~~~~~~~~~~~2 Fig.5a-f Stalkingrolesoccu-
Z
1 1.
by thesixpridelionesses
~~~~~~~~~~~~pied
0T* of the Ombika pride; n=23
A BC D EFG A BC D EFG A BCD EF G
Stalking roles

60

U) 50 Table2. Stalkingroles occupied by lionesses Nos. 34 and 39 during


145 observationsof them hunting together
40
No. 34
30

Stalking A B C D E F G
roles
20 cnA 0 0 0 0 0 0 0
C; B 0 0 0 0 5 4 0
10 ZC 0 0 0 0 5 9 1
D 34 17 0 0 0 3 2
0 - - E 11 35 8 0 0 0 0
A B C D E F G F 2 4 4 0 0 0 0
G 0 0 1 0 0 0 0
Stalking roles
Fig. 6. Stalking roles occupied by two lionesses of the Okondeka Pearson x2 =205.1; P<0.001
pride, Nos. 34 (openbars)and 39 (shadedbars),during 145 observa- 99% Confidence interval - P = 0.0011
tions of this group composition P values based on exact permutationaldistributions

the "right wing" side, but when No. 34 was not present by No. 73 had a non-significant tendency to differ from
No. 40 hunted more often on the left hand side (D = the 4"4center" roles frequented by her mother, No. 4 (D =
0.588; P < 0.01). This suggests that No. 34 was dominant 0.19; P>0.1). The variations in No. 73's stalking roles
in occupying stalking roles A and B while No. 40 altered were independent of the roles occupied by her mother
her stalking behaviour accordingly. (Pearsons Z2=40.61; df=30; P>0.05). During 70 ob-
Not all females based their stalking patterns on the served hunts, No. 73 was part of a group which included
behaviour of other females. This was particularly true her mother, and Nos. 27 and 46. As with No. 4 the
for a young female, No. 73, daughter of the "centre" stalking roles played by No. 73 were not related with
lioness No. 4 (Fig. 2a), who was observed hunting with the positions of either No. 27 or No. 46 (P>0.1).
the Okondeka pride from the age of 18 months (Fig. 8).
Observations on this lioness lasted until the study was Characteristics and success
terminated when she was 30 months old. No. 73 preferred
wing roles, especially " right wing", although she seldom When comparing lionesses that frequently occupied
occupied the outer roles, A and G. The roles occupied ".centre" and "wing" stalk-categories in terms of physi-
450

30

20

(0
20
0

E
z 10
E
10
z

A B C D E F G A B C D E F G

Stalking roles Stalking roles


Fig. 7. Stalkingroles occupied by the Okondekapride lioness, No. Fig. 8. Stalking roles occupied by the young lioness, No. 73 of
40, duringhunts when "left wing" No. 34 was presentin the hunt- the Okondekapride, between the age of 18 and 30 months, during
ing group (open bars), and when No. 34 was not present (shaded 58 observations
bars),regardlessof group composition; n = 89

Table 3. The age, mass, and body mea- Stalking % of hunts Agea Mass Body measurements
surementsof individual lionesses that fre- roles individual (Years) (kg)
quently occupied either "wing" or
Nos. played Shoulder Heart Body
"centre" stalk catagories indicated height girth length
role (cm) (cm) (cm)

"Wings"
34 88 6-9 145 93 105 167
37 79 6 139 91 101 169
40 71 5-8 130 85 96 152
45 88 6 1i8 84 103 157
46 87 7-10 132 91 95 157
49 85 5-8 134 85 96 152
611 91 6 130 95 103 173
612 89 5 117 83 90 150

Mean 130.6 88.4 98.6 159.6


SD 8.9 4.3 4.8 8.2

"Centres"
4 68 8-11 138 94 106 166
27 83 9-12 152 92 111 168
39 83 7-11 147 92 110 164
47 79 5-6 136 90 104 161

