Evolutionary ecology meets wildlife management: artificial
group augmentation in the re-introduction of endangered
African wild dogs (
Lycaon pictus
)
J. A. Graf
1
, M. Gusset
1,2
, C. Reid
3
, S. Janse van Rensburg
3
, R. Slotow
1
& M. J. Somers
2,4
1 School of Biological and Conservation Sciences, University of KwaZulu-Natal, Durban, South Africa
2 Centre for Wildlife Management, University of Pretoria, Pretoria, South Africa
3 Ezemvelo KZN Wildlife, Hluhluwe-iMfolozi Park, Mtubatuba, South Africa
4 DST–NRF Centre of Excellence for Invasion Biology, University of Pretoria, Pretoria, South Africa
Keywords
group augmentation; adoption; African wild
dog; Lycaon pictus; re-introduction.
Correspondence
Michael J. Somers, Centre for Wildlife
Management, University of Pretoria,
Pretoria 0002, South Africa.
Email: michael.somers@up.ac.za
Received 3 February 2006; accepted
10 May 2006
doi:10.1111/j.1469-1795.2006.00048.x
Abstract
As an alternative to kin selection, group augmentation theory provides a frame-
work for evolutionary mechanisms maintaining cooperative breeding when
individual fitness is positively related to group size. It is expected that a cooperator
group would accept or adopt unrelated foreigners when it is below a critical
threshold size and group members could thus benefit from recruiting additional
helpers. In re-introduction attempts, this would allow for a group to be augmented
artificially before release, which would enhance its chance to establish itself
successfully in the release area. This possibility was tested using endangered
African wild dogs Lycaon pictus studied in Hluhluwe-iMfolozi Park, South Africa.
Here, we report on the first successful artificial integration of an unrelated adult
female with her three male pups into an existing pack. In addition, post-release
monitoring data are presented, including how a yearling male displaced the
dominant male that adopted him as a pup, adding to the controversy over the
evolutionary stability of group augmentation as a route to cooperative breeding.
This study thus demonstrates how theory from evolutionary ecology can be
applied to practical wildlife management, and vice versa.
Introduction
Why individuals assist in the rearing of offspring other than
their own has traditionally been explained with inclusive
fitness benefits accruing to helpers through kin selection (for
critical reviews, see Jennions & Macdonald, 1994; Clutton-
Brock, 2002; Griffin & West, 2002). Kin selection, however,
cannot account for a number of observed phenomena (most
notably, the involvement of unrelated animals in coopera-
tive activities), and thus additional explanations for the
evolution and maintenance of cooperative breeding have
been sought. Group augmentation theory (e.g. Kokko,
Johnstone & Clutton-Brock, 2001) provides a framework
that can be helpful in explaining cases of alloparental care
in which groups benefit directly from a larger size
(e.g. improved vigilance or thermoregulation), irrespective
of relatedness. This has mainly been discussed regarding the
recruitment of young born into a cooperator group to
benefit from delayed reciprocity, but it can easily be ex-
panded to include incidences where unrelated foreigners are
accepted into a group as helpers. A striking example in this
regard is provided by Heinsohn (1991), where white-winged
choughs Corcorax melanorhamphos ‘kidnapped’ and raised
unrelated young from neighbouring groups, presumably to
augment the size of their own group, as group size
is positively related to individual fitness in this obligate
cooperative breeder.
This poses the question of whether group augmentation
theory could be applied to the conservation management of
species that live in cooperator groups and which rely on the
presence of a critical number of helpers. We tested this
possibility using African wild dogs Lycaon pictus, which
owing to their number declining to less than 6000 animals in
the wild (Woodroffe, McNutt & Mills, 2004) are considered
to be under severe threat and are thus classified as ‘endan-
gered’ by the IUCN.
The current status and distribution of wild dogs as well as
the intensity of the different threats to their existence and re-
establishment vary locally across their range in sub-Saharan
Africa (Woodroffe et al., 2004). In the context of South
Africa, this has mainly meant the absence of sufficiently
large conservation areas containing suitable wild dog habi-
tat aside from Kruger National Park. This situation resulted
in the proposal (Mills et al., 1998) and subsequent creation
of an actively managed meta-population comprising a
number of re-introduced sub-populations in several small,
geographically isolated conservation areas (also see Moeh-
renschlager & Somers, 2004). The initial target size of nine
Animal Conservation (2006)
c
2006 The Authors. Journal compilation
c
2006 The Zoological Society of London 1
Animal Conservation. Print ISSN 1367-9430
packs for this meta-population (Mills et al., 1998) was
achieved in half of the allotted 10 years (Lindsey et al.,
2005).
