Content uploaded by Simon Hedges
Author content
All content in this area was uploaded by Simon Hedges on Mar 23, 2018
Content may be subject to copyright.
Molecular Ecology (2005)
14
, 2281–2297 doi: 10.1111/j.1365-294X.2005.02582.x
© 2005 Blackwell Publishing Ltd
Blackwell Publishing, Ltd.
Phylogeography, genetic structure, and diversity in
the dhole (
Cuon alpinus
)
A. IYENGAR,
*†
V. N. BABU,
‡
S. HEDGES,
§
A. B. VENKATARAMAN,
¶
N. MACLEAN
†
and
P. A. MORIN
* **
*
Laboratory for Conservation Genetics, Max Planck Institute for Evolutionary Anthropology, Leipzig, Germany,
†
School of Biological
Sciences, University of Southampton, Bassett Crescent East, Southampton SO16 7PX, United Kingdom,
‡
Centre for Ecological
Sciences, Indian Institute of Science, Bangalore, India,
§
Wildlife Conservation Society — International Programs, 2300 Southern
Boulevard, Bronx, New York, NY 10460, USA,
¶
Asian Elephant Research and Conservation Centre, Centre for Ecological Sciences,
Indian Institute of Science, Bangalore, India
Abstract
The Asiatic wild dog or dhole was once very widely distributed across Asia but now has
a very fragmented range. In this first genetic study of this little-known species, we
obtained information on genetic diversity, phylogeography, and social structure using both
mitochondrial control region sequencing and microsatellite genotyping of noninvasive
faecal samples from wild populations, as well as from museum and captive samples.
A pattern largely consistent with isolation by distance across the Asian mainland was
observed, with no clear subspecies distinctions. However, two major phylogeographical
groupings were found across the mainland, one extending from South, Central, and North
India (south of the Ganges) into Myanmar, and the other extending from India north of the
Ganges into northeastern India, Myanmar, Thailand and the Malaysian Peninsula. We
propose a scenario involving glaciation events that could explain this pattern. The origin
of the dhole populations in Sumatra and Java is enigmatic and requires further study. Very
low levels of genetic diversity were observed among wild dholes from Baluran National
Park in Java, Indonesia, but in contrast, high levels were observed in Mudumalai Wildlife
Sanctuary in South India.
Keywords
:Asia, Asiatic wild dog, Canidae, conservation genetics, microsatellite, mitochondrial DNA
Received 1 December 2004; revision accepted 14 March 2005
Introduction
The dhole (
Cuon alpinus
), also called the Asiatic wild dog,
red dog, red wolf, or whistling dog, was widespread across
North America, Europe and Asia during the Pleistocene.
The species’ range, however, became restricted to Asia after
the late Pleistocene mass extinctions
c
. 12 000–18 000
bp
,
when it became extinct across North America and Europe,
along with several other large species such as mammoths
and dire wolves (Thenius 1954; Dundas 1999). Dholes are
adapted to life in very different environments ranging from
mountainous alpine regions in Russia to open steppes in
Tibet to scrubland and dense forests in South Asia. Unfor-
tunately in recent times, due to habitat fragmentation,
collapse of prey populations, and increased disturbance
and persecution by humans, the range of the dhole has
been much reduced. The dhole’s range was known to have
encompassed most of South, East, and Southeast Asia
(Fig. 1). The present range of the species is imperfectly
known, but dholes are thought to occur in Bhutan, Cam-
bodia, China (including Tibet), India, Indonesia (Sumatra
and Java), Lao PDR, the Malaysian Peninsula, Myanmar,
Thailand, and Vietnam. Dholes are also thought to likely
still remain in Bangladesh and Nepal, and there may be
relic populations in the Central Asian states of the former
USSR, Korea, Mongolia, Pakistan, and Russia (Durbin
et al
. 2004b). The species, given protected status in most
countries only in the 1970s, is at high risk of extinction in
many parts of its range and is listed as endangered in the
Correspondence: Arati Iyengar, Fax: + 44 2380594459; E-mail:
a.iyengar@soton.ac.uk.
**Present address: Protected Resources Division, Southwest Fisheries
Science Center (SWFSC), 8604 La Jolla Shores Drive, La Jolla CA
92037, USA.
2282
A. IYENGAR
ET AL.
© 2005 Blackwell Publishing Ltd,
Molecular Ecology
, 14, 2281–2297
2004 International Union for Conservation of Nature and
Natural Resources (IUCN) Red List of Threatened Species
and the IUCN Canid Action Plan (Durbin
et al
. 2004a, b).
Molecular phylogenetic analyses have placed the dhole
as an early divergent within the wolf-like canid group
consisting of the grey wolf (
Canis lupus
), the coyote (
Canis
latrans
), the Ethiopian wolf (
Canis simensis
), some jackals
(
Canis aureus
,
Canis mesomelas
) and the dhole (Wayne
et al
.
1997). The dhole is the sole species in the genus
Cuon
.
Eleven subspecies are sometimes recognized (see appro-
ximate distribution ranges in Fig. 1). Differences in pelage
length and colour have been noted among these putative
subspecies, with a general increase in pelage length with
latitude and a much darker and brighter pelage in the high-
rainfall areas of Southeast Asia (Pocock 1936; Cohen 1978;
Durbin
et al
. 2004b). Although dholes have been well
studied in terms of their behavioural ecology within India
(e.g. Cohen
et al
. 1978; Johnsingh 1982, 1992; Fox 1984;
Karanth & Sunquist 1995, 2000; Venkataraman
et al
. 1995;
Venkataraman 1998), aside from a few brief surveys
(Johnsingh 1985; Stewart 1993, 1994), little is known about
the status, distribution, and biology of this species in the
rest of its extensive range.
Studies in India have revealed that similar to other social
canids such as grey wolves and painted hunting dogs
(
Lycaon pictus
), dholes are highly social animals with a
rigid structure of fixed dominance hierarchies. They also
exhibit a high level of cooperative behaviour and live in
packs of 3–20 individuals. Circumstantial evidence sug-
gests that packs have a sex ratio biased towards males as
a consequence of higher female dispersal (Johnsingh 1982;
Venkataraman 1998). In contrast to grey wolves from
North America and Europe, and painted hunting dogs
from Africa, for which species extensive genetic studies
have been conducted (e.g. Girman
et al
. 1997; Vilà
et al
.
1999; Girman
et al
. 2001; Lucchini
et al
. 2002, 2004), this is
the first genetic study of the dhole. Using both microsatellite
markers and mitochondrial DNA sequencing, we examine
the extent and phylogenetic patterns of DNA variation
across a number of regions in Asia. Since dholes are highly
elusive carnivores and, moreover, a protected species, we
relied on noninvasive faecal samples from extant wild
populations. Faecal samples are widely used in the field
of conservation genetics for these reasons (e.g. Adams
et al
. 2003). We also obtained samples from dhole skulls
collected over the period 1846 –1939 and held in the Natural
Fig. 1 Map of Asia, showing approximate
distribution ranges of the 11 putative sub-
species of dholes, with sampling locations,
sample numbers, and observed haplotypes.
—samples from extant wild populations,
—museum samples, —captive samples.
Original template map obtained from the
BYU Geography Department (http://www.
geog.byu.edu/maps/outline/outline.html)
GENETIC STRUCTURE AND DIVERSITY IN THE DHOLE
2283
© 2005 Blackwell Publishing Ltd,
Molecular Ecology
, 14, 2281–2297
History Museum in London. Museum samples have proved
to be very useful in studies involving endangered species
with fragmented and restricted distributions, since they
allow a reconstruction of recent evolutionary history across
the historical range of the species (e.g. Girman
et al
. 2001;
Godoy
et al
. 2004).
Materials and methods
Samples
The two main collections of faecal samples were made in
protected areas in East Java (Indonesia) and South India
(Fig. 1). We are confident that these faecal samples
originated from dholes and not domestic dogs based on
the very low chance of encountering domestic dog scats in
these areas, the appearance of the scats, and the congruence
in results obtained with DNA sequencing. Fifty-one fresh
(< 12 h old) faecal samples were collected in 1998 from
Baluran National Park (BNP), an area 250 km
2
situated
in East Java, by placing a small amount into 2-mL tubes
containing 20% dimethyl sulfoxide (DMSO), 0.25
m
ethylenediamene tetra-acetic acid (EDTA), 100 m
m
Tris,
pH 7.5, saturated with NaCl DET buffer — Frantzen
et al
.
1998. A much larger set of 143 samples was collected
during 2001 from Mudumalai Wildlife Sanctuary (MWS)
in South India. Samples were placed into 50-mL tubes
containing
c
. 30 g silica gel (type III indicating, Sigma) with
a small piece of filter paper separating the faecal material
from the silica gel (Wasser
et al
. 1997). The tubes were then
held at ambient temperature for several weeks prior to
being stored at 4
°
C. MWS is an area of 326 km
2
within the
much larger Nilgiri Biosphere Reserve where the home
ranges of five packs of dholes have been estimated by
authors VNB and ABV. Minimum convex polygons (MCPs)
were constructed with the Animal Movement extension
for
arcview
3.2® using GPS locations of pack sightings and
GPS locations of fresh scats attributed to a pack based on
the number of scats found and similarity to the number of
individuals within the pack. Between 30 and 60 location
points were used for the estimation of each home range
(except in the case of the Gamehut pack where only 15 such
points were available). Site fidelity tests and mitigation of
outlier effects were both carried out within
arcview
3.2®.
