Navigation – Plan du site

AccueilNuméros35 (1)ArticlesThe prehistoric origins of the do...

Articles

The prehistoric origins of the domestic horse and horseback riding

Les origines préhistoriques du cheval domestique et de l’équitation
Markku Niskanen

Résumés

L’étude de Librado et al. (2021) montre que les chevaux domestiques modernes (DOM2) ont émergé dans la région inférieure du Don-Volga et impliquent que l’équitation a conduit à la sélection aboutissant à ces chevaux, a alimenté leur dispersion initiale. Cette étude implique que les chevaux DOM2 ont remplacé d’autres chevaux en raison de leurs meilleures aptitudes à l’équitation en raison d’un tempérament plus docile et d’un dos durable. Dans cet article, il est avancé que l’élevage en captivité des chevaux menant à leur domestication a commencé vers 4500-3000 av. J.-C. dans la steppe pontique-caspienne et a rendu l’équitation nécessaire car la gestion et surtout le transport des chevaux sur de longues distances nécessitaient un élevage monté. L’équitation était expérimentée depuis la seconde moitié du Ve millénaire av. J.-C., devenue courante vers 3100 av. J.-C. au début de la culture Yamnaya, et nécessaire au milieu du IIIe millénaire av. J.-C. pour la plus récente. Au fur et à mesure que l’équitation devenait plus courante, la sélection des tempéraments malléables et des dos durables s’est intensifiée, ce qui a donné des chevaux DOM2 vers 2300-2200 av. J.-C. dans la région inférieure du Don-Volga. La taille corporelle et la capacité de portage de charges n’étaient pas des facteurs limitants pour l’équitation chez les chevaux ancestraux et les premiers DOM2. La dispersion initiale des chevaux DOM2 a été facilitée par l’équitation et a commencé vers 2300±150 av. J.-C. L’utilisation du char à deux roues s’est dispersée avec les chevaux DOM2 après 2000 av. J.-C., mais une grande visibilité archéologique peut avoir gonflé l’importance du char, qui a une utilité pratique quelque peu limitée dans l’élevage et d’autres tâches quotidiennes.

Haut de page

Notes de la rédaction

Cet article fait suite à une communication invitée présentée lors des 1847es journées de la Société d’Anthropologie de Paris dans le cadre de la session "Les relations entre les humains et les animaux"

Texte intégral

Introduction

1The domestic horse revolutionized human mobility and warfare, changed economic and socio-political systems, ideologies, human gene pools and the spread of languages (Kalekna, 2009), and gave rise to an exceptional human-animal relationship that includes emotional transfers (Scopa et al., 2021). The last few years have seen a huge increase in knowledge on the origin of our modern domestic horse, commonly referred to as the DOM2 horse (the second domestic lineage). Because findings that demand reviews of earlier hypotheses are published almost yearly (e.g., Gaunitz et al., 2018; Fages et al., 2019; Librado et al., 2021), important pieces of the puzzle can still be found. However, perhaps the most important piece of this puzzle was discovered by Librado et al. (2021), who established that DOM2 horses first emerged in the lower Don-Volga region of the Pontic-Caspian steppe and dispersed elsewhere from about 2200 BC.

2The horse domestication process in this steppe region naturally began considerably earlier than the initial dispersal, but we do not know how much earlier. Morphology is of limited use for differentiating wild from domestic horses, with the obvious exception of pathologies resulting from riding and driving (Olsen, 2006b). Domestication of horses may also have started in more than one area, for example on the Kazakh steppe (Outram et al., 2009; 2021) and in the Carpathian Basin (Kanne, 2022), although the stage of full domestication, resulting in the DOM2 horse, was reached only on the Pontic-Caspian steppe.

  • 1 A chariot refers to a type of two-wheeled cart suitable for warfare, hunting and other high-speed h (...)

3The findings of Librado et al. (2021) that DOM2 horses appeared outside the Pontic-Caspian Steppe before chariots1 were invented, and the genetic risk factors behind chronic back problems, strongly suggest that horseback riding influenced the emergence of these DOM2 horses and "fuelled" their initial dispersal from this steppe region. These findings thus support the "horse before the cart" view and an early emergence of horseback riding (e.g., Levine, 1999; 2003; 2005; Olsen, 2006a; Anthony, 2007) more than the "cart before the horse" view with riding emerging later (e.g., Dietz, 2003; Drews, 2004; Taylor et al., 2020; 2021). A bioanthropological study by Trautmann et al. (2023) provides additional strong support for the "horse before the cart" view by finding diagnostic traits associated with habitual horseback riding in human skeletons that considerably predate the earliest wheeled vehicles pulled by horses, which are dated to the late 3rd millennium BC (Chechushkov and Epimakhov, 2018; Lindner, 2020).

4This article, based on an invited keynote given at the Journées de la Société d’Anthropologie de Paris 2022 in Toulouse on "the diversity of human-animal relationships" (Niskanen, 2022), does not provide new evidence on horse domestication and the origin of riding, but offers a partial review of the existing evidence as a further contribution. More specifically, it reviews the archaeological and genetic evidence for the emergence and dispersal of DOM2 horses, arguing that horseback riding became common due to the need for mounted herding of horses and drove selection resulting in DOM2 horses. It also assesses the suitability of ancestral and early DOM2 horse for riding and investigates the practical utility of chariots.

5Although it does not provide new evidence, the article does offer some new estimations. Because there is a gap in genomic sampling of about six centuries between the ancestral DOM2 horses and the earliest DOM2 horses, I propose estimated times for the emergence of DOM2 horses based on available genetic evidence presented in Librado et al. (2021), and for their dispersal based on my own estimations of the rates at which horses spread across the Great Plains and adjacent regions of North America. Because the size and strength of horses matter in riding and traction, I introduce and apply new methods to estimate their body size and weight-carrying ability. Materials (including raw data) and methods are presented as supplementary information. Dates are calibrated 95.4% ranges unless stated otherwise. These estimates are intended to augment the discussion in the context of the present invited keynote article. They can be tested later through more sophisticated approaches such as modelling-simulation and by including additional information provided by demographic, geographic and palaeo-environmental data.

A review of archaeological and genetic evidence

From hunted wild horses to a "cult of the horse"

6Wild horses were important game animals for hunter-gatherers of the Orlovskaya culture (6200-5300 BC) between the lower Volga and Ural rivers. They represented about one-fifth of hunted mammals according to the combined faunal data from Varfalomeevka, Oroshaemoe and Algay (Vybornov et al., 2018: Table 1). These equids had both dietary and symbolic importance for the hunter-gatherers of the Syezzheye stage of the Samara culture (5250/5300-4800 BC, Morgunova, 2015) in the Samara River valley. Skulls and distal limb bones as well as a bone amulet depicting a horse (figure 1) have been found in graves (Anthony, 2007: 189-190).

Figure 1

Figure 1

A drawing of a carved bone amulet depicting a horse from the Syezzheye cemetery |
Dessin d’une amulette en os sculpté représentant un cheval du cimetière de Syezzheye

after Anthony, 2007: Fig. 9.8 / d’après Anthony, 2007: Fig. 9.8

7Acquiring livestock may have affected how the steppe people viewed horses. Livestock arrived on the western Pontic-Caspian steppe in about 5800-5700 BC with farmers of the Starčevo-Körös-Criș culture (6500-5300 BC), who gave rise to the Tripolye culture (5100-2800 BC). The local foragers of the Bug-Dniester culture (6500-5300 BC, Motuzaitė Matuzevičiūtė, 2013: Table 1) became assimilated into the farming societies, while livestock spread via cultural dissemination to the foragers of the Dnieper-Donets II culture (5200-4200 BC) from about 5200 BC (Anthony, 2007: 149-154, 174-175). Livestock arrived on the eastern Pontic-Caspian steppe from the eastern Caucasus from about 5200-5000 BC with the Cis-Caspian culture that replaced the Orlovskaya culture, and was in turn replaced by the Khvalynsk culture (4700-3300 BC) that grouped the inhabitants of the lower and middle Volga into a single culture (Anthony, 2007:275; 2016: Table 1.1; Vybornov et al., 2018).

