RESEARCH ARTICLE
Phylogeographic analysis of nuclear and mtDNA supports
subspecies designations in the ostrich (Struthio camelus)
Joshua M. Miller ? Sara Hallager ? Steven L. Monfort ?
John Newby ? Kelley Bishop ? Scott A. Tidmus ? Peter Black ?
Bill Houston ? Conrad A. Matthee ? Robert C. Fleischer
Received: 12 September 2010 / Accepted: 28 September 2010 / Published online: 15 October 2010
 US Government 2010
Abstract We investigated the phylogeography and sub-
species classification of the ostrich (Struthio camelus) by
assessing patterns of variation in mitochondrial DNA con-
trol region (mtDNA-CR) sequence and across fourteen
nuclear microsatellite loci. The current consensus taxonomy
of S. camelus names five subspecies based on morphology,
geographic range, mtDNA restriction fragment length
polymorphism and mtDNA-CR sequence analysis: S. c.
camelus, S. c. syriacus, S. c. molybdephanes, S. c. massaicus
and S. c. australis. We expanded a previous mtDNA dataset
from 18 individual mtDNA-CR sequences to 123 sequences,
including sequences from all five subspecies. Importantly,
these additional sequences included 43 novel sequences of
the red-necked ostrich, S. c. camelus, obtained from birds
from Niger. Phylogeographic reconstruction of these
sequences matches previous results, with three well-sup-
ported clades containing S. c. camelus/syriacus, S. c.
molybdophanes, and S. c. massaicus/australis, respectively.
The 14 microsatellite loci assessed for 119 individuals of
four subspecies (all but S. c. syriacus) showed considerable
variation, with an average of 13.4 (?2.0) alleles per locus
and a mean observed heterozygosity of 55.7 (?5.3)%. These
data revealed high levels of variation within most subspe-
cies, and a structure analysis revealed strong separation
between each of the four subspecies. The level of divergence
across both marker types suggests the consideration of
separate species status for S. c. molybdophanes, and perhaps
also for S. c. camelus/syriacus. Both the mtDNA-CR and
microsatellite analyzes also suggest that there has been no
recent hybridization between the subspecies. These findings
are of importance for management of the highly endangered
red-necked subspecies (S. c. camelus) and may warrant its
placement onto the IUCN red list of threatened animals.
Keywords Struthio camelus  Ostrich  Africa 
Mitochondrial DNA control region  Phylogeography 
Microsatellites
Electronic supplementary material The online version of this
article (doi:10.1007/s10592-010-0149-x) contains supplementary
material, which is available to authorized users.
J. M. Miller  R. C. Fleischer (&)
Center for Conservation and Evolutionary Genetics, Smithsonian
Conservation Biology Institute, National Zoological Park,
PO BOX 37012 MRC 5503, Washington, DC 20013-7012, USA
e-mail: fleischerr@si.edu
S. Hallager
Animal Care Directorate, Smithsonian National Zoological Park,
PO BOX 37012 MRC 5507, Washington, DC 20013-7012, USA
S. L. Monfort
Smithsonian Conservation Biology Institute, National Zoological
Park, 1500 Remount Road, Front Royal, VA 22630, USA
J. Newby  K. Bishop
Sahara Conservation Fund, Rue des Tigneuses 2, 1148 L?Isle,
Switzerland
S. A. Tidmus
Animal Programs and Environmental Initiatives, Disney?s
Animal Kingdom, Lake Buena Vista, FL 32830, USA
P. Black  B. Houston
Saint Louis Zoo, One Government Drive, St Louis,
MO 63110, USA
C. A. Matthee
Department of Botany and Zoology, Stellenbosch University,
Private Bag X1, Matieland, Stellenbosch 7602, South Africa
123
Conserv Genet (2011) 12:423?431
DOI 10.1007/s10592-010-0149-x
Introduction
Historically, the ostrich, Struthio camelus, was widely
distributed throughout Africa and the Arabian Peninsula,
but recently has been in decline, especially in the northern
extent of its range. A number of subspecies have been
described by different authors, but currently five subspecies
are recognized based on morphology and geographic range
(del Hoyo et al. 1992): three in northern and eastern Africa:
S. c. camelus, S. c. massaicus, and S. c. molybdophanes;
one in southern Africa: S. c. australis; and lastly, the
Arabian ostrich: S. c. syriacus. However, human expansion
and exploitation, as well as local extinctions have reduced
this distribution significantly on the African continent and
led to the recent, cir. 1940 (Seddon and Soorae 1999),
extinction of S. c. syriacus. Of particular concern today is
the Saharan red-necked subspecies (S. c. camelus), which
has declined to critically low numbers (Ostrowski et al.