Although individual dholes can only be identified with
difficulty, pack identification was possible because overlap
between home ranges was not always extensive and pack
sizes were different from one another. Faecal sample
collection was carried out mainly from communal latrines
(Johnsingh 1982) within estimated home ranges (Fig. 2).
Collection was frequently closely preceded by or followed
by actual visual sighting of the pack, allowing us to allocate
samples to packs with a good level of certainty, even when
collections were made in regions of overlap. Repeat
collections were carried out within home ranges in order to
attempt to sample as many individuals as possible (see
Fig. 2). We also obtained a few faecal samples from Phu
Khieo Wildlife Sanctuary (PKWS) in northeastern Thailand
and Taman Negara National Park (TNNP) in Peninsular
Malaysia. All sampling sites are shown in Fig. 1. In addition,
we obtained blood samples from captive dholes from three
zoos within Germany and scrapings of internal tissue from
several dhole skulls held at the Natural History Museum in
London (see Table 3 for details).
DNA extractions
DNA was extracted from the faecal samples in a dedicated
area using the QIAamp® DNA stool minikit (QIAGEN)
according to the manufacturers instructions but with the
following modifications: approximately 100 mg dried
faecal material was separated from the bolus on a sterile
Petri dish, placed in an Eppendorf tube with 1.8 mL ASL
buffer and allowed to incubate at 37
°
C for 12–24 h, and
the final postextraction elution step was carried out for
30 min. A maximum of 15 samples were processed at
one time with 1–2 negative controls. DNA from blood
samples was extracted using a QIAamp® DNA blood
mini kit following manufacturer’s instructions. DNA from
museum samples was extracted in a dedicated ‘ancient
DNA’ lab, by placing the entire material (no more than
c
. 200 mg) into 5 mL of extraction buffer (0.45
m
EDTA
pH 8, 1% sarcosyl, 0.4 mg/mL proteinase K), followed
by the procedure described in Vigilant
et al
. (2001).
Quantification of faecal DNA samples using qPCR
Quantitative polymerase chain reaction (qPCR) was used
to quantify the amount of amplifiable genomic DNA in all
faecal DNA samples, and to determine the number of
replications required when genotyping in order to obtain
high levels of accuracy. PCRs were carried out in dedicated
areas with appropriate negative controls. Heterozygous
genotypes were accepted once confirmed in two separate
amplifications, while in the case of homozygous genotypes,
repeats were carried out between four and seven times (for
template amounts of 26–200 pg/reaction, as described in
Morin
et al
. 2001).
Microsatellite analyses
A set of 13 microsatellite loci was used, consisting of both
dinucleotide (CXX436, CXX466, AHT130, CXX608, CXX250,
CXX279, CXX253, CXX434, CXX374) and tetranucleotide
repeats (2140, 2010, 2001, 2096), all of which have been
isolated in dogs (Ostrander
et al
. 1993, 1995; Holmes
et al
.
1995; Francisco
et al
. 1996). PCRs were carried out in a
15-
µ
L volume containing 2
µ
L of DNA extract from faeces,
2284
A. IYENGAR
ET AL.
© 2005 Blackwell Publishing Ltd,
Molecular Ecology
, 14, 2281–2297
1
×
PCR buffer [ABgene, 75 m
m
Tris-HCl, pH 8.8, 20 m
m
(NH
4
)
2
SO
4
, 0.01% (v/v) Tween 20®], 0.8–3.0 m
m
MgCl
2
,
12
µ
g bovine serum albumin (BSA) (Roche), 200
µ
m
each
dNTP, 200 n
m
each primer and 0.4 U
Taq
DNA polymerase
(ABgene). Amplification conditions consisted of initial
denaturation for 4 min followed by 45 cycles at 94
°
C for
30 s, 50–63
°
C annealing temperature for 30 s, and 72
°
C for
30 s followed by a final extension at 72
°
C for 10 min. The
5
′
end of the forward primer was fluorescently labelled and
the products were separated using capillary electrophoresis
(ABI Prism 310). Alleles were sized relative to an internal
standard (HD400 with Rox label) and scored using
genescan
3.0 and
genotyper
software (Applied Biosystems).
Mitochondrial DNA sequencing
A 650-bp fragment of the 5
′
region of the dhole control
region was initially amplified on DNA from fresh blood
using a set of primers used in wolf (Randi
et al
. 2000). The pro-
duct was then sequenced and dhole-specific primers designed
to amplify 310 bp of the dhole 5
′
control region sequence
(DHDLOOPFOR 5
′
-CTACCATCAACCCCCAAAGC and
DHDLOOPR310 5
′
-GCAAGGATTGATGGTTTCTCG). PCR
amplification of the 310-bp product was carried out in a
30-
µ
L volume containing 2
µ
L template DNA, 1
×
PCR buffer
(ABgene), 2.5 m
m
MgCl
2
, 24
µ
g BSA, 200
µ
m
each dNTP,
200 n
m
each primer and 0.5 U
Taq
DNA polymerase
(ABgene). Amplification conditions were as follows: initial
denaturation for 4 min, followed by 35– 40 cycles of 94
°
C
for 30 s, 57
°
C for 30 s and 72
°
C for 30 s followed by a final
extension at 72
°
C for 10 min. PCR products were electro-
phoresed on a 1% agarose gel and the 310-bp PCR product
cut out and purified using QIAQuick gel purification columns
(QIAGEN). Sequencing was carried out using an ABI3700
or an ABI377 (Applied Biosystems). A grey wolf faecal
sample obtained from Leipzig Zoo was also used for PCR
amplification of the same 310-bp region.
A 321-bp fragment of the mitochondrial cytochrome
b
region was also amplified using universal primers CB1
and CB2 described in Palumbi (1996) using amplification
conditions as described above, except with an annealing
temperature of 50
°
C.
Since the occurrence of several nuclear copies of mito-
chondrial sequences including the control region has been
reported in the canine genome (Ishiguro
et al
. 2002), and
this appears to be a common problem across many taxa
(e.g. Thalmann
et al
. 2004), a number of procedures were
followed in order to exclude all sequences of nuclear origin
from our analyses. First, dhole-specific primers were
designed from a larger mitochondrial product from DNA
obtained from fresh blood. Second, every sequence was
put through
blast
in GenBank to ensure high homology
Fig. 2 Estimated home ranges of five packs of dholes within Mudumalai Wildlife Sanctuary in Tamilnadu, South India, with locations o
f
faecal sample collections. South of the MWS boundary lies the Nilgiri North Reserve Forest. Numbers refer to individual collections. Repea
t
collections were carried out within most packs (10 within Vazhaithotam, 6 within Manraddiar, 4 within Moyar, 4 within Gamehut).
GENETIC STRUCTURE AND DIVERSITY IN THE DHOLE
2285
© 2005 Blackwell Publishing Ltd,
Molecular Ecology
, 14, 2281–2297
to existing canid mitochondrial D-loop sequences and
no homology to existing nuclear copies of mitochondrial
sequences. Finally, every variable haplotype was confirmed
by cloning the PCR product into pDrive vector (PCR cloning
kit, QIAGEN) and sequencing five to six clones in each case
(12 clones each for haplotypes B and C) and comparing the
sequences obtained to the one obtained with direct PCR
sequencing. Although a number of single substitutions
(singletons) were seen in some sequences and attributed
to polymerase error, every substitution seen upon direct
sequencing of the PCR product was also confirmed across
every one of the clones. Only one museum sample, which
was not satisfactorily confirmed, was removed from the
analyses.
Sexing of dholes
Individuals were initially sexed using primers for the
Y-chromosome-based SRY gene (
SRYfor
5
′
-CTCGCGATC-
AAAGGCGCAAGAT and
SRYrev
5
′
-TTCGGCTTCTGT-
AAGCATTTTC, Meyers-Wallen
et al
. 1995) generating a
product of 104 bp, and as an internal PCR control, the
cmyc proto-oncogene (CMYC_E3_F1U1 5′-GCCAGAGG-
AGGAACGAGCT and CMYC_E3_R1U1 5′-GGGCCTTTT-
CATTGTTTTCCA, Morin et al. 2001) generating a product
of 81 bp. Since these gene sequences are highly conserved
across mammals, in order to eliminate the possibility of
obtaining incorrect results from male human handler con-
tamination, we also designed primers within the Canis SRY
sequence in regions of dissimilarity to the human SRY
sequence so as to generate a product of 199 bp (CanisSRYfor
5′-ATGGCTCTAGAGAATCCCCA-3′ and CanisSRYrev
5′-GCAATTTGTGACTTTTCTGTGC-3′) and as the con-
trol, to the dhole cmyc gene (GenBank Accession no.
AF519448) generating a product of 127 bp (Dholecmycfor
5′-GAGGAGGAACGAGCTG-3′ and Dholecmycrev 5′-
TTGGACGGACAGGATGTAC). Both sets of primers were
successfully used for sexing dholes in a volume of 15 µl
containing 2 µl DNA extract, 1× PCR buffer (ABgene),
2.5 mm MgCl2, 12 µg BSA, 200 µM each dNTP, 200 nm each
of the four primers and 0.4 U Taq DNA polymerase (ABgene).
Amplification conditions were as follows: initial denatura-
tion for 4 min, followed by 35– 40 cycles of 94 °C for 30 s,
57 °C for 30 s and 72 °C for 30 s and a final extension at 72 °C
for 10 min. Every PCR was repeated three times in the case of
apparent females to ensure accuracy, and the final results
of the two assays were 100% concordant (data not shown).