Mean 143.3 92 107.8 164.8


SD 6.4 1.4 2.9 2.6

Mann-WhitneyU= 5.5 4.0 9.0 1.0 11.0


n.s. P<0.05 n.s. P<0.01 n.s.
a
The age estimatesfor some individualsare broad becausethey were observedcontinuously
over a period of up to 4 years

cal characteristics(Table3), therewereno statisticaldif- larger heart-girth measurements (P <0.01) than


ferences between the age, shoulder height and body " wings ".
lengthmeasurementsof the two groups.Lionessesreach On the plains of Etosha N.P., lions had a low average
the asymptotes of their growth curves at roughly the capture success of 15% for all species and hunts ob-
age of 5 years (Smuts et al. 1980). "Centre" lionesses served in both the dry and wet seasons (Stander 1992).
did, however,appearsomewhatolder (P = 0.069), heavi- With single lionesses mostly unsuccessful in capturing
er (P <0.05) and largerthan "wings". The formerhave large and fleet-footed prey (2.3%) they are expected to
451

Table 4. The frequencythat "centre" and


"wing" lionesses of the Okondeka pride Attack prey Capture prey
attacked preya; by initiating a chase or Initiate Participate Ambush Rush
participatingin a chase, and captured chase in chase
prey' either by " ambush" or " rush" chase_In_chase
"Centre" lionesses 28 48 29 5
(Nos. 4, 27, 39)
"Wing" lionesses 66 41 12 24
(No. 34, 40, 46, 49)
Fishersexact P=0.0003 P<0.0001
probability test n = 183 n = 70

a
Initiatechase= The first lioness to charge at prey
Participatein chase= Subsequentlionessjoins the chase after it has been initiated
b
" Ambush"= The position from which a lioness capturesprey fleeingfrom anotherlioness
" Rush"= Lioness initiates an attack and capturesprey after a chase of up to 150 m

be cooperativehunters(Packerand Ruttan1988).Coop-
erativehunts had a high successrate of 27%, increasing
with hunting group size (r,=0.821; Stander 1992; 0.4 x
Standerand Albon in press).This cooperationis further 5

reflectedby the methods of prey capture. During suc-


cessful cooperativehunts on the large and fleet-footed
prey, lions captured73% (n=92) of the prey from an
"ambush"position. The lioness in "ambush"captured .C
0.3

the prey by leaping at it or rushing a short distance, 0


c 4
frequently< 10 m (n= 57), but seldom > 10 m (Binomial 30x
test: n=81; x=24; P<0.001). The remaining27% of o_
03
28
the large and fleet-footed prey were pursued and cap- >% x 221
0.2
turedby lionessesduringa "rush'"of 20-150 m.
The rate of injury to lionesses during prey capture
was low. No sign of injuryto the hunter was detected 28
duringthe 63 observedkills on adult springbok.Of the
25 large prey kills observed, the lioness capturingthe 0.1 /
prey was injuredon two (8%) occasions. The wounds 0~~~~~ 39/