2001).
Previous studies elucidated the genetic relationships
between the five subspecies and showed significant struc-
turing among them (Freitag and Robinson 1993; Robinson
and Matthee 1999). These studies, based on variation in
mitochondrial restriction fragment length (RFLP) and
mitochondrial DNA control region (mtDNA-CR) sequence,
showed S. c. camelus and S. c. syriacus to fall into one
clade, S. c. massaicus and S. c. australis into a second, and
S. c. molybdophanes to form a third clade that was highly
divergent from either of the others.
Here we present an expanded phylogeny and population
genetic assessment of all extant subspecies, with special
emphasis on the critically endangered S. c camelus. Our
study includes 43 novel samples of S.c camelus from
Niger, and also a greatly increased sample size of the other
subspecies from what was included in previous work
(Robinson and Matthee 1999). In addition, we assess var-
iability and relationships using both mtDNA-CR sequences
and fourteen nuclear microsatellite loci. The inclusion of
nuclear microsatellites can give greater resolution to pop-
ulation genetic studies as they are bi-parentally inherited,
evolve rapidly (Goldstein and Pollock 1997), and are often
able to detect structuring at a finer scale than other markers
(Coltman et al. 2003; Freeman-Gallant 1996). Inclusion of
nuclear microsatellites also helps to address any incorrect
inferences that can arise when relying solely on mtDNA for
phylogenetic analysis (Ballard and Whitlock 2004).
Mitochondrial DNA sequences showed much the same
pattern as those previously published, with the presence of
three distinct clades. The microsatellite data revealed
higher levels of population structure, showing a distinction
between S. c. massaicus and S.c. australis, corresponding
to the established subspecies classification system. Nota-
bly, all individuals thought to represent S. c. camelus in
Niger show no evidence of mixed ancestry. This is
important not only for proposed recovery plans within
northern Africa but also for potential use of S. c. camelus to
reintroduce ostriches back to the Arabian Peninsula (Sed-
don and Soorae 1999). Our results also lend credence to the
suggestion of separate species status for S. c. molybdo-
phanes, and perhaps for S. c. camelus/syriacus and S. c.
massaicus/australis as well.
Methods and materials
Sample collection and DNA extraction
Samples were taken from 43 red-necked birds (S. c.
camelus) in private collections at ten localities throughout
Niger. These included 31 blood samples, eight dung sam-
ples, and four feathers. An additional sample was derived
from an egg found in the Tin Toumma desert of Eastern
Niger at N 16.20 E 12.00 on 21 February 2004, and esti-
mated to be *50 years old. The egg is presumed to be
representative of S. c. camelus as it falls within the his-
torical distribution of the subspecies.
Total genomic DNA was extracted from Niger ostriches
using various protocols based on sample type at the Center
for Conservation and Evolutionary Genetics lab at the
National Zoological Park. Blood samples were extracted
using the Qiagen Biosprint 96 DNA Extraction Worksta-
tion, which utilizes magnetic particles to separate DNA.
Fecal DNA was extracted using the QIAamp DNA Stool
Mini Kit following manufacturer?s protocol. For feather
samples the extraction protocol described by Horva?th et al.
(2005) was used. Additionally, DNA was extracted from
the *50 years old egg following a procedure outlined for
ancient samples (Fleischer et al. 2000).
In addition, 76 samples representing the three other
extant ostrich subspecies were extracted at Pretoria Uni-
versity using standard phenol?chloroform isolation meth-
ods (Robinson and Mathee 1999). These samples represent
23 localities from throughout most of the ostrich?s range in
Africa (Table 1) and correspond to those used in a previous
RFLP study (Freitag and Robinson 1993).
Mitochondrial DNA amplification
A 445 base pair (bp) segment of the mitochondrial control
region was amplified, corresponding to previously published
sequences for the ostrich (Robinson and Matthee 1999).
However, we were unable to optimize the previously pub-
lished primer set because of the presence of a large mono-
nucleotide repeat near the end of the reverse primer.
Therefore, a new primer set was developed and used, OS-
TPCRF (50-CGGATATTACCCGAGACCTACA-30) and
424 Conserv Genet (2011) 12:423?431
123
OSTPCRR (50-TAGGCTCTCTGGGGTTCACT-30) that
extends beyond each of the previous primers and amplifies a
product of 691 bp.