Statistical analyses
Microsatellites. Polymorphism within populations, meas-
ured as the total number of alleles, mean number of alleles
per locus (A), mean observed heterozygosity (HO) and
mean expected heterozygosity (HE) was calculated using
genepop (Raymond & Rousset 1995). Departures from
Hardy–Weinberg equilibrium for each locus and
population were assessed using the Markov chain method
(Guo & Thompson 1992) implemented in genepop and
a sequential Bonferroni test was applied to the results.
Deviation from linkage equilibrium for all pairwise locus
combinations was tested using the exact probability test in
genepop.
The probability of identity PID (the probability that two
individuals drawn at random from a population will have
the same genotype at multiple loci) was calculated using
the program api-calc version 1.0 (Ayres & Overall 2004).
This program estimates PID (or PIave, as denoted by these
authors) while incorporating the effects of population sub-
structure and inbreeding and also allowing for different
proportions of close relatives to be tested.
Evidence for the occurrence of genetic bottlenecks was
investigated using bottleneck (Cornuet & Luikart 1996).
Pairwise estimates of the coefficient of relatedness (r) for
all individuals were calculated using both Queller and
Goodnight (rxyQG) (Queller & Goodnight 1989) and Lynch
and Ritland (rxyLR) (Lynch & Ritland 1999) measures within
the program identix using empirical allele frequencies
(Belkhir et al. 2002) and used in comparisons for within and
between pack relatedness levels. The problem of pseudo-
replication exists in using every pairwise r value in such a
manner with an associated artificial increase in the number
of degrees of freedom (Prugnolle & de Meeus 2002). Con-
sequently, in a comparison of a regression of r values with
distance (d) values between packs, we used ranked r values
as described in Knight et al. (1999), with d values allocated
as follows: 0 within a pack, 1 when neighbouring packs, 2
when separated by one intervening pack, 3 when separated
by two intervening packs and a maximum of 4 when
separated by three packs (e.g. Vazhaithotam and Gamehut).
To further evaluate the patterns of relatedness, a Monte
Carlo resampling procedure implemented in identix was
used with 1000 permutations in order to compare the
observed distribution of r to that expected in non-kin-
structured populations.
Mitochondrial DNA phylogenetic analyses. All sequences were
checked by eye, edited, and then aligned using bioedit
version 5.0.9 (Hall 1999). The model of DNA substitution
that best fitted the data was selected using modeltest,
version 3.06 (Posada & Crandall 1998). The model HKY + G
(ti/tv ratio = 28.6459, alpha = 0.0874) was selected by both
the hierarchical likelihood ratio test and the Akaike informa-
tion criterion (AIC).
Phylogenetic relationships were analysed using maximum-
likelihood (ML) and maximum-parsimony (MP) approaches
in pau p, version 4.0b10 (Swofford 2002), using a heuristic
search with the tree-bisection–reconnection (TBR) swap-
ping algorithm. A homologous sequence obtained from a
2286 A. IYENGAR ET AL.
© 2005 Blackwell Publishing Ltd, Molecular Ecology, 14, 2281–2297
grey wolf faecal sample (found to be identical to GenBank
sequence AF098117) was used as the outgroup for the ana-
lyses. Node support was assessed using 1000 bootstrap
replicates. Mean HKY + G and uncorrected P distances
between groups of haplotypes were measured in paup and
mega version 2.1 (Kumar et al. 2001), respectively.
An analysis of molecular variance (amova) was employed
to assess the significance of genetic differentiation between
various possible groupings using the program arlequin
version 2.001 (Schneider et al. 2000). In an amova, where cor-
relations among haplotype distances at various hierarchical
levels are used as F statistic analogues designated Φ statistics,
ΦCT measures the proportion of genetic variation among
groups of populations (Excoffier et al. 1992). Since arlequin
does not contain the HKY model, we used the Tamura–Nei
model + G (alpha = 0.0874) in addition to total pairwise
differences between haplotypes to estimate ΦCT values.
A mismatch distribution of pairwise substitutional
differences among haplotypes and a range of neutrality
statistics which are capable of detecting the genetic traces
of population growth, decline or stability were examined
using dnasp (version 4.0, Rozas et al. 2003). Values for Fu’s
F statistic (FS, which specifically tests for population
growth and detects excesses of low-frequency alleles in an
expanding population) and Fu and Li’s F* and D* statistics
were obtained and compared. Median-joining networks were
estimated using the median-joining method described
by Bandelt et al. (1999) and the software network 3.1.1.1
(http://www.fluxus-engineering.com), assigning equal
weights to all variable sites and with default values for the
epsilon parameter (epsilon = 0).
Results
Microsatellite analyses
Rates of amplification from faecal DNA samples. Only 35 of
the 143 silica-dried faecal samples (c. 25%) from MWS were
found to contain DNA in adequate quantities to allow
accurate genotyping based on a qPCR assay. Of these 35
samples, nine multilocus genotypes were found to be from
the same two individuals (six samples from one and three
from another), necessitating the removal of repeat geno-
types from the analyses. A further seven samples had to
be eliminated from the final analyses due to inconsistent
amplification across all loci. Thus, 21 individual genotypes
were used for all subsequent analyses. In the case of the
BNP samples, 30 of the 51 (c. 60%) DET buffer-preserved
samples were found to contain adequate amounts of DNA
by qPCR. Eleven of these were found to be repeat sampling
of individuals, and two were found to show inconsistent
amplification with resulting missing information. Upon
deleting these individuals, 17 individuals with unique
multilocus genotypes remained, which were used for all
subsequent analyses.
Genetic diversity. The dholes from BNP were found to have
fixed alleles at 8 of the 13 loci (all except CXX466, 2140,
2010, CXX608, CXX253), with extremely low overall allelic
diversity and mean observed and expected heterozygosity
values. The samples from MWS, in contrast, had high allelic
diversity and high observed and expected heterozygosity
levels (Table 1). One locus (CXX608) was fixed in the MWS
samples, so only 12 loci were used in subsequent analyses.
One locus (CXX250) was found to deviate significantly from
Hardy–Weinberg equilibrium in this population after Bon-
ferroni correction for multiple comparisons (P = 0.018).
Since this locus demonstrated a heterozygote deficit which
may be attributed to the existence of null alleles, we
eliminated this locus from subsequent pack structure
analyses. Tests for linkage disequilibrium after Bonferroni
correction did not reveal any that were significant.
Dholes have a history of human persecution, and in India,
dholes were widely shot or poisoned and bounties paid for
carcasses right up to 1972, when they were given legal pro-
tection (Durbin et al. 2004b). Hence the possible existence of
a genetic bottleneck in the MWS population was investigated.
No such evidence was detected using a two-phase muta-
tion (TPM) model (90% stepwise-mutation model) and the
Wilcoxon’s test (P = 0.38, normal L-shaped distribution).
The average probability of identity PIave was calculated
in the MWS population by incorporating the FST value
Location N#loci #AA R #UA HO (± SE) HE (± SE)
MWS 21 13 47 3.81 4.0 37 0.54 (± 0.07) 0.52 (± 0.06)
BNP 17 13 18 1.38 0.54 8 0.14 (± 0.06) 0.15 (± 0.06)
MWS, Mudumalai Wildlife Sanctuary, South India; BNP, Baluran National Park, Java;
N, number of individuals with unique multilocus genotypes; #loci, number of microsatellite
loci typed; #A, total number of alleles across all loci; A, mean allele number per locus; R,
range of allele size expansion in repeat motif number; #UA, number of unique alleles found;
H
O, mean observed heterozygosity; HE, mean expected heterozygosity (SE, standard error).
Table 1 Genetic diversity in dholes from
southern India and Java
GENETIC STRUCTURE AND DIVERSITY IN THE DHOLE 2287
© 2005 Blackwell Publishing Ltd, Molecular Ecology, 14, 2281–2297
(estimated at 0.0744) and allowing for a range of sib
proportions within the population (Ayres & Overall 2004).
Values of 3.6 × 10−8, 8.5 × 10−6, and 3.3 × 10−4 were obtained
for sib proportions of 0, 0.5 and 1, respectively. Thus, the
probability of two individuals within the MWS population
having the same multilocus genotype is extremely small,
since even the very conservative estimate of PIave obtained
upon assuming that every individual was a full sibling
(i.e. the P(ID)sib), was very low, at 3.3 × 10−4. So we can be
confident that only unique individuals were used in
subsequent analyses.
Pack size and structure in MWS. The number of unique multi-
locus genotypes detected among faecal samples from indi-
vidual home ranges and comparison with pack number
estimates from field observations (Table 2), reveals that
despite repeat sampling within home ranges between 2
to 10 times (except Kargudi, with only one sampling), it
was not possible to sample every individual within a pack.
There was disagreement between the field-based estimate
of pack size and the number of unique multilocus geno-
types found in Vazhaithotam, with five unique multilocus
genotypes, but only three dholes observed. In the Gamehut
pack with a large number of individuals (nine) most of which
(eight) have been sampled, a male bias was not evident even
if it was assumed that the unsampled individual was a male.