were minor and superficialand did not appearto affect


the lioness' behaviour thereafter.On seven occasions 0 14 x
when an adult zebra was able to free itself from the
grip of a lioness, by kicking or falling onto her, no no- 1
0.0 v b , .
ticeableinjuriesto the lionesswere observed. 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
During observationsof attacks on prey animals by
the Okondekapride, the behaviourof individuallion- Preference rating
esses was recorded(Table4). "Wing" lionessesinitiated Fig. 9. A logistic curve representingthe probabilityof hunting suc-
attacks on prey significantlymore often than "centre" cess at differentpreferenceratings. Each hunt was assigneda pref-
lionesses (n=183). "Centre" lionesses participatedin erence rating (p/N); wherep = the numberof lionesses in their pre-
subsequentattacks on prey more often than "wings", ferredhunt-catagories;and N= the total numberof lionesses parti-
cipating in the hunt. Thus, if all the lionesses presentwere in their
by chasing prey after another lioness had initiated the preferredstalk categories the preferencerating would be 1. Since
attack. During 70 observations of prey capture, raw data are binary (either 0 or 1) the success rates for hunts
"centre" lionesses were observedto captureprey from in each 0.1 intervalof the preferenceratingare shown, with sample
an ambush position significantly more often than sizes
"wings (Table4). Lionessesthat occupied"'wing" po-
sitions capturedprey most often during a rush. When
all successful hunts by the Okondeka pride are com- > 0.5 (more than half the lionesses present were in their
bined; "centre" lionesses captured prey slightly more preferred stalk category) had a success rate of 17% (n =
often than "wings" (" centres"= 34; "wings"= 26) 337), whereas hunts with a preference rating <0.5 had
(X2=0.313; n.s.). a lower success rate of 7% (n=61). The binary depen-
To assess whetherhunts were more successfulwhen dent variable (hunting success) was related to the prefer-
lionesses played their preferredroles, their attendance ence rating, and the probability of hunting success calcu-
in these stalkingroleswas ratedon a relativeproportion lated using a generalized linear model (Albon et al.
scale (Fig. 9). Hunts where the preferencerating was 1986). The logistic model suggests that hunts where most
452

lionessespresentoccupiedtheir preferredstalk category ing preyfrom an "ambush" positionwhichis the prima-


("wing" or "centre") had a higherprobabilityof success ry means of capturinglargeand fleet-footedprey. Lion-
(Z= 8.5; df= 1; P <0.01) (Fig. 9). By includinglioness esses that occupy "wing" positions, on the other hand,
group size in the model, a significantincrease in the stalk furtherand often initiatean attackon prey.
deviancewas found (x2= 10.7; df= 1; P<0.01) and pro- According to body measurementsit would appear
vided a significantlybetter fit (x2 =19.2; df= 2; P< that "centres" are of a stockier build. This apparent
0.001). This suggests that group size is an important physical difference,although speculative,could be ex-
variablewhilepreferenceratingremainsa significantfac- plained by behaviouralidiosyncrasiesbetween the two
tor influencinghunting success. Most likely to succeed categories of hunters. Although no accuratemeasures
are hunts in large groups where most individualsare of the distancesstalkedby individuallionesseswerepos-
occupyingtheirpreferredstalkingcategories. sible, it is, however,quite clear that "wings" stalk sub-
During 156 observedprey captures,the lioness that stantiallyfurtherthan "centres". The stockier"centres"
capturedthe prey alwayskilledit. Killingwas the proce- participatein most chasesinitiatedby "wings" and cap-
dure whereby a lioness suffocated the prey (Schaller ture prey more frequentlyfrom the ambush position.
1972). Other lionesses present often startedfeeding be- Their apparentlyheavierbuild may furthermorebe ad-
fore the prey was dead. During only four kills (2.6%) vantageousin negotiatingthe captureof oncominglarge
did a secondlioness assist in the killingprocedure. and fleet-footedprey.
The stalking roles occupied by individuallionesses
appearsensitiveto a numberof variables.Samplesizes
Discussion of hunts by the Okondekapride were sufficientto look
at only a few variations.Firstly,during145 hunts where
Lionessesin all four pridesrepeatedlyoccupiedparticu- the same two lionesses (Nos. 34 and 39) were foraging
lar stalkingroles duringcooperativehunts. The unequal together,the stalkingrole occupiedby one was depen-
participationby individualssuggestsindividualvariabili- dent on the role occupiedby the other. It would appear
ty, exemplifiedby lionessesin the same pride occupying that the two lionessescoordinatedtheir positions so as
differentstalkingroles. Lionessesin this study, showed to keep the prey between them; therefore,keeping the
distinctbehaviouraldifferencesin two majorcategories: thirdpoint (the prey) on an imaginarystraightline con-
"wings" and "centres". Lionesses that occupied the nected between the two lionesses. The importanceof
"centre" stalking roles usually stalked very short dis- this behaviouris illustratedby the followingdescription:
tancesand mostly lay watchingthe preyand the stalking If one lionesscharges,the preycould flee in the direction
activitiesof the "wings". Thesedata show that lionesses of the second lioness,who will then be in an "ambush"
hunt cooperativelyby division of labor with the "f position to capturethe prey. Secondly,the stalkingroles
and right wings" circlingthe prey and "centres" lying occupiedby lioness No. 40, in hunts of all group sizes,
in a position from which they may catch prey fleeing variedsubstantiallybetweenhunts in which No. 34 was
from the "wings". Within the context of this hunting presentor not present in the hunting group. From the
behaviour,I do not presupposeconsciousdecision-mak- limited sampleavailableit appearsthat lionessesadjust
ing by individuals. their tactics in each hunt, in response to the presence,
Cooperativehuntingwith "drivers"and "catchers" absenceand position of other individuals.
has beendescribedby Guggisberg(1962),Schaller(1972) A young Okondekapride lioness was observedfrom
and Griffin(1984)but has not previouslybeensupported the age of 18 months when she was first seen hunting,
with sufficient data. The individuallydistinct hunting and the process by which her hunting skills were ac-
strategiesby lionesses observed on the open plains of quired exemplifiesthe complexity of the development
Etosha N.P. could not be due to constraintsof the ter- of cooperativehuntingstrategies.Ratherthan imitating
rain because of the homogeneity and general lack of the older lionesses, No. 73 occupiedstalkingroles very
cover. Scheel and Packer (1991) found no relation be- differentfromthose of hermother(No. 4) and two other
tween stalkingpatternsand availablecover duringcom- lionesses(Nos. 27 and 46) who werepresentduringmost
munalhunts by lions in the Serengeti. of the observedhunts. The stalkingactivitiesof this lio-
Scheeland Packer(1991)show that a substantialpro- ness, and four large cubs in another pride, appeared
portion of lionesses "refrain" from participation in to be an acquired behaviour (Alcock 1989), through
communal hunts by stalking very short distances and watchingthe huntingactivitiesof all the adult lionesses
not participatingin the final stages of the hunt. This of the pride(Eaton 1970).
behaviour was common in males and observed in fe-
males during hunts on species where the success rate
for single lionesses was high (Scheel and Packer 1991). Cooperative hunting in a semi-arid environment
The Etosha lionesses are expectedto be more coopera-
tive than those in the Serengeti, and less likely to In parts of East Africa, where ungulate and carnivore
"cheat" (Packerand Ruttan 1988). Repeatedobserva- biomassis high (East 1984), the huntingsuccessof soli-
tions of the behaviourof individuallyknown lionesses tary lions ranges between 11 and 29%(Schaller1972;
at Etosha supportthese hypotheses,and show that even Elliot et al. 1977; Van Orsdol 1981). On the open and
though "centres" stalk very short distancesin relation flat terrainof Etosha N.P. where small pockets of prey
to "wings", they are most often responsiblefor captur- occur at a low density (East 1984) the hunting success
453

of solitaries was much lower, suggesting that Etosha during various stages of the field work. The National Geographic
lions shouldmore often hunt in groups(Packerand Rut- Society kindly provided an infra-redvideo camera and Mrs Holly
tan 1988).Cooperativehuntingby divisionof labormay Standeris thankedfor her contributionin recordinghuntingbehav-
iour on video tape. The Cornell Laboratory of Ornithology pro-
have developedwithin the context of the semi-aridcon- vided logistical support during data analysis, and I am especially
ditions of EtoshaN.P. grateful to Mrs Katy Payne and Dr. C. Walcott. Mr. D. Andrews
The apparentadvantagesof grouphuntingat Etosha provided advice on statistical problems. Drs. Craig Packer and
are discussed briefly. During the wet season lionesses D.R. Griffin, Dave Scheel, Karen Laurensonand two anonymous
in all group sizes, including solitary females, acquired refereesprovidedvaluablecriticismson earlierdrafts.
daily food intakeabove the estimatedminimumrequire-
mentsof 5-8.5 kg (Packeret al. 1990).The 8-month-long
dry season, however,proved to be a period of food de-
privation and lionesses acquiredsignificantlyless food References
per day than in the wet season (Stander1992).Solitaries
did not meet the estimateddaily food requirementsand Ablon SD, Mitchell B, Huby BJ, Brown D (1986) Fertilityin female
infrequentlyforagedalone. Lionesseshuntedin groups, red deer (Cervuselaphus):the effects of body composition, age
and reproductivestatus. J Zool (Lond) 209:447-460
wherethe huntingsuccessincreasedwith largernumbers Alcock J (1989) Animal behavior: an evolutionary approach. 4th
of lionesses. Cooperative hunts were more successful edn. SinauerAssociates, Sunderland,MA
than non-coordinatedgroup hunts (Stander1992) and Anonymous (1989) Laws of the game of rugby football. Freedman
hunting success was further improved when lionesses & Rossi, Cape Town, South Africa
hunted in their preferredstalk categories. The advan- BertramBCR (1979) Serengetipredatorsand their social systems.
In: SinclairARE, Norton-GriffithsM (eds) Serengeti:dynam-
tages of cooperativestalking are illustratedby the fact ics of an ecosystem. University of Chicago Press, Chicago, pp
that most prey were capturedfrom an "ambush" posi- 221-248
tion, which requiredcoordinated stalking and pursuit Caraco T, Wolf LL (1975) Ecological determinantsof group sizes
of largeand fleet-footedpreyin an open habitat. of foraginglions. Am Nat 109:343-352
Whileit has been suggestedthat lions hunt in groups East R (1984) Rainfall, soil nutrient status and biomass of large
to subdue and kill large prey (Schaller1972; Bertram Africa savannamammals.Afr J Ecol 22:245-270
1979; Lamprecht1981),evidencefrom the presentstudy Eaton RL (1970) The predatorysequencewith emphasison killing
behavior and its ontogeny, in the cheetah. Z Tierpsychol
reveals that the lioness that captured prey up to the 27:492-504
size of an adult zebra also subduedand killed the prey Elliot JP, Cowan I McT, Holling CS (1977) Prey capture by the
(also see Stander1992).The rateof injuryto lions during Africanlion. Can J Zool 56:1726-1734
prey capture was low. Nonetheless,cooperationunder Griffin DR (1984) Animal thinking. Harvard University Press,
the describedconditions, was essentialin order to cap- Cambridge,MA
GuggisbergCAW (1962) Simba. Bailey Bros & Swinfen,London
ture large and fleet-footed prey, which formed >80% Kruuk H (1972) The spotted hyena. University of Chicago Press,
of their diet. Larger prey such as buffalo (Syncerus Chicago
caffer) and giraffe,however,may requirea cooperative LamprechtJ (1981) The function of social hunting in largerterres-
effort for lions to subdueand kill them (Packer1986). trial carnivores.Mammal Rev 11: 169-179
Group hunting and cooperation,in the Etosha N.P. Le Roux CJG, Grunow JO, Morris JW, Bredenkamp GJ,
context, suggestclearbenefitsto the individuals'feeding ScheepersJC (1988) A classification of the vegetation of the
Ethosha National Park. S Afr J Bot 54:1-10
efficiencyand that cooperativehunts wherelionessesoc- MaynardSmith J (1982) Evolution and the theory of games. Cam-
cupied their preferredpositions were more successful. bridgeUniversityPress, New York
The complex huntingstrategiesinvolvingdivision of la- Mech LD (1970) The wolf: the ecology and behaviourof an endan-
bor in lionesses, described in this paper, may have gered species. The Natural History Press, New York
evolved from the fundamentaladvantageof higherper Mills MGL (1990) Kalahari hyaenas: comparative behavioural
capita food intake derived from cooperative hunting ecology of two species. Unwin Hyman, London
Norris KS, Schilt CR (1988) Cooperativesocieties in three-dimen-
duringthe dry season (Stander1992). sional space: on the originsof aggregations,flocks, and schools,
A controversyexists as to the extent to which non- with special reference to dolphins and fish. Ethol Sociobiol
human animals are capable of conscious decision-mak- 9:149-179
ing (Griffin 1984). In a single example of cooperative Orford HJL, Perrin MR, Berry HH (1988) Contraception,repro-
hunting,Griffin(1984)speculatesthat lionessescoordin- duction and demographyof free-rangingEtosha lions (Panthera
ated their stalkingpatternsby taking conscious advan- leo). J Zool (Lond) 216:717-733
Packer C (1986) The ecology of sociality in felids. In: Rubenstein
tage of the locationsof others.Whetheror not the pres- DI, WranghamRW (eds) Ecologicalaspects of social evolution.
ent study reveals a level of mental awarenessnot pre- PrincetonUniversityPress, Princeton,pp 429-451
viouslydocumentedin non-humananimals,it does indi- Packer C, Pusey AE (1982) Cooperation and competition within
cate that measurablebenefits come from cooperative coalitions of male lions: kin selection or game theory? Nature
huntingin this population. 296:740-742
Packer C, Pusey AE (1985) Asymmetriccontests in social mam-
mals: respect, manipulation and age-specificaspects. In: Har-
vey PH, Slatkin M (eds) Evolution: essays in honour of John
Acknowledgements.The Ministry of Wildlife, Conservation and Maynard Smith. CambridgeUniversity Press, Cambridge,pp
Tourismprovidedlogisticaland financialsupportduringthe study. 173-186
I am grateful to Drs. B. Joubert, H.H. Berry, M. Lindeque, W.C. Packer C, Ruttan L (1988) The evolution of cooperative hunting.
Gasaway, Mr. T.B. Nott and Mr. J.L. Scheepersfor their support Am Nat 132:159-198
454

Packer C, Scheel D, Pusey AE (1990) Why lions form groups: StanderPE (1992) Foragingdynamics of lions in a semi-aridenvi-
food is not enough. Am Nat 136:1-19 ronment.Can J Zool (in press)
SchallerGB (1972) The Serengetilion. Universityof Chicago Press, Stander PE, Albon SD (in press) Hunting success of lions in a
Chicago semi-acidenvironment.Symp Zool Soc Lond, vol 65
Scheel D, Packer C (1991) Group hunting behaviour of lions: a Stander PE, Morkel PvdB (1991) Field immobilization of lions
search for cooperation. Anim Behav 41:697-709 using disassociativeanaestheticsin combinationwith sedatives.
Siegel S (1956) Nonparametric statistics for the behavioural sci- Afr J Ecol 29:137-148
ences. McGraw-Hill,New York Van Orsdol KG (1981) Lion predationin RwenzoriNational Park.
Smuts GL, Anderson JL, Austin JC (1978) Age determinationof PhD thesis, Universityof Cambridge,UK
the African lion (Pantheraleo). J Zool (Lond) 185:115-146 Van Wyk TC, Berry HH (1986) Tolazoline as an antagonist in
Smuts GL, Robinson GA, Whyte IJ (1980) Comparativegrowth free-livinglions immobilizedwith a ketamine-xylazinecombina-
of wild male and female lions (Panthera leo). J Zool (Lond) tion. J S Afr Vet Ass 57:99-104
190:365-373 Wilson EO (1975) Sociobiology: the new synthesis. HarvardUni-
Stander PE (1991) Demography of lions in the Etosha National versity Press, Cambridge,MA
Park. Madoqua 18:1-9

You might also like