Domestication of the horse: Genetic relationships
between domestic and wild horses
⁎, Peter Dovč
Agricultural Institute of Slovenia, Crop and Seed Science Department, Hacquetova 17, SI-1000 Ljubljana, Slovenia
University of Ljubljana, Biotechnical faculty, Animal Science Department, Groblje 3, SI-1230 Domžale, Slovenia
Received 13 March 2007; received in revised form 3 March 2008; accepted 5 March 2008
To date, a large amount of equine genetic data has been obtained regarding (i) extant domestic horses of various breeds from all over
the world, (ii) ancient domestic horses, (iii) the extant Przewalski's wild horse, and (iv) the late Pleistocene wild horse from Eurasia and
North America. Here, a review of mtDNA and Y chromosome marker analyses is presented in the context of horse domestication. High
matrilineal (mtDNA) diversity, which can be found in both extant and ancient (domestic and wild) horses, has suggested that a high
number of wild (and tamed) mares were domesticated. Alternatively, Y chromosome marker analysis revealed a single haplotype in all
domestic horses analyzed; interestingly even a small population of extant Przewalski's wild horses showed two different Y chromosome
haplotypes. It seems that an extreme male population bottleneck occurred due to domestication, while reduction in the female population
was only moderate,leaving about 100 distinct haplotypes. For this reason, we speculate that domestication might have started when the
appropriate stallion was found or was obtained by selection. Perhaps it had some unusual but special characteristics which could have
accelerated the process of domestication. We doubt that only a single Y chromosome haplotype will be found in present-day domestic
horses if there are no important differences between the founder stallion/s and the other stallions that were not included in the
domestication. In the Eneolithic, tamed and wild mares have probably been spread all over Eurasia, although the number of animals was
most likely very low and the populations were limited to a restricted area (e.g., taming centers). Only two subspecies of wild horses
(Tarpan and Przewalski's wild horse) have survived up to recently. During the further process of domestication, mares (tamed or wild)
were preferentially crossed to stallions having more desirable characteristics. We assume that mares from differentregions varied in their
morphology due toadaptation to their local environmental conditions. These datamight explain rapid expansion of horse populations, as
well as their rapid differentiation into various phenotypes during the early phase of domestication.
© 2008 Elsevier B.V. All rights reserved.
Keywords: Equus caballus; Domestication; Mitochondrial DNA (mtDNA); Y chromosome markers
1. Introduction ....................................................... 2
1.1. Wild and feral horse populations ......................................... 2
1.2. Tamed and domesticated horses.......................................... 3
vailable online at www.sciencedirect.com
Livestock Science 116 (2008) 1–14
⁎Corresponding author. Tel.: +386 12805261; fax: +386 2805255.
E-mail address: firstname.lastname@example.org (T. Kavar).
1871-1413/$ - see front matter © 2008 Elsevier B.V. All rights reserved.
1.3. Morphology of wild and early domesticated horses ................................3
1.4. Models of domestication..............................................4
2. Genetic data........................................................4
3. mtDNA markers......................................................5
3.1. mtDNA haplotypes of extant domestic horses ................................... 5
3.2. mtDNA haplotypes of ancient domestic horses ..................................8
3.3. mtDNA haplotypes of wild horses.........................................8
3.4. Phylogeographic structure ............................................ 10
4. Y chromosome markers ................................................. 11
5. Reconstruction of the process of horse domestication ................................. 12
Acknowledgements ...................................................... 12
References .......................................................... 12
1.1. Wild and feral horse populations
During the late Pleistocene, large herds of wild horse
were common on the open plains of Europe, Asia, and
North America (Clutton-Brock, 1999). Evidence for the
existence of wild horses in North Africa has also been
found (Bagtache et al., 1984). By the end of the ice age,
the range of all wild horses was very much reduced
(Clutton-Brock, 1999). In North America, due to the
climatic/vegetational shift, horses underwent a rapid
decline in body size and became extinct around 10,500
BC (Guthrie, 2003). If humans played a role in their
extinction, it occurred on top of other ecological changes
(Guthrie, 2006). In Eurasia, only two subspecies of wild
horses survived to historic times:
•The Tarpan. Equus ferus ferus Boddaert, 1785.
Synonyms: sylvestris Brincken, 1826; gmelini Anto-
nius, 1912; silvaticus Vetulani, 1928 (Groves, 1986).
•Przewalski's horse. Equus ferus przewalskii Polia-
kov, 1881. Synonyms: hagenbecki Matschie, 1903;
typicus Hilzheimer, 1909; probably also equuleus
Smith, 1841 (Groves, 1986).
The Tarpan (Fig. 1) was described in numerous ref-
erences from the eighteen and nineteenth centuries as a
small animal, having a mouse-dun coat with a light
underbelly, sooty to black limbs fromthe knees and hocks
down, a short, frizzled mane, and a tail with short darkhair
(Olsen, 2006). Although the last Tarpan went extinct in
Poland in 1918 or 1919, the only available skeletal ma-
terial consists of one complete skeleton and a cranium
lacking a mandible from another individual. There are
many accounts of Tarpans stealing domestic mares and
forming harems, thus many Tarpans may simply have
been feral horses or hybrids (Olsen, 2006). Exmoor ponies
of Britain which are often claimed to be directly de-
scended from British Pleistocene stocks may also repre-
sent a population of feral ponies (Clutton-Brock, 1999).
Przewalski's horse (Fig. 2) would also be extinct
today if a small number had not been captured in the
Mongolian steppes in the first years of the 20th century
and brought to Europe, where they have multiplied in
zoos and wildlife parks (Clutton-Brock, 1999). Like the
Tarpan, Przewalski's stallions were notorious for steal-
ing mares from Mongol horse herds (Olsen, 2006). The
karyotypes of a Przewalski's horse (2n= 66) and a do-
mestic horse (2n=64) differed by one Robertsonian
translocation (ECA 5= EPR23 + EPR 24); the direction
of evolution –fusion or a fission –cannot be concluded
(Benirschke et al., 1965; Ryder et al., 1978; Ahrens and
Stranzinger, 2005). The hybrids (2n=65), however, are
fertile (Koulischer and Frechkop, 1966). Przewalski's
horse is considered as a sister taxon of wild progenitors
of domestic horses, mainly due to the differences in
chromosome number. The existing populations of
Przewalski's horses, however, are derived from very
limited number of individuals, and only descendants
Fig. 1. Drawing of a Tarpan horse (Borisov, 1841).