Pairwise relatedness (r) values for both Queller and
Goodnight (rxyQG) and Lynch and Ritland (rxyLR) estimators
were calculated among individuals within MWS to look
for differences in values within and between packs. Within
pack values were significantly higher than between pack
values in males [mean rxyQG within packs = 0.43 (range =
0.12–0.64), mean rxyQG between packs = −0.09 (range =
−0.58– 0.51), t-test, P = 0.006, d.f. = 5; mean rxyLR within
packs = 0.30 (range = 0.11–0.46), mean rxyLR between packs =
−0.14 (range = −0.49–0.27), t-test, P = 0.004]. A similar
trend was also seen in females [mean rxyQG within packs =
0.22 (range = −0.51 –0.77), mean rxyQG between packs = −0.06
(range = −0.60–0.72), t-test, P = 0.007, d.f. = 24; mean rxyLR
within packs = 0.12 (range = −0.42– 0.74), mean rxyLR between
packs = −0.09 (range = −0.53–0.51), t-test, P = 0.01]. A com-
parison of the regressions of ranked rxyQG and rxyLR values
and corresponding distance values (in pack units) for both
sexes revealed a clear negative trend of mean r values with
distance in both sexes (rxyLR regression shown in Fig. 3a). In
addition, the slopes of the two regression lines were found
to be different (P < 0.05). The slope in the case of females
is steeper than in the case of males, which suggests that
females may be more likely to disperse close to their natal
packs and conversely, males may be dispersing longer
distances. A permutation test using rxyLR values clearly sup-
ported the presence of kin structure within the population,
with 99% of the permuted populations showing mean r
scores higher than the mean observed value (Fig. 3b). The
use of rxyQG values however, did not show significant results
with the permutation test (data not shown). The rxyLR meas-
ure has been found to outperform the rxyQG measure in a
recent study (Russello & Amato 2004).
Mitochondrial DNA diversity and phylogeography. Approxi-
mately 50% and 80% of faecal DNA samples from MWS
and BNP, respectively, was found to amplify the 310-bp
control region product, of which 246 bp were successfully
used for all phylogenetic analyses. Despite this higher
success rate with mitochondrial DNA amplification,
sequences from the 21 and 17 individuals with unique
multilocus genotypes in MWS and BNP, respectively, were
used to avoid repeated sampling. A total of 19 haplotypes
were found among all samples (faecal, museum, and
captive) with a total of 37 polymorphic sites (Table 3). Only
one haplotype was detected among all the BNP samples
(haplotype P), while four haplotypes were detected among
the MWS samples (haplotype A in 18 out of 21 samples and
haplotypes B, C and D in one individual each). These
haplotypes were also found among museum samples from
these regions: haplotype P was detected in a 1935 sample
from East Java; haplotype A was detected in a 1937 sample
from the MWS region, as well as in a 1911 sample from
Central India; and haplotype C was detected in an 1888
museum sample from the Anamalai Hills region in South
India (Fig. 1, Table 3).
Haplotypes were found to be shared between samples
across vast distances on the Asian mainland, e.g. haplotype
E, found in a 1922 museum sample from Assam in north-
eastern India was also found in two individuals from an
extant population within PKWS in Thailand and haplotype
J was found in museum samples from Central India, North
India (south of the Ganges river) and North Myanmar
(Fig. 1, Table 3).
Both maximum-likelihood (ML) and maximum-
parsimony (MP) analyses produced identical trees. One tree
was obtained with ML and a strict consensus of six most
parsimonious trees was constructed. The bootstrap consensus
Table 2 Estimates of pack size and results of molecular sexing
Sample
Pack size estimates
Sex
From field
observations
# unique multilocus
genotypes
BNP, Java 17
MWS, South India 21
Vazhaitotham 3 5 2M, 3F
Kargudi 4 1 1M
Gamehut 9 8 3M, 5F
Manraddiar 12 3 1M, 2F
Moyar 10 4 1M, 3F
2288 A. IYENGAR ET AL.
© 2005 Blackwell Publishing Ltd, Molecular Ecology, 14, 2281–2297
tree (1000 replicates) shows two sister clades with moderate
bootstrap support (Fig. 4). Mean uncorrected P distance
was 3.4% and 1.9% within clades I and II, respectively, and
6.7% between the two clades. All haplotypes from South,
Central and North India (south of the Ganges River) were
found within clade I. A strongly supported grouping of haplo-
types M, P, and R from Sumatra, Java and captive zoo
animals (presumed to belong to the lepturus subspecies),
was placed within a separate clade with weak/moderate
support as part of the larger clade I. Another strong grouping
consisted of haplotypes C, L, O, all from South India. Thus,
high divergence of haplotypes in samples from South
India was evident. Clade II consisted solely of haplotypes
from North India (north of the Ganges), northeastern India,
Myanmar, Thailand and Malaysia.
Pairwise mean uncorrected p and HKY + G distances
between haplotypes and haplotype groups are shown in
Table 4. Haplotype F, seen in a museum sample from north
of the Ganges in North India, was found to be very distinct
from the other North Indian and Central Indian haplo-
types. Haplotype K (grouping in clade II) was found to be
very distinct from the other haplotype detected in Myanmar
(haplotype J which grouped within clade I). The two
clades are shown pictorially in Fig. 5 and a median-joining
network is shown in Fig. 6. Results from an amova show
that significant genetic differentiation of populations is
seen in several cases (Table 5). However, rather than in
the case of a straightforward grouping by geographical
location (groups 3, 5 and 6), high ΦCT values were obtained
when haplotypes from south of the Ganges in India were
grouped separately from haplotypes from north of the
Ganges and from other clade II haplotypes, and also when
haplotypes from Sumatra and Java were grouped separ-
ately from haplotypes from Malaysia and other clade II
haplotypes (groups 7, 8, 9 and 10). The highest ΦCT values
were obtained when all clade I haplotypes were grouped
separately from clade II haplotypes, with haplotypes from
Sumatra, Java and captive samples either contained within
the clade I group, or as a separate third group (groups 12
and 11, respectively).
Fig. 3
(
a
)
Regression of mean ranked rxyLR
values and corresponding mean distance
values in males and females from MWS.
and solid line, females (r2 = 0.33 P = 0.05)
and dotted line, males (r2 = 0.94, P =
0.0003). (b) Results from the permutation
test comparing the observed mean pairwise
rxyLR value to those expected in randomized
unstructured populations.
GENETIC STRUCTURE AND DIVERSITY IN THE DHOLE 2289
© 2005 Blackwell Publishing Ltd, Molecular Ecology, 14, 2281–2297
To attempt further resolution of these results, a 321-bp
fragment of the mitochondrial cytochrome b region was
also amplified from a number of samples. MWS and BNP
faecal samples (n = 5 and 6, respectively), museum samples
from East Java (1935), South India (1937, 1888, 1919, 1935,
1929), Myanmar (1937 — control region haplotype J), and
captive samples from Allwetter Zoo, Germany, were
found to have identical sequences (haplotype 1). However,
a single substitution (C-T transition, resulting in an amino
acid change from arginine to cysteine; haplotype 2) was
seen in the PKWS and TNNP faecal samples (n = 5 and 2,
respectively), and in the museum sample from north of the
Ganges (control region haplotype F) (see Fig. 5 for distribu-
tion pattern). This result supports the finding that samples
grouping in clade II are very distinct from samples in clade
I. The existence of this substitution within the sample from
Myanmar and northeastern India (control region haplo-
types K and E, respectively), which also grouped in clade
II, could not be confirmed since these samples, along with
several other museum samples, were found to amplify
contaminating human cytochrome b sequences.
A mismatch distribution of pairwise differences among
haplotypes did not show a unimodal pattern (generally
interpreted as a signature of an ancient population explosion;
Rogers & Harpending 1992). Instead, it showed a multimodal,
erratic (ragged) pattern even when samples from each
clade were analysed independently (data not shown).
However, Fu’s FS was found to be significant when either
all haplotypes [FS = −10.3 (95% CI −3.35, 4.56)] or just clade
I haplotypes [FS = −7.56 (95% CI −3.22, 5.81)] were analysed,
Table 3 Variable sites within the various mtDNA control region haplotypes. Extant samples are faecal samples from current wild
populations in MWS (South India), PKWS (northeastern Thailand), BNP (East Java) and TNNP (Malaysia). Numbers of samples are listed
in parentheses. Museum samples are listed with museum accession numbers, date of collection and location details. Captive samples fro
m
zoos within Germany are listed with information on origin
Hap
11111
2299901223
6913477053
1111111111
4455555666
2902349046
1111111111
6667777777
7890134567
1111222
7899123
8617613
GenBank
accession
numbers
Sample number, date and location
information
AGCGGGAGCTC GTTACTTTTA TTATCGGCTA GAACGAT AY682699 Extant (MWS, S. India) (18), ZD1937.1.10.24
(1937, Masinagudi, S. India), ZD1934.10.4.3a
(1911, Sipna valley, Berar, C. India)
B.T........ ..CG...... ......ATC. .G..A.. AY682700 Extant (MWS, S. India) (1)
C..A.....C. ..CGT....G ...C..AT.. .G..A.. AY682701 Extant (MWS, S. India) (1), ZD1888.2.5.22
(1888, Anamalai Hills, S. India)
D.......... .......... .......... .G..... AY682702 Extant (MWS, S. India) (1)
E.........T T.C.....C. C.CC.AATCG AG...G. AY682703 Extant (PKWS, Thailand) (2), ZD1922.12.22.4
(1922, Naga Hills, Assam, E. India)
F.........T T.C....CC. C.CC.AA.CG AG...G. AY682704 ZD1929.10.19.5 (1909, Oudh, U.P., N. India)
G.........T T.C.....CG C.CC.AATC. AG...G. AY682705 Extant (PKWS, Thailand) (3)
H ........C. ..CG...... ......ATC. .G..A.. AY682706 ZD1919.6.2.23 (1919, Kotagiri, Nilgiri Hills, S. India),
I.......... ..CG.....G ......ATC. .G..A.. AY682707 ZD1935.1.1.3 (1935, B.R. Hills, Coimbatore, S. India),
ZD1930.5.24.125 (1929, Salem, E. Ghats, S. India)
J
.......... ......C... .......... ....... AY682708 ZE1952.4.7.17 (1923, Nagpur, C. India),
ZD1937.12.3.33 (1937, Upper Burma),
ZD1907.10.18.3 (1907, Mirzapur, U.P., N. India)
K....AG.T.T TCC.....C. C.CC.AATC. AG...GC AY682709 ZD1937.12.3.34 (1937, Katha Dist., Upper Burma)
L..A...A.C. ..CGT....G ...C..AT.. .G..A.. AY682710 ZD1934.8.3.2 (1934, B.R. Hills, Coimbatore, S. India)
M...AA..... ..C..CC... C...T..... ...TA.. AY682711 ZD1845.3.19.5 (1839, Sumatra)
N.......... ..CG...... ......ATC. .G..A.. AY682712 ZD1934.10.4.2 (1912, Hoshangabad, C. India)
O..A.....C. ..C.T....G ...C.AAT.. .G..A.. AY682713 ZD1911.5.31.4 (1910, Kanara, S. India)
P.......... ..C..CC... C...T..... ...TA.. AY682714 Extant (BNP, E. Java) (17), ZD1935.2.26.1
(1935, Modjokerto E. Java)
QA......... ..C....... .......... ..G.A.. AY682715 Captive (Dresden zoo, Germany, origin Winnipeg
zoo, Canada, wild origin presumed to be China) (3)
R ..A....... ..C..CC... .C..T..... ...TA.. AY682716 Captive (Allwetter & Hodenhagen zoos, Germany
origin Novosibirsk zoo, Russia & European zoos,
wild origin presumed to be China) (4)
S........CT T.C.....C. C.CCTAATC. AG...G. AY682717 Extant (TNNP, Malaysia) (2), ZD 1846.5.13.2
(1846, Malacca)