Amplification reactions were conducted in 10 ll vol-
umes consisting of 1.0 ll genomic DNA extract, 1.0 ll
59 PCR buffer, 1.0 ll 2 lM dNTPs, 0.6 ll MgCl2, 0.5 ll
of each primer diluted to 10 lM, 0.1 ll AmpliTaq DNA
polymerase, and 5.3 ll DNAse/RNAse free water. The
thermocyler profile began with 95C for 2 min followed by
thirty cycles of 95C for 30 s, 59C for 30 s, 72C for 30 s
and a final extension step of 72C for 2 min.
In cases where DNA quality was poor, such as with the
egg, scat, and feather extracts, a set of nested internal
primers, OSTINTF (50CTAGGACTACAACTGTCACAA
TCTCT-30) and OSTINTR (50-CGCATACTGGGTATGT
GGTG-30), were used to obtain the entire 445 bp region in
two overlapping products. The primer sets were OS-
TPCRF/OSTINTR (445 bp) and OSTINTF/OSTPCRR
(368 bp) and amplification reactions were conducted in
20 ll volumes consisting of 2.0 ll DNA extract, 2.0 ll
59 PCR buffer gold, 2.0 ll 2 lM dNTPs, 2.0 ll MgCl2,
0.8 ll of each primer, 2.0 ll BSA, 0.2 ll AmpliTaq DNA
polymerase gold, and 8.2 ll DNAse/RNAse free water.
In all cases PCR reactions were purified using ExoSAP-
IT (Amersham Biosciences) and then subjected to cyclese-
quencing with ABI PRISM BigDye Terminator v3.1 Cycle
Sequencing Kit. Sequences were resolved on an ABI 3130xl
genetic analyzer, initially aligned in Sequencher v4.8 (Gene
Codes), exported as Nexus files, and finally aligned by hand
in MacClade v4.06 OS X (Maddison and Maddison 2001).
Additionally, eighteen previously published sequences,
including two representing S. c. syriacus, were obtained
from GenBank (accession numbers AF073001?AF073018),
used as references, and included in subsequent analyzes.
Sequence analysis
A maximum likelihood (ML) tree with 500 bootstrap rep-
licates was generated using RAxML Black Box (Stamatkis
et al. 2008) under a GTR?I?G model, and imported into
PAUP (Swofford 2003) for visualization. Relationships
among haplotypes were also analyzed through a median-
joining haplotype network created using Arlequin version
3.0 (Excoffier et al. 2005).
Genetic diversity was further analyzed using Arlequin
version 3.0 (Excoffier et al. 2005) through calculations of
Fst and AMOVAs. Initially, groupings corresponded to
subspecies classification, regardless of sampling locality.
Subsequently, nested AMOVAs were generated to look at
partitioning of variance within subspecies, based on sam-
pling localities. In cases where a sampling locality has only
one or two individuals, several localities were pooled based
on geographic proximity. AMOVAs were generated using
each of the four available distance methods in Arlequin,
with no difference among the results.
Microsatellite amplification/genotyping
Fourteen microsatellite loci designed for the ostrich (Tang
et al. 2003) were used in this study. Only six of these loci
were mapped on a microsatellite based genetic map by
Huang et al. (2008), but each of the six were assigned to
different linkage groups (of 13 linkage groups: CAU17 on
1, CAU69 on 11, CAU97 on 4, CAU84 on 9, CAU3 on 2,
Table 1 Geographic origin of samples used in the present study
Subspecies Country Locality Number of
samples
S. c. camelus Niger 43
S. c. australis Botswana
Mabuasehube game
reserve
1
Makgadikgadi pans 10
Namibia
Escourt game farm 3
Gorrasis game farm 2
Ongombeanavita farm 1
South
Africa De hoop nature
reserve
1
Doornkraal, de rust 3
Joubertina 2
Kalahari gemsbok
national park
1
Kleinsee nature
reserve
1
Kruger national park 3
Ladismith 2
Langjan nature reserve 3
Matjiesrivier 2
Oude muragie, de rust 3
Oudtshoorn proefplaas 2
Prince Albert 4
Oudtshoorn 3
Zimbabwe
Centeral estate 5
Hwange national park 5
Northern
matabeleland
3
S. c. massaicus Kenya
Kajiado 7
Tanzania
Arusha 3
S. c.
molybdophanes
Kenya
Mt Kenya 6
Conserv Genet (2011) 12:423?431 425
123
and CAU14 on 6). The loci were arranged into five mul-
tiplexes based on common annealing temperature (Ta) and
size ranges (Table 2). PCR reactions were carried out with
dye labeled primers (the forward primer for each primer
set) in 10 ll volumes with 1.0 ll DNA extract, 1.0 ll PCR
buffer, 1.0 ll 20 lM dNTPs, 1.0 ll BSA, and 0.1 ll
AmpliTaq DNA polymerase. Thermocycler profile began
with 94C for 5 min, followed by thirty-five cycles of 94C
for 40 s, Ta for 1 min, and 72C for 1 min. A final
extension step of 72C for 30 min completed the reactions.