2T. Kavar, P. Dovč/ Livestock Science 116 (2008) 1–14
of these animals were karyotyped. If a translocation
occurred recently (i.e., up to several thousand years ago),
some other Asian populations of Holocene wild horses
might still have 64 chromosomes.
1.2. Tamed and domesticated horses
The questions of when, where and why horses were
first domesticated, are still hotly debated. It is generally
accepted that domestication should be considered a
process that flows through our arbitrary classes of
wild, captive, tame, and domesticated horses (Olsen,
2006). Horse-keeping, perhaps just for food, might
have begun in a limited way in the European steppes
during the Early Eneolithic, about 5000 BC, when
horses were first included with cattle and sheep in
graveside ritual deposits like those at Khvalynsk on the
Vol g a r i ve r (Brown and Anthony, 1998; Anthony and
Horseback riding, which might be a good indicator of
horse domestication, first appeared in the steppes east of
the Ural Mountains. Here, bit wear (a dental pathology
that occurs regularly among bitted horses) has been
found on the lower second premolars (P
's) of at least
four horses at the sites of Botai and Kozhai 1 in northern
Kazakhstan, both dated around 3500-3000 BC (Brown
and Anthony, 1998). It is possible that 85-90% of the
horses butchered at Botai were never bitted. Perhaps the
Botai hunters rode horses to hunt wild horses (Brown
and Anthony, 1998; Anthony and Brown, 2000). An-
other indirect evidence for horse domestication was
found in the way of circle and semicircular horse corrals
that have been uncovered in Eneolithic Botai settle-
ments (Olsen, 2006; Stiff et al., 2006). However, popu-
lation structure analysis for Botai (the age and sex
structure of horse populations) fit the hunting model
(Levine, 2005). Levine (2005) suggested that horse
taming probably first arose as a by-product of horse
hunting for meat. Wild horses, particularly as foals, can
be captured and tamed and, as such, ridden or harnessed
and, at the end of their lives, if necessary, slaughtered
and eaten. Considering the problems encountered by
modern collectors trying to breed Przewalski's horses, it
seems likely that horse-keeping would have had to have
been relatively advanced before controlled breeding
over successive generations, and thus domestication,
would have been possible (Levine, 2005).
The earliest direct evidence for horse domestication –
using dateable textual and artistic evidence –only dates
back to the end of the third millennium BC (Levine,
2005). Evidence of horses in graves, accompanied by
artifacts unambiguously associated with riding or trac-
tion, is even more recent, dating to no later than the
beginning of the second millennium BC. By the middle
of the second millennium, horses were widely used to
pull chariots, e.g., in the Near East, Greece and on the
Eurasian steppe (Levine, 2005).
1.3. Morphology of wild and early domesticated horses
Reduction in overall animal size on the one hand and
an increase in variability on the other are classic indica-
tors of domestication (Uerpmann, 1990). It seems, how-
ever, that domestication has little impact on the equine
anatomy; there are no indisputable osteological differ-
ences between wild and domesticated horses (Levine,
Fig. 2. Przewalski's wild horse (copyright Mark Kostich).
3T. Kavar, P. Dovč/ Livestock Science 116 (2008) 1–14
2006). In addition, horse populations exhibited high
variability throughout the Pleistocene into the post-
Pleistocene (Levine, 2006). The systematic classifica-
tion of the remains of wild horses from an enormous
geographic area, say from Western Europe to Siberia,
that includes a wide range of habitats and a long tem-
poral span, is very difficult to perform. This is due in
large part to the fact that in Quaternary horses, adaptive
and non-adaptive traits combine to form a reticular
pattern, a mosaic of characters that represent changes in
different places and times (Eisenmann, 1996). Even
within an assemblage from a single locality, multiple
forms have been detected (Olsen, 2006).
The wild horses of the mid-Holocene varied naturally
in size. The horses of the central Eurasian steppes in
Kazakhstan were somewhat larger than those of the
western steppes in central Ukraine, which were larger
than those of the steppe/forest-steppe border in Western
Ukraine and Romania. All the horses of the steppes were
significantly larger than the pony-sized wild horses of
central and Western Europe (Anthony, 2007). There is a
range of variation in both wild and domestic animals,
and concerning most characteristics, there is consider-
able overlap (Olsen, 2006). Nevertheless, the increase
in variation that began about 2500 BC and continued
thereafter has often been taken as an indicator of do-
mestication, although increased variation is sensitive to
sample size (Anthony, 2007).
By 1500 BC during what would have been the early
phase in the domestication of the horse, there seems
already to be definition into northern pony and Arabian
types (Clutton-Brock, 1999). One group consisted of the
small “Celtic”ponies of Britain, Western Europe and
Greece, while the second group consisted of larger horses
from Scythia and the Russian steppes (Clutton-Brock,
1999). “Celtic”ponies were very small, some being less
than a meter high atthe withers; the Scythian horses were
characteristically large and resembled Arabian horses in
their conformation (Clutton-Brock, 1999).
1.4. Models of domestication
Two models of horse domestication have been sug-
gested (Clutton-Brock, 1999):
The wild stock, from which all domestic horses were
bred, inhabited the plains of southern Russia, from
the Ukraine to the region of Turkestan. The earliest
domesticated horses spread out from this arc and
all the different types and breeds of horses that are
known today were developed as a result of artificial
selection in combination with natural selection for
adaptation to local environmental conditions.
The alternative model is based on the possibility that
there was a geographical cline in the population of
wild horses: those in the northern part of the range
being smaller and sturdier than those found in the
south. The reason for believing this is that even in the
earliest findings of the domestic horse there are con-
siderable differences in the size and proportions of
the bones from different regions.
Recent research suggests, however, that the natural
distribution of the Holocene horse was much wider than
what has been formerly believed, particularly in Western
and central Europe (Levine, 2005). Neolithic remains
from putatively wild horses have, for example, been
found in Sweden, Denmark, the Netherlands, France,
Spain, Italy, Germany, Switzerland, Hungary and Serbia –
in addition to Ukraine, Russia and Kazakhstan. These
results are underlying the assumption that the earliest
domestication must have taken place in Eastern Europe or
central Asia. The fact that wild horses were more common
in those regions does not definitely prove that they were
first domesticated there (Levine, 2005).