2290 A. IYENGAR ET AL.
© 2005 Blackwell Publishing Ltd, Molecular Ecology, 14, 2281–2297
but not when just clade II haplotypes were analysed [FS =
−1.48 (95% CI −3.64, 4.85)]. A significantly large negative
value of Fu’s FS rejects population stasis/neutrality, indi-
cating an excess of recent mutations and thus population
expansion and/or selection. Range expansion can be dis-
tinguished from the effects of selection by the patterns of
significance of FS, F* and D* (e.g. Peck & Congdon 2004).
Thus a range expansion is indicated when FS is significant
and F* and D* are not, while the reverse suggests selection.
Nonsignificant results were obtained for Fu and Li’s F* and
D* in all cases.
Discussion
Genetic diversity
The single mitochondrial haplotype and extremely
low microsatellite diversity seen in the BNP samples is
suggestive of a highly genetically impoverished group of
dholes in this part of Java. Such low levels of microsatellite
diversity were reported in the endangered Florida puma
(Puma concolor), which showed a sevenfold depletion of
microsatellite variation compared to other pumas (Driscoll
et al. 2002) and displayed stark phenotypic signs of inbreed-
ing (Roelke et al. 1993). Upon isolation of the islands of
Sumatra and Java from the mainland of Asia c. 10 000 bp
(Meijaard 2003), various species most likely underwent a
genetic bottleneck in comparison to their Asian mainland
counterparts. Island species are known to retain less
molecular diversity than mainland species (Frankham
1998). Massive human population growth and large-scale
habitat destruction and fragmentation over the past cen-
tury or so in Indonesia in general, and in Java in particular,
may have led to this highly genetically impoverished
group of dholes. BNP, which is situated in the northeastern
corner of East Java is surrounded by areas of scrubby
terrain with teak plantations, agricultural land and human
settlements, which are unlikely to support dholes for any
length of time. It is unclear whether two other national
parks on the southern coast of East Java (Alas Purwo and
Meru Betiri) which have been known to contain dholes in
the recent past, retain some interconnectivity with BNP for
Fig. 4 Bootstrap consensus tree depicting
haplotype relationships based on maximum-
likelihood (HKY + G) and maximum-
parsimony analysis. Numbers above nodes
are bootstrap support for the node in 1000
replicates under likelihood on the left and
parsimony on the right.
GENETIC STRUCTURE AND DIVERSITY IN THE DHOLE 2291
© 2005 Blackwell Publishing Ltd, Molecular Ecology, 14, 2281–2297
dholes through such terrain. What is also not clear at this
stage is whether there are fitness costs associated with this
genetic impoverishment.
In contrast to the BNP samples, in the MWS samples
from South India, we observed high levels of microsatellite
diversity and four mitochondrial control region haplotypes
with a 3% mean uncorrected p genetic distance. One hap-
lotype (A) was found to be the prevailing haplotype across
the majority of individuals in the MWS packs. It has been
suggested that dholes undergo periodic population ‘boom
and busts’, the reasons for which are unclear (e.g. Davidar
1975). It is possible that such ‘busts’ result in a temporary
loss/skew of genetic variability in dhole populations, but
that this is stabilized over time with immigration and
genetic exchange. The overall high genetic variability
detected within this small area suggests that any ‘busts’
have not resulted in a big loss of genetic diversity in this
region, perhaps because large tracts of interconnected
habitat suitable for dholes still remain, facilitating the
movement of animals. In addition, despite a history of
human persecution for centuries in India, no evidence for
a recent genetic bottleneck was detected in the MWS
samples using microsatellite data.
Pack size and structure in MWS
The 25% microsatellite amplification from silica-dried
faecal samples obtained in this study is much lower than
what has been reported in other carnivores, e.g. 65% and
45% from alcohol-preserved winter and summer scat
samples from wolves in Italy (Lucchini et al. 2002), and
36.5% from silica-dried faecal samples from brown bears
(Murphy et al. 2002). The DET buffer preservation method
resulted in higher success rates (60%). However, for
reasons of safety in the field and during transporta-
tion, it remains desirable to evaluate alternative methods
for dhole faecal preservation (e.g. Nsubuga et al. 2004;
Roeder et al. 2004). Low amplification rates and repeat
identification of individuals in MWS precluded the geno-
typing of every individual within a pack. However, our
results allow us to draw a number of conclusions about
dhole pack structure. First, within-pack relatedness is
significantly greater than between-pack relatedness in
both males and females. This supports delayed dispersal
in both sexes as previously proposed by Venkataraman
(1998). These results are not likely to be affected by the
sampling of predispersal-age juveniles because samples
were pooled from across a number of separate packs and
high numbers of such individuals across every pack is not
likely. The clear trend in both sexes for a decline in mean
relatedness values with an increase in distance values
suggests that both dhole males and females are more
closely related to individuals from the same or from
neighbouring packs than more distant packs. Clear evidence
for the existence of kin structure was also obtained using
a permutation method. Such a kin-based social structure
has been observed in other pack-living canids practising
cooperative breeding and hunting, e.g. painted hunting
dogs (Girman et al. 1997). Our results also suggest that
females may be more likely to disperse close to their natal
packs while males may disperse longer distances, along
similar lines to what has been observed in painted hunting
dogs (Girman et al. 2001). A male-biased pack structure
previously reported from field observations (Johnsingh
Table 4 × 100 pairwise uncorrected p (below diagonal) and HKY + G distances (above diagonal) between haplotypes and haplotype groups
S. India C.N. India
N. India
N. Ganges
Myanmar
E. India Thailand Malaysia
Captive,
Germany
Sumatra Java JK QR
S. India (8,2.7) 5.5 16.7 6.7 23.6 13.4 12.5 13.4 6.1 10.2 13.9 10.1
C.N. India (3, 2.2) 2.8 14.5 1.9 23.6 13.4 13.4 15.5 3.0 6.2 7.4 5.1
N. India (N. Ganges) 6.4 6.1 16.6 5.3 0.9 1.5 2.7 19.1 33.2 29.1 21.9
Myanmar (2, 8.1) J 3.4 1.2 6.5 16.6 16.6 19.1 2.7 4.3 5.3 3.5
K7.5 7.4 3.2 3.5 3.5 4.3 33.2 57.4 38.5 38.0
E. India 5.8 5.8 0.8 6.5 2.4 1.4 19.1 33.2 29.1 22.0
Thailand (2, 0.8) 5.6 5.8 1.2 6.5 2.4 0.4 1.4 19.1 33.2 29.1 22.0
Malaysia (1, 0) 5.8 6.2 2.0 6.9 2.8 1.2 1.2 22.1 29.1 25.6 19.2
Captive Q 3.4 2.2 6.9 2.0 8.6 6.9 6.9 7.3 6.4 4.3
R4.7 3.5 8.5 2.8 10.2 8.5 8.5 8.1 2.7 1.4
Sumatra 5.4 3.9 8.1 3.2 8.9 8.1 8.1 7.7 3.7 2.0 0.9
J
ava (1, 0) 4.6 3.1 7.3 2.4 8.9 7.3 7.3 6.9 2.8 1.2 0.8
Haplotypes from Myanmar and captive animals are shown separately. Values in parentheses in column 1 denote number of haplotypes
and within-group p distance. S. India, South India; C.N. India, Central and North India; N. India, North India; N. Ganges, North of Ganges
River; E. India, East India. Only one sample was available from N. India (North of Ganges), East India and Sumatra.
2292 A. IYENGAR ET AL.
© 2005 Blackwell Publishing Ltd, Molecular Ecology, 14, 2281–2297
1982; Venkataraman 1998) was not supported by the results
obtained for the Gamehut pack using molecular sexing.
