A number of hypotheses exist on many of the key issues regarding the domestication of the horse. Although horses appeared in Paleolithic cave art as early as 30,000 BCE, these were wild horses and were probably hunted for meat.
How and when horses became domesticated is disputed. The clearest evidence of early use of the horse as a means of transport is from chariot burials dated c. 2000 BCE. However, an increasing amount of evidence supports the hypothesis that horses were domesticated in the Eurasian Steppes approximately 3500 BCE; recent discoveries in the context of the Botai culture suggest that Botai settlements in the Akmola Province of Kazakhstan are the location of the earliest domestication of the horse.
The date of the domestication of the horse depends to some degree upon the definition of "domestication". Some zoologists define "domestication" as human control over breeding, which can be detected in ancient skeletal samples by changes in the size and variability of ancient horse populations. Other researchers look at broader evidence, including skeletal and dental evidence of working activity; weapons, art, and spiritual artifacts; and lifestyle patterns of human cultures. There is also evidence that horses were kept as meat animals prior to being trained as working animals.
Attempts to date domestication by genetic study or analysis of physical remains rests on the assumption that there was a separation of the genotypes of domesticated and wild populations. Such a separation appears to have taken place, but dates based on such methods can only produce an estimate of the latest possible date for domestication without excluding the possibility of an unknown period of earlier gene-flow between wild and domestic populations (which will occur naturally as long as the domesticated population is kept within the habitat of the wild population). Further, all modern horse populations retain the ability to revert to a feral state, and all feral horses are of domestic types; that is, they descend from ancestors that escaped from captivity.
Whether one adopts the narrower zoological definition of domestication or the broader cultural definition that rests on an array of zoological and archaeological evidence affects the time frame chosen for domestication of the horse. The date of 4000 BCE is based on evidence that includes the appearance of dental pathologies associated with bitting, changes in butchering practices, changes in human economies and settlement patterns, the depiction of horses as symbols of power in artifacts, and the appearance of horse bones in human graves. On the other hand, measurable changes in size and increases in variability associated with domestication occurred later, about 2500–2000 BCE, as seen in horse remains found at the site of Csepel-Haros in Hungary, a settlement of the Bell Beaker culture.
Use of horses spread across Eurasia for transportation, agricultural work and warfare. Horses and mules in agriculture used a breastplate type harness or a yoke more suitable for oxen, which was not as efficient at utilizing the full strength of the animals as the later-invented padded horse collar that arose several millennia later.
A 2005 study analyzed the mitochondrial DNA (mtDNA) of a worldwide range of equids, from 53,000-year-old fossils to contemporary horses. Their analysis placed all equids into a single clade, or group with a single common ancestor, consisting of three genetically divergent species: Hippidion, the New World stilt-legged horse, and the true horse. The true horse, which ranged from western Europe to eastern Beringia, included prehistoric horses and the Przewalski's Horse, as well as what is now the modern domestic horse, belonged to a single Holarctic species. A more detailed analysis of the true horses grouped them into two major clades. One of these clades, which seemed to have been restricted to North America, is now extinct. The other clade was broadly distributed from North America to central Europe, north and south of Pleistocene ice sheets. It became extinct in Beringia around 14,200 years ago, and in the rest of the Americas around 10,000 years ago. This clade survived in Eurasia, however, and it is from these horses which all domestic horses appear to have descended. These horses showed little phylogeographic structure, probably reflecting their high degree of mobility and adaptability.
Therefore, the domestic horse today is classified as Equus ferus caballus. No genetic originals of native wild horses currently exist, other than the never-domesticated Przewalski's Horse. The Przewalski has 66 chromosomes, however, as opposed to 64 among modern domesticated horses, and their Mitochondrial DNA (mtDNA) forms a distinct cluster. Genetic evidence suggests that modern Przewalski's horses are descended from a distinct regional gene pool in the eastern part of the Eurasian steppes, not from the same genetic group that gave rise to modern domesticated horses. Nevertheless, evidence such as the cave paintings of Lascaux suggests that the ancient wild horses that some researchers now label the "Tarpan subtype" probably resembled Przewalski horses in their general appearance: big heads, dun coloration, thick necks, stiff upright manes, and relatively short, stout legs.
