Was the kouprey a feral hybrid? A response to Galbreath
et al. (2006)
S. Hedges1, C. P. Groves2, J. W. Duckworth1, E. Meijaard2,3, R. J. Timmins4 & J. A. Burton5
1Wildlife Conservation Society – Asia Program, Wildlife Conservation Society, Bronx, NY, USA
2 School of Archaeology & Anthropology, Australian National University, Canberra, Australia
3 The Nature Conservancy, Kalimantan Timur, Indonesia
4Willard Avenue, Madison, WI, USA
5 Veterinary Biomedical Sciences, Royal (Dick) School of Veterinary Studies, University of Edinburgh, Summerhall, Edinburgh, Scotland, UK
Correspondence
S. Hedges, Wildlife Conservation Society – Asia Program, Wildlife Conservation Society, 2300 Southern Boulevard, Bronx, NY 10460, USA.
Email: shedges@wcs.org
doi:10.1111/j.1469-7998.2006.00293.x
In a recent paper, Galbreath, Mordacq & Weiler (2006)
argue that the kouprey Bos sauveli – currently considered to
be a valid if possibly extinct species of wild cattle – is or was
most likely a feral hybrid form with banteng Bos javanicus
and zebu Bos indicus ancestry.
Here, we examine that conclusion and argue that it is
premature and thus unhelpful, from the conservation point
of view at least, to suggest that the kouprey is a feral hybrid,
which would downgrade its status from that of a critically
endangered (albeit possibly extinct) animal of high conservation
importance (Hedges & Duckworth, 2000) to an entity
of questionable conservation significance. While these genetic
results are extremely interesting and warrant swift
follow-up with further samples and analyses, there is a
danger, if taken on their own, of the premature dismissal of
the kouprey from conservation considerations.
The kouprey did not come to the attention of Western
science until the 1930s and it was not until 1937 that a
specimen (a young male) became available for scientific
study. The first formal description was that of Urbain
(1937), who named it Bos (Bibos) sauveli. This animal was
subsequently designated as the holotype (Urbain, 1939), and
is the single specimen of kouprey included in the analysis of
Galbreath et al. In 1940, Coolidge published a description of
an adult male specimen, which had been shot in Cambodia
in March 1939 and was subsequently added to the collection
of the Museum of Comparative Zoology at Harvard in the
United States. Interestingly, given the present debate, Coolidge
was of the opinion that the kouprey was sufficiently
distinct to warrant the creation of a new genus, which he
called Novibos. The validity of Novibos has, however, been
challenged and most subsequent taxonomists have placed
kouprey in the genus Bos (Bohlken, 1961).
The kouprey has always been enigmatic and the theory
that it may have had a hybrid origin advanced by Galbreath
et al. is not new (as they acknowledge). Edmond-Blanc
(1947) suggested that it was a hybrid between banteng and
(1) gaur Bos gaurus, (2) water buffalo Bubalus arnee or (3)
domestic oxen. Gray (1972) reported that crosses between
buffalo Bubalus and members of the genus Bos have invariably
proved unsuccessful, whereas crosses between species
within Bos are easily obtained. However, the hypothesis that
the kouprey is a hybrid receives little or no support from
either comparative anatomy or field observations of the
behaviour of kouprey (Coolidge, 1940; Wharton, 1957;
Lekagul & McNeely, 1977). Bohlken (1958) revived the
hybrid hypothesis; he specifically proposed that kouprey
were hybrids between banteng and domestic cattle (zebu).
Later, however, Bohlken, (1961), he revised his opinion,
recognizing it as a valid species and assigning it to the
subgenus Bibos.
On hybridization
Before discussing the arguments of Galbreath et al., we must
be clear about what could be meant by the kouprey being
‘a hybrid’. There are three different meanings of ‘being a
hybrid’, and the kouprey must conform, if it is indeed a
hybrid, to one of these:
(1) It may be a first-generation cross between two species.