However, inadequate overall sampling precludes a definitive
conclusion on this issue. The discrepancy seen in the
estimates of individuals within the Vazhaithotam pack
based on field observations and multilocus genotyping
may be due to temporary influx of individuals into the
Vazhaithotam pack or single individuals or ‘floaters’
passing through the home range.
Phylogeography
The phylogenetic analyses reveal two major geograph-
ically abutting clades. The deeper phylogenetic structure
and longer branches within clade I when compared to
clade II suggest older coalescence times and/or higher
historical effective population sizes for these populations.
This is also reflected by the fact that several of the clade I
haplotypes are positioned internally within the network
(Posada & Crandall 2001). In the case of clade II, shorter
branch lengths and more terminal nodes within the
network suggest a more recent coalescence. While we are
reluctant to estimate times of divergence using a short
control region sequence, we propose a scenario that could
explain the observed patterns. There is evidence for con-
siderable effects of glaciation events on the climate and
biogeography within South Asia. During glacial periods,
the climate is thought to have been drier, colder, and more
seasonal, with lowland areas suffering from increased
desertification, and reduction of moist tropical forests into
savannah and patchy deciduous forests (e.g. Brandon-
Jones 1996; Meijaard 2003, 2004). Much of northern and
western India is thought to have been desert at glacial
maxima (Fleischer et al. 2001). Also, recurrent patterns of
disjunct distribution of closely related, forest dependent
taxa are observed across Asia that are often best explained
in light of the effects of glaciation events where they are
thought to have become isolated within ice age refugia
and represent relics of a former continuous population
(Brandon-Jones 1996; Karanth 2003; Meijaard 2003). For
example, a number of taxa found in the wet evergreen
forests of southwest India and Sri Lanka are absent from
the rest of the Indian subcontinent and then found again in
the wet evergreen forests of northeast India and Southeast
Asia (Karanth 2003). High levels of diversity were found
in dholes from South India, parts of which (the Western
Fig. 5 A diagrammatic representation of the two major phylogeographical groupings seen with mitochondrial control region haplotypes.
Haplotypes J and K from Myanmar, and Q and R from captive animals presumed to originate from China are shown separately. Dotted
lines represent possible extension of clade I. Mitochondrial cytochrome b haplotypes identified from the various locations are also shown.
GENETIC STRUCTURE AND DIVERSITY IN THE DHOLE 2293
© 2005 Blackwell Publishing Ltd, Molecular Ecology, 14, 2281–2297
Ghats) are considered to be among the most biologically
diverse regions within South Asia. Thus, since South India
represents a region that is likely to have acted as a re-
fugium during glaciation, we propose that ancestral clade
I haplotypes from such a refugium expanded northwards
and eastwards, with a barrier to gene flow in the north in
the form of the Ganges River system. Haplotype F from
north of the Ganges in India was very distinct from the other
North and Central Indian haplotypes from south of the
Ganges. Results from cytochrome b sequencing and amovas
also support this finding. The distinct group of clade II
haplotypes are likely to have originated more recently from
one or more refugia within this region, separately from
clade I haplotypes. A number of studies have suggested the
existence of glacial refugia in northern Indochina and the
Malay Peninsula (Brandon-Jones 1996; Gorog et al. 2004). Our
sampling is not sufficiently extensive to confirm or reject the
apparent monophyly of clades I and II, and to determine
whether Myanmar, where haplotypes from both clades were
found, may represent a zone of more recent secondary contact.
Significantly large negative values for Fu’s FS were
obtained both when all haplotypes were analysed simul-
taneously and when clade I haplotypes were analysed sep-
arately. In addition, nonsignificant values for Fu and Li’s
F* and D* were seen in both these cases, a pattern that is
highly suggestive of an ancient population expansion as
expected from a refugial expansion scenario. Fu’s FS is a
powerful test for population expansion and has been
found to be considerably more sensitive than other such
statistics (Fu 1997). However, a mismatch distribution did
not show a unimodal pattern characteristic of such an
ancient population explosion. Interpretation of mismatch
distribution results have on occasion been considered
tentative since a number of factors have been found to
affect the results, such as time of population expansion,
population size before expansion, and subdivision of
populations (Marjoram & Donnelly 1994).
Haplotypes M and P from Sumatra and Java, which
grouped within clade I, showed high genetic distance
values to the haplotype found in Peninsular Malaysia,
and were very distinct from the group of haplotypes within
clade II. Cytochrome b haplotypes from these samples
were also identical to those found in samples from clade I
and different to that found in Peninsular Malaysia and
Thailand. Sumatra and Java are known to have been con-
nected to the Asian mainland via the Malaysian Peninsula
as recently as 10 000 years bp when the Sunda Shelf was
exposed as a consequence of low sea levels during glaciation
(Meijaard 2003). The existence of this land bridge has led
some researchers to conclude widespread dispersal of
animals across the Sunda Shelf (e.g. Heaney 1986). However,
others have suggested that dispersal of terrestrial species
Fig. 6 Median-joining network of haplotypes. The curved dividing line separates the two clades seen in the bootstrapped consensus tree.
Open circles represent haplotypes obtained in this study, while closed circles represent median vectors, i.e. presumed unsampled or missing
intermediates. Numbers on branches represent substitution positions.
2294 A. IYENGAR ET AL.
© 2005 Blackwell Publishing Ltd, Molecular Ecology, 14, 2281–2297
between the land masses of Sundaland was limited due to
the presence of a number of ecological and physical barriers
(Brandon-Jones 1996; Meijaard 2003, 2004). In this study,
dhole samples from Sumatra and Java were found to be
very distinct from the samples from Malaysia, supporting
the existence of such barriers. Lower genetic distances
were seen between these haplotypes and other haplotypes
within clade I, i.e. South, Central and North India (south of
the Ganges), Myanmar (haplotype J) and possibly China
(haplotypes Q and R). Further studies are required to clarify
these results but in the absence of alternative explanations,
these results may be suggestive of human translocation of
dholes from one of these regions into Sumatra and/or Java.
But since there is no documented evidence for such trans-
location(s), and given that dholes have been long considered
vermin (and not hunted for sport, cf. red fox introductions
to Australasia), this hypothesis must remain highly specu-
lative. Nonetheless, the Sumatran and Javan populations
of dholes are of high conservation importance given the
relatively few populations known from the Southeast
Asian mainland (Durbin et al. 2004b), making further
study into the origins of these dholes a high priority for
conservation. The low genetic distances displayed by both
haplotypes (Q and R) from captive animals thought to
have their origins in China and belonging to the lepturus
subspecies (H. Maisch, personal communication) to haplo-
type J within clade I and high distances to haplotypes
within clade II, suggest that they may be part of the east-
wards expansion of clade I.
Haplotypes were found to be shared between individuals
across vast distances, reflecting the highly mobile nature
of these carnivores. No clear distinction between sub-
species was noted, with haplotypes being shared between
Cuon alpinus dukhunensis & Cuon alpinus adjustus and
C. alpinusadjustus & Cuon alpinus infuscus, and very low genetic
distances between several subspecies. A similar lack of
distinctness between subspecies of widely distributed Asian
mammals has been reported for tigers (Panthera tigris) by
Cracraft et al. (1998). However, the presence of two distinct
phylogeographical groups and the grouping of Sumatran
and Javan haplotypes with those from India (south of the
Ganges) and Myanmar, as opposed to those from Malaysia
and Thailand, should be noted.
Conservation implications
Our results do not support the recognition of 11 subspecies
of dholes and instead, show admixture over vast areas.
However, conservationists should recognize the existence of
the two distinct phylogeographical groupings that we have
Table 5 Results from amova on various mtDNA haplotype groupings
No.
Pairwise difference Tamura Nei + G
Grouping ΦCT PΦCT P
10.245 0.057 0.046 0.211 [S. India, C. India, N. India] [N. India N. Ganges, E. India, Myanmar
(J and K), captive, Thailand, Malaysia] [Sumatra, Java]
20.382 0.007 0.446 0.007 [S. India, C. India, N. India, Myanmar J] [N. India N. Ganges, E.
India] [Myanmar K, Thailand] [Malaysia, Sumatra, Java] [captive]
30.507 0 0.602 0.002 [S. India, C. India, N. India, N. India N. Ganges] [E. India, Myanmar
(J and K)] [Thailand, Malaysia] [Sumatra, Java] [captive]
40.510 0 0.661 0 [S. India, C. India, N. India, Myanmar J, Sumatra, Java] [N. India N.
Ganges, E. India, Thailand, Myanmar K, Malaysia] [captive]
50.534 0 0.522 0.003 [S. India, C. India, N. India] [N. India N. Ganges, E. India],
[Myanmar (J and K), captive] [Thailand, Malaysia] [Sumatra, Java]
60.563 0 0.672 0 [S. India, C. India, N. India] [N. India N. Ganges, E. India,
Myanmar (J and K)] [Thailand, Malaysia] [Sumatra, Java] [captive]
70.590 0 0.704 0 [S. India, C. India, N. India, Myanmar J, captive] [N. India N.
Ganges, E. India, Myanmar K, Thailand, Malaysia] [Sumatra, Java]
80.608 0 0.716 0 [S. India, C. India, N. India, Myanmar J] [N. India N. Ganges, E
India] [Myanmar K, Thailand, Malaysia] [Sumatra, Java] [captive]
90.609 0 0.734 0 [S. India, C. India, N. India, Myanmar J] [N. India N. Ganges, E.