8The gene pools of steppe horses were not affected by the above events because the horses were still wild. All horses from the sites discussed above, from the earliest (a horse from Varfalomeevka dated to 5616-5482 BC) to the most recent (a horse from Oroshaemoe dated to 4673-4498 BC), belong to a genetic cluster named NEO-NCAS (Librado et al., 2021: SI Table 1, Extended Data Fig. 4).

  • 2 Horses can reach grass through snow cover of up to 50-60 cm by pawing with their hooves (Khazanov, (...)

9The horse became better represented in faunal assemblages after the arrival of livestock, reaching peak percentages in 4000-3000 BC while the representation of other game species decreased (Anthony and Brown, 2014: Fig. 6.2). Starting from about 4500-4400 BC, horse-head shaped mace heads (figure 2) were placed in graves, while horses – but no other wild game animals – were included with cattle and sheep in funerary feasts and/or sacrifices throughout the Pontic-Caspian steppe. This elevated symbolic status, referred to here as the "horse cult" – following Anthony and Brown (2003) and Kuzmina (2003) – may simply reflect the high dietary importance of horses, which survive winter better than cattle and sheep as they are better able to feed on grass under snow.2 However, the "horse cult" is also thought to be associated with the early stages of horse domestication (Anthony and Brown, 2000; 2003; 2011; 2014; Anthony, 2007:256-258; Anthony et al., 2022). Training horses for riding has been suggested (Anthony et al., 2022). Domestication would have begun if horses were bred in captivity.

Figure 2

Figure 2

Eneolithic mace heads from Novoorsk (A) and Suvorovo (B) |
Têtes de masse énéolithiques de Novoorsk (A) et Suvorovo (B)

redrawn after Anthony (2007: Fig. 11.5) / redessinées d’après Anthony (2007 : Fig. 11.5)

10Whether or not there was any captive breeding and riding of horses, this "horse cult" arrived from the Volga-Ural steppe into the Dnieper-Don steppe with Khvalynsk immigrants, who with the local Dnieper-Donets II people formed the Sredni Stog culture dating to about 4750/4400-3400 BC (Anthony, 2007:244; Kotova, 2008:122; Rassamakin, 2012: Table 1; Anthony et al., 2022). Increased aridity drove some early Sredni Stog groups westward (Kotova and Makhortykh, 2010), giving rise to the Suvorovo-Novodanilovka group. Graves in the Danube delta and the westernmost steppe represent the Suvorovo group and those further east the Novodanilovka group (Anthony, 2007:251). These were graves of high-status individuals, who controlled long-distance exchanges of prestige goods including copper and gold from the "Old European" settlements in the lower Danube Valley, pottery from Tripolye settlements and the horse-head shaped mace heads of the Sredni Stog and Khvalynsk cultures (Rassamakin 1999:100-112; Anthony 2007:249-258).

11The "horse cult" had arrived in the lower Danube Valley by about 4400 (4457-4350) BC if Giurgiuleşti in Romania indeed represents the earliest Suvorovo sites. Skulls of one male horse, one female horse and one foal were recovered from Grave 4 at Giurgiuleşti (Govedarica and Manzura, 2019). Mace-heads shaped like horse heads have been recovered not only from other Suvorovo sites such as Suvorovo and Casimcea, but also from local Old European settlements. David Anthony found it difficult to explain this "sudden symbolic importance" of horses in these local settlements "if horses were not ridden into the Danube valley" (Anthony, 2007: Fig. 11.5, 254). It is indeed possible that these "Suvorovos" rode horses. A human male skeleton from Csongrád-Kettőshalom, Hungary, dated to 4442-4243 BC, exhibits skeletal evidence of habitual riding (Trautmann et al., 2023). Suvorovo migrants from the lower Danube region arrived in this area towards the end of the Tiszapolgár culture of the Early Copper Age (Anthony, 2007:255).

12Although the "Suvorovos" may have ridden horses, we do not know if they used horses brought from the steppe or local horses. Horses from the local Gumelnița culture sites of Pietrele and Căscioarele in Romania have been dated to 4494-4251 BC and cluster in a Neighbour Joining Tree with a horse from Semenovka 1 in Ukraine, dated to 4315-4054 BC (Librado et al, 2021: SI Table 1, Extended Data Fig. 4). If these Gumelnița horses were indeed local wild horses (Bréhard and Bălăşescu, 2012), this genetic affinity reflects population events among wild horses rather than captive horses arriving with immigrants from the steppe.

13Two pre-Yamnaya or very early Yamnaya individuals from Blejoi in the northern parts of the lower Danube Valley in Romania, dated respectively to 3331-2927 BC and 3338-2939 BC, display at least three of the six diagnostic skeletal traits of horseback riding. These diagnostic traits are also found in Yamnaya individuals from Romania, Bulgaria and Hungary dated to 3021-2501 BC, strongly suggesting that horseback riding was a common and widespread activity in this culture (Trautmann et al., 2023: Table 2).

The emergence of DOM2 horses on the Pontic-Caspian Steppe

14Horse milk peptides in calculus from two human individuals (KR19 K.4 21 A and KRI9 K.2 N-2) from the Krivyanski 9 site in the lower Don River basin (Wilkin et al., 2021: SI Tables S4-S5), representing the Yamnaya culture (3300-2600 BC, Morgunova and Khokhlova, 2013) and dated to 3345-3096 BC and 2881-2633 BC respectively, is evidence of foaling in captivity and thus of captive breeding. Dates corrected for the reservoir effect are about 200 years more recent if the same correction is applied to these Yamnaya as to those from the Caspian Steppe (Shishlina et al., 2009:496).

15These "Yamnaya horses", already being milked and bred in captivity in the Don River basin around 3000 BC (200 years subtracted from the date of KR19 K.4 21 A), probably represented a group named C-PONT, whose representatives are divided between the Maikop culture, the Repin stage of the Yamnaya culture and the Poltavka culture. They are dated to 3526-2631 BC and lived between the lower Don and Volga rivers. Horses from Turganik, about halfway between the Samara bend and the southern Ural Mountains and referred to as TURG horses, are associated with the Yamnaya culture and dated to 2897-2636 BC. These horses had diverged less from the earlier NEO-NCAS horses than the C-PONT horses and had acquired some admixture from horses on the Kazakh steppe (Librado et al., 2021: SI Table 1).

16These early captive horses ancestral to DOM2 horses must have required management of some kind. Pasturing small groups of horses close to the settlement would have been possible by herders on foot if the horses were not free-ranging and there was no need to move them over long distances (Khazanov, 1994:92), and especially if corralling, hobbling or other forms of restraint were used (Taylor et al., 2020). It has even been argued that early domestic horses could have been kept ranging freely half a mile from the village without any need for herding (Drews, 2004:44-47 in E-book). However, mounted herding is generally thought necessary if the herds are large and the horses range freely and/or need to be moved over long distances (e.g., Barclay, 1982:246; Clutton-Brock, 1992:12; Azzaroli, 1985:6; 1998:41; Khazanov, 1994:92; Levine, 1999; 2005; Taylor et al., 2020). Horses are much faster and more excitable than cattle (Robinson, 2001:17), so that herders on foot cannot cope even if helped by dogs (Azzaroli, 1985: 6; 1998:41). As already mentioned, some horseback riding was apparently already taking place during the Copper Age and was unquestionably practiced during the Yamnaya period (Trautmann et al., 2023).

  • 3 The ZFPM1 gene is associated with docility and stress tolerance (Tikker et al., 2020; Taira and Des (...)