Samples were then run on an ABI 3130xl genetic analyzer
using ABI ROX 500 as an internal size standard. Geno-
types were assigned using Genemapper v4.0 (Applied
Biosystems). Note that DNA samples were not available
for S. c. syriacus, and thus it was not represented in the
microsatellite analysis.
Microsatellite analysis
We used Genepop (Raymond and Rousset 1995) to esti-
mate heterozygosity, test for linkage disequilibrium, and
conduct Hardy?Weinberg exact tests using the default
parameters. Arlequin version 3.0 (Excoffier et al. 2005)
was used to calculate Fst and Rst (pair-wise differences). As
with the mtDNA analysis, initially the data were parti-
tioned by subspecies alone then into subgroups based on
sampling locality. A second nested AMOVA was calcu-
lated to investigate the effect of domestic stock designation
on genetic structure in S. c. australis. In this case, samples
were organized into three groups: domestic stocks, game
reserves, and nature reserves.
The Bayesian clustering program Structure v2.2 was run
to assess genetic structure among the subspecies (Pritchard
et al. 2000). We evaluated K = 1?7 possible population
groupings, using a burn in of 50,000 iterations followed by
500,000 iterations for each K. We allowed for admixture
within individuals, allowed for independent allele fre-
quencies between populations, and initiated each run with
popinfo. Each K value was run with four replicates to
assess stability.
In addition, clustering of individuals was assessed using
the program Populations v 1.2.19 (Langella 1999). Popu-
lations generated a distance matrix, based on the shared
allele distance DAS (Jin and Chakraborty 1994), which
was used to generate a neighbor-joining network with
1,000 bootstrap replicates that was visualized in TreeView
v1.6.6 (Page 1996).
Results
Analysis of the mtDNA sequences
We successfully amplified the entire 445 bp control region
segment for 106 individuals: 43 Niger birds and 63 of the
samples representing the other extant subspecies. Addi-
tionally, we were able to amplify a 76 bp region from the
degraded DNA isolated from the old egg using the internal
primers (OSTINTF/OSTINTR). Combined with the 18
sequences from Genbank (Robinson and Matthee 1999), we
had sequences for 125 individuals. We detected 80 haplo-
types, characterized by 73 polymorphic sites. The haplo-
types generally segregated according to subspecies, with
only two haplotypes shared between subspecies: one
between S.c camelus and S. c. syriacus and the second
between S. c. australis and S. c. massaicus. As expected the
egg fell within the S. c. camelus/S. c. syriacus clade. (Fig. 1).
Phylogenetic relationships Based on mtDNA
Three well-defined clades were found with the control
region data from both ML analysis (Fig. 2) and construction
of a minimum spanning network (Fig. 3). These correspond
to those found in previous studies with S. c. camelus and
S. c. syriacus in one clade, S. c. massaicus and S. c. australis
in another, and S. c. molybdophanes monophyletic.
Pair-wise Fst comparisons between all subspecies indi-
cated substantial divergences between subspecies, except
S. c. camelus and S. c. syriacus (Table 3). For all Fst cal-
culations, S. c. massaicus was found to be distinct from the
other subspecies, including S. c. australis, despite the tree
topology.
Table 2 Summary statistics for microsatellite data for entire sample
Locus Number
of alleles
Expected
heterozygosity
Observed
heterozygosity
CAU28 4 0.64 0.56
CAU17 22 0.892 0.667
CAU36 2 0.19 0.144
CAU69 13 0.801 0.576
CAU97 10 0.83 0.706
CAU76 27 0.906 0.788
CAU5 11 0.859 0.709
CAU23 20 0.856 0.607
CAU84 11 0.746 0.667
CAU68 13 0.832 0.627
CAU92 6 0.414 0.259
CAU46 23 0.874 0.278
CAU3 9 0.689 0.485
CAU14 16 0.875 0.729
Primers and other specifics for the loci are available in Tang et al.
(2003)
426 Conserv Genet (2011) 12:423?431
123
When samples were grouped by locality, nested
AMOVA showed that over 64% of genetic variation was
due to differences among subspecies groups, 5% was due
to populations within groups, and 30% was due to variation
within populations. In most cases, pair-wise Fst compari-
sons showed no significant differences between sampling
localities within a subspecies.