Nevertheless, the sourcing of sufficient numbers of suita-
ble wild dogs has been identified as an obstacle to the South
African re-introduction programme (WAG–SA, 2005) as
well as to potential re-introduction efforts elsewhere in
Africa (Woodroffe & Ginsberg, 1999). This problem is
further compounded by wild dogs being nearly obligate
cooperative breeders, with resultant reliance on helpers for
rearing their young, and a positive relationship between
pack size and successful reproduction as well as pack
and population persistence (Courchamp, Clutton-Brock &
Grenfell, 2000). Following on their proposal of a critical
minimum pack size of five animals, Courchamp & Macdo-
nald (2001) recommended a novel approach for the con-
servation of wild dogs by artificially augmenting the size of
small packs. This could be achieved through the integration
of unrelated animals into an existing pack by means of
adoption of pups and/or immigration of adults (see Gusset,
Slotow & Somers, 2006a).
Adoption of pups by unrelated adults has been recorded
in both free-ranging and temporarily confined wild dogs
(McNutt, 1996a; Hartwig & Rasmussen, 1999), and would
thus seem to be a feasible strategy by which to augment pack
size. Concerning the use of unrelated adults to increase the
size of an existing pack, Courchamp & Macdonald (2001)
stated that ‘wild dogs readily accept unrelated individuals’.
The majority of literature regarding this topic reports the
opposite though (Frame et al., 1979; Burrows, 1995; McNutt,
1996b;Girmanet al., 1997; McCreery & Robbins, 2001; Creel
& Creel, 2002). In established packs, the immigration of
adults unrelated to adults of the same sex generally results in
supplanting same-sex lineage; therefore, there is an under-
standing that only related animals within a sex should be used
for re-introduction attempts (WAG–SA, 2005).
According to group augmentation theory, however, we
expect a cooperator group to accept or adopt unrelated
foreigners when it is below a critical threshold and group
members could thus benefit from recruiting additional
helpers. In re-introduction and translocation attempts, this
would allow for a group to be augmented artificially before
release, which would enhance its chance to establish itself
successfully in the release area. Here, we report on the first
successful artificial integration of an unrelated adult female
with her three male pups into an existing pack. In addition,
we present post-release monitoring data, including how a
yearling male displaced the dominant male that adopted
him as a pup, adding to the controversy over the evolu-
tionary stability of group augmentation as a route to
cooperative breeding. Our study thus demonstrates how
theory from evolutionary ecology can be applied to practical
wildlife management, and vice versa.
Methods
Study area
The study was conducted in Hluhluwe-iMfolozi Park (HiP),
which is located in KwaZulu-Natal Province, South Africa,
just 281 south and around 321 east. The park is c. 900 km
2
in
size and is enclosed by an electrified fence. The predominant
vegetation is Acacia-dominated savannah thornveld, with
forest, grassland and broad-leaved thicket communities also
present (Whateley & Porter, 1983).
Wild dogs were originally widely distributed in KwaZulu-
Natal (Pringle, 1977), with breeding packs being found in
the Zululand region, where HiP is located, still at the
beginning of the 1900s (Vaughan-Kirby, 1916). The last
pack was recorded in Zululand in the 1930s, and after that
only stragglers were encountered (Pringle, 1977). In
1980/1981, 22 wild dogs were re-introduced into HiP by the
then Natal Parks Board (now Ezemvelo KZN Wildlife). This
re-introduction was considered successful based on the survi-
val and breeding of a single pack (Maddock, 1995, 1999; also
see Andreka et al., 1999; Kr
¨
uger, Lawes & Maddock, 1999).
After a population and habitat viability assessment for
wild dogs in southern Africa was conducted in 1997, HiP
became part of the proposed South African meta-popula-
tion (Mills et al ., 1998). Four wild dogs were subsequently
translocated to HiP in 1997 and two animals in 2001
(Somers & Maddock, 1999; Gusset, Graf & Somers,
2006b). The translocation reported on here took place
in 2003.