Kouprey appear to be homogeneous: they have a certain
predictable appearance and size range, and they change
colour predictably over their lifespans. They do not possess
the variability in morphology expected of F2 hybrids or
backcrosses; in each generation, therefore, there must be
new crossing between the parent species in order to produce
a new set of kouprey. Moreover, given that interspecific
hybrids commonly vary according to which species is the
sire and which the dam, the homogeneity of kouprey would
seem to require that they are always the product of, say,
male banteng mating with female zebu.
(2) The population that is of hybrid origin may be the result
of initial hybridization, followed by repeated one-way backcrossing.
For this to occur, we have to have two species coexisting,
one of them dominant to the other. Males of the
dominant species approach herds of the other species where
they drive off the males and mate with the females. Each
generation this occurs regularly; in each generation, the
proportion of genes of the maternal species halves, so that
in rather few generations the repeatedly backcrossed hybrids
are to all intents and purposes indistinguishable phenotypically
from the paternal species, and of course possess its Y
chromosome DNA, while still carrying the mitochondrial
DNA (mtDNA) of the maternal species. In this way, we
have what in effect is one species with the mtDNA of
another species. A well-studied example of this is a population
of mule deer Odocoileus hemionus in Texas, which bears
the mtDNA of white-tailed deer Odocoileus virginianus;
Cathey, Bickham & Patton, 1998).
(3) It may be a species of hybrid origin. Such species are
known; indeed, within the Bovini, we have the example of
the wisent or European bison Bos bonasus (Verkaar et al.,
2004). Verkaar et al. (2004) proposed initial matings between
a male proto-bison with a herd of female protoaurochsen,
followed by repeated backcrossing of the hybrids
with the male parent species as in hybrid model two above;
an alternative might be where populations of the two species
have become isolated from others of their species, and after
the first generation of one-way crossing there is panmixis. In
the first case (that proposed by Verkaar et al.), little or no
nuclear DNA (nDNA) of the maternal parent species will be
detectable in the species that is of hybrid origin; in the
alternative case, there will be roughly equal amounts of
nDNA descended from both parent species.
None of this is new; indeed, we have understated the
complexity of hybridization (Barton, 2001; Hewitt, 2001;
Bell & Travis, 2005). This complexity requires an appropriately
comprehensive methodological and analytical approach
if one wishes to make strong inferences about the
history of taxa – an approach we shall show is lacking in the
study of kouprey reported by Galbreath et al. (2006). In an
influential review, Ballard & Whitlock (2004) argue that
mtDNA is not a sufficient marker for phylogeographic
studies if the focus of the investigation is the species and
not the organelle. They state, ‘[it] is difficult to gather an
unbiased dataset about the relative impact introgression has
on phylogenies from mtDNA in contrast to nDNA’ and go
on to say, ‘in taxa with some level of hybridization or
migration, there is a non-negligible probability of introgression
of mtDNA from one taxon into another. This can
happen just by chance (because of the low effective size of
mtDNA), by selective pressure (because of local adaptation
of the mitochondria) or by selective introgression following
mutational meltdown in small populations. In each case, the
mtDNA is equally or more sensitive to introgression than
nDNA, and this introgression can occur despite the levels of
gene flow or hybridization between the populations being
very low. Moreover, the introgression of mtDNA can reflect
chance events, either in the form of genetic drift or the
variance of selection. As a result, the rest of the genome may
tell a very different story, and the results from mtDNA may
not reflect the typical history of the taxa involved’.
Critique of Galbreath et al.
Galbreath et al. do not present any genetic evidence for
kouprey having zebu ancestry. Indeed, they note that, ‘[the]
available kouprey haplotype is quite different, and phylogenetically
far removed, from both zebu and Bos taurus
(humpless domestic ox) haplotypes (Hassanin & Ropiquet,
2004)’. Their argument for zebu ancestry apparently relies
on Bohlken’s morphometric analyses, which was presumed
to demonstrate that ‘the kouprey was suspiciously intermediate
between banteng . . . and domestic zebu’. Bohlken
in fact, as noted above, subsequently changed his mind
about the origin of kouprey, recognizing it as a valid species
and assigning it to the subgenus Bibos. No subsequent
published study of kouprey morphology has suggested a
zebu ancestry (Hassanin & Ropiquet, 2004).