17Captive breeding had resulted in large herds of horses by the middle of the 3rd millennium BC, when some funerals included sacrifices of dozens of horses. For example, the skulls of forty horses were placed above a Catacomb-culture grave near Tsatsa, Russia, dated to about 2500 BC (Anthony, 2007:325). Due to a gap in genomic samples from the Pontic-Caspian steppe between 2871-2631 BC and 1929-1773 BC, it is not known whether or not these horses at Tsatsa were DOM2 horses, which are modelled as 90.7% Eneolithic Ukrainian (ancestral to C-PONT) and 9.3% TURG or, alternatively, 69.9% C-PONT, 24.4% TURG and 5.8% Eneolithic Hungarian (Librado et al., 2021; SI Tables 1 & 3, Extended Data Fig. 4). The TURG ancestry explains why 2.7% of DOM2 ancestry is from horses related to those hunted and possibly bred in captivity by people of the Botai culture on the Kazakh steppe (Gaunitz et al., 2018). Furthermore, DOM2 horses are distinguished from non-DOM2 horses by differences in the ZFPM1 gene associated with docility and stress tolerance and the GSDMC gene associated with a lower risk of developing chronic back problems3 (Librado et al., 2021).

The dispersal of DOM2 horses from the Pontic-Caspian steppe

18Genomic data reveals that large numbers of DOM2 horses were produced from around 2200 BC, when the initial dispersal of DOM2 horses, presumably "fuelled" by horseback riding, also began. This dispersal reached very large proportions in "the late third millennium and early second millennium" (Librado et al., 2021:3) and therefore around the time of the earliest chariots, which appeared "…in the decades around the turn of the second millennium BC…" (Lindner, 2020:377).

19The initial dispersal of DOM2 horses first resulted in a temporal overlap and genetic admixture with non-DOM2 horses, but the ancestry of these non-DOM2 horses was later almost totally erased. Due to the temporal gap in genomic samples from the Pontic-Caspian steppe, a horse from Acemhöyük, Turkey, dated to 2205-2044 BC, is the earliest known DOM2 horse. This horse had about 15-20% of non-DOM2 ancestry, based on my estimation from its ancestry profile (Librado et al., 2021: SI Table 1; Extended Data Fig. 4) and a maternal lineage P of Anatolian wild horse (Guimares et al., 2020: Table S1), perhaps acquired three horse generations earlier (about 30 years) from an Anatolian great-granddam.

20DOM2 horses arrived in Mesopotamia during either the Akkadian period (2334-2154 BC) or the Ur III period (2112-2004 BC) and based on artistic evidence, they were already being ridden on their arrival. An Akkadian seal impression from Kish depicts a man riding an equid with a horse’s mane and tail but long ears (figure 3a). An Ur III seal impression from Abbakalla dated to 2037-2029 BC clearly depicts a man riding a horse (figure 3b). Equids referred to as "anše-zi-zi" (equids of the mountain) in written documents from the Ur III period are thought to have been horses (Bennett et al., 2022 and references therein). Clay plaques from the Early Dynastic period (2900-2350 BC) depicting human figures riding donkeys, onagers or donkey-onager hybrids (Moorey, 1970) show that equids were ridden in Mesopotamia before DOM2 horses arrived.

Figure 3

Figure 3

A) Akkadian (2334-2154 BC) seal impression from Kish; B) Ur III (2112-2004 BC) seal impression from Abbakalla dated to 2037-2029 BC |
A) Impression de sceau de la période akkadienne (2334-2154 av. J.-C.) de Kish; B) Impression de sceau de la période Ur III (2112-2004 av. J.-C.) d’Abbakalla datée de 2037-2029 av. J.-C.

A: redrawn from Buchanan, 1966 / redessinée d’après Buchanan, 1966;
B: redrawn from Owen, 1991 / redessiné d’après Owen, 1991

21DOM2 horses also overlapped with non-DOM2 horses over time and geography in Central Europe. A horse from Gordinesti II in Moldova, dated to 2140-1985 BC, and a horse from Holubice in the Czech Republic, dated to 2137-1936 BC, are the earliest known DOM2 horses west of the steppe. These horses were contemporary with local non-DOM2 horses because a non-DOM2 horse from Dunaújváros-Kosziderpadlás in Hungary dates to 2140-1977 BC. Hungarian non-DOM2 horses preceding DOM2 horses had derived 51.5% of their ancestry from Eneolithic Czech horses and 48.5% from horses from Turkish Thrace, based on a horse from Kaposújlak-Várdomb dated to 2571-2344 BC (Librado et al., 2021: SI Tables 1 & 3).

22These Hungarian Early Bronze Age non-DOM2 horses preceding DOM2 horses and partly contemporary with them are "suspected to have been domesticated" (Kanne, 2022). Evidence for this domestic status is now stronger. Some horses preceding DOM2 horses must have already been ridden in this region during the Early Copper Age, based on diagnostic traits associated with riding in a human male individual from Csongrád dated to 4442-4243 BC (Trautmann et al., 2023).

23Early Bronze Age non-DOM2 horses from the site of the Kirklareli-Kanligecit settlement in Turkish Thrace are also thought to represent domestic horses (Benecke, 2009). A horse from this site dated to 2466-2306 BC is related to Eneolithic horses from Ukraine and the lower Danube, based on a Neighbour Joining tree. It had derived 69.7% of its ancestry from Eneolithic horses of the lower Danube, 20.1% from the non-DOM2 horses from Hungary discussed above and the rest (10.2%) from an unknown (or untested?) population (Librado et al. 2021: Extended Data Fig. 4, SI Tables 1 & 3), perhaps from Anatolia.

24DOM2 horses dispersed eastwards onto the Kazakh steppe during the late 3rd millennium BC. A horse from the Early Bronze Age Yelunin culture site of Michuruno in north-eastern Kazakhstan, dated to 2109-1928 BC, has ancestry from both DOM2 and non-DOM2 horses. This horse slightly predates the earliest directly dated DOM2 horse (2026-1884 BC) associated with the Sintashta-Petrovka culture in the Ural-Tobol interfluve (Librado et al., 2021: SI Table 1, Extended Data Fig. 4). The non-DOM2 ancestry detected in the Michuruno horse is from horses related to those that were hunted, tamed and possibly partly domesticated by people of the Botai culture (3700-3100 BC), based on evidence of corralling, milking and bitting (Outram et al., 2009; 2021, but see Taylor and Barrón-Ortiz, 2021), and on the presence of an allele associated with the Leopard spotting complex and night blindness (Gaunitz et al., 2018; Librado et al., 2021).

25This eastward dispersal of DOM2 horses is thus roughly contemporary with the early stages of the Sintashta-Petrovka culture (2040-1730 BC) and the earliest spoke-wheeled chariots dated to a few decades around 2000 BC (Lindner, 2020) and preceded by the two-wheeled carts already being used in the Don-Volga interfluve in 2400-2200 BC (Chechushkov and Epimakhov, 2018). The Sintashta-Petrovka chariots represent the earliest indisputable evidence for the use of horses for traction and have been thought to represent the earliest use of horses for transportation. Supporters of the "cart before horse" view argued that riding before the introduction of chariots at most involved hitching rides on the backs of packhorses or recreational riding by "daredevils" and "show-offs" (Drews, 2004:42-44, 44-47, 47-50 in E-book; Taylor et al., 2020, 2021). This view conflicts with very recently reported diagnostic traits associated with habitual riding in pre-Yamnaya and Yamnaya individuals (Trautmann et al., 2023).

26Chariotry was introduced in central and eastern Kazakhstan by people of the Andronovo culture about 1800 BC (Anthony, 2007:448-451), but the Michuruno horse of mixed DOM2 and non-DOM2 ancestry (see above) is evidence that DOM2 horses arrived in eastern Kazakhstan a century or two earlier than chariots. The earliest DOM2 horse from the Kent Bronze Age settlement site of the Begazy-Dandybaev culture dated to 1535-1452 BC (Librado et al., 2021: SI Table 1) thus postdates the earliest DOM2 horses in this region by about half a millennium. Drews (2004:136-139 in E-book) presumes that more horses were used for riding than for driving (i.e., for traction) on the steppes during the 2nd millennium BC. Outram et al. (2021) believe that riding was important for herding and mobility in the Begazy-Dandybaev culture.