Microsatellite amplification and diversity
All 119 birds were genotyped, with 97% of all microsat-
ellite genotypes attempted attained. The number of alleles
across subspecies ranged from 2 to 27 per locus, with a
mean of 13.4 (SD = 2.0) alleles per locus and average
observed heterozygosity of 55.7 (SD = 5.3) % (Table 2;
Table S1, online supplementary information). Several loci
(CAU 28, 36, 69, 97, 5, and 92) were observed to be fixed
within S. c. molybdophanes (the subspecies with the
smallest sample size), while in the other subspecies allele
number ranged from one (CAU 36 in S. c. camelus) to
twenty-two (CAU 76 in S. c. australis). No linkage dis-
equilibrium was detected after performing sequential
Bonferroni corrections (P [ 0.05). Pooling all subspecies
results in a strong heterozygote deficiency across all the
loci (Table 2). However, all loci conformed to Hardy?
Weinberg equilibrium within subspecies after sequential
Bonferroni correction except CAU 46 (P \ 0.05, hetero-
zygote deficit in three subspecies), CAU23 and CAU3
(P \ 0.05 heterozygote deficit in one subspecies each).
Phylogenetic relationships based on microsatellites
As with mtDNA analysis, pair-wise distance analysis
comparing subspecies found significant differences
between the subspecies, as measured both by Fst and Rst
(Table 4). Each subspecies was found to be distinct in a
neighbor-joining network based on DAS (not shown), as
well as though Bayesian clustering (Fig. 4). While we
supposed that the optimal number of populations, K, would
be four we calculated DK, a measure of the second order
rate of change in the likelihood of K (Evanno et al. 2005) to
test this assumption. These calculations gave high weight
to the presence of four populations, but also to three or five
populations. When K = 3, S. c. molybdophanes and S. c.
massaicus are clustered in a single population, while when
K = 5, each subspecies is recovered with additional sub-
structure in S. c. australis. In all cases the individuals held
in Niger were assigned to a single cluster, presumably
representing S. c. camelus; they show no evidence of
genetic introgression from other subspecies or sub-
structuring.
Discussion
Subspecies classification and taxonomy implications
Our analysis of mtDNA recovered three major clades
consistent with those found in pervious studies (Freitag and
Robinson 1993; Robinson and Matthee 1999) and largely
consistent with current subspecies taxonomy. Though
S. c australis and S. c. massaicus are nested together in the
tree and share a haplotype, they were found to be signifi-
cantly differentiated from one another as measured by Fst
(Table 3). Original analysis of mitochondrial RFLPs (Fre-
itag and Robinson 1993) found very limited diversity
within and between subspecies of ostriches, save for
S. c. molybdophanes, and it was postulated that a recent
population bottleneck might have been the cause. The high
degree of diversity seen in the results of Robinson and
Matthee (1999) and our mitochondrial tree provides no
evidence for such a bottleneck.
Nuclear microsatellite variation recovered finer levels of
structure than mtDNA, showing each subspecies to be
distinct. Tree building, AMOVAs, and Bayesian cluster
analysis showed a clear distinction between S. c australis
and S. c. massaicus (Fig. 3) not seen in mtDNA. This
distinction suggests that the current subspecies classifica-
tion system for each of these should be upheld. Though
there appears to be strong heterozygote deficiency across
loci (Table 2) this is most likely due to a Wahlund effect
from comparing across differentiated subspecies.
S. c. camelus
S. c. syriacus
S. c. massaicus S. c. 
molybdophanes
S. c. australis
Fig. 1 Map of the historic range of the ostrich (Struthio camelus) in
Africa, with each of the estimated subspecies ranges delineated by
color (adapted from Robinson and Matthee 1999)
Conserv Genet (2011) 12:423?431 427
123
While calculation of DK gives high weight to K = 3,
which groups S. c. molybdophanes and S. c. massaicus
together, this makes little biological sense as it goes against
the large genetic distances seen in mtDNA and microsatel-
lites, as well as our DAS neighbor-joining tree and a prin-
cipal components analysis (not shown). On the other hand,
Fig. 2 Maximum likelihood tree based on 445 base pairs of mtDNA-
CR sequence from 119 ostriches (this study) and 18 reference
sequences (Robinson and Matthee 1999). Values above the bars are
bootstrap support values (500 replicates). Taxa include S. c. camelus,
S. c. syriacus, S. c. massaicus, S. c. australis, and S. c. molybdophanes
428 Conserv Genet (2011) 12:423?431
123
our STRUCTURE results for K = 4 and 5 both reflect the
accepted subspecies classification, with the latter K showing
additional structure within S. c. australis. This substructure
does not appear to be the result of inclusion of domestic
stocks in our analysis, nor does it correspond to specific
geographic regions. Thus, we are unsure of the exact nature
or cause of this substructure.