According to Levine (2005), horse domestication
would have taken a relatively long time to develop and
might well have been dependent upon arbitrary genetic
changes that would have predisposed some horses to
breed in captivity. Even if the earliest assays into do-
mestication had been relatively restricted –either tem-
porally or spatially –until their extinction, wild horses
could, and probably would have been introduced into
domestic herds (Levine, 2006). Historical records of
horses, described as wild, in Eurasia –in the post Iron
Age –do suggest that they were widely distributed at
least until the medieval period. Thus wild genes could
have been introduced into domestic stock until relatively
recent times (Levine, 2006).
2. Genetic data
In horses, three types of genetic data have been
studied extensively during the last decade:
•Nucleotide sequences of mitochondrial DNA (mtDNA),
which can be used for the reconstruction of past
events regarding female populations
•Y chromosome markers (nucleotide sequences,
microsatellites and single-nucleotide polymorphisms
(SNP)) - for reconstruction of past events in male
4T. Kavar, P. Dovč/ Livestock Science 116 (2008) 1–14
•Microsatellites of nuclear autosomal DNA - for esti-
mation of the relationship (or the differentiation)
among horse breeds.
3. mtDNA markers
Due to the maternal inheritance of mtDNA and lack
of recombination, mtDNA has been widely used for
studying the history of maternal lines. If no mutations
occur, progeny show the same mtDNA nucleotide se-
quence (haplotype or mtDNA type) as their mother.
Occasionally, mutations occur and become fixed in the
maternal (female) lines. By examining these mutations,
we can follow distinct maternal lines back into the
population history (Fig. 3). The majority of mtDNA
genetic data concern the nucleotide sequences of the
most variable part of the mtDNA: that found in the
control (or D-loop) region, which represents a noncod-
ing region that is about 1200 bp long (Fig. 4)(Xu and
Arnason, 1994; Ishida et al., 1994).
3.1. mtDNA haplotypes of extant domestic horses
To date, from various breeds all over the world, over
100 distinct nucleotide sequences (mtDNA haplotypes)
have been obtained. Among the first breeds most tho-
roughly investigated were Lipizzans (Kavar et al., 1999;
Kavar et al., 2002) and Arabians (Bowling et al., 2000;
Jansen et al., 2002). Among both breeds, over 30 distinct
haplotypes were detected. Haplotypes clustered to the
haplogroups C1-C4 (Kavar et al., 2002) or to clusters A-
F(Fig. 5). The bootstrap values were high for groups C2,
C3a, C3b and C4 but values for the C1 group were
slightly lower. The integrity of the C1 group, however, is
supported by high sequence identity in the downstream
part (HVR2) of the control region (Kavar et al., 2002).
High intra-breed diversity and roughly the same
clusters were detected in other breeds too; the differ-
entiation among clusters was unclear. In native or more
primitive breeds, or in a breed with a small population
size, the extent of mtDNA variation might be smaller;
Fig. 4. Scheme of the equine mtDNA control region (Xu and Arnason, 1994; Ishida et al., 1994). The 1200 bp control region (out of about 16600 bp
mtDNA) is located between tRNA
; it consists of two highly variable parts (HVR1 and HVR2), four conserved sequence blocks
(CSB) and repeats of the 8 bp motif GTGCACCT. The upstream part (HVR1) of the control region is usually more variable (and thus more
informative) than the downstream part (HVR2); many analyses are therefore based just on the sequences of the HVR1 fragment.
Fig. 3. An example of an accumulation of mutations in the maternal lines.
5T. Kavar, P. Dovč/ Livestock Science 116 (2008) 1–14
e.g., seven distinct haplotypes were detected by se-
quence analysis of 20 Exmoor ponies while the variation
in the Sorraia breed (a primitive Iberian equine type,
recovered from 12 founders in 1937) is limited to two
distinct haplotypes (Luís et al., 2002, 2006a).
Modern horse lineages coalesce at about 320,000 to
630,000 years ago, long before the first domestic horse
appeared (Vila et al., 2001). The majority of nucleotide
differences were therefore present before domestication
(Lister et al., 1998). The majority of haplotypes are
shared by different breeds (e.g., Lipizzan haplotype Ca-
priola was found in Pura Raza Española, Lusitano,
Kerry bog pony, Arabian, Norvegian Fjord, Holsteiner,
Rhineland Heavy draft, Chillean Criollo, Zemaitukai,
Trakehner, Duelmener and Barb). It has been reported
that most of the 17 very frequent mtDNA haplotypes
were old enough to have developed a star-like branching
10905/F2)(Jansen et al., 2002). Therefore these an-
cestral haplotypes are characterized by the lack of
unique polymorphic sites in comparison to their evolu-
tionary derivatives. Derivative haplotypes are usually
Fig. 5. Relationships among 39 mtDNA haplotypes of the Lipizzan horse breed (a historically well-documented cultural breed originating from many
different breeds including Karst-, Spanish-, Italian-, Kladruber- and Arabian horses (Nürnberg, 1993)). Nucleotide sequences of Lipizzan haplotypes
were retrieved from the GenBank database (AF168689–705, AY057408–36, DQ233731–2) (Kavar et al., 1999, 2002, 2004). Composite sequences
of HVR1 (nt 15450-15834) and HVR2 (nt 16351-16660) of the control region were aligned using ClustalW (Higgins et al., 1994). Kimura two-
parameter distances were calculated (Kimura, 1980) and a neighbour-joining tree (Saitou and Nei, 1987) was constructed with Phylip (ver. 3.66)
(Felsenstein, 2005). The same program was used to perform bootstrap analysis on 1000 data sets. Bootstrap values higher than 50 were entered into
the tree. Haplogroups according to the classification of Kavar et al. (2002) are represented by green letters ( ); haplogroups and subgroups according
to the classifications of Vila et al. (2001) and Jansen et al. (2002) are represented by orange letters in parenthesis ( and ). Lipizzan haplotypes are
represented by the black circles and by name ( ).
6T. Kavar, P. Dovč/ Livestock Science 116 (2008) 1–14
less frequent and differ from the ancestral haplotype by
one or a few nucleotides (Kavar et al., 2002). However,
the boundary among “ancestral”haplotypes and their
“derivatives”is often unclear because many very fre-
quent haplotypes are closely related and the possibility
of recurrent mutations is not negligible.