27The earliest DOM2 horse of the Bronze Age Deer Stone-Khirigsuur Complex (DSK), and therefore in Mongolia, dates to 1273-1016 BC (Librado et al., 2021: SI Table 1). This culture dated to 1200-700 BC is named after anthropomorphic stelae known as deer stones and burial mounds called khirigsuurs, which are surrounded by sacrificial offerings of horses to honour departed leaders (Fitzhugh, 2009). Chariotry seems to have been glorified, because the carvings in the deer stones depict chariots but not riders. Also, chariotry spread together with DOM2 horses beyond the steppe in East Asia and parts of South Asia (Taylor et al., 2020; 2021). For example, horses and chariots arrived together, and abruptly, in Anyang, the capital of the Shang dynasty, at about 1180 BC from the steppe, together with grooms, horse trainers, veterinarians and wheelwrights (Kalekna, 2009:136-137 and references therein).

28Although chariotry certainly spread eastwards beyond the steppe together with DOM2 horses, it had little or no role in the dispersal of DOM2 horses across Central and Western Europe, where it arrived hundreds of years after DOM2 horses (Kanne, 2022 and references therein). However, horses selectively bred for chariotry spread together with the chariot complex, genetically affecting populations of DOM2 horses that had arrived earlier. A relatively large number of these early chariot horses must have been chestnuts, because this colour became more common (Librado et al., 2021).

29The dissemination of the chariot complex, also known as the chariot package, starting in about 1950 BC along interregional trade routes and connecting complex societies in Eurasia, was concurrent with the transformation of the Eurasian Bronze Age societies (Kristiansen, 2018: Fig. 3). This "package", which included chariots, chariot builders, chariot horses and their trainers and grooms, had emerged with complex societies (e.g., chiefdoms) on the steppe because a complex set of technologies, skills, and resources was required (Chechushkov and Epimakhov, 2018; Kristiansen, 2018). Teams of horsemen, trainers and grooms accompanied chariots and chariot horses in the dissemination of this complex and the associated horse industry, as indicated for example by 3800-year-old clay tablets from north-eastern Syria (Klecel and Martyniuk, 2021).

30The role and importance of chariotry may have become inflated due to the high archaeological visibility of burials and art associated with high-status members of complex societies (e.g., Chechushkov and Epimakhov, 2018; Kristiansen, 2018; Maran, 2020; Metzner-Nebelsick, 2021). Because chariots were prestige items (e.g., Chechushkov and Epimakhov, 2018; Kristiansen, 2018; Maran, 2020; Metzner-Nebelsick, 2021), only high-status individuals were charioteers even in charioteering societies (Chechushkov and Epimakhov, 2018). The practical utility of chariotry is doubted by many (e.g., Barclay, 1982: Littauer and Crouwel, 1996; Levine, 1999; 2005; Chechushkov and Epimakhov, 2018; Maran, 2020; Metzner-Nebelsick, 2021; Kanne, 2022), but Taylor et al. (2021: 1491) suggest "…an expanded practical role for chariots in early pastoral lifeways of the eastern Eurasian Steppe…"

31Chariots were widely used in large numbers in warfare in South and East Asia. The most common use of chariots in warfare was for transportation to the battleground and/or as mobile and elevated firing platforms for archers (Crouwel, 2013:87-88). Interestingly, there is no evidence of chariot warfare on the steppes (Drews, 2004:136-139 in E-book). Chariots may have been used more in ritual races than in warfare in Europe and Central Asia (Maran, 2020).

32Sufficiently competent riding to herd livestock and fight on horseback, for example, is not thought to have emerged until around the turn of the 2nd millennium BC for several reasons. First, the horsemanship of early riders has been questioned due to many Near Eastern artistic depictions showing riders sitting too far back on the horse as if riding a donkey (Sherratt, 1997:217; Drews, 2004: Chapter 3). Secondly, early domestic horses were presumably too small for proper riding (Sherratt, 1997:217). Third, adequate control of horses was thought unlikely before the invention of metal bits. Bronze bits were introduced in about 1300-1100 BC but became common only after about 1000 BC (Drews, 2004:136-139 in E-book; Taylor et al., 2020; 2021). Therefore, horseback riding for military purposes presumably emerged during the second half of the 2nd millennium BC and became common only after 1000 BC (Sherratt, 2003:242; Drews, 2004: Chapters 3 & 4; Turchin et al., 2016).

The replacement of non-DOM2 horses by DOM2 horses

33DOM2 horses had replaced all local populations of horses except the ancestors of some local wild horses (e.g., the Tarpan and Przewalski’s horse) by the end of the 2nd millennium BC (Librado et al., 2021). This replacement presumably occurred through large-scale genetic introgression, inbreeding among local populations or feralization (Outram et al., 2021). Some genetic ancestry from local non-DOM2 horses has been detected in the earliest DOM2 horses of Anatolia (Guimares et al., 2020: Table S1) and some ancestry of the native Iberian horse is carried even by present-day DOM2 horses (Fages et al., 2019). It has been argued that the Przewalski’s horse may be a feral descendant of horses hunted and managed by the people of the Botai culture (Gaunitz et al., 2018).

Estimations made for this study

34There is a gap of about six centuries in genomic sampling between the most recent C-PONT horses and the earliest DOM2 horses anywhere. Therefore, other means are needed to estimate when DOM2 horses emerged and dispersed.

Estimating when DOM2 horses emerged

35Dating the emergence of DOM2 horses is based on changes over time in an ancestry component (k=6, six ancestral populations assumed) which is maximized in DOM2 horses. This ancestry component (taken from Librado et al., 2021: Extended Data Fig. 4), which is typical for DOM2 horses and referred to here as the "DOM2 component", was used in Librado et al. (2021: Fig. 2c) to visualize the C-PONT ancestry in DOM2 horses and their later dispersal.

36The "DOM2 component" is plotted against calibrated dates in NEO-NCAS, C-PONT and DOM2 horses dated before 1000 BC to estimate when the distinct ancestry profile of DOM2 horses emerged (figure 4). This component represents at least 80% of ancestry in all DOM2 horses except in the horse from Acemhöyük. The lower percentage (76.6%) in this horse reflects admixture with non-DOM2 horses as indicated by mtDNA from local Anatolian horses (see Guimares et al., 2020). It is also possible that the distinctive ancestry profile of DOM2 horses had not yet fully emerged when this earliest DOM2 horse included in genomic sampling was alive. A Loess curve shows a linear increase in this "DOM2 component" from NEO-NCAS to C-PONT, reaching 80% at about 2500 BC. The rate of increase from C-PONT to DOM2 accelerates from about 2300 BC to 2200 BC, when this component reaches the 90% typical for DOM2 horses.