Previous publications have suggested that S. c. molyb-
dophanes might be a distinct species (Freitag and Robinson
1993; Lewis and Pomeroy 1989; Brown et al. 1982). The
high levels of differentiation between S. c. molybdophanes
and the other subspecies, in both mtDNA and microsatellite
analysis, support this assertion. In addition, the high level
of differentiation between S. c. camelus and the other
subspecies (except S. c. syriacus) suggests this clade may
also represent a species level divergence.
Historical causes of diversification
The topology of the mtDNA tree roughly groups subspe-
cies into a southern clade (S. c. australis and S. c. mas-
saicus), a northern clade (S. c. camelus and S. c. syriacus),
and a highly divergent eastern clade (S. c. molybdophanes).
This type of eastern/southern split is also seen in many
mammal species, including carnivores (Rohland et al.
2005; Freeman et al. 2001; Dubach et al. 2005; Barnett
et al. 2006), and large ungulates (Flagstad et al. 2001;
Moodley and Bruford 2007; Muwanika et al. 2003; Arct-
ander et al. 1999; Brown et al. 2007).
In most cases the split is attributed to glacial refugia
present in eastern, western and southern Africa separating
populations during periods of climate change and glacia-
tion. These refugia were then sources from which most
Fig. 3 Median-joining
haplotype network of mtDNA-
CR sequences from 119
ostriches (this study) plus 18
sequences from Robinson and
Matthee (1999). Circles
represent haplotypes, and their
diameter is proportional to
number of individuals with that
haplotype. Branch lengths are
proportional to the number of
mutations, except for the dashed
or dotted lines, where the
number of mutations is noted
below the line. Same color
scheme as in Fig. 2
Table 3 Subspecies pairwise Fst based on mtDNA-CR sequence
S. c. camelus S. c. australis S. c. molybdophanes S. c. syriacus S. c. massaicus
S. c. camelus 0
S. c. australis 0.47828 0
S. c. molybdophanes 0.53198 0.78684 0
S. c. syriacus 20.00176 0.71151 0.88444 0
S. c. massaicus 0.39993 0.2139 0.74462 0.68119 0
Value in bold is not significant (P [ 0.05); all others significantly different from zero at P \ 0.05
Table 4 Subspecies pair-wise divergence based on microsatellites
S. c. australis S. c. massaicus S. c. molybdophanes S. c. camelus
S. c. australis 0 0.160 0.330 0.108
S. c. massaicus 0.127 0 0.398 0.208
S. c. molybdophanes 0.432 0.563 0 0.353
S. c. camelus 0.306 0.304 0.637 0
Values above the diagonal are calculated as traditional Fst; values below the line are calculated as Rst. All values are significantly different from
zero (P \ 0.05)
Conserv Genet (2011) 12:423?431 429
123
populations expanded their range and subsequently diver-
sified (Hewitt 2004). Tectonic activity associated with the
rift valley system has also been posed as a diversification
mechanism (Faulkes et al. 2004; Partridge 1995). We did
not attempt to date diversification among the ostrich clades
because of the uncertainty of rate estimates for mtDNA-CR
(Ruokonen and Kvist, 2002), and no appropriate calibration
point being available for ostriches.
The abrupt disjunction in range boundary and large
genetic divergence between S. c. molybdophanes and S. c.
massaicus is seen between subspecies of other vertebrates
from the same area (Faulkes et al. 2004; Brown et al.
2007). Such disjunctions are attributed to past paleocli-
matic events initially separating the populations, thus
leading to the evolution of differences in behavior, habitat
preference, or other isolating mechanisms. These differ-
ences could continue to drive speciation even in zones of
secondary contact. Reports of difficulties in mating
between subspecies of ostrich, particularly S. c. molybdo-
phanes and S. c. massaicus (Lewis and Pomeroy 1989),
provide some evidence for the existence of such isolating
mechanisms.