Study animals
All wild dogs mentioned below were parts of free-ranging
packs unwanted on the properties where they were captured
(Table 1) and subsequently translocated to HiP to
Table 1 Details of the animals involved in the 2003 African wild dog Lycaon pictus translocation to Hluhluwe-iMfolozi Park
ID Age
a
Sex Date of capture Place of capture
F1 Adult Female July 2002 Pragtig Farm, Musina, Limpopo Province
P1 Pup Male July 2002 Pragtig Farm, Musina, Limpopo Province
P2 Pup Male July 2002 Pragtig Farm, Musina, Limpopo Province
P3 Pup Male July 2002 Pragtig Farm, Musina, Limpopo Province
M1 Adult Male November 2002 Crocodile Ranch, Phalaborwa, Limpopo Province
M2 Adult Male November 2002 Crocodile Ranch, Phalaborwa, Limpopo Province
F2 Adult Female November 2002 Crocodile Ranch, Phalaborwa, Limpopo Province
F3 Adult Female November 2002 Crocodile Ranch, Phalaborwa, Limpopo Province
a
Pup, o1 year old; adult, 42 years old.
Animal Conservation (2006)
c
2006 The Authors. Journal compilation
c
2006 The Zoological Society of London2
Artificial group augmentation in African wild dogs J. A. Graf et al.
supplement the existing wild dog sub-population (Gusset
et al., 2006a). After transfer to HiP, wild dogs were kept
under semi-natural conditions in an electrified fenced en-
closure (boma) of c.80 65 m in size. Within this boma was
another triangular fenced enclosure (Fig. 1). Animals were
fed every second to fourth day on a whole or halved, freshly
killed adult impala Aepyceros melampus and water was
provided ad infinitum in both sections of the boma.
On 15 September 2002, a wild-caught adult female (F1)
with her three c. 3-month-old male pups (P1, P2 and P3)
were transferred to HiP (Table 1). F1 was fitted with a radio
collar and released into section B of the enclosure. The three
pups were released into section A on 16 September, but they
broke through the separation fence and joined their mother
in section B on the same day.
On 30 November 2002, four wild-caught adults (M1, M2,
F2 and F3) arrived in HiP (Table 1). Relatedness among
these animals was unknown, but from behavioural observa-
tions on spatial relationships and social interactions they
appeared to constitute a cohesive pack. M2 and F3 were
fitted with radio collars and the four wild dogs were released
into section A of the enclosure.
We took advantage of this opportunity to test the predic-
tion that, according to group augmentation theory, an
existing pack below threshold size should integrate unre-
lated foreigners, with subsequent benefits of an augmented
pack size upon release.
Data collection and analysis
While in the boma, animals were observed from a vehicle or
from behind a shade-cloth screen at a distance of 5–50 m.
Individual wild dogs were identified by their unique coat
patterns. Detailed behavioural data were collected, using
ad libitum sampling (Altmann, 1974), from 3 March to
2 April 2003. This was carried out from c. 30 min before
sunrise for 2–3 h (mean =2.60 h during 13 observation
sessions) and again in the afternoon for 2–3 h (mean=
2.37 h during 13 observation sessions) until last light
(total= 64.60 h during 26 observation sessions), in order to
cover the wild dogs’ crepuscular activity pattern. Occasional
behavioural observations were made on an additional
16 days during the boma period.
A female dominance hierarchy was determined through
the construction of a matrix based on the direction of dyadic
interactions involving aggressive and submissive behaviours
in all social contexts (De Villiers et al., 1997).
After release, the pack was located using standard radio-
tracking techniques and behavioural data were collected
opportunistically upon sighting of the pack.
Two criteria for integration success were used: (1) while in
the boma, formation of a cohesive pack by the two initially
separate groups, with an established dominance hierarchy
among the females, and (2) reproduction of the pack upon
release.
Results
Integration process
On the morning of 4 December 2002, it was discovered that
M1 and F2 had broken through the separation fence into
section B of the enclosure during the previous night. One
pup (P2) had several bite wounds, and was sedated and
treated (one lesion on the neck sutured). F1, P1, P2 and P3
were then moved into section A, and M1, M2, F2 and F3
into section B of the enclosure. During subsequent occa-
sional behavioural observations made of the animals inter-
acting across the separation fence, all three pups acted
submissively towards the four adults in the opposite section.
Around 10 February 2003, a resident pack of eight wild
dogs discovered the boma. This pack was subsequently seen
at the boma, or located in its vicinity, on a total of 12 days
before release (on 2 April 2003; see below).