Fig. 6. Relationships among mtDNA haplotypes of ancient domestic horses. Nucleotide sequences of ancient domestic horses were retrieved from the
GenBank database. Samples consist of horses from: ( ) the Viking age from archeological sites in southern Sweden and Estonia dating from 1000 to
2000 years ago (AF326674–86) (Vila et al., 2001); ( ) Kwakji (Korea) dating from AD 700 to AD 800 (AY049720) (Jung et al., 2002); ( ) Pompeii
and Herculaneum dating to 79 BC (AY129545–6, AY129530–532) (Di Bernardo et al., 2004); ( ) a Scythian princely tomb in Kazakhstan dating from
the beginning of the 3rd century BC (AJ876883–90) (Keyser-Tracquiet al., 2005); ( ) Yakutia (Russia) from the 17th or 18th century AD (AJ876891–
2) (Keyser-Tracqui et al., 2005); ( ) Siberia and Ural dating to 2200 BP (DQ007573, DQ007571) (Weinstock et al., 2005); ( ) Ireland and England
dating from 1675 to 1314 BP (DQ327848–9, DQ327851) (McGahern et al., 2006b); ( ) Chifeng region 'Inner Mongolia' China dating to 4000-2000
BP (DQ900922–30) (Cai et al., 2007). For the comparison, ( ) mtDNA haplotypes of the Lipizzan horse breed (see Fig. 5) were added as a sample of
extant domestic horses. For the majority of ancient horses only sequences of the most variable (HVR1) part of the control region (about 228-348 bp)
were available; therefore, whenever sequences of ancient horses were identical to some of the sequences of the Lipizzan horse, they are presented in the
tree by circles composed of two or more different colors ( , , ). Otherwise, the neighbour-joining tree was constructed as described in Fig. 5.
7T. Kavar, P. Dovč/ Livestock Science 116 (2008) 1–14
3.2. mtDNA haplotypes of ancient domestic horses
mtDNA control regions were sequenced from sam-
ples of ancient domestic horses:
•from the Viking age from archeological sites in
southern Sweden and Estonia dating to 1000 to
2000 years ago (Vila et al., 2001)
•from Kwakji (Korea) dating to AD 700 to AD 800
(Jung et al., 2002)
•from the Pompeii and Herculaneum dating to 79 BC
(Di Bernardo et al., 2004)
•from a Scythian princely tomb in Kazakhstan dating
from the beginning of the 3rd century BC (Keyser-
Tracqui et al., 2005)
•from Yakutia (Russia) from the 17th or 18th century
AD (Keyser-Tracqui et al., 2005)
•from Siberia and Ural dating to 2200 BP (Weinstock
et al., 2005)
•from Ireland and England dating to 1244 to 1595 BP
(McGahern et al., 2006b)
•from Bronze Age horses recovered from the Chifeng
region of the ‘Inner Mongolia’, China dating to 4000-
2000 BP (Cai et al., 2007).
Zero or very low numbers of mutations are expected
in the mtDNA control region during the last few thou-
sands years. Therefore it is not surprising that haplo-
types of ancient domestic horses are identical or highly
similar to haplotypes of extant domestic horses (Fig. 6).
Six out of eight mtDNA haplotypes of Scythian horses
(Ber01, Ber02, Ber05, Ber06, Ber07 and Ber10) were
identical to the Lipizzan haplotypes.
In addition, it seems that the frequent haplotypes in
extant domestic horses (Jansen et al., 2002) were also
very frequent in the early phase after horse domestica-
tion. Several such haplotypes were found by analysis of
ancient samples: haplotypes A5 and D1 in domestic
horses from Ireland; haplotypes A1, B2, C1 and D1 in
the horses from the Viking Age; haplotypes D1, B2 and
C1 in Scythian horses; and haplotype F1 in Bronze Age
horses from Chifeng, ‘Inner Mongolia’, China. Only a
very low level of sequence variation (due to novel
mutations) is expected since the domestication; there-
fore we could assume that the main reason for the
presence of very frequent haplotypes in extant domestic
horses is that first domesticated mares (founder mares)
had such haplotypes. This does not mean that mares
with other haplotypes were not domesticated. Many
maternal lines died out, due to random genetic drift and
selection (natural and artificial). mtDNA (matrilineal)
diversity observed in ancient domestic horses is higher
(up to 7%) than in the extant domestic horses (up to
3.5%). The main reasons for higher maximal genetic
distances in ancient horses are due mainly to (i) the few
“outliers”(e.g., two samples from the Roman Age from,
both Pompeii and Herculaneum, and the sample from
Ural from about the same period) (Fig. 6), and (ii) the
fact that sequences of ancient horses were shorter and
from the most variable part of the control region
(HVR1) while calculations for the modern samples is
based on larger sequences including less variable HVR2
3.3. mtDNA haplotypes of wild horses
Recently, many new samples from wild horses have
been sequenced. Besides haplotypes of the extant wild
horse (Przewalski's horse) (Oakenfull and Ryder, 1998;
Kim et al., 1999), sequences from the late Pleistocene
horse are also available, including samples from:
•eight horses from Alaska, near Fairbanks, dating to
12,000 to 28,000 BP (Vila et al., 2001)
•one horse from Ireland dating to 27,630 BP
(McGahern et al., 2006b)
•four horses from Germany dating to 12,545 to 47,100
BP (Weinstock et al., 2005)
•seven horses from Siberia dating from 20,100 to
53,100 BP (Weinstock et al., 2005)
•five horses from America (Alaska, Yukon, Alberta,
Nevada) dating from 11,200 to 43,900 BP (Wein-
stock et al., 2005) (only samples for which the ap-
proximate age was determined, are listed)
Some haplotypes of wild horses are very similar to
extant domestic horse haplotypes (Fig. 7); this is true for
both haplotypes detected in Przewalski's wild horse,
haplotypes of European wild horses (from Germany and
Ireland), and some haplotypes of Siberian horses. Other
Siberian haplotypes clustered in the haplogroups of the
North American late Pleistocene horses. Haplotypes of
domestic horse and haplotypes of North American late
Pleistocene horses can thus be found in separate clusters
(Fig. 7), although some subgroups are not very distant to
subgroups of extant domestic horses, such as group C3c
(cluster C according to Vila et al., 2001). Group C3c
differed by as little as 1.2% from modern counterparts
(Vila et al., 2001).
It seems that the majority of North American
haplotypes did not survive the process of domestication,
because such haplotypes have not been detected in do-
mestic horses. This is in agreement with the hypothesis
that American populations of wild horses were not
8T. Kavar, P. Dovč/ Livestock Science 116 (2008) 1–14
Fig. 7. NJ tree showing the relationships among mtDNA haplotypes of ( ) wild and ( ) domestic horses. Samples consist of haplotypes of extant
wild horses: (PR) Przewalski's horse (AF055878–9; AF014409) (Oakenfull and Ryder, 1998; Kim et al., 1999) and of extinct wild horses from: (AK)
Alaska dating from 12,000 to 28,000 BP (Vila et al., 2001) (AF326668–75), (IR) Ireland dating to 27,630 BP (DQ327850) (McGahern et al., 2006b);
(D) Germany dating from 12,545 to 47,100 BP (DQ007556, DQ007558, DQ007609, DQ007611, DQ007590–1), (SB) Siberia dating from 20,100 to
53,100 BP (DQ007552–3, DQ007574–83, DQ007606–7), (AK, YK, AB and WY) Alaska, Yukon, Alberta and Wyoming, dating from 11,200 to
43,900 BP (DQ007555, DQ007557, DQ007559, DQ007584–89, DQ007592–602, DQ007608, DQ007610, DQ007612), and (CN and UR) China
and Ural (DQ007604, DQ007572) (Weinstock et al., 2005). For all extinct wild horses, sequences of the HVR1 part of the control region were
available (about 348 bp), and for seven of them the HVR2 fragment (133 bp) was available as well. Samples of domestic horses included those with
known sequences at the HVR1 (nt 15469-15834) and HVR2 (nt 16351-16660) part of the control region from various breeds from all over the world
(AY246174–271, AF064627–29, AF011411–115, AF011405, AF011407, AF056071, AF168689–705, AY057408–36, DQ233731–2). Sequences
of donkey (Xu et al., 1996) (X97337) and of Late Pleistocene “Stilt-legged”horses from North America (DQ007620–1, DQ007569–70) were used
for the outgroup. The neighbour-joining tree was constructed as described in Fig. 5. Haplogroups according to the classification of Kavar et al. (2002)
are represented by green letters ( ); haplogroups and subgroups according to the classifications of Vila et al. (2001) and Jansen et al. (2002) are
represented by blue letters ( ).
9T. Kavar, P. Dovč/ Livestock Science 116 (2008) 1–14
involved in horse domestication, because they were
extinct (Guthrie, 2006) before the process of domestica-
tion had begun.
The overall matrilineal diversity, of domestic- and
wild horses, is higher than the diversity of domestic
horses (up to 7.8%) (Fig. 7). This shows that many lines
did not survive two recent population bottlenecks: one
was the mass extinction of large-bodied animals around
8000 BC (due to the drastic climatic changes and over-
killing as reported by Clutton-Brock, 1999; Macfadden,
1992), and the second was horse domestication which
took place several thousand years later (Fig. 8). Only the
haplotypes (and their derivatives) which survived both
population bottlenecks can be found in domestic horses
(Kavar et al., 2002).
3.4. Phylogeographic structure
Two previously discussed models (I and II) of do-
mestication were suggested (Clutton-Brock, 1999). Ac-
cording to model I, wild horse populations from which
all domestic horses originated were highly variable but
not phylogeographically structured (Fig. 9). Model II
assumes that the wild horse populations were differ-
entiated due to the geographical cline or because they
belong to different subspecies.
Migrations (gene flow) are the main factor for the
lack of phylogeographic structure. Horses are very mo-
bile animals, so a high level of migration can be ex-
pected. Identical haplotypes that were found in two
wild horses from the late Pleistocene, one was from
Germany (DQ007590) and the other one from Siberia
(DQ007573), demonstrate good evidence of long-dis-
tance migrations. Evidence of long-distance migrations
after horse domestication can also be found in numerous
historical records. On the contrary, the absence of large
migrations of wild horses between the North and the
South of Western Europe during the Late Glacial period
(Bignon et al., 2005) has been suggested according to
the geometric morphometric study of metacarpals
and metatarsals from horse bone collections from three
distinct areas (Switzerland Plateau, Paris Basin, and
Charente, France). Specifically, multivariate analysis
of shape revealed that the between-group variability
dominated the within one, suggesting regional frag-
mentation of Late Glacial horses (Equus caballus
arcelini) in Western Europe rather than long-distance
Fig. 9. Models of horse domestication. The intermediate model of domestication (model IIa) is probably the best explanation of horse domestication.
Domestic horses (mares) having arisen from different wild stock distributed over a moderately extensive geographical region, large enough to have
contained within it considerable pre-existing haplotype diversity (Lister et al., 1998).
Fig. 8. Two recent population bottlenecks are the main reasons for
reduced matrilineal diversity in extant horses.
10 T. Kavar, P. Dovč/ Livestock Science 116 (2008) 1–14
The first evidence in favor of model II was presented
by Vila et al. (2001). Their observation that the se-
quences obtained from the late Pleistocene Alaskan
horses cluster in two discrete groups indicates that the
diversity of mtDNA lineages in single, natural popula-
tions might have been limited and single geographically
restricted populations would not suffice as founding
stock for the diversity observed in extant domestic
horses. Both Przewalski's horse haplotypes might
provide more evidence for phylogeographic structure
of wild horse populations, because both haplotypes
clustered in the C1 group (Kavar et al., 2002). In
addition, neither of the European wild horses (four from
Germany and one from Ireland) clustered in the North
American group haplotypes (Fig. 7). Therefore, wild
horse populations were phylogeographically structured,
at least on the level of continents. As expected, due to
the location of Siberia, haplotypes of Siberian horses
can be found among Eurasian haplotypes and North
American haplotypes (Fig. 7).
Phylogeographic structure among the Eurasian wild
horse populations has also been suggested. Jansen et al.
(2002) reported association between cluster C1 (C3a in
this paper) and European ponies (Exmoor, Fjord, Ice-
landic, and Scottish Highland); this cluster was geo-
graphically restricted to central Europe, the British Isles,
and Scandinavia, including Iceland. Jansen et al. (2002)
also observed that D1 (C4 group in this paper) is another
geographically striking mtDNA cluster with a frequency
maximum in Iberian (Andalusian and Lusitano) and
North African horses (Barbs). Only a small proportion
of the Arabs belong in this cluster (around 5%). They
concluded that this genetic result is in accordance with
the phenotype: Barb conformation is very similar to that
of Iberians, whereas the phenotype of the Arabs is quite
distinct from that of Iberians and Barbs. Further studies
of Iberian horse breeds (from both native groups:
Northern Iberian ponies or Celtic ponies, and Southern
Iberian horses such as Pura Raza Española, Andalusian,
Marismeño, Lusitano and Sorraia) and South American
horses, confirmed a high incidence of D1 haplogroup
haplotypes (Mirol et al., 2002; Lopes et al., 2005; Royo
et al., 2005; Luís et al., 2006b). High frequency of
Iberian origin haplotypes in the New World Horse
breeds, is consistent with historical documentation that
horses were brought to the American continent in 1493,
with the navigator Christopher Columbus and during the
subsequent Spanish colonization period. High incidence
of D1 haplogroup haplotypes also supports the histori-
cally well documented gene flow between horse popu-
lations of Iberian Peninsula and North Africa. However,
because the same haplotypes were detected in ancient
horses from Ural (DQ007571), the high incidence of D1
group in North African horses might be merely a result
of the majority of founder mares of Barbs having such
haplotypes, and not that the wild horses from North
Africa (perhaps the descendents of Caballine horse
(Equus algericus n. sp.), which was found in Algeria
(Bagtache et al., 1984)) had such haplotypes.