Conservation implications
The red-necked subspecies S. c. camelus is of critical
conservation concern: very few individuals are currently
seen in the wild (Ostrowski et al. 2001) and all 46 of the
samples in this study come from birds held in private
collections (with the exception of the old egg). While the
original provenance of the 46 S. c. camelus individuals of
our study is uncertain (probably arising from Chad), the old
egg found in the Tin Toumma desert of Eastern Niger is
very likely from a wild, resident Niger ostrich, and its
clustering in the mtDNA-CR analysis with our Niger
samples (and with the previous samples from Robinson and
Matthee 1999) suggest that these are also birds from the
region and S. c. camelus subspecies. In addition, the
Bayesian clustering analysis of 14 microsatellites showed
no indication of genetic structuring within the captive S. c.
camelus populations or introgression from other subspe-
cies. Thus these 46 individuals should make a good
founding population for captive propagation efforts cur-
rently underway. The large numbers of haplotypes in
mtDNA and alleles in microsatellites suggest no major loss
of genetic diversity despite reports of drastic population
decline.
Our analyzes of mtDNA and nuclear microsatellites
support the designation of at least four described subspecies
(four of the five subspecies previously based on morphol-
ogy, range and other characteristics). The data also would
perhaps support species status for three taxa: S. c. molyb-
dophanes, S. c. camelus/syriacus, and S. c. australis/mas-
saicus. Regardless of taxonomic status, we believe that the
level of genetic divergence of S. c. camelus from other extant
taxa, coupled with its current rarity and vulnerability, should
Fig. 4 Genetic structure of
ostriches based on a Bayesian
cluster analysis of 14
microsatellite loci using
structure v2.2. (Pritchard et al.
2000). We assessed values of K
from 1 to 7, and show results for
K = 3, 4, and 5. Numbers
below the axis correspond to
subspecies, 1 S. c. australis, 2
S. c. massaicus, 3
S. c. molybdophanes, and 4
S. c. camelus
430 Conserv Genet (2011) 12:423?431
123
warrant its placement on the IUCN Red List of threatened
species.
Acknowledgments Fieldwork was funded by the Sahara Con-
servation Fund, Saint Louis Zoo, Smithsonian National Zoological
Park, Niger?s Ministry of Environment, San Diego Zoo?s Wild
Animal Park, AZA Conservation Endowment Fund, and Disney?s
Animal Kingdom; and laboratory analyses by Friends of the National
Zoo and the Center for Conservation and Evolutionary Genetics. We
thank Niger?s Ministry, the Directorate General of Environment,
T. J. Robinson, and Dr. Maikano for assistance in the field, and Nancy
Rotzel, Laura Morse, Frank Hailer and Emily Latch for help with
logistics and/or analysis.
References
Arctander P, Johansen C, Coutellec-Vreto M (1999) Phylogeography
of three closely related African bovids (tribe Alcelaphini). Mol
Biol 16:1724?1739
Ballard JWO, Whitlock MC (2004) The incomplete natural history of
mitochondria. Mol Ecol 13:729?744
Barnett R, Yamaguchi N, Barnes I, Cooper A (2006) The origin,
current diversity and future conservation of the modern lion
(Panthera leo). Proc Royal Soc B 273:2119?2125
Brown L, Urban E, Newman K (1982) The Birds of Africa, vol 1.
Academic Press, London
Brown D, Brenneman R, Koepfli K-P et al (2007) Extensive
population genetic structure in the giraffe. BMC Biology 5:57
Coltman DW, Pilkington JG, Pemberton JM (2003) Fine-scale genetic
structure in a free-living ungulate population. Mol Ecol 12:
733?742
Del-Hoyo J, Elliott A, Sargatal J (eds) (1992) Handbook of the birds
of the world. volume 1: Ostrich to Ducks. Lynx Edicions,
Barcelona
Dubach J, Patterson BD, Briggs MB et al (2005) Molecular genetic
variation across the southern and eastern geographic ranges of
the African lion, Panthera leo. Conserv Genet 6:15?24
Evanno G, Regnaut S, Goudet J (2005) Detecting the number of
clusters of individuals using the software structure: a simulation
study. Mol Ecol 14:2611?2620
Excoffier L, Laval G, Schneider S (2005) Arlequin (version 3.0): an
integrated software package for population genetics data anal-
ysis. Evol Bioinformatics 1:47?50
Faulkes GC, Verheyen E, Verheyen W, Jarvis JUM, Bennett NC
(2004) Phylogeographical patterns of genetic divergence and
speciation in African mole-rats (Family: Bathyergidae). Mol
Ecol 13:613?629
Flagstad O, Syvertsen PO, Stenseth NC, Jakobsen KS (2001)
Environmental change and rates of evolution: the phylogeo-
graphic pattern within the hartebeest complex as related to
climatic variation. Proc Royal Soc B 268:667?677
Fleischer RC, Olson S, James HF, Cooper AC (2000) The identity of
the extinct Hawaiian eagle (Haliaeetus) as determined by
mitochondrial DNA sequence. Auk 117:1051?1056
Freeman AR, Machugh DE, McKeown S et al (2001) Sequence
variation in the mitochondrial DNA control region of wild
African cheetahs (Acinonyx jubatus). Heredity 86:355?362
Freeman-Gallant CR (1996) Microgeographic patterns of genetic and
morphological variation in savannah sparrows (Passerculus
sandwichensis). Evolution 50:1631?1637
Freitag S, Robinson TJ (1993) Phylogeographic patterns in mito-
chondrial DNA of the ostrich (Struthio camelus). Auk
110:614?622
Goldstein DB, Pollock DD (1997) Launching microsatellites: a
review of mutation processes and methods of phylogenetic
interference. Heredity 88:335?342
Hewitt GM (2004) The structure of biodiversity?insights from
molecular phylogeography. Front Zool 1(4). doi:10.1186/1742-
9994-1-4
Horva?th MB, Mart??nez-Cruz B, Negro JJ, Kalma?r L, Godoy JA
(2005) An overlooked DNA source for non-invasive genetic
analysis in birds. J Avian Biol 36:84?88
Huang Y, Liu Q, Tang B, Lin L, Liu W, Zhang L, Li N, Hu X (2008)
A preliminary microsatellite genetic map of the ostrich (Struthio
camelus). Cytogenet Genome Res 121:130?136
Jin L, Chakraborty R (1994) Estimation of genetic distance and
coefficient of gene diversity from single-probe multilocus DNA
fingerprinting data. Mol Biol Evol 11:120?127
Langella O (1999) Populations. 1.2.19. [http://bioinformatics.org/
*tryphon/populations/]
Lewis A, Pomeroy DE (1989) A bird atlas of Kenya. Balkema,
Rotterdam
Maddison DR, Maddison WP (2001) MacClade 4: analysis of
phylogeny and character evolution. Version 4.02. Sinauer
Associates, Sunderland
Moodley Y, Bruford MW (2007) Molecular biogeography: towards
an integrated framework for conserving pan-african biodiversity.
PLoS ONE 2:e454
Muwanika VB, Nyakaana S, Siegismund HR, Arctander P (2003)
Phylogeography and population structure of the common
warthog (Phacochoerus africanus) inferred from variation in
mitochondrial DNA sequences and microsatellite loci. Heredity
91:361?372
Ostrowski S, Massalatchi MS, Mamane M (2001) Evidence of a
dramatic decline of the red-necked ostrich Struthio camelus
camelus in the A??r and Te?ne?re? National Nature Reserve, Niger.
Oryx 35:349?352
Page R (1996) TreeView: an application to display phylogenetic trees
on personal computers. Comput Appl Biosci 12:357?358
Partridge T, Wood B, de Menocal P (1995) The influence of global
climate change and regional uplift on large-mammalian evolu-
tion in eastern and southern Africa. In: Vrba ES, Denton GH,
Partridge TC (eds) Paleoclimate and evolution with emphasis on
human origins. Yale University Press, London, pp 331?355
Pritchard JK, Stephens M, Donnelly P (2000) Inference of population
structure using multilocus genotype data. Genetics 155:945?959
Raymond M, Rousset F (1995) GENEPOP (version 1.2): population
genetics software for exact tests and ecumenicism. J Hered
86:248?249
Robinson TJ, Matthee CA (1999) Molecular genetic relationships of
the extinct ostrich, Struthio camelus syriacus: consequences for
ostrich introductions into Saudi Arabia. Anim Conserv
2:165?171
Rohland N, Pollack JL, Nagel D et al (2005) The population history
of extant and extinct hyenas. Mol Biol Evol 22:2435?2443
Ruokonen M, Kvist L (2002) Structure and evolution of the avian
mitochondrial control region. Mol Phylogenet Evol 23:422?432
Seddon PJ, Soorae PS (1999) Guidelines for subspecific substitutions
in wildlife restoration projects. Conserv Biol 13:177?184
Stamatakis A, Hoover P, Rougemont J (2008) A rapid bootstrap
algorithm for the RAxML web servers. Syst Biol 57:758?771
Swofford DL (2003) PAUP*. Phylogenetic analysis using parsimony
(* and other methods). version 4.0. Sinauer Associates,
Sunderland
Tang B, Huang Y, Lin L et al (2003) Isolation and characterization of
70 novel microsatellite markers from ostrich (Struthio camelus)
genome. Genome 46:833?840
Conserv Genet (2011) 12:423?431 431
123