The horses of the Ice Age were hunted for meat in Europe and across the Eurasian steppes and in North America by early modern humans. Numerous kill sites exist and many cave paintings in Europe indicate what they looked like. Many of these Ice Age subspecies died out during the rapid climate changes associated with the end of the last Ice Age or were hunted out by humans, particularly in North America, where the horse became completely extinct.
Classification based on body types and conformation, absent the availability of DNA for research, once suggested that there were roughly four basic wild prototypes, thought to have developed with adaptations to their environment prior to domestication. There were competing theories; some argued that the four prototypes were separate species or subspecies, while others suggested that the prototypes were physically different manifestations of the same species. However, more recent study indicates that there was only one wild species and all different body types were entirely a result of selective breeding or landrace adaptation after domestication. Either way, the most common theories of prototypes from which all modern breeds are thought to have developed suggests than in addition to the so-called Tarpan subtype, there were the following base prototypes:
Only two never-domesticated "wild" groups survived into historic times, Przewalski's horse (Equus ferus przewalski), and the Tarpan (Equus ferus ferus). The Tarpan became extinct in the late 19th century and Przewalski's horse is endangered; it became extinct in the wild during the 1960s, but was re-introduced in the late 1980s to two preserves in Mongolia. Although researchers such as Marija Gimbutas theorized that the horses of the Chalcolithic period were Przewalski's, more recent genetic studies indicate that Przewalski's horse is not an ancestor to modern domesticated horses. Other subspecies of Equus ferus appear to have existed and could have been the stock from which domesticated horses are descended.
A 2014 study compared DNA from ancient horse bones that predated domestication and compared them to DNA of modern horses, discovering 125 genes that correlated to domestication. Some were physical, affecting muscle and limb development, cardiac strength and balance. Others were linked to cognitive function and most likely were critical to the taming of the horse, including social behavior, learning capabilities, fear response, and agreeableness. The DNA used in this study came from horse bones 16,000 to 43,000 years ago, and therefore the precise changes that occurred at the time of domestication have yet to be sequenced.
The domestication of stallions and mares can be analyzed separately by looking at those portions of the DNA that are passed on exclusively along the maternal (mitochondrial DNA or mtDNA) or paternal line (Y-chromosome or Y-DNA). DNA studies indicate that there may have been multiple domestication events for mares, as the number of female lines required to account for the genetic diversity of the modern horse suggests a minimum of 77 different ancestral mares, divided into 17 distinct lineages. On the other hand, genetic evidence with regard to the domestication of stallions points at a single domestication event for a limited number of stallions combined with repeated restocking of wild females into the domesticated herds.
A study published in 2012 that performed genomic sampling on 300 work horses from local areas as well as a review of previous studies of archaeology, mitochondrial DNA, and Y-DNA suggested that horses were originally domesticated in the western part of the Eurasian steppe. Both domesticated stallions and mares spread out from this area, and then additional wild mares were added from local herds; wild mares were easier to handle than wild stallions. Most other parts of the world were ruled out as sites for horse domestication, either due to climate unsuitable for an indigenous wild horse population or no evidence of domestication.
Genes located on the Y-chromosome are inherited only from sire to its male offspring and these lines show a very reduced degree of genetic variation (aka genetic homogeneity) in modern domestic horses, far less than expected based on the overall genetic variation in the remaining genetic material. This indicates that a relatively few stallions were domesticated, and that it is unlikely that many male offspring originating from unions between wild stallions and domestic mares were included in early domesticated breeding stock.
Genes located in the mitochondrial DNA are passed on along the maternal line from the mother to her offspring. Multiple analyses of the mitochondrial DNA obtained from modern horses as well as from horse bones and teeth from archaeological and palaeological finds consistently shows an increased genetic diversity in the mitochondrial DNA compared to the remaining DNA, showing that a large number of mares has been included into the breeding stock of the originally domesticated horse. Variation in the mitochondrial DNA is used to determine so-called haplogroups. A haplogroup is a group of closely related haplotypes that share the same common ancestor. In horses, seven main haplogroups are recognized (A-G), each with several subgroups. Several haplogroups are unequally distributed around the world, indicating the addition of local wild mares to the domesticated stock. One of these haplotypes (Lusitano group C) is exclusively found on in Iberian Peninsula, leading to a hypothesis that the Iberian peninsula or North Africa was an independent origin for domestication of the horse. However, until there is additional analysis of nuclear DNA and a better understanding of the genetic structure of the earliest domestic herds, this theory cannot be confirmed or refuted. It remains possible that a second, independent, domestication site might exist but, as of 2012, research has neither confirmed nor disproven that hypothesis.