India] [Myanmar K, Thailand, Malaysia] [Sumatra, Java, captive]
10 0.613 0 0.722 0 [S. India, C. India, N. India, Myanmar J] [N. India N. Ganges, E.
India, Myanmar K, Thailand, Malaysia] [Sumatra, Java] [captive]
11 0.615 0 0.740 0 [S. India, C. India, N. India, Myanmar J] [N. India N. Ganges, E.
India, Myanmar K, Thailand, Malaysia] [Sumatra, Java, captive]
12 0.627 0 0.760 0 [S. India, C. India, N. India, Myanmar J, Sumatra, Java, captive]
[N. India N. Ganges, E India, Thailand, Myanmar K, Malaysia]
GENETIC STRUCTURE AND DIVERSITY IN THE DHOLE 2295
© 2005 Blackwell Publishing Ltd, Molecular Ecology, 14, 2281–2297
observed. The unexpected results, which indicate that dholes
may have been introduced to Sumatra and Java, need further
investigation, given the likely conservation importance of
dhole populations on these two large islands. Dholes from
the Baluran National Park in East Java were found to be
highly genetically impoverished, stressing the requirement
for additional genetic studies within this park, and other
protected areas in Java in order to better assess the status of
these populations and the possible need for genetic mana-
gement. Cuon alpinus primaevus, reported as ‘very rare’ from
a survey in the early 1980s ( Johnsingh 1985), and found to be
very distinct from C. alpinus dukhunensis in this study, must
be accorded particularly high priority for conservation
action. Within South India, the continued maintenance of
habitat interconnectivity is important to retain maximal
amounts of the high genetic diversity seen.
Acknowledgements
We thank the Max Planck Institute for Evolutionary Anthropology,
Leipzig, Germany, for funding this project. We also thank the
Institute for Bioarchaeology, USA, and the Marwell Preservation
Trust, UK, for providing us with additional financial support. We
are grateful to the chief wildlife warden of Tamilnadu Forest
Department for giving permission to carry out the ecological
study in MWS and to the deputy director of the Wildlife Preserva-
tion (Southern Region), Government of India, for permission to
carry faecal samples abroad. We thank Katrin Nowak for excellent
laboratory assistance in Leipzig and the field assistants in India for
collection of faecal samples. We thank Lon Grassman Jr. (Texas
A&M University, Kingsville, USA), Martin Tyson (WCS) and George
Amato (WCS) for providing us with faecal samples from Phu Khieo
Wildlife Sanctuary, Thailand, Baluran National Park, Java, and
Taman Negara National Park, Malaysia, respectively. We thank
Paula Jenkins, Daphne Hills, and Richard Sabin from the Natural
History Museum in London, for providing us with museum
samples. We also thank Heike Maisch (Schwerin zoo), Dr M. Bõer
(Hodenhagen zoo), Drs K. Schaller and E. Dag (Allwetter Zoo) and
Drs H. Lücker and W. Ludwig (Dresden Zoo) for information and
blood samples from captive dholes in Germany, and the keepers
at Leipzig Zoo for providing us with wolf faecal samples. We are
very grateful to Stephan Funk (Institute of Zoology, London) and
Vittorio Lucchini (University of Bologna, Italy) for advice with
microsatellite and mitochondrial primers. Finally, we thank Olaf
Thalmann, Heike Siedel, Karen Chambers, and Michael Hofreiter
(MPI, Leipzig) for advice with protocols, Erik Meijaard (Australian
National University), Fabio Diniz, Mairi Knight, Gagan Lushai and
Patrick Doncaster (University of Southampton) for helpful discus-
sion, and anonymous referees for useful comments.
References
Adams JR, Kelly BT, Waits LP (2003) Using faecal DNA sampling
and GIS to monitor hybridization between red wolves (Canis rufus)
and Coyotes (Canis latrans). Molecular Ecology, 12, 2175–2186.
Ayres KL, Overall ADJ (2004) api-calc 1.0: a computer program
for calculating the average probability of identity allowing for
substructure, inbreeding and the presence of close relatives.
Molecular Ecology Notes, 4, 315–318.
Bandelt H-J, Forster P, Röhl A (1999) Median-joining networks for
inferring intraspecific phylogenies. Molecular Biology and Evolu-
tion, 16, 37–48.
Belkhir K, Castric V, Bonhomme F (2002) identix, a software to
test for relatedness in a population using permutation methods.
Molecular Ecology Notes, 2, 611–614.
Brandon-Jones D (1996) The Asian Colobinae (Mammalia: Cercop-
ithecidae) as indicators of Quaternary climatic change. Biological
Journal of the Linnean Society, 59, 327–350.
Cohen JA (1978) Cuon alpinus. Mammalian Species, 100, 1–3.
Cohen JA, Fox MW, Johnsingh AJT, Barnett BD (1978) Food habits of
the dhole in south India. Journal of Wildlife Management, 42, 933– 936.
Cornuet JM, Luikart G (1996) Description and power analysis of
two tests for detecting recent population bottlenecks from allele
frequency data. Genetics, 144, 2001–2014.
Cracraft J, Feinstein J, Vaughn J, Helm-Bychowski K (1998) Sorting
out tigers (Panthera tigris): mitochondrial sequences, nuclear
inserts, systematics, and conservation genetics. Animal Conser-
vation, 1, 139–150.
Davidar ERC (1975) Ecology and behavior of the dhole or Indian
wild dog Cuon alpinus (Pallas). In: The Wild Canids: Their System-
atics, Behavioral Ecology and Evolution (ed. Fox MW), pp. 109–119.
Van Nostrand Reinhold Company, New York.
Driscoll CA, Menotti-Raymond M, Nelson G, Goldstein D, O’Brien
SJ (2002) Genomic microsatellites as evolutionary chronometers:
a test in wild cats. Genome Research, 12, 414–423.
Dundas RG (1999) Quaternary records of the dire wolf, Canis dirus,
in North and South America. Boreas, 28, 375–385.
Durbin LS, Hedges S, Duckworth JW, Tyson M, Iyengar A,
Venkataraman A (2004a) Cuon alpinus. In: 2004 IUCN Red List of
Threatened Species. http://www.redlist.org.
Durbin L, Venkataraman A, Hedges S, Duckworth JW (2004b)
Dhole (Cuon alpinus). In: Status Survery and Conservation Action
Plan. Canids: Foxes, Wolves, Jackals and Dogs (eds Sillero-Zubiri C,
Hoffman M, Macdonald DW), pp. 210–219. IUCN, Gland,
Switzerland and Cambridge, UK.
Excoffier L, Smouse PE, Quattro JM (1992) Analysis of molecular
variance inferred from metric distances among DNA haplo-
types: application to human mitochondrial DNA restriction
data. Genetics, 131, 479–491.
Fleischer RC, Perry EA, Muralidharan K, Stevens EE, Wemmer
CM (2001) Phylogeography of the Asian elephant (Elephas
maximus) based on mitochondrial DNA. Evolution, 55, 1882–
1892.
Fox MW (1984) The Whistling Hunters. Field Studies of the Asiatic
Wild Dog (Cuon alpinus). State University of New York Press,
New York.
Francisco LV, Langston AA, Mellersh CS, Neal CL, Ostrander EA
(1996) A class of highly polymorphic tetranucleotide repeats for
canine genetic mapping. Mammalian Genome, 7, 359–362.
Frankham R (1998) Inbreeding and extinction: island populations.
Conservation Biology, 12, 665–675.
Frantzen MA, Silk JB, Ferguson JW, Wayne RK, Kohn MH (1998)
Empirical evaluation of preservation methods for faecal DNA.
Molecular Ecology, 7, 1423–1428.
Fu Y-X (1997) Statistical tests of neutrality of mutations against
population growth, hitchhiking and background selection.
Genetics, 147, 915–925.
Girman DJ, Mills MGL, Geffen E, Wayne RK (1997) A molecular
genetic analysis of social structure, dispersal, and interpack
relationships of the African wild dog (Lycaon pictus) Behavioral
Ecology & Sociobiology, 40, 187–198.
2296 A. IYENGAR ET AL.
© 2005 Blackwell Publishing Ltd, Molecular Ecology, 14, 2281–2297
Girman DJ, Vilà C, Geffen E et al. (2001) Patterns of population
subdivision, gene flow and genetic variability in the African
wild dog (Lycaon pictus) Molecular Ecology, 10, 1703–1723.
Godoy JA, Negro JJ, Hiraldo F, Donázar JA (2004) Phylogeo-
graphy, genetic structure and diversity in the endangered bearded
vulture (Gypaetus barbatus L.) as revealed by mitochondrial
DNA. Molecular Ecology, 13, 371–390.
Gorog AJ, Sinaga MH, Engstrom MD (2004) Vicariance or dispersal?
Historical biogeography of three Sunda shelf murine rodents
(Maxomys surifer, Leopoldamys sabanus and Maxomys whiteheadi).
Biological Journal of the Linnean Society, 81, 91–109.
Guo SW, Thompson EA (1992) Performing the exact test of
Hardy–Weinberg proportions for multiple alleles. Biometrics,
48, 361–372.
Hall TA (1999) bioedit: a user-friendly biological sequence align-
ment editor and analysis program for Windows 95/98/NT.
Nucleic Acids Symposium Series, 41, 95–98.
Heaney LR (1986) Biogeography of mammals in SE Asia: estimates
of rates of colonization, extinction and speciation. Biological
Journal of the Linnean Society, 28, 127–165.
Holmes NG, Dickens HF, Parker HL, Binns MM, Mellersh CS,
Sampson J (1995) Eighteen canine microsatellites. Animal Genetics,
25, 132–133.
Ishiguro N, Nakajima A, Horiucji M, Shinagawa M (2002) Multiple
nuclear pseudogenes of mitochondrial DNA exist in the canine
genome. Mammalian Genome, 13, 365–372.
Johnsingh AJT (1982) Reproductive and social behaviour of the
dhole Cuon alpinus (Canidae). Journal of Zoology, London, 198,
443–463.
Johnsingh AJT (1985) Distribution and status of dhole Cuon alpinus
Pallas, 1811 in South Asia. Mammalia, 49, 203–208.
Johnsingh AJT (1992) Prey selection in three large sympatric
carnivores in Bandipur. Mammalia, 56, 517–526.
Karanth PK (2003) Evolution of disjunct distributions among
wet–zone species of the Indian subcontinent: testing various
hypotheses using a phylogenetic approach. Current Science, 85,
1276–1283.
Karanth KU, Sunquist ME (1995) Prey selection by tiger, leopard and
dhole in tropical forests. Journal of Animal Ecology, 64, 439–450.
Karanth KU, Sunquist ME (2000) Behavioural correlates of preda-
tion by tiger (Panthera tigris), leopard (Panthera pardus) and
dhole (Cuon alpinus) in Nagarahole, India. Journal of Zoology,
London, 250, 255–265.
Knight ME, Van Oppen MJH, Smith HL, Rico C, Hewitt GM, Turner
GF (1999) Evidence for male-biased dispersal in Lake Malawi
cichlids from microsatellites. Molecular Ecology, 8, 1521–1527.
Kumar S, Tamura K, Jakobsen IB, Nei M (2001) mega2: Molecular
evolutionary genetics analysis software. Bioinformatics, 17,
1244–1245.
Lucchini V, Fabbri E, Marucco F, Ricci S, Boitani L, Randi E (2002)
Noninvasive molecular tracking of colonising wolf (Canis lupus)
packs in the western Italian Alps. Molecular Ecology, 11, 857–
868.
Lucchini V, Galov A, Randi E (2004) Evidence of genetic distinc-
tion and long-term population decline in wolves (Canis lupus) in
the Italian Apennines Molecular Ecology, 13, 523–536.
Lynch M, Ritland K (1999) Estimation of pairwise relatedness with
molecular markers. Genetics, 152, 1753–1766.
Marjoram P, Donnelly P (1994) Pairwise comparisons of mito-
chondrial DNA sequences in subdivided populations and
implications for early human evolution. Genetics, 136, 673–683.
Meijaard E (2003) Mammals of South-east Asian islands and
their late Pleistocene environments. Journal of Biogeography,
30, 1245–1257.
Meijaard E (2004) Biogeographic history of the Javan leopard
Panthera pardus based on a craniometric analysis. Journal of
Mammalogy, 85, 302–310.
Meyers-Wallen VN, Palmer VL, Acland GM, Hershfiled B (1995)
Sry-negative XX sex reversal in the American cocker spaniel
dog. Molecular Reproduction and Development, 41, 300–305.
Morin PA, Chambers KE, Boesch C, Vigilant L (2001) Quantitative
PCR analysis of DNA from non-invasive samples for accurate
microsatellite genotyping of wild chimpanzees (Pan troglodytes
verus). Molecular Ecology, 10, 1835–1844.
Murphy MA, Waits LP, Kendall KC, Wasser SK, Higbee JA,
Bogden R (2002) An evaluation of long-term preservation
methods for brown bear (Ursus arctos) faecal DNA samples.
Conservation Genetics, 3, 435–440.
Nsubuga AM, Robbins MM, Roeder AD, Morin PA, Boesch C,
Vigilant L (2004) Factors affecting the amount of genomic DNA
extracted from ape faeces and the identification of an improved
sample storage method. Molecular Ecology, 13, 2089–2094.
Ostrander EA, Sprague GF, Rine J (1993) Identification and char-
acterisation of dinucleotide repeat (CA) n markers for genetic
mapping in dog. Genomics, 16, 207–213.
Ostrander EA, Mapa FA, Yee M, Rine J (1995) 101 new simple
sequence repeat-based markers for the canine genome. Mamma-
lian Genome, 6, 192–195.
Palumbi SR (1996) Nucleic acids II: the polymerase chain reaction.
In: Molecular Systematics (eds Hillis DM, Moritz C, Mable BK),
pp. 205–247. Sinauer Associates, Sunderland, Massachusetts.
Peck DR, Congdon BC (2004) Reconciling historical processes and
population structure in the sooty tern Sterna fuscata. Journal of
Avian Biology, 35, 327–335.
Pocock RI (1936) The Asiatic wild dog or dhole (Cuon javanicus).
Proceedings of the Royal Society of London. Series B, Biological Sci-
ences, 33–55.
Posada D, Crandall KA (1998) modeltest: testing the model of
DNA substitution. Bioinformatics, 14, 817–818.
Posada D, Crandall KA (2001) Intraspecific gene genealogies:
trees grafting into networks. Trends in Ecology & Evolution, 16,
37–45.
Prugnolle F, de Meeus T (2002) Inferring sex-biased dispersal from
population genetic tools: a review. Heredity, 88, 161–165.
Queller DC, Goodnight KF (1989) Estimating relatedness using
genetic markers. Evolution, 43, 258–275.
Randi E, Lucchini V, Christensen MF et al. (2000) Mitochondrial
DNA variability in Italian and East European wolves: detecting
the consequences of small population size and hybridization.
Conservation Biology, 14, 464–473.
Raymond M, Rousset R (1995) genepop (version 1.2): population
genetics software for exact tests and ecumenicism. Journal of
Heredity, 86, 248–249.
Roeder AD, Archer FI, Poinar HN, Morin PA (2004) A novel
method for collection and preservation of faeces for genetic
studies. Molecular Ecology Notes, 4, 761–764.
Roelke ME, Martenson JS, O’Brien SJ (1993) The consequences of
demographic reduction and genetic depletion in the endan-
gered Florida panther. Current Biology, 3, 340–350.
Rogers AR, Harpending H (1992) Population growth makes waves
in the distribution of pairwise genetic differences. Molecular
Biology and Evolution, 9, 552–569.
GENETIC STRUCTURE AND DIVERSITY IN THE DHOLE 2297
© 2005 Blackwell Publishing Ltd, Molecular Ecology, 14, 2281–2297
Rozas J, Sánchez-DeI, Barrio JC, Messeguer X, Rozas R (2003)
dnasp, DNA polymorphism analyses by the coalescent and
other methods. Bioinformatics, 19, 2496–2497.
Russello MA, Amato G (2004) Ex situ population management in
the absence of pedigree information. Molecular Ecology, 13,
2829–2840.
Schneider S, Roessli D, Excoffier L (2000) ARLEQUIN, ver 2.001. A soft-
ware for population genetics data analysis. URL:
http://lgb.unige.ch/arlequin/.
Stewart P (1993) Mapping the dhole. Canid News, 1, 18–21.
Stewart P (1994) Mapping the dhole update. Canid News, 2, 35–
36.
Swofford DL (2002) PAUP*: Phylogenetic Analysis Using Parsimony (and
Other Methods). Sinauer Associates, Sunderland, Massachusetts.
Thalmann O, Hebler J, Poinar HN, Pääbo S, Vigilant L (2004)
Unreliable mtDNA data due to nuclear insertions: a cautionary
tale from analysis of humans and other great apes. Molecular
Ecology, 13, 321–335.
Thenius E (1954) On the origins of the dhole. Osterr Zoological
Zeitsch, 5, 377–388.
Venkataraman A (1998) Male-biased adult sex ratios and their sig-
nificance for cooperative breeding in dhole, Cuon alpinus packs.
Ethology, 104, 671–684.
Venkataraman A, Arumugam R, Sukumar R (1995) The foraging
ecology of dhole (Cuon alpinus) in Mudumalai Sanctuary, South-
ern India. Journal of Zoology, London, 237, 543–561.
Vigilant L, Hofreiter M, Siedel H, Boesch C (2001) Paternity and
relatedness in wild chimpanzee communities. Proceedings of the
National Academy of Sciences, USA, 98, 12890–12895.
Vilà C, Amorim IR, Leonard JA et al. (1999) Mitochondrial DNA
phylogeography and population history of the grey wolf Canis
lupus. Molecular Ecology, 8, 2089–2103.
Wasser SK, Houston CS, Koehiler GM, Cadd GG, Fain SR (1997)
Techniques for the application of faecal DNA methods to field
studies of Ursids. Molecular Ecology, 6, 1091–1097.
Wayne RK, Geffen E, Girman DJ, Koepfli KP, Lau LM, Marshall
CR (1997) Molecular systematics of the Canidae. Systematic
Biology, 46, 622–653.
A. Iyengar is interested in the application of molecular techniques
to address issues important to the conservation of endangered
species. V. N. Babu is interested in canid ecology and behaviour,
and carnivore community ecology. S. Hedges currently works on
Asian mammal conservation projects focusing on elephants, wild
cattle, and dholes. A. B. Venkataraman is interested in applying
knowledge of behavioural ecology in the design of landscapes for
conservation. N. Maclean is professor of genetics at the University
of Southampton, with interests in genetic manipulation of fish and
molecular ecology. P. A. Morin currently directs the Molecular
Ecology Laboratory at the Southwest Fisheries Science Centre,
conducting basic and applied genetics research for marine
mammal conservation.