Figure 4

Figure 4

Changes over time in the ancestry component (k=6) maximized in DOM2 horses from the earliest NEO-NCAS horses to DOM2 horses dated no earlier than 1000 BC (based on Librado et al., 2021: Extended Data Fig. 4). The blue horizontal bars represent time frames for the earliest milking of horses and their initial dispersal. The numbers refer to sites: 1) Varfolomeevka, 2) Algay, 3) Turganik, 4) Oroshaeomoe I, 5) Semenovka 1, 6) Aygurskiy 2, 7) Repin, 8) Sosnovka. The dashed line is the Loess curve through NEO-NCAS, C-PONT and DOM2 horses |
Les changements temporels de la composante d’ascendance (k=6) sont maximisés chez les chevaux DOM2, des premiers chevaux NEO-NCAS aux chevaux DOM2 datés au plus tôt de 1000 av. J.-C. (basé sur Librado et al., 2021 : Données étendues Fig. 4). Les barres horizontales bleues représentent les délais pour la première traite des chevaux et la dispersion initiale. Les numéros renvoient aux sites : 1) Varfolomeevka, 2) Algay, 3) Turganik, 4) Oroshaeomoe I, 5) Semenovka 1, 6) Aygurskiy 2, 7) Repin, 8) Sosnovka. La ligne pointillée est la courbe de lœss passant par les chevaux NEO-NCAS, C-PONT et DOM2

37Interestingly, this component had increased in C-PONT horses but not in TURG horses whose domestication status is unknown. Also, one early Yamnaya horse from Repin is a C-PONT horse, whereas another one is similar to TURG horses, perhaps due to genetic admixture (Librado et al., 2021). Additional genomic samples will hopefully one day allow us to replace the estimated emergence time of DOM2 horses with the actual time.

Estimating the time of initial dispersal of DOM2 horses

38Dating the initial dispersal of DOM2 horses is based on dates for the earliest DOM2 horses outside the Pontic-
Caspian Steppe and rates of dissemination of horses based on ethnohistorical and archaeological information from the Great Plains and the adjacent regions of North America (see supplementary information A). This information from North America is used because it is the only material that provides sufficiently accurate dating information on the spread of horses and equestrianism.

39The rates of dissemination were calculated by dividing the shortest walking distances between locations provided by Google Maps (https://maps.google.com) by the differences in years of the earliest known horse acquisitions to derive dissemination rates in kilometres per year. Figure 5 shows the locations used, the years when horses were first acquired and the calculated dissemination rates between locations. The resulting rates over total distances ranged from 11.8 km/year to 19.4 km/year, whereas those between adjacent points on the map ranged from 7.8 km/year to 24.2 km/year.

Figure 5

Figure 5

Years when horses were first acquired by selected groups of Native Americans (A). Distances and rates of dissemination of horses (B). The distances are walking distances |
Années d’acquisition des chevaux de groupes sélectionnés d’Amérindiens (A) Distances de diffusion et vitesses de diffusion des chevaux (B). Les distances sont des distances à pied

40The initial dispersal of DOM2 horses began from their area of origin between the Dnieper steppes in the west and Turganik in the east (Librado et al., 2021). Therefore, the southern edge of this area of origin would be the current steppe boundary along the lower Kuban River close to the present-day city of Krasnodar to the north of the western Caucasus Mountains (45.03603 N, 38.97457 E), its western edge at the intersection of the current northern boundary of the steppe with the Dnieper River (48.747566 N, 34.305448 E), which may have been a population boundary, and its eastern edge along the Ural River (51.22782 N, 51.38654 E), also a possible population boundary, around the present-day city of Uralsk. The southern edge is likely to be more correct than the other two because it is better defined geographically.

  • 4 This effect is the reverse of the Signor-Lipps effect (Spllil-Rongis is Signor-Lipps spelled backwa (...)

41Walking distances from the southern edge, the western edge and the eastern edge of the DOM2 area of origin to Acemhöyük, Holubice and Michuruno divided by dissemination rates provide estimates of dissemination times, which when added to the range of dates for DOM2 horses from these sites provide estimated dates for the beginning of their dispersal. Thirty years are added in the case of the Acemhöyük horse, which had acquired ancestry from local Anatolian horses perhaps three generations back. The resulting estimations presented in figure 6 indicate that southward dispersal began in about 2450-2150 BC and dispersal both westward and eastward in 2400-2050 BC. Dates are rounded to the closest half century to allow for the Spllil-Rongis effect, which can cause the oldest evidence to post-date the earliest occurrence.4 I will use 2300±150 BC as the start date for this dispersal.

Figure 6

Figure 6

Dispersal of DOM2 horses from the Pontic-Caspian steppe. Dispersal times based on distances between adjacent points on the map (see figure 5) are given in parentheses to differentiate them from those based on total distances |
Dispersion des chevaux DOM2 de la steppe pontique-caspienne. Les temps de dispersion basés sur les distances entre points adjacents sur la carte (voir figure 5) sont donnés entre parenthèses pour les différencier de ceux basés sur les distances totales

Estimating the body size and weight-carrying ability of horses

42Because body size and strength matter in horse-riding and traction, new methods were applied to estimate the size and weight-carrying ability of horses from their skeletal dimensions. For instance, horses in the size category of small ponies (127 cm and under) generally do not have enough strength to carry adult male riders well, but they can be used as meat and milk animals, as packhorses and for traction to pull small carts. Paired teams can pull chariots well enough in all gaits. The methods used to estimate the body size and weight-carrying ability of horses are described in supplementary information B, but they are briefly summarised here.

43Withers heights were reconstructed anatomically in the case of complete specimens (most present-day domestic horses and Przewalski’s horses) using the conversion factors provided in Tables S6 and S7, but from metacarpal dimensions in the case of prehistoric specimens because that was the only bone available. In these cases, a regression equation (SJH=4.389×MC2+9.098×MC11/MC2x100 – 180.036, r=0.984, N=155) provided in table S10 was first applied to estimate shoulder joint height (SJH) from metacarpal lateral length (MC2) and the percentage of distal articular breadth (MC11) from lateral length (MC11/MC2x100). This ratio is included because relatively thick metapodials tend to be short for withers height and vice versa (Niskanen, 2022). Next, the resulting estimate was multiplied by 1.3927 to derive a withers height estimation for horses with low (i.e., non-prominent) withers typical of most native horses and ponies (Table S11).

44Body mass is estimated from available distal metapodial dimensions with regression equations (Table S12). Maximum rider weight is estimated from metapodial dimensions with equations provided in supplementary information B. Maximum rider weight estimated from metacarpal dimensions is heavier than that estimated from both metacarpal and metatarsal dimensions in horses with heavy forequarters relative to their hindquarters, as exemplified by Przewalski’s horses and to a lesser extent by native horses and ponies not selectively bred for riding. Therefore, estimates derived from the metacarpal dimensions of all horses preceding DOM2 horses are multiplied by 0.962 and those of all early DOM2 horses are multiplied by 0.9909 to make these estimates comparable with those for current horses estimated from both metacarpal and metatarsal dimensions.

45Published information on withers heights of Bronze Age horses is included in comparisons with the reconstructions derived in this study. Descriptive statistics of withers height, body mass and maximum and adjusted maximum rider weight are provided in table 1.

  • 5 Gaunitz et al. (2018: Supplementary Material) thought that the Dereivka sample dates to the Iron Ag (...)

46Horses from the Pontic-Caspian steppe preceding DOM2 horses, represented here by a pooled sample of presumably NEO-NCAS horses from Mirnoie and Tripolie and late NEO-NCAS or early C-PONT horses from Dereivka,5 were not small ponies. These steppe horses dated to 7500-3700 BC averaged ~135 cm in height, ~364 kg in body mass and ~104 kg in maximum rider weight. Horses hunted and possibly managed by people of the Botai and Tersek cultures (~3600-3000 BC) on the Kazakh steppe averaged ~137-139 cm in height, ~363-369 kg in body mass and ~99-102 kg in maximum rider weight. Early DOM2 horses of the Sintashta-Petrovka (2026-1749 BC, Librado et al., 2021: SI Table 1) and Begazy-Dandybaev (1535-1426 BC; Librado et al., 2021: SI Table 1) cultures of Kazakhstan averaged ~140-141 cm in height, ~380 kg in body mass and ~106 kg in maximum rider weight. Eneolithic and Bronze Age horses from Western Europe, Central Europe and Italy were smaller on average than contemporary steppe horses. Their average heights were in the range of 128-137 cm, but even those averaging about 129 cm in height had mean body masses and maximum rider weights of about 300 kg and 89 kg, respectively (table 1).