During the night of 23/24 February 2003, F1 broke
through the separation fence, joining M1, M2, F2 and F3
in section B. All three females were observed feeding
together on the same carcass with M1 and M2 during the
afternoon of 24 February. On 26 February, F1 broke back
into section A, re-joining her three pups. As the separation
fence clearly could not prevent the animals from uninten-
tional break-throughs, it was opened on 3 March. Within
10 min, F1, P1, P2 and P3 moved over into section B.
From removal of the separation fence until release of the
pack, a stable linear female dominance hierarchy emerged,
with F3 ranking highest and F1 lowest (Fig. 2). During the
same period, no aggressive interactions among these females
resulted in injury or escalated fighting. The two initially
separate groups formed a cohesive pack and the first
criterion for successful integration was thus met.
Post-release monitoring
On 2 April 2003, the pack was released from the boma. At
the end of this month, both F2 and F3 appeared to be
pregnant and thus must have conceived during the boma
period in March, when both females came into oestrus and
~80 m
~65 mSection BSection A
Figure 1 Layout of the pre-release holding enclosure used in the 2003
African wild dog Lycaon pictus translocation to Hluhluwe-iMfolozi Park.
Animal Conservation (2006)
c
2006 The Authors. Journal compilation
c
2006 The Zoological Society of London 3
Artificial group augmentation in African wild dogsJ. A. Graf et al.
mating attempts were observed. By mid-May, both M1 and
F2 had disappeared. As neither individual was radio-
collared, the cause of their disappearance could not be
established. F3 lost her pups sometime during the second
or third week of June. It was not clear whether F3 had
whelped or the pack had started denning at all; a den site
was never discovered.
One year after release, in April 2004, P3 was found to
have displaced the previously dominant male M2. It is
however suspected that the change in the dominant position
from M2 to P3 already occurred during the breeding season
in February/March 2004. M2 stayed in the pack after
displacement. Subsequently the pack denned and F3 pro-
duced a litter in May. Four out of nine pups survived until
the end of the year and reproduction upon release as the
second criterion for integration success was thus also met.
Finally, during November 2004, F1 dispersed and joined
a dispersing group of two resident males to form a new pack,
whereas the remaining four animals (M2, F3, P1 and P3; P2
had disappeared for unknown reasons during September
2004) and their pups continued to constitute the introduced
pack.
Discussion
The evolutionary stability of group augmentation as a route
to cooperative breeding has been disputed, despite the
modelling efforts of Kokko et al. (2001). Our findings add
to this controversy. In free-ranging wild dogs, the immigra-
tion of adults related to their same-sex adults into an
established pack, with resulting augmented pack size, has
been documented (Frame et al., 1979; Burrows, 1995; Gir-
man et al., 1997; McCreery & Robbins, 2001; Creel & Creel,
2002). In contrast, the immigration of unrelated adults,
resulting in a cohesive and successfully reproducing pack,
has been observed only rarely (Burrows, 1995; McNutt,
1996a; Creel & Creel, 2002). This behavioural trend of
within-sex intolerance among unrelated wild dogs seems to
be further underscored by the tendency for aggression to be
particularly focused within a sex during inter-pack clashes
(Creel & Creel, 2002).
Contrary to previous management recommendations
(WAG–SA, 2005) but in line with group augmentation
theory, the pack accepted F1 and her three pups into the
group. Presumably, the pack allowed the four unrelated
animals to join as group size was below the proposed critical
threshold of five animals, and the augmented pack thus
potentially had benefits in terms of improved foraging
efficiency and individual survival (Courchamp & Mac-
donald, 2001). Illustrating this, other research has shown
that inter-pack clashes may result in injuries or even fatal-
ities, with smaller packs generally losing during such con-
tests (Creel & Creel, 2002). In this regard, the regular visits
of a resident pack to the boma may have promoted the
integration process. Further in accordance, all five docu-
mented cases of adoption of unrelated pups in free-ranging
wild dogs come from smaller than average packs containing
six or less animals (McNutt, 1996a). F1 and her three pups
behaved submissively and initially appeared to serve the
purpose of group augmentation. F1 also seemed to benefit,
as it would have been unlikely for her to manage raising her
offspring alone.
However, M2 did not gain any reciprocal fitness benefits
from accepting F1 and her three pups, as the first reproduc-
tion attempt after release failed, where the newly accepted
animals could have acted as helpers, thereby increasing the
survival probability of the offspring (Courchamp, Rasmus-
sen & Macdonald, 2002). To the contrary, the adoption of
the pups turned out to be detrimental, as M2 was displaced
from the dominant position by one of the unrelated pups
that he adopted, before gaining a second opportunity for
reproduction. A similar case was reported by McNutt
(1996a). Interestingly, M2 stayed in the pack as a helper
after displacement, most probably without being related
to the newly dominant pair, whereas F1, the mother of the
newly dominant male P3, did not help raise the offspring
of her son, but dispersed instead to avail herself of
the opportunity of reproduction (in a pack below
threshold size).
From a management perspective, applying group aug-
mentation theory to endangered species recovery clearly
paid, as the augmented pack established itself and success-
fully reproduced in the second breeding season after release.
It also paid in that a disperser from the introduced pack
formed a new pack with dispersing animals from a resident
pack. As wild dog re-introductions and translocations are
costly events (Lindsey et al., 2005), any aspects enhancing
the efficiency of such efforts are clearly of importance.
The long acclimatization period before and after the
removal of the separation fence may have allowed the wild
dogs to become sufficiently familiar with each other and to
induce the establishment of social bonds (McCreery, 2000).
Keeping artificially formed packs in a boma for a period of
time before release was found to be crucial for re-introduc-
tion success in wild dogs (Gusset et al., 2006a) as well as
other canids (Moehrenschlager & Somers, 2004). More
detailed and in-depth behavioural studies in this regard
may elucidate the frequency of this phenomenon and its
possible underlying mechanisms. It would be enlightening to
carry out a comparative analysis regarding factors such as
relatedness, immigrant group size and initial pack size,
which potentially govern the outcome of immigration
events. Once this process is better understood, it could
35/37 won F2 23/23 won
F3 + F2
8/8 won (in coalition)
F1
48/49 won
Figure 2 Outcome of social interactions between three adult females
in the 2003 African wild dog Lycaon pictus translocation to Hluhluwe-
iMfolozi Park (whenever F2 displayed dominance over F1, F3 subse-
quently prompted submission from F2; F, female).
Animal Conservation (2006)
c
2006 The Authors. Journal compilation
c
2006 The Zoological Society of London4
Artificial group augmentation in African wild dogs J. A. Graf et al.
possibly be applied in situ, as proposed by Courchamp &
Macdonald (2001), by capturing small packs and augment-
ing them in a boma through the integration of unrelated
translocees, including captive-bred animals (Hartwig &
Rasmussen, 1999; Van Dyk & Slotow, 2003). A related
approach was successfully used in the 2001 wild dog trans-
location to HiP, where two dispersing resident males were
bonded with two translocated females in a boma before
release (Gusset et al., 2006a).
To our knowledge, the reported observations represent
the first attempt to apply group augmentation theory to
conservation. Similar approaches to the management of
other cooperatively breeding species are encouraged. Our
study also demonstrates how findings from conservation
practice can add to our understanding of fundamental
biological processes, in this case the evolutionary mechan-
isms maintaining cooperative breeding.
Acknowledgements
We thank The Green Trust (WWF–SA), The Wildlands
Conservation Trust, The Endangered Wildlife Trust,
DST–NRF Centre of Excellence for Invasion Biology,
Ezemvelo KZN Wildlife, National Research Foundation
and the University of KwaZulu-Natal for funding. We
thank Ezemvelo KZN Wildlife, its staff (especially veteri-
narian Dave Cooper) and the iMfolozi Honorary Officers
for logistical and material support. We thank the members
of the Wild Dog Advisory Group of South Africa (WAG–
SA), especially Markus Hofmeyr, for instructive discus-
sions. We are grateful to two anonymous referees for helpful
comments on this paper.
References
Altmann, J. (1974). Observational study of behavior:
sampling methods. Behaviour 49, 227–267.
Andreka, G., Linn, I.J., Perrin, M.R. & Maddock, A.H.
(1999). Range use by the wild dog in the Hluhluwe-
Umfolozi Park, South Africa. S. Afr. J. Wildl. Res. 29, 1–9.
Burrows, R. (1995). Demographic changes and social con-
sequences in wild dogs, 1964–1992. In Serengeti II:
dynamics, management, and conservation of an ecosystem:
400–420. Sinclair, A.R.E. & Arcese, P. (Eds). Chicago:
University of Chicago Press.
Clutton-Brock, T. (2002). Breeding together: kin selection and
mutualism in cooperative vertebrates. Science 296, 69–72.
Courchamp, F., Clutton-Brock, T. & Grenfell, B. (2000).
Multipack dynamics and the Allee effect in the African
wild dog, Lycaon pictus. Anim. Conserv. 3, 277–285.
Courchamp, F. & Macdonald, D.W. (2001). Crucial impor-
tance of pack size in the African wild dog Lycaon pictus.
Anim. Conserv. 4, 169–174.
Courchamp, F., Rasmussen, G.S.A. & Macdonald, D.W.
(2002). Small pack size imposes a trade-off between hunt-
ing and pup-guarding in the painted hunting dog Lycaon
pictus. Behav. Ecol. 13, 20–27.
Creel, S. & Creel, N.M. (2002). The African wild dog: behavior,
ecology, and conservation. Princeton: Princeton
University Press.
De Villiers, M.S., Van Jaarsveld, A.S., Meltzer, D.G.A. &
Richardson, P.R.K. (1997). Social dynamics and the corti-
sol response to immobilization stress of the African wild
dog, Lycaon pictus. Horm. Behav. 31, 3–14.
Frame, L.H., Malcolm, J.R., Frame, G.W. & Van Lawick, H.
(1979). Social organization of African wild dogs (Lycaon
pictus) on the Serengeti plains, Tanzania 1967–1978.
Z. Tierpsychol. 50, 225–249.
Girman, D.J., Mills, M.G.L., Geffen, E. & Wayne, R.K.
(1997). A molecular genetic analysis of social structure,
dispersal, and interpack relationships of the African wild
dog (Lycaon pictus). Behav. Ecol. Sociobiol. 40, 187–198.
Griffin, A.S. & West, S.A. (2002). Kin selection: fact and
fiction. Trends Ecol. Evol. 17, 15–21.
Gusset, M., Slotow, R. & Somers, M.J. (2006a). Divided we
fail: the importance of social integration for the re-intro-
duction of endangered African wild dogs (Lycaon pictus).
J. Zool. (Lond.) 270, in press.
Gusset, M., Graf, J.A. & Somers, M.J. (2006b). The re-
introduction of endangered wild dogs into Hluhluwe-
iMfolozi Park, South Africa: an update on the first 25
years. Re-introduct. News 25, 31–33.
Hartwig, S. & Rasmussen, G.S.A. (1999). Beobachtungen zur
Integration eines Afrikanischen Wildhundes (Lycaon pic-
tus) aus Zoohaltung in eine Wildfang-Gruppe und deren
gemeinsame Auswilderung. Zool. Garten N. F. 69, 324–334.
Heinsohn, R.G. (1991). Kidnapping and reciprocity in coop-
eratively breeding white-winged choughs. Anim. Behav. 41,
1097–1100.
Jennions, M.D. & Macdonald, D.W. (1994). Cooperative
breeding in mammals. Trends Ecol. Evol. 9, 89–93.
Kokko, H., Johnstone, R.A. & Clutton-Brock, T.H. (2001).
The evolution of cooperative breeding through group
augmentation. Proc. Roy. Soc. Lond. Ser. B 268, 187–196.
Kr
¨
uger, S.C., Lawes, M.J. & Maddock, A.H. (1999). Diet
choice and capture success of wild dog (Lycaon pictus)in
Hluhluwe-Umfolozi Park, South Africa. J. Zool. (Lond.)
248, 543–551.
Lindsey, P.A., Alexander, R., Du Toit, J.T. & Mills, M.G.L.
(2005). The cost efficiency of wild dog conservation
in South Africa. Conserv. Biol. 19, 1205–1214.
Maddock, A. (1995). Wild dogs in Hluhluwe-Umfolozi Park.
Re-introduct. News 11, 16–17.
Maddock, A. (1999). Wild dog demography in Hluhluwe-
Umfolozi Park, South Africa. Conserv. Biol. 13, 412–417.
McCreery, E.K. (2000). Spatial relationships as an indicator
of successful pack formation in free-ranging African wild
dogs. Behaviour 137, 579–590.
McCreery, E.K. & Robbins, R.L. (2001). Proximate explana-
tions for failed pack formation in Lycaon pictus. Behaviour
138, 1467–1479.
Animal Conservation (2006)
c
2006 The Authors. Journal compilation
c
2006 The Zoological Society of London 5
Artificial group augmentation in African wild dogsJ. A. Graf et al.