Even though horse domestication became widespread in a short period of time, it is still possible that domestication began with a single culture, which passed on techniques and breeding stock. It is possible that the two "wild" subspecies remained when all other groups of once-"wild" horses died out because all others had been, perhaps, more suitable for taming by humans and the selective breeding that gave rise to the modern domestic horse.
Evidence for the domestication of the horse comes from three kinds of sources: 1) changes in the skeletons and teeth of ancient horses; 2) changes in the geographic distribution of ancient horses, particularly the introduction of horses into regions where no wild horses had existed; and 3) archaeological sites containing artifacts, images, or evidence of changes in human behavior connected with horses.
Archaeological evidence includes horse remains interred in human graves; changes in the ages and sexes of the horses killed by humans; the appearance of horse corrals; equipment such as bits or other types of horse tack; horses interred with equipment intended for use by horses, such as chariots; and depictions of horses used for riding, driving, draught work, or symbols of human power.
Few of these categories, taken alone, provide irrefutable evidence of domestication, but combined add up to a persuasive argument.
The least ancient, but most persuasive, evidence of domestication comes from sites where horse leg bones and skulls, probably originally attached to hides, were interred with the remains of chariots in at least 16 graves of the Sintashta and Petrovka cultures. These were located in the steppes southeast of the Ural Mountains, between the upper Ural and upper Tobol Rivers, a region today divided between southern Russia and northern Kazakhstan. Petrovka was a little later than and probably grew out of Sintashta, and the two complexes together spanned about 2100–1700 BCE. A few of these graves contained the remains of as many as eight sacrificed horses placed in, above, and beside the grave.
In all of the dated chariot graves, the heads and hooves of a pair of horses were placed in a grave that once contained a chariot. Evidence of chariots in these graves was inferred from the impressions of two spoked wheels set in grave floors 1.2–1.6m apart; in most cases the rest of the vehicle left no trace. In addition a pair of disk-shaped antler "cheekpieces," an ancient predecessor to a modern bit shank or bit ring, were placed in pairs beside each horse head-and-hoof sacrifice. The inner faces of the disks had protruding prongs or studs that would have pressed against the horse's lips when the reins were pulled on the opposite side. Studded cheekpieces were a new and fairly severe kind of control device that appeared simultaneously with chariots.
All of the dated chariot graves contained wheel impressions, horse bones, weapons (arrow and javelin points, axes, daggers, or stone mace-heads), human skeletal remains, and cheekpieces. Because they were buried in teams of two with chariots and studded cheekpieces, the evidence is extremely persuasive that these steppe horses of 2100–1700 BCE were domesticated. Shortly after the period of these burials, the expansion of the domestic horse throughout Europe was little short of explosive. In the space of possibly 500 years, there is evidence of horse-drawn chariots in Greece, Egypt, and Mesopotamia. By another 500 years, the horse-drawn chariot had spread to China.
Some researchers do not consider an animal to be "domesticated" until it exhibits physical changes consistent with selective breeding, or at least having been born and raised entirely in captivity. Until that point, they classify captive animals as merely "tamed". Those who hold to this theory of domestication point to a change in skeletal measurements was detected among horse bones recovered from middens dated about 2500 BCE in eastern Hungary in Bell-Beaker sites, and in later Bronze Age sites in the Russian steppes, Spain, and eastern Europe. Horse bones from these contexts exhibited an increase in variability, thought to reflect the survival under human care of both larger and smaller individuals than appeared in the wild; and a decrease in average size, thought to reflect penning and restriction in diet. Horse populations that showed this combination of skeletal changes probably were domesticated. Most evidence suggests that horses were increasingly controlled by humans after about 2500 BCE. However, more recently there have been skeletal remains found at a site in Kazakhstan which display the smaller, more slender limbs characteristic of corralled animals, dated to 3500 BCE.
Some of the most intriguing evidence of early domestication comes from the Botai culture, found in northern Kazakhstan. The Botai culture was a culture of foragers who seem to have adopted horseback riding in order to hunt the abundant wild horses of northern Kazakhstan between 3500–3000 BCE. Botai sites had no cattle or sheep bones; the only domesticated animals, in addition to horses, were dogs. Botai settlements in this period contained between 50–150 pit houses. Garbage deposits contained tens to hundreds of thousands of discarded animal bones, 65% to 99% of which had come from horses. Also, there has been evidence found of horse milking at these sites, with horse milk fats soaked into pottery shards dating to 3500 BCE. Earlier hunter-gatherers who lived in the same region had not hunted wild horses with such success, and lived for millennia in smaller, more shifting settlements, often containing less than 200 wild animal bones.
Entire herds of horses were slaughtered by the Botai hunters, apparently in hunting drives. The adoption of horseback riding might explain the emergence of specialized horse-hunting techniques and larger, more permanent settlements. Domesticated horses could have been adopted from neighboring herding societies in the steppes west of the Ural Mountains, where the Khvalynsk culture had herds of cattle and sheep, and perhaps had domesticated horses, as early as 4800 BCE.
Other researchers have argued that all of the Botai horses were wild, and that the horse-hunters of Botai hunted wild horses on foot. As evidence, they note that zoologists have found no skeletal changes in the Botai horses that indicate domestication. Moreover, because they were hunted for food, the majority of the horse remains found in Botai-culture settlements indeed probably were wild. On the other hand, any domesticated riding horses were probably the same size as their wild cousins and cannot now be distinguished by bone measurements. They also note that the age structure of the horses slaughtered at Botai represents a natural demographic profile for hunted animals, not the pattern expected if they were domesticated and selected for slaughter. However, these arguments were published prior to the discovery of a corral at Krasnyi Yar and mats of horse-dung at two other Botai sites.
The presence of bit wear suggest that a horse was ridden or driven, and the earliest of such evidence from a site in Kazakhstan dates to 3500 BCE. Because horses can be ridden and controlled without bits by using a noseband or a hackamore, and such tools are used even today, the absence of bit wear on horse teeth is not conclusive evidence against domestication, but such materials do not produce significant physiological changes nor are they apt to be preserved for millennia.
The regular use of a bit to control a horse can create wear facets or bevels on the anterior corners of the lower second premolars. The corners of the horse's mouth normally keep the bit on the "bars" of the mouth, an interdental space where there are no teeth, forward of the premolars. The bit must be manipulated by a human or the horse must move it with its tongue for it to touch the teeth. Wear can be caused by the bit abrading the front corners of the premolars if the horse grasps and releases the bit between its teeth; other wear can be created by the bit striking the vertical front edge of the lower premolars, due to very strong pressure from a human handler.
Modern experiments showed that even organic bits of rope or leather can create significant wear facets, and also showed that facets 3 mm deep or more do not appear on the premolars of wild horses. However, other researchers disputed both conclusions.
Wear facets of 3 mm or more also were found on seven horse premolars in two sites of the Botai, Botai and Kozhai 1, dated about 3500–3000 BCE. The Botai culture premolars are the earliest reported multiple examples of this dental pathology in any archaeological site, and preceded any skeletal change indicators by 1,000 years. While wear facets more than 3 mm deep were discovered on the lower second premolars of a single stallion from Dereivka in Ukraine, an Eneolithic settlement dated about 4000 BCE, dental material from one of the worn teeth later produced a radiocarbon date of 700–200 BCE, indicating that this stallion was actually deposited in a pit dug into the older Eneolithic site during the Iron Age.
Soil scientists working with Sandra Olsen of the Carnegie Museum of Natural History at the Chalcolithic (also called Eneolithic, or "Copper Age") settlements of Botai and Krasnyi Yar in northern Kazakhstan found layers of horse dung, discarded in unused house pits in both settlements. The collection and disposal of horse dung suggests that horses were confined in corrals or stables. An actual corral, dated to 3500–3000 BCE was identified at Krasnyi Yar by a pattern of post holes for a circular fence, with the soils inside the fence yielding ten times more phosphorus than the soils outside. The phosphorus could represent the remains of manure.
The appearance of horse remains in human settlements in regions where they had not previously been present is another indicator of domestication. Although images of horses appear as early as the Upper Paleolithic period in places such as the caves of Lascaux, France, suggesting that wild horses lived in regions outside of the Eurasian steppes prior to domestication and may have even been hunted by early humans, concentration of remains suggests animals being deliberately captured and contained, an indicator of domestication, at least for food, if not necessarily use as a working animal.
Around 3500–3000 BCE, horse bones began to appear more frequently in archaeological sites beyond their center of distribution in the Eurasian steppes and were seen in central Europe, the middle and lower Danube valley, and the North Caucasus and Transcaucasia. Evidence of horses in these areas had been rare before, and as numbers increased, larger animals also began to appear in horse remains. This expansion in range was contemporary with the Botai culture, where there are indications that horses were corralled and ridden. This does not necessarily mean that horses were first domesticated in the steppes, but the horse-hunters of the steppes certainly pursued wild horses more than in any other region. This geographic expansion is interpreted by many zoologists as an early phase in the spread of domesticated horses.
European wild horses were hunted for up to 10% of the animal bones in a handful of Mesolithic and Neolithic settlements scattered across Spain, France, and the marshlands of northern Germany, but in many other parts of Europe, including Greece, the Balkans, the British Isles, and much of central Europe, horse bones do not occur or occur very rarely in Mesolithic, Neolithic or Chalcolithic sites. In contrast, wild horse bones regularly exceeded 40% of the identified animal bones in Mesolithic and Neolithic camps in the Eurasian steppes, west of the Ural Mountains.
Horse bones were rare or absent in Neolithic and Chalcolithic kitchen garbage in western Turkey, Mesopotamia, most of Iran, South and Central Asia, and much of Europe. While horse bones have been identified in Neolithic sites in central Turkey, all equids together totaled less than 3% of the animal bones. Within this three percent, horses were less than 10%, with 90% or more of the equids represented by onagers (Equus hemionus) or another ass-like equid that later became extinct, Equus hydruntinus. Onagers were the most common native wild equids of the Near East. They were hunted in Syria, Anatolia, Mesopotamia, Iran, and Central Asia; and domesticated asses (Equus asinus) were imported into Mesopotamia, probably from Egypt, but wild horses apparently did not live there.
In Northern Caucasus, the Maikop culture settlements and burials of c. 3300 BC contain both horse bones and images of horses. A frieze of nineteen horses painted in black and red colors is found in one of the Maikop graves. The widespread appearance of horse bones and images in Maikop sites suggest to some observers that horseback riding began in the Maikop period.
Later, images of horses, identified by their short ears, flowing manes, and tails that bushed out at the dock, began to appear in artistic media in Mesopotamia during the Akkadian period, 2300-2100 BCE. The word for "horse", literally translated as ass of the mountains, first appeared in Sumerian documents during the Third dynasty of Ur, about 2100-2000 BCE. The kings of the Third Dynasty of Ur apparently fed horses to lions for royal entertainment, perhaps indicating that horses were still regarded as more exotic than useful, but King Shulgi, about 2050 BCE, compared himself to "a horse of the highway that swishes its tail", and one image from his reign showed a man apparently riding a horse at full gallop. Horses were imported into Mesopotamia and the lowland Near East in larger numbers after 2000 BCE in connection with the beginning of chariot warfare.
A further expansion, into the lowland Near East and northwestern China, also happened around 2000 BCE, again apparently in conjunction with the chariot. Although Equus bones of uncertain species are found in some Late Neolithic sites in China dated before 2000 BCE, Equus caballus or Equus ferus bones first appeared in multiple sites and in significant numbers in sites of the Qijia and Siba cultures, 2000-1600 BCE, in Gansu and the northwestern provinces of China. The Qijia culture was in contact with cultures of the Eurasian steppes, as shown through similarities between Qijia and Late Bronze Age steppe metallurgy, so it was probably through these contacts that domesticated horses first became frequent in northwestern China.
In 2008, archaeologists announced the discovery of rock art in Somalia's northern Dhambalin region, which the researchers suggest is one of the earliest known depictions of a hunter on horseback. The rock art is in the Ethiopian-Arabian style, dated to 1000 to 3000 BCE.
About 4200-4000 BCE, more than 500 years before the geographic expansion evidenced by the presence of horse bones, new kinds of graves, named after a grave at Suvorovo, appeared north of the Danube delta in the coastal steppes of Ukraine near Izmail. Suvorovo graves were similar to and probably derived from earlier funeral traditions in the steppes around the Dnieper River. Some Suvorovo graves contained polished stone mace-heads shaped like horse heads and horse tooth beads. Earlier steppe graves also had contained polished stone mace-heads, some of them carved in the shape of animal heads. Settlements in the steppes contemporary with Suvorovo, such as Sredni Stog II and Dereivka on the Dnieper River, contained 12%-52% horse bones.
When Suvorovo graves appeared in the Danube delta grasslands, horse-head maces also appeared in some of the indigenous farming towns of the Tripolye and Gumelnitsa cultures in present-day Romania and Moldova, near the Suvorovo graves. These agricultural cultures had not previously used polished-stone maces, and horse bones were rare or absent in their settlement sites. Probably their horse-head maces came from the Suvorovo immigrants. The Suvorovo people in turn acquired many copper ornaments from the Tripolye and Gumelnitsa towns. After this episode of contact and trade, but still during the period 4200-4000 BCE, about 600 agricultural towns in the Balkans and the lower Danube valley, some of which had been occupied for 2000 years, were abandoned. Copper mining ceased in the Balkan copper mines, and the cultural traditions associated with the agricultural towns were terminated in the Balkans and the lower Danube valley. This collapse of "Old Europe" has been attributed to the immigration of mounted Indo-European warriors. The collapse could have been caused by intensified warfare, for which there is some evidence; and warfare could have been worsened by mounted raiding; and the horse-head maces have been interpreted as indicating the introduction of domesticated horses and riding just before the collapse.
However, mounted raiding is just one possible explanation for this complex event. Environmental deterioration, ecological degradation from millennia of farming, and the exhaustion of easily mined oxide copper ores also are cited as causal factors.
Perforated antler objects discovered at Dereivka and other sites contemporary with Suvorovo have been identified as cheekpieces or ‘’psalia’’ for horse bits. This identification is no longer widely accepted, as the objects in question have not been found associated with horse bones, and could have had a variety of other functions. However, through studies of microscopic wear, it has been established that many of the bone tools at Botai were used to smooth rawhide thongs, and rawhide thongs might have been used to manufacture of rawhide cords and ropes, useful for horse tack. Similar bone thong-smoothers are known from many other steppe settlements, but it cannot be known how the thongs were used. The oldest artifacts clearly identified as horse tack—bits, bridles, cheekpieces, or any other kind of horse gear—are the antler disk-shaped cheekpieces associated with the invention of the chariot, at the Sintashta-Petrovka sites.
The oldest possible archaeological indicator of a changed relationship between horses and humans is the appearance about 4800-4400 BCE of horse bones and carved images of horses in Chalcolithic graves of the early Khvalynsk culture and the Samara culture in the middle Volga region of Russia. At the Khvalynsk cemetery near the town of Khvalynsk, 158 graves of this period were excavated. Of these, 26 graves contained parts of sacrificed domestic animals, and additional sacrifices occurred in ritual deposits on the original ground surface above the graves. Ten graves contained parts of lower horse legs; two of these also contained the bones of domesticated cattle and sheep. At least 52 domesticated sheep or goats, 23 domesticated cattle, and 11 horses were sacrificed at Khvalynsk. The inclusion of horses with cattle and sheep and the exclusion of obviously wild animals together suggest that horses were categorized symbolically with domesticated animals.
At S’yezzhe, a contemporary cemetery of the Samara culture, parts of two horses were placed above a group of human graves. The pair of horses here was represented by the head and hooves, probably originally attached to hides. The same ritual—using the hide with the head and lower leg bones as a symbol for the whole animal—was used for many domesticated cattle and sheep sacrifices at Khvalynsk. Horse images carved from bone were placed in the above-ground ochre deposit at S’yezzhe and occurred at several other sites of the same period in the middle and lower Volga region. Together these archaeological clues suggest that horses had a symbolic importance in the Khvalynsk and Samara cultures that they had lacked earlier, and that they were associated with humans, domesticated cattle, and domesticated sheep. Thus, the earliest phase in the domestication of the horse might have begun during the period 4800-4400 BCE.
Equidae died out in the Western Hemisphere at the end of the last glacial period. A question raised is why and how horses avoided this fate on the Eurasian continent. It has been theorized that domestication saved the species. While the environmental conditions for equine survival in Europe were somewhat more favorable in Eurasia than in the Americas, the same stressors that led to extinction for the Mammoth had an effect upon horse populations. Thus, some time after 8000 BCE, the approximate date of extinction in the Americas, humans in Eurasia may have begun to keep horses as a livestock food source, and by keeping them in captivity, may have helped to preserve the species. Horses also fit the six core criteria for livestock domestication, and thus, it could be argued, "chose" to live in close proximity to humans.
One model of horse domestication starts with individual foals being kept as pets while the adult horses were slaughtered for meat. Foals are relatively small and easy to handle. Horses behave as herd animals and need companionship to thrive. Both historic and modern data shows that foals can and will bond to humans and other domestic animals to meet their social needs. Thus domestication may have started with young horses being repeatedly made into pets over time, preceding the great discovery that these pets could be ridden or otherwise put to work.
However, there is disagreement over the definition of the term domestication. One interpretation of domestication is that it must include physiological changes associated with being selectively bred in captivity, and not merely "tamed." It has been noted that traditional peoples worldwide (both hunter-gatherers and horticulturists) routinely tame individuals from wild species, typically by hand-rearing infants whose parents have been killed, and these animals are not necessarily "domesticated."
On the other hand, some researchers look to examples from historical times to hypothesize how domestication occurred. For example, while Native American cultures captured and rode horses from the 16th century on, most tribes did not exert significant control over their breeding, thus their horses developed a genotype and phenotype adapted to the uses and climatological conditions in which they were kept, making them more of a landrace than a planned breed as defined by modern standards, but nonetheless "domesticated".
A difficult question is if domesticated horses were first ridden or driven. While the most unequivocal evidence shows horses first being used to pull chariots in warfare, there is strong, though indirect, evidence for riding occurring first, particularly by the Botai. Bit wear may correlate to riding, though, as the modern hackamore demonstrates, horses can be ridden without a bit by using rope and other evanescent materials to make equipment that fastens around the nose. So the absence of unequivocal evidence of early riding in the record does not settle the question.
Thus, on one hand, logic suggests that horses would have been ridden long before they were driven. But it is also far more difficult to gather evidence of this, as the materials required for riding—simple hackamores or blankets—would not survive as artifacts, and other than tooth wear from a bit, the skeletal changes in an animal that was ridden would not necessarily be particularly noticeable. Direct evidence of horses being driven is much stronger.
On the other hand, others argue that evidence of bit wear does not necessarily correlate to riding. Some theorists speculate that a horse could have been controlled from the ground by placing a bit in the mouth, connected to a lead rope, and leading the animal while pulling a primitive wagon or plow. Since oxen were usually relegated to this duty in Mesopotamia, it is possible that early plows might have been attempted with the horse, and a bit may indeed have been significant as part of agrarian development rather than as warfare technology.
While riding may have been practiced during the 4th and 3rd millennia BCE, and the disappearance of "Old European" settlements may be related to attacks by horseback-mounted warriors, the clearest influence by horses on ancient warfare was by pulling chariots, introduced c. 2000 BCE.
Horses in the Bronze Age were relatively small by modern standards, which led some theorists to believe the ancient horses were too small to be ridden and so must have been driven. Herodotus' description of the Sigynnae, a steppe people who bred horses too small to ride but extremely efficient at drawing chariots, illustrates this stage. However, as horses remained generally smaller than modern equines well into the Middle Ages, this theory is highly questionable.
The Iron Age in Mesopotamia saw the rise of mounted cavalry as a tool of war, as evidenced by the notable successes of mounted archer tactics used by various invading equestrian nomads such as the Parthians. Over time, the chariot gradually became obsolete.
The horse of the Iron Age was still relatively small, perhaps 12.2 to 14.2 hands (50 to 58 inches, 127 to 147 cm) high (measured at the withers.) This was shorter overall than the average height of modern riding horses, which range from about 14.2 to 17.2 hands (58 to 70 inches, 147 to 178 cm). However, small horses were used successfully as light cavalry for many centuries. For example, Fell ponies, believed to be descended from Roman cavalry horses, are comfortably able to carry fully grown adults (although with rather limited ground clearance) at an average height of 13.2 hands (54 inches, 137 cm) Likewise, the Arabian horse is noted for a short back and dense bone, and the successes of the Muslims against the heavy mounted knights of Europe demonstrated that a horse standing 14.2 hands (58 inches, 147 cm) can easily carry a full-grown human adult into battle.
Mounted warriors such as the Scythians, Huns and Vandals of late Roman antiquity, the Mongols who invaded eastern Europe in the 7th century through 14th centuries CE, the Muslim warriors of the 8th through 14th centuries CE, and the American Indians in the 16th through 19th centuries each demonstrated effective forms of light cavalry.