We further suggest that, in the light of our earlier
discussion on the nature of hybridization, Galbreath et al.
are overly reliant on mtDNA sequences and that additional
analyses involving Y chromosome DNA and multiple nuclear
markers are needed before strong conclusions can be
drawn about the origin of kouprey. It is essential to include
multiple molecules with distinct evolutionary histories if one
wants to make strong inferences about the history of taxa
(Ballard & Whitlock, 2004).
So what do the results of Galbreath et al. tell us? They
show a close relationship between the mtDNA sequences of
a single kouprey (the holotype) and two probably Cambodian
banteng, and both banteng sequences showed the
unique derived transversion pair that Hassanin & Ropiquet
(2004) considered potentially diagnostic for the kouprey.
Galbreath et al. conclude from this that most likely ‘the
kouprey was not a natural species, but rather a self-perpetuating
feral form with banteng and zebu (and conceivably
B. taurus) ancestry’ (we shall call this hypothesis one, H1).
As we demonstrated above, this conclusion does not in fact
follow from the data; not only were no Y chromosome
DNA or other nDNA examined, but several explanations
are possible for their findings concerning mtDNA, and they
do not specify precisely what form of mixed ancestry
(hybridization) they have in mind. Presumably, they are
implying our meaning (3) of hybridization, above.
Galbreath et al. do, however, acknowledge two alternative
hypotheses to explain their results (we shall call these
H2 and H3), and these are much more specific about the
actual nature of hybridization:
H2: ‘The kouprey could have originated hundreds of
thousands of years ago via massive genetic introgression
involving a banteng population and a zebu-like wild ox
population’. This would seem to differ from their favoured
hypothesis H1 only in that the zebu-like ancestor is envisaged
as a wild species (perhaps Bos namadicus, the presumed
ancestor of the zebu) and in the timing of the
hybridization event.
H3: Kouprey are/were a true species but ‘[as they] became progressively rarer in recent years, there could have been
modest introgression of kouprey mtDNA into banteng
population(s). If so, [Galbreath et al’s two] banteng specimens
could conceivably possess mtDNA that originated
with kouprey females’. Note, however, that it would not be
necessary to suppose that this integration is a recent phenomenon;
it does not depend on the increasing rarity of the
kouprey. It is in fact a special case of our meaning (2) of
hybridization, above.
We suggest that there is a fourth hypothesis:
H4: Kouprey are/were a true species and there was some
level of interbreeding between kouprey and banteng in the
past, so remnant banteng mtDNA is present in some
kouprey (including the holotype studied by Hassanin &
Ropiquet (2004) and by Galbreath et al.) as a result of
incomplete lineage sorting.
Furthermore, animals are only considered to be ‘feral’ if
they or their ancestors were formerly domestic but they
subsequently live(d) independently of humans. The term
‘feral hybrid’ only therefore applies to H1 and arguably only
then if the putative banteng ancestors were domestic animals;
if H2, H3 or H4 reflect the true origin of kouprey then
use of the term ‘feral’ would be incorrect.
Hypothesis number two (H2) suffers the same flaw as the
hypothesis favoured by Galbreath et al. (H1), that is it
merely assumes that kouprey have a zebu ancestry, which
has nowhere been shown to be the case, relying instead
entirely on the original hypothesis of Bohlken (1958), which
he himself later repudiated (Bohlken, 1961). We think H3 is
the most likely explanation. We prefer it over our own H4,
because the mtDNA in question is almost equally different
from gaur and (Javan) banteng, implying that it is indeed
that of a third distinct species, rather than that of Cambodian
banteng, which, in morphology, are barely distinguishable
from Javan banteng (E. Meijaard, P. Grubb, & C. P.
Groves, pers. obs.).
Conclusion
On the basis of the data and analyses presented by Galbreath
et al., it is premature to decide that any (let alone all)
kouprey were hybrids, feral or otherwise. Further analyses
of banteng and kouprey specimens are needed to decide
between the alternative hypotheses discussed above. Large
numbers of kouprey trophies (skulls or partial skulls with
horns) exist in private collections in Thailand and every
effort should be made to obtain samples from these specimens.
Zebu specimens should also be included in the
analyses. Both nuclear and mitochondrial markers, including
Y-chromosome DNA, should be used and the issues of
sample size and the problem of finding the pure-bred
banteng needed for meaningful comparisons across the
species’ range will also need to be addressed. We would also
like to emphasize the need to have independently verifiable
nongenetic-based identifications for all specimens included
in the analyses because there are multiple cases of serious
erroneous genetic conclusions being reached through misidentification
of source individuals, as shown by Ruedas
et al. (2000). Furthermore, zoo animals should not be used
for these analyses as the origin of zoo specimens, particularly
in south-east Asia, will in most cases be open to doubt
as animal traders often fail to keep adequate records of the
source of their specimens, specimens are misidentified/
labelled and hybridization is a known issue especially among
the Bovidae (e.g. Namikawa & Widodo, 1978; Davis &
Read, 1985; Burton, Hedges & Mustari, 2005). The last of
these problems was highlighted, in a study of particular
relevance to the present case, when ‘Bali cattle’-like animals
that appeared phenotypically to be banteng were found to
have haemoglobin profiles that revealed a genetic history
similar to known hybrid cattle specimens (Namikawa &
Widodo, 1978; Davis & Read, 1985).
Galbreath et al. suggest that banteng from outside the
range of the kouprey should be sampled in order to test
hypothesis three (H3) because that hypothesis unambiguously
predicts that banteng from outside the kouprey’s
range will not possess mtDNA phylogenetically related to
that of the kouprey. They suggested south-west Cambodia
as a suitable site but we argue that the uncertainties
surrounding the original range of the kouprey require specimens
of wild banteng from further afield than south-western
Cambodia and from known localities to be used for the
additional tests; specimens from Java should be included as
these will be particularly helpful for the test of H3.
We would also like to suggest that, as a matter of urgency,
those domestic cattle breeds that have been variously
suggested as being either domestic kouprey, or at least as
having some kouprey ancestry, be sought out and carefully
studied genetically. These include the Thanh-hoa cattle
(Wharton, 1957), the peculiar animal studied by Hassanin
et al. (2006) and the various other breeds referred to therein,
and the Stieng ox, of which there is a skull in the Paris
Museum’s Laboratory of Comparative Anatomy, which is a
typical kouprey (C. P. Groves, pers. obs.).
Finally, we believe that one of the key messages arising
from this debate about the origin of the kouprey should be
recognition that the forces that caused the likely extinction
of the kouprey are still extant in much of south-east Asia
and the challenge is therefore to stop other species such as
the banteng suffering the same fate as the kouprey (whatever
it is/was).
Acknowledgements
We thank G. Galbreath, A. Hassanin, J. Walston and
H.Weiler for fruitful discussions.We also thank C. Bradshaw
and an anonymous reviewer for their helpful comments on an
earlier version of this paper.
References
Ballard, J.W.O. & Whitlock, M.C. (2004). The incomplete
natural history of mitochondria. Mol. Ecol. 13, 729–744.
Barton, N.H. (2001). The role of hybridization in evolution.
Mol. Ecol. 10, 551–568.
Bell, M.A. & Travis, M.P. (2005). Hybridization, transgressive
segregation, genetic covariation, and adaptive radiation.
Trends Ecol. Evol. 20, 358–361.
Bohlken, H. (1958). Vergleichende Untersuchungen an Wildrindern
(Tribus Bovini Simpson, 1945). Z. Jahrb. Allg.
Zool. Physiol. 68, 113–202.
Bohlken, H. (1961). Der Kouprey, Bos (Bibos) sauveli Urbain,
1937. Z. S ¨ augetierk. 26, 193–254.
Burton, J., Hedges, S. & Mustari, A.H. (2005). The anoas
(Bubalus depressicornis & B. quarlesi): taxonomic status,
distribution, and conservation needs. Mammal Rev. 35,
25–50.
Cathey, J.C., Bickham, J.W. & Patton, J.C. (1998). Introgressive
hybridization and nonconcordant evolutionary
history of maternal and paternal lineages in North American
deer. Evolution 52, 1224–1229.
Coolidge, H.J. (1940). The Indo-Chinese forest ox or Kouprey.
Mem. Mus. Compar. Zool. Harvard College 54,
421–531.
Davis, S. & Read, B. (1985). The status of the North American
captive herds of the Banteng, Bos javanicus d’Alton.
Zoo Biol. 4, 269–279.
Edmond-Blanc, F. (1947). A contribution to the knowledge of
the Cambodian wild ox or kouproh. J. Mammal. 28,
245–248.
Galbreath, G.J.,Mordacq, J.C. & Weiler, F.H. (2006). Genetically
solving a zoological mystery: was the kouprey (Bos
sauveli) a feral hybrid? J. Zool. (Lond.) 270, 561–564.
Gray, A.P. (1972). Mammalian hybrids: a checklist with
bibliography. Technical communication animal breeding
and genetics, Edinburgh no. 10 (revised), Commonwealth
Agricultural Bureau, Farnham Royal, Bucks, UK.
Hassanin, A. & Ropiquet, A. (2004). Molecular phylogeny of
the tribe Bovini (Bovidae, Bovinae) and the taxonomic
status of the Kouprey, Bos sauveli Urbain, 1937. Mol.
Phylogenet. Evol. 33, 896–907.
Hassanin, A., Ropiquet, A., Cornette, R., Tranier, M.,
Pfeffer, P., Candegabe, P. & Lemaire, M. (2006). Has the
kouprey (Bos sauveli Urbain, 1937) been domesticated in
Cambodia? Comptes Rendues Biologies 329, 124–135.
Hedges, S. & Duckworth, J.W. (2000). Bos sauveli. In IUCN
2006. 2006 IUCN Red List of Threatened Species. IUCN
2006: hwww.iucnredlist.orgi. Downloaded on 11 October
2006.
Hewitt, G.M. (2001). Speciation, hybrid zones and phylogeography–
or seeing genes in space and time. Mol. Ecol. 10,
537–549.
Lekagul, B. & McNeely, J.A. (1977). Mammals of Thailand.
Bangkok, Thailand: Association for the Conservation of
Wildlife.
Namikawa, T. & Widodo, W. (1978). Electrophoretic variation
of hemoglobin and serum albumin in the Indonesian
cattle including Bali cattle (Bos banteng). Japanese J.
Zootechn. Sci. 49, 817–827.
Ruedas, L.A., Salazar-Bravo, J., Dragoo, J.W. & Yates, T.L.
(2000). The importance of being earnest: what, if anything,
constitutes a ‘‘specimen examined?’’. Mol. Phylogenet.
Evol. 17, 129–132.
Urbain, A. (1937). Le Kou-Prey ou Boeuf Sauvage Cambodgien.
Bull. Soc. Zool. Fr. 62, 305–307.
Urbain, A. (1939). Note comple´mentaire sur le Boeuf Sauvage
du Cambodge (Bos (Bibos) sauveli Urbain). Bulletin du
Muse´ um 2e s., t. XI, No 6.
Verkaar, E.L.C., Nijman, I.J., Beeke, M., Handkamp, E. &
Lenstra, J.A. (2004). Maternal and paternal lineages in
cross-breeding bovine species. Has Wisent a hybrid origin?
Mol. Biol. Evol. 21, 1165–1170.
Wharton, C.H. (1957). An ecological study of the Kouprey.
Novibos sauveli (Urbain).Monographs of the Institute of
Science and Technology,Monograph 5,Manila, Philippines.
Source:
arts.anu.edu.au/grovco/Hedges%20et%20al%20Kouprey.pdf