Recently, more evidence which supports biogeographic
pattering has been reported. McGahern et al. (2006a)
found highly significant associations of Eastern popula-
tions with the F haplogroup (C2 group in this paper). In
addition, 3 out of 9 samples of Bronze Age horses from the
Chifeng region ‘Inner Mongolia’, China clustered in this
haplogroup (two samples in F2 and one in F1) (Cai et al.,
2007). We could therefore assume that there was at least a
weak phylogeographic structure before domestication, and
that the extension of migration was very high before and
after the horse domestication. One of the explanations for
lack of consistency between mtDNA sequences and
breeds and/or geographical regions is also the use of the
same founder mares in different breeds (Luís et al., 2006a).
In their study Luís et al. (2006a) demonstrated how the use
of the same founders in different breeds might be one of
the explanations for the lack of consistency between
mtDNA, extant breeds and/or geographic regions.
There is also not any obvious correlation between
haplogroup and phenotypic groups of horses such as
native ponies, cold-blooded horses, hot-blooded breeds
etc. (Lister et al., 1998). Genetic data clearly confirm that
all the Pleistocene heavy horses (e.g., Equus germani-
cus) became extinct by the end of the last Ice Age and
therefore do not contribute to the gene pool of domestic
horses (Olsen, 2006).
The first publication concerning horse domestication
predicated on genetic data (Lister et al., 1998)suggested
an intermediate model of domestication (model IIa in
Fig. 9) as the most likely explanation. Domestic horses,
having arisen from wild stock distributed over a mode-
rately extensive geographical region, were at a population
size large enough to have contained within it considerable
pre-existing haplotype diversity. This remains as the best
explanation to date. The intermediate model is also in
agreement with the suggestions of Levine (2006),which
assume incorporation of wild horses from different popu-
lations until recently (see last paragraph in the chapter
1.4). At the end of this paper, a slightly different scenario
of the process of horse domestication is suggested.
4. Y chromosome markers
Y chromosome markers are specific for males and
therefore have been widely used to examine patrilineal
11T. Kavar, P. Dovč/ Livestock Science 116 (2008) 1–14
relationships. In domestic horses, stallion lines have
been studied by three, highly polymorphic sets of
markers: Y-specific sequence (Wallner et al., 2003),
microsatellite loci (Wallner et al., 2004) and SNPs
(Lindgren et al., 2004). A single haplotype was detected
by examining Y-specific sequences (Wallner et al.,
2003) and by genotyping six Y chromosome-specific
microsatellite loci in 49 male horses of 32 different
breeds (Wallner et al., 2004). A single haplotype was
also detected by examining SNP mutations in 14.3 kb of
noncoding Y chromosome sequence (Lindgren et al.,
2004) in 52 male horses from 15 different breeds. These
results suggest that a limited number of patrilines were
involved in horse domestication (Lindgren et al., 2004).
Przewalski's wild horse had different haplotypes
than the domestic horse. Y-specific sequences revealed a
single haplotype, which differed by two nucleotides
from the domestic horse haplotype (Wallner et al.,
2003). Genotyping of six microsatellite loci revealed
two haplotypes in Przewalski's wild horse, both of
which were distinct from the domestic horse haplotype,
although closely related, differing only at one locus
(Wallner et al., 2004). Lastly, the examination of SNPs
revealed a single haplotype, which differed from the
domestic horse haplotype by six nucleotides (Lindgren
et al., 2004).
Interestingly, in Przewalski's wild horse, there are no
marked differences between mtDNA and Y chromosome
marker results: two distinct haplotypes were found by
mtDNA analysis and two haplotypes by Y chromosome
marker analysis. The number of haplotypes detected in
the small population of this extant wild horse is low, but
relative to the population size is relatively high. It might
indicate high Y chromosome variability in the wild horse
population (approximately as high as the mtDNA vari-
ability). Due to domestication, the extension of a male
population bottleneck was therefore much greater than
the extension of female population bottleneck.
5. Reconstruction of the process of horse
All extant domestic horses can be traced back to one
founder stallion, or to closely related stallions having the
same Y haplotype (Lindgren et al., 2004). Therefore the
presence of low patrilineal and high matrilineal diversity
might suggest that the process of horse domestication (not
taming) started when the “appropriate male”was found or
obtained by selection. Perhaps it had some unusual special
characteristics which could accelerate the process of
domestication. We doubt that only one Y haplotype will
be found in present-day domestic horses, if there are no
important differences between the founder stallion/s and
the other stallions, which were not included in the
domestication. In the Eneolithic, tamed (and wild) mares
have been probably spread all over Eurasia, although the
number of animals was likely very low and populations
were limited to a restricted area (e.g., taming centers).
Both subspecies of wild horses (Tarpan and Przewalski's
wild horse) survived until recently (Groves, 1986).
During the further process of domestication, (tamed or
wild) mares from different regions (centers) were crossed
to stallions having more desirable characteristics. This
might have allowed rapid expansion of horse populations
as we find horse remains throughout Europe around 2000
BC (Clutton-Brock, 1999). We assume that mares from
different regions varied in their morphology due to the
adaptation to local environmental conditions. This could
be the main reason why differentiation into different
phenotypes such as small “Celtic”ponies of Britain,
Western Europe and Greece, and larger Arabian types of
horses from Scythia and the Russian steppes (Clutton-
Brock, 1999), occurred so rapidly.
The authors would like to acknowledge three anon-
ymous reviewers and Dr. Simona Sušnik for useful
comments. Thanks to the late Dr. Ann T. Bowling, who
encourage us at the beginning of our work on horse
Ahrens, E., Stranzinger, G., 2005. Comparative chromosomal studies
of E. caballus (ECA) and E. przewalskii (EPR) in a female F1
hybrid. J. Anim. Breed. Genet. 122, 97–102.
Anthony, D.W., 2007. The Horse, the Wheel, and Language: How
Bronze-Age Riders from the Eurasian Steppes Shaped the Modern
World. Princeton University Press, Princeton.
Anthony, D.W., Brown, D.R., 2000. Eneolithic horse exploitation in
the Eurasian steppes: diet, ritual and riding. Antiquity 74, 75–86.
Bagtache, B., Hadjouis, D., Eisenmann, V., Piveteau, J., 1984.
Présence d'un Equus caballin (E. algericus n. sp.) et d'une autre
espèce nouvelle d'Equus (E. melkiensis n. sp.) dans l'Atérien des
Allobroges, Algérie. Cr. Acad. Sci. II 298, 609–612.
Benirschke, K., Malouf, N., Low, R.J., Heck, H., 1965. Chromosome
Complement: Differences between Equus caballus and Equus
przewalskii, Poliakoff. Science 148, 382–383.
Bignon, O., Baylac, M., Vigne, J.-D., Eisenmann, V., 2005. Geometric
morphometrics and the population diversity of Late Glacial horses
in Western Europe (Equus caballus arcelini): phylogeographic and
anthropological implications. J. Archaeol. Sci. 32, 375–391.
Bowling, A.T., Del Valle, A., Bowling, M., 2000. A pedigree-based
study of mitochondrial D-loop DNA sequence variation among
Arabian horses. Anim. Genet. 31, 1–7.
Brown, D.R., Anthony, D.W., 1998. Bit wear, horseback riding, and
the Botai site in Kazakstan. J. Archaeol. Sci. 25, 331–347.
12 T. Kavar, P. Dovč/ Livestock Science 116 (2008) 1–14
Cai, D.W., Han, L., Xie, C.Z., Li, S.N., Zhou, H., Zhu, H., 2007.
Mitochondrial DNA analysis of Bronze Age horses recovered
from Chifeng region ‘Inner Mongolia’China. Prog. Nat. Sci. 17,
Clutton-Brock, J., 1999. A natural history of domesticated mammals.
Cambridge University Press, Cambridge.
Di Bernardo, G., Galderisi, U., Del Gaudio, S., D'Aniello, A., Lanave,
C., De Robertis, M.T., Cascino, A., Cipollaro, M., 2004. Genetic
characterization of Pompeii and Herculaneum Equidae buried by
Vesuvius in 79 AD. J. Cell. Physiol. 199, 200–205.
Eisenmann, V., 1996. Quaternary horses: possible candidates to domes-
tication. In: Peretto, C., Giunchi, C. (Eds.), Proc. XIII Congr. Inter-
national Union of Prehistoric and Protohistoric Sciences, pp. 27–36.
Felsenstein, J., 2005. PHYLIP (Phylogeny Inference Package) version
Groves, C.P., 1986. The taxonomy, distribution and adaptations of
recent Equids. In: Meadow, R.H., Uerpmann, H.-P. (Eds.), Equids
in the Ancient World. Dr Ludwig Reichert. Verlag, Wiesbaden,
Guthrie, D., 2003. Rapid body size decline in Alaskan Pleistocene
horses before extinction. Nature 426, 169–171.
Guthrie, D., 2006. New carbon dates link climatic change with human
colonization and Pleistocene extinctions. Nature 441, 207–209.
Higgins, D., Thompson, J., Gibson, T., Thompson, J.D., Higgins, D.G.,
Gibson, T.J., 1994. CLUSTAL W: improving the sensitivity of
progressive multiple sequence a lignment through sequence weighting,
position-specific gap penalties and weight matrix choice. Nucleic
Acids Res. 22, 4673–4680 (ClustalW WWW Service at the European
Bioinformatics Institute) http://www.ebi.ac.uk/clustalw.
Ishida, N., Hasegawa, T., Takeda, K., Sakagami, M., Onishi, A.,
Inumaru, S., Komatsu, M., Mukoyama, H., 1994. Polymorphic
sequence in the D-loop region of equine mitochondrial DNA.
Anim. Genet. 25, 215–221.
Jansen, T., Forster, P., Levine, M.A., Oelke,H., Hurles, M., Renfrew, C.,
Weber, J., Olek, K., 2002. Mitochondrial DNA and the Origins of the
Domestic Horse. Proc. Natl. Acad. Sci. U.S.A. 99, 10905–10910.
Jung, Y.H., Han, S.H., Shin, T., Oh, M.Y., 2002. Genetic Character-
ization of Horse Bone Excavated from the Kwakji Archaeological
Site, Jeju, Korea. Mol. Cells 14, 224–230.
Kavar, T., Habe, F., Brem, G., Dovč, P., 1999. Mitochondrial D-loop
sequence variation among the 16 maternal lines of the Lipizzan
horse breed. Anim. Genet. 30, 423–430.
Kavar, T., Brem, G., Habe, F., Sölkner, J., Dovč, P., 2002. History of
Lipizzan horse maternal lines as revealed by mtDNA analysis.
Genet. Sel. Evol. 34, 1–14.
Kavar, T., Habe, F., Dovč, P., 2004. Rodovi lipicancev slovenske reje
glede na haplotip mitohondrijske DNK. Acta agr. slov. 84, 131–139
(in Slovene, with English abstract).
Keyser-Tracqui, C., Blandin-Frappin, P., Francfort, H.P., Ricaut, F.X.,
Lepetz, S., Crubezy, E., Samashev, Z., Ludes, B., 2005. Mitochondrial
DNA analysis of horses recovered from a frozen tomb (Berel site,
Kazakhstan, 3rd Century BC). Anim. Genet. 36, 203–209.
Kim, K.I., Yang, Y.H., Lee, S.S., Park, C., Ma, R., Bouzat, J.L., Lewin,
H.A., 1999. Phylogenetic relationships of Cheju horses to other
horse breeds as determined by mtDNA D-loop sequence
polymorphism. Anim. Genet. 30, 102–108.
Kimura, M., 1980. A simple method for estimating evolutionary rates
of base substitutions through comparative studies of nucleotide
sequences. J. Mol. Evol. 16, 111–120.
Koulischer, L., Frechkop, S., 1966. Chromosome complement: a
fertile hybrid between Equus priewalskii and Equus caballus.
Science 151, 93–95.
Levine, M.A., 2005. Domestication and early history of the horse. In:
Mills, D.M., McDonnell, S.M. (Eds.), The Domestic Horse: The
Origins, Development, and Management of Its Behaviour. Cam-
bridge University Press, Cambridge, pp. 5–22.
Levine, M.A., 2006. mtDNA and horse domestication: the archae-
ologist's cut. In: Mashkour, M. (Ed.), Equids in Time and Space.
Proc. 9th Int. Conf. of Archaeozoology. Oxbow Books, Oxford,
Lindgren, G., Backstrom, N., Swinburne, J., Hellborg, L., Einarsson,
A., Sandberg, K., Cothran, G., Vila, C., Binns, M., Ellegren, H.,
2004. Limited number of patrilines in horse domestication. Nat.
Genet. 36, 335–336.
Lister, A.M., Kadwell, M., Kaagan, L.M., Jordan, W.C., Richards,
M.B., Stanley, H.E., 1998. Ancient and modern DNA in a study
of horse domestication. Anc. Biomol. 2, 267–280.
Lopes, M.S., Mendonça, D., Cymbron, T., Valera, M., Costa-Ferreira,
J., da Câmara Machado, A., 2005. The Lusitano horse maternal
lineage based on mitochondrial D-loop sequence variation. Anim.
Genet. 36, 196–202.
Luís, C., Bastos-Silveira, C., Cothran, E.G., Oom, M.M., 2002. Variation
in the mitochondrial control region sequence between the two ma-
ternal linesof the Sorraia horse breed. Genet.Mol. Biol. 25, 309–311.
Luís, C., Bastos-Silveira, C., Costa-Ferreira, J., Cothran, E.G., Oom,
M.M., 2006a. A lost Sorraia maternal lineage found in the Lusitano
horse breed. J. Anim. Breed. Genet. 123, 399–402.
Luís, C., Bastos-Silveira, C., Cothran, E.G., Oom, M.M., 2006b. Iberian
Origins of New World Horse Breeds. J. Heredity 97, 107–113.
MacFadden, B.J., 1992. Fossil horses: systematics, paleobiology, and
evolution of the family Equidae. Cambridge University Press,
McGahern, A., Bower, M.A., Edwards, C.J., Brophy, P.O., Sulimova,
MacHugh, D.E., Hill, E.W., 2006a. Evidence for biogeographic
patterning of mitochondrial DNA sequences in Eastern horse
populations. Anim. Genet. 37, 494–497.
McGahern, A.M., Edwards, C.J., Bower, M.A., Heffernan, A., Park, S.D.,
Brophy, P.O., Bradley, D.G., MacHugh, D.E., Hill, E.W., 2006b.
Mitochondrial DNA sequence diversity in extant Irish horse
populations and in ancient horses. Anim. Genet. 37, 498–502.
Mirol, P.M., Peral García, P., Vega-Pla, J.L., Dulot, F.N., 2002.
Phylogenetic relationships of Argentinean Creole horses and other
South American and Spanish breeds inferred from mitochondrial
DNA sequences. Anim. Genet. 33, 356–363.
Nürnberg, H., 1993. Der Lipizzaner: mit einem Anhang über den
Kladruber. Westarp Wissenschaften, Magdenburg.
Oakenfull, E.A., Ryder, O.A., 1998. Mitochondrial control region and
12S rRNA variation in Przewalski's horse (Equus przewalskii).
Anim. Genet. 29, 456–459.
Olsen, S.L., 2006. EarlyHorse Domesticationon the Eurasian Steppe.In:
Zeder, M.A., Bradley, D.G., Emshwiller, E., Smith, B.D. (Eds.),
Documenting Domestication. New Genetic and Archaeological
Paradigms. University Pressesof California, Princeton, pp.245–272.
Royo, L.J., Álvarez, I., Beja-Pereira, A., Molina, A., Fernández, I.,
Jordana, J., Gómez, E., Gutiérrez, J.P., Goyache, F., 2005. The
origins of the Iberian horses assessed via mitochondrial DNA.
J. Heredity 96, 663–669.
Ryder, O.A., Epel, N.C., Benirschke, K., 1978. Chromosome banding
studies of the equidae. Cytogenet. Cell Genet. 20, 323–350.
Saitou, N., Nei, M., 1987. The neighbor-joining method: a new method
for reconstructing phylogenetic trees. Mol. Biol. Evol. 4, 406–425.
Stiff, A.R., Capo, R.C., Gardiner, J.B., Olsen, S.L., Rosenmeier, M.F.,
2006. Geochemical evidence of possible horse domestication at the
13T. Kavar, P. Dovč/ Livestock Science 116 (2008) 1–14
copper age Botai settlement of Krasnyi yar, Kazakhstan. The
Geological Society of America. 2006 Philadelphia Annu. Meet.
Uerpmann, H.-P., 1990. Die Domestikation des Pferdes im Chalko-
lithikum West-und Mitteleuropas. Madrid. Mitt. 31, 109–153.
Vila, C., Leonard, J.A., Gotherstrom, A., Marklund, S., Sandberg, K.,
Liden, K., Wayne,R.K., Ellegren, H., 2001. Widespread origins of
domestic horse lineages. Science 291, 474–477.
Wallner, B., Brem, G., Muller, M., Achmann, R., 2003. Fixed nucleotide
differences on the Y chromosome indicate clear divergence between
Equus przewalskii and Equus caballus. Anim. Genet. 34, 453–456.
Wallner, B., Piumi, F., Brem, G., Muller, M., Achmann, R., 2004.
Isolation of Y chromosome-specific microsatellites in the horse
and cross-species amplification in the genus Equus. J. Heredity 95,
Weinstock, J., Willerslev, E., Sher, A., Tong, W., Ho, S.Y., Rubenstein,
D., Storer, J., Burns, J., Martin, L., Bravi, C., Prieto, A., Froese, D.,
Scott, E., Xulong, L., Cooper, A., 2005. Evolution, systematics,
and phylogeography of pleistocene horses in the new world: a
molecular perspective. PLoS Biol. 3, e241.
Xu, X., Arnason, U., 1994. The complete mitochondrial DNA
sequence of the horse, Equus caballus: extensive heteroplasmy
of the control region. Gene 148, 357–362.
Xu, X., Gullberg, A., Arnason, U., 1996. The complete mitochondrial
DNA (mtDNA) of the donkey and mtDNA comparisons among
four closely related mammalian species-pairs. J. Mol. Evol. 43,
14 T. Kavar, P. Dovč/ Livestock Science 116 (2008) 1–14