  • 6 Khalkha Mongol men average 66.5 kg in body weight (Eveleth and Tanner, 1976: Table S1), to which cl (...)

47For comparison with the above early horses, Mongolian horses averaging ~127-135 cm in height (Peilieu, 1980) and ~272 kg in weight (Smith, 2000) typically carry rider weights of ~80 kg; Icelandic horses averaging 141 cm and 339 kg carry ~92-93 kg (Stefánsdóttir et al., 2014);6 the Indian ponies in the North-western Plains averaged 140-145 cm and 300-350 kg (Ewers, 2001:33). The ancestral DOM2 horses and early DOM2 horses were therefore bigger and stronger than present-day Przewalski’s horses and many native ponies and horses, but smaller and weaker than present-day riding horses (hotbloods and warmbloods) as shown in table 1.

Discussion

From captive wild horse pets to captive breeding of horses

48Domestication is an evolutionary process, and thus naturally involves genetic separation of the domestic population from the wild ancestral population (Clutton-Brock, 2012:5-7). Horse domestication must have started from keeping wild horses as pets which, given the bison and moose kept as pets by some Native Americans (Serpell, 1989), would not have been extraordinary. Most pet wild horses surviving in captivity were likely foals at least two months old to be able to survive without mare’s milk (Berger, 1986:116) but, based on information on feral horses, not older than yearlings to be manageable (Ewers, 2001:59). Unfortunately, the archaeological record does not provide any evidence of wild horse pets.

49Most wild horse pets probably escaped once they matured, but some mares in foal may have returned to humans as did some domestic mares stolen by Tarpan stallions (see Heptner et al., 1988:1052). Although this would have resulted in captive-born horses, it was not captive breeding because not all stages of reproduction occurred in captivity. At first, captive breeding of horses may have been as difficult as for zebras (see Lasley et al., 1994), so it may have taken some time before a captive breeding population large enough for domestication emerged.

50The growing dietary importance of horses combined with the "horse cult" suggest that captive breeding of presumably NEO-NCAS horses may have begun about 4500 BC. Having captive horses would have been convenient because capturing wild horses alive and moving them over long distances to places of exchange and/or sacrifice would have been challenging. Horseback riding, already experimented with during the second half the 5th millennium BC (see Trautmann et al., 2023), would also have helped. Horses ridden during the pre-Yamnaya periods must have been at least well-tamed captive wild horses but whether they were bred in captivity is unknown. Therefore, milking of presumably C-PONT horses in the Don River basin by the Yamnaya by about 3000 BC (see Wilkin et al., 2021) is the earliest concrete evidence of captive breeding because milking mares requires them to foal in captivity.

51Captive breeding, starting with a small number of wild horse pets making up the founders of the domestic population, may explain the initial genetic differentiation of C-PONT horses dated to 3526-2631 BC from the preceding NEO-NCAS horses dated to 4315-4054 BC (if an Eneolithic Ukrainian horse is included in this sample) and from contemporary TURG horses dated to 2897-2636 BC (Librado et al., 2021: Extended Data Fig. 4, SI Table 1). A change in the "DOM2 component" over time suggests that this differentiation was gradual, but this may be due to a gap in the genetic sampling sequence (figure 4).

52Through natural selection, the wild ancestors of DOM2 horses had become fast, agile and enduring runners with strong flight-or-fight reactions and intolerance of restraint, but not good at carrying a weight on their backs or tolerating human-induced stress. The weak backs, untameable character, poor rideability and low survival in captivity of historic wild horses have indeed been noted (e.g., Stella, 1518:21; Gmelin, 1770:47; Smith, 1866:165). Captive Przewalski’s horses and zebras are also more aggressive, panicky, unpredictable and intolerant of restraint than domestic horses (see Bowling and Ruvinsky, 2000:30; Brubaker and Coss, 2015).

53However, this does not mean that taming and riding the wild ancestors of DOM2 horses was impossible. A Przewalski’s stallion named "Vaska", one of the earliest Przewalski’s horses brought to Europe, was ridden (figure 7). A six-month-old Przewalski’s colt had tolerated a young boy on its back (Mohr, 1971:69). Individual zebras have also been trained for riding or traction (Brubaker and Coss, 2015). For example, in 1891 Captain M.H. Hayes trained an old mountain zebra stallion for riding in two days for his wife, Alice M. Hayes, to ride and trained a young plains (Burchell’s) zebra to accept a rider in one hour in 1892. Although these zebras were captive animals used to being handled before their training for riding began (Hayes, 1893:311-313), this demonstrates that the natural aversion of wild equids against a rider, which these prey animals easily perceive as an attacking predator (see Dietz, 2003:190), can be overcome.

Figure 7

Figure 7

A Przewalski’s stallion named "Vaska" photographed in 1904. This image is in the public domain|
L’étalon de Przewalski nommé "Vaska" photographié en 1904. Cette image est dans le domaine public

https://upload.wikimedia.org/​wikipedia/​commons/​c/​cf/​Equus_ferus_przewalskii_2.jpg

54Selection for docility and tolerance of restraint must have already started during the first generations of captive breeding. Horses naturally dislike being restrained. Early captive horses probably objected to being haltered and led, just like current Przewalski’s horses still do after a century of captive breeding, according to Bowling and Ruvinsky (2000:30). Selection would have intensified when aggressive and/or panicky mares intolerant of milking were culled. In addition, milking resulted in closer human – horse relationships that possibly encouraged more individuals to try horseback riding. Riding captive wild horses on the steppe was not likely to be more difficult than riding captive wild onagers, which, based on artistic depictions, occurred in Mesopotamia during the 3rd millennium BC (see Moorey, 1970).

Horseback riding became necessary when managing horses required mounted herding

55Riding horses became more common over time on the steppe most probably because there were no good alternatives to mounted herding if the horse herds were large and the horses ranged freely and/or needed to be moved over long distances (e.g., Clutton-Brock, 1992:12; Azzaroli, 1985:6, 1998:41; Khazanov, 1994:92; Levine, 1999; 2005; Taylor et al., 2020). Mounted herding would have been very useful or even necessary with large herds of horses even when they were not being moved over long distances, because the alternatives, discussed below, were either impractical or just temporary solutions.

56Drews (2004:44-47 in E-book) thought that free-ranging horses grazing half a mile from their keepers’ settlement would have galloped towards that settlement when in danger, stayed within calling range and immediately obeyed food calls, and that penning or tethering the foals would have ensured that the older horses stayed close by. In fact, these horses would have galloped in any direction they regarded as safe. They would have strayed beyond the calling range in search of grazing unless the only fodder available was kept by their keepers, which was unlikely even in winter because horses are good at reaching grass under snow (Khazanov, 1994:50; Ewers, 2001:42). Separating dams from their unweaned foals by penning the latter is very stressful for both and easily arouses aggression from both dams (Crowell-Davies and Weeks, 2005) and sires. Harem stallions protecting their foals were so dangerous that they had to be killed during the large-scale operations conducted to capture Przewalski’s foals for zoos at the turn of the 19th and 20th centuries (Bökönyi, 2008:92-93). The Mongolian horse herders manage to keep mares to be milked and other horses successfully nearby by tethering foals to a line (Taylor et al., 2023: Supplementary Material p. 22), but this method is only applicable until foals are weaned and may not be equally applicable on horses not yet fully domesticated.

57Keeping horses corralled all the time is impractical because they need water and fodder. Moreover, wood for constructing corrals is scarce in steppe environments, which restricts corralling to riparian forests. The Plains Indians took advantage of rolling hills, dips and coulees forming natural enclosures to prevent horses from straying, to shelter them against the elements and to hide them from horse raiders when on the move (Ewers, 2001:44; Bethke, 2017:804).

58Picketing horses by tying one end of a rope around the neck or one of the fetlocks and the other end to a picket pin driven into the ground can only be a temporary solution. The Plains Indians generally only picketed a particular horse or a small group for the night within the camp as a precaution against theft (Mandelbaum 1940:196; Ewers, 2001:39-40). Picketed horses must be frequently relocated and/or brought fodder and water because they quickly consume all the fodder within their reach. Supervision is also necessary because serious injuries can easily result when horses take fright and try to bolt. I once witnessed a picketed wild-born but tamed mustang mare somersaulting when she attempted to bolt.

59Hobbling by tying a rope restraint around the distal limbs can also prevent horses from moving too far and/or too fast (Ewers, 2001:38-39) but only as a temporary solution. Because hobbled forelegs prevent the horse from pawing snow with its hooves to access grass, the Blackfoot did not hobble their horses when there was snow on the ground (Ewers, 2001:42) and the Plains Cree hobbled hind legs instead of forelegs (Mandelbaum, 1940:196). As hobbling an entire herd is impractical, only the lead mare or a small group of horses would be hobbled for the night (Ewers, 2001:38-39).

60For lack of good alternatives, mounted herding of horses began when horse herds became too large for herders to cope with on foot. Large herds certainly existed by about 2500 BC, as indicated by the large numbers of horses sacrificed in burial rituals by the Catacomb culture (Anthony, 2007:325). A shift to mounted herding also allowed herders to manage larger herds of other livestock. A herder on foot can manage 150-200 sheep, but 500 sheep or 150 horses when mounted (Khazanov, 1994:32 and references therein). This advantage of mounted over pedestrian herding would have motivated horseless groups to become equestrian.

Did horseback riding drive selection resulting in DOM2 horses?

61Although horses from the Pontic-Caspian steppe ancestral to DOM2 horses may have exhibited a front-heavy conformation (see Dietz, 2003: 191), comparisons with Mongolian horses, Icelandic horses and horses of the Plains Indians show that these were also big and strong enough to carry adult male riders for herding, hunting and warfare (table 1, figure 8). There was therefore no reason to use them only as packhorses and/or for meat and milk. Size and strength were even more unlikely to be limiting factors for riding early DOM2 horses (e.g., those of the Sintashta-Petrovka and Begazy-Dandybaev cultures), which were of about the same size and strength, but were less front-heavy in conformation (table 1).

Figure 8

Figure 8

Average ancestral DOM2 horse (about 135 cm tall) and average Eneolithic / Bronze Age man (about 175 cm tall) from the Pontic-Caspian steppe |
Cheval ancestral moyen DOM2 (environ 135 cm) et homme moyen de l’Énéolithique / de l’âge du bronze (environ 175 cm) de la steppe pontique-caspienne

Table 1

Table 1

Mean, standard deviation, minimum and maximum values of estimated withers heights (WH), body masses (BM), and maximum rider weights (XRW). Sources are as follows: 1) Table S1; 2) Kosintsev (1995: 6 referenced by Kuzmina 2008: 44); 3) Kanne (2018: Table 10.5 and references therein); 4) De Grossi Mazzorin et al. (1998: Table 1) |
Moyenne, écart type, valeurs minimales et maximales des hauteurs au garrot estimées (WH), des masses corporelles (BM), et des poids maximaux du cavalier (XRW). Les sources sont les suivantes : 1) Tableau S1 ; 2) Kosintsev (1995 : 6 référencé par Kuzmina 2008 : 44) ; 3) Kanne (2018 : Tableau 10.5 et références y figurant) ; 4) De Grossi Mazzorin et al. (1998 : Tableau 1)

62However, ancestral DOM2 horses must have been more challenging to ride than most DOM2 horses if they were like the zebra ridden by Alice M. Hayes. She described this zebra as obstinate and intolerant of being touched, and exhibiting a tendency to kick, strong flight-or-fight reactions and a neck and throatlatch conformation unsuitable for reining (Hayes, 1903: 456, Fig. 146). Early mounted herders therefore probably had no more control over their still semi-wild horses than Mongolian and Siberian reindeer herders have when riding their reindeer (Stépanoff, 2012; Davydov, 2017). But, like these reindeer herders, they would still have found mounted herding easier than herding on foot. Their obstinate horses would at least have stubbornly kept up with the horses being herded and, if they were dominant individuals, would have insisted on herding them in any case. The usefulness of these horses for riding may have been somewhat limited by a higher likelihood than in DOM2 horses of developing chronic back problems if ridden a lot, because their evolution had not prepared them to carry weights on their backs.

63Selection for more rideable horses intensified as riding became more common, perhaps from around 3100 BC based on the associated diagnostic traits in human skeletons in very late pre-Yamnaya or very early Yamnaya individuals from Romania reported by Trautmann et al. (2023). Excitable, exuberant and/or fearful horses and those suffering from pain in the spinal region are prone to buck (Dyson and Thomson, 2022). Culling the most violently bucking horses, as well as generally fearful and aggressive ones, would have reduced genetic predispositions for panic and aggressiveness, resulting in changes in the ZFPM1 gene, and also chronic back problems, resulting in changes in the GSDMC gene as noted by Librado et al. (2121). Specific variants of these genes in DOM2 horses must have reached high frequencies when the distinctive ancestry profile of DOM2 horses emerged in about 2300-2200 BC (figure 4).

64A significant increase in carrying loads on the back during the centuries separating the most recent C-PONT horses and the earliest DOM2 horses, around the middle of the 3rd millennium BC, is the most logical explanation of why selection strongly eliminated genetic risk factors for chronic back problems. Use of horses as packhorses probably began at about the same time as riding and would also have contributed to this selection. In any event, even the earliest DOM2 horses would have been temperamentally and physically about as suitable for competent riding as more recent DOM2 horses thanks to these genetic changes.

65Horseback riding facilitated mounted herding and thus made it possible to move horses over long distances. This probably aided the initial dispersal of DOM2 horses, which started at about 2300±150 BC (this study) or 2200 BC (Librado et al., 2021) when these horses were still emerging. Librado et al. (2021) believes that horseback riding "fuelled" this initial dispersal because it predates the earliest evidence of chariots, and thus puts "the horse before the cart" rather than "the cart before the horse".

66Because changes in the ZFPM1 and GSDMC genes had made the earliest DOM2 horses temperamentally and physically about as suitable for proper riding as more recent DOM2 horses, proper riding of these horses could only have been prevented by early riders being physically and/or intellectually inferior to more recent riders, which could hardly have been the case. Due to their more physical lifestyle, they were fitter on average than industrial populations, making them more suited to the physical aspects of riding. They must also have been able to understand and, if necessary, manhandle horses and the other large-sized and potentially dangerous species they herded. They did not have examples and instructions to begin with, but it would hardly have required centuries to learn how horses react to a rider’s hand, body and leg movements. Because horses respond negatively to heavy-handed riding (see König Von Borstel and Glißman, 2014; Christensen et al., 2021), these early riders must have soon learned to ride more with their seat than with their hands.

67Bits made of bronze or other metals are not prerequisites for proper riding of DOM2 horses. For example, the Plains and Plateau Indians managed superbly with only a rawhide "war bridle" tied around the lower jaw of the horse. If necessary, they rode without any kind of bridle, aided only by their legs and weight. Some even hunted deer and pronghorn antelope on horseback (Ewers, 2001:70, 153, 170). A Bronze Age seal impression from the Bactria-Margiana Archaeological Complex (2100-1800 BC) in Afghanistan shows proper riding at a gallop without a metal bit during the period when DOM2 horses were spreading, with the correct posture and seat of the rider close behind the withers (figure 9).

Figure 9

Figure 9

Part of a seal impression from the Bactria-Margiana Archaeological Complex (BMAC), Afghanistan, dated to 2100-1800 BC|
Partie d’une empreinte de sceau du complexe archéologique Bactria-Margiana (BMAC), Afghanistan, datée de 2100-1800 av. J.-C.

redrawn after Anthony, 2007: Fig. 16.3 / redessinée d’après Anthony, 2007 : Fig. 16.3

Chariotry and other forms of traction

68Although horseback riding became common from about 3100 BC (see Trautmann et al., 2023), the role of traction should not be dismissed. Some horses may have already been harnessed for traction during the initial dispersal of DOM2 horses if two-wheeled carts were pulled by horses in the Don-Volga interfluve in 2400-2200 BC (Chechushkov and Epimakhov, 2018) and thus before chariots were introduced around 2000 BC (Lindner, 2020). The Sintashta-Petrovka chariots and harness were so highly developed that lengthy experimentation must have been required. Also, these chariots could hardly have been conceived in the minds of their inventors before horses had demonstrated their usefulness for traction by pulling earlier wheeled vehicles. They are thus unlikely to represent the earliest wheeled vehicles pulled by horses.

69Although the dispersal of the chariot complex during the first half of the 2nd millennium BC was concurrent with the transformation of the Eurasian Bronze Age societies (Kristiansen, 2018), the role and importance of chariotry may have become inflated due to its strong association with high status. The practical utility of chariots is also questionable (e.g., Barclay, 1982: Littauer and Crouwel, 1996; Levine, 1999; 2005; Chechushkov and Epimakhov, 2018; Kristiansen, 2018; Maran, 2020; Metzner-Nebelsick, 2021; Kanne, 2022).

70It is obvious that chariots are inferior to other wheeled vehicles for transporting people and goods. Moreover, the cross-country performance of all wheeled vehicles pulled by horses is inadequate for many herding and hunting situations. A chariot can break or capsize if turns are taken too fast, if a wheel hits a large rock at high speed or when going down a slope (hillsides or banks of rivers and creeks) of more than 26.8 degrees, which is also too steep for uphill traction (see Efkleidou, 2019: Table 1). Waterlogged ground, deep snow, trees and other obstructions less than two metres apart will stop chariots, as will low obstacles (e.g., fallen tree trunks) that are easily jumped over by horses. Chariot teams cannot match the speed of horses with equal physical attributes but unencumbered by chariots. Nor can they separate individual horses from a herd because they cannot stop, turn and change direction fast enough. Chariots could have even scared herds of horses into stampedes (Barclay, 1982; Littauer and Crouwel, 1996; Chechushkov and Epimakhov, 2018).

71Even if chariots were suited to herding, there would not have been enough chariot teams because, as noted by Chechushkov and Epimakhov (2018), charioteers were a small numerical minority even in charioteering societies. Also, more than twice the number of horses would have been needed for chariot teams to equal mounted herders in number if each chariot is pulled by two horses, because packhorses would have to carry spare parts (e.g., wheels) and dismantled chariots across unsuitable terrain. Each chariot team would have included a driver and a herder, thus doubling the number of people required. Additional manpower would have been required to handle packhorses and to manage a double-sized herd because of the spare horses. Herding on horseback and/or on foot would have been required whenever chariots were carried by packhorses over unsuitable terrain, making chariots both redundant and an extra burden on long drives of livestock.

72The unsuitability of chariots for high-speed cross-country chases limited their usefulness in hunting and warfare. They may have been primarily used to transport an aristocratic hunter to a hunting location or battleground (Crouwel, 2013:86; Chechushkov and Epimakhov, 2018). They would have been useful as elevated and mobile firing platforms in both hunting and warfare, but beaters were likely to have been needed to drive game towards the aristocratic hunters waiting in their chariots (Crouwel, 2013:86-88).

73Due to the above, I interpret a Mongolian petroglyph (see Littauer et al., 2002:113) depicting a high-status individual viewing his herd of horses from his chariot rather than hunting or herding them. Even Robert Drews, a prominent supporter of the "cart before the horse" view, thought that on the steppes "…the chariot was a means of transportation, a recreational vehicle, and a status symbol…". He believed this was because there were no rulers who could have assembled and maintained chariot armies (Drews, 2004:136-139 in E-book), but the limited practical utility of wheeled vehicles in herding and hunting was perhaps a more important factor.

74Controlling horses harnessed to pull wheeled vehicles is not necessarily easier than controlling horses when riding. The driver can communicate with the horse only via the reins, but the rider can also communicate via weight and legs as well as by touch with the hand (Hančar, 1956:552). Also, horses used for traction need to be more docile, tolerant of restraint, obedient and better trained than for riding. The Plains Indians selected mature mares with a gentle and trainable disposition, and previously ridden to train them to pull travois (Ewers, 2001:64-65); the Khalkha selected mature and previously saddle-broken horses to be trained to pull carts (Levine, 1990:730); Captain M.H. Hayes advised that a horse should be first trained for riding before training it for traction (Hayes, 1889:212).

75Longer training is especially needed in the case of paired horses (Chechushkov and Epimakhov, 2018:470). Although harnessing horses in pairs may make them feel more secure and less excitable (Dietz, 2003:190), two bolting horses would be harder to control and cause more injuries and wreckage than just one bolting horse. It is therefore not surprising that chariot horses required at least seven months of training, based on Kikkuli’s text from around 1400 BC (Raulwing, 2009), but recent cavalry remounts only two months (Nolan, 1852:1-2). Furthermore, training chariot horses included training for riding in case a chariot broke down or one horse of the pair became disabled (see Drews, 2004:89-91 in E-book).

76The Plains Indians used their first horses for packing and/or pulling travois before riding. They had no previous experience with horses and thus learned to use trained and docile horses acquired via trade, with appropriate tack for packing and traction, sooner than they learned to ride (Ewers, 2001:5, 14; Hämäläinen, 2019:59). The situation on the Eurasian Steppe was different because horses had not yet been used for any of these purposes and no tack existed.

77In any event, chariotry and other forms of traction did not replace or exclude riding, which, based on skeletal evidence reported by Trautmann et al. (2023), had been experimented with since the second half of the 5th millennium BC. Due to the clear advantages of mounted herding, the chariot complex would have been imperfect or even incomplete if it had not included mounted herders in addition to chariot horses, chariots, chariot builders, grooms, horse trainers and veterinarians. For example, bringing the first chariot horses to north-eastern Syria around 1800 BC (Klecel and Martyniuk, 2021) and Anyang, the Shang capital, about 1180 BC (Kalekna, 2009:136-137 and referenced therein) would have been difficult without mounted herders. Ensuring that prestigious and valuable chariots arrived in brand new condition would have also required their transportation using carts, wagons or packhorses.

The earliest mounted warfare

78Mounted raiding and skirmishing between tribal steppe societies almost certainly preceded the cavalry warfare of the second half of the 2nd millennium BC and later periods (Sidnell, 2006:3; Anthony et al., 2006; Anthony and Brown, 2011). Mounted raiding probably began as early as mounted herding because it required no more riding tack and even less skill. If horses were only used as a means of travel, raiders would only have needed to stay on horseback and have reasonable control over speed and direction. Mounted raiders can easily evade pursuers on foot after hitting their target because they can travel at least twice as fast (Ewers, 2001:184-189). The earliest skirmishes on horseback may have been between mounted herders and mounted livestock raiders.

79More elaborate mounted combat and hunting large, fast-moving game species require more skilful riders and more immediate obedience from the horse. DOM2 horses as obedient as present-day horses may have been required, but horsemanship matters here more than metal bits, spurs, etc. As already mentioned, the Plains and Plateau Indians only used a rawhide "war bridle" tied around the lower jaw of the horse. If necessary, they rode without any bridle, aided only by their legs and weight (Ewers, 2001:70, 153, 170), as Numidian cavalrymen customarily did still during the Second Punic War (Sidnell, 2006:172). The lack of evidence of chariot warfare from the steppes (Drews, 2004:136-139 in E-book) suggests that horseback riding may have had a more important role in warfare than generally thought in this horse-rich ecoregion during the 2nd millennium BC.

The Przewalski’s horse is the only remaining wild horse and non-DOM2 horse

80Malleable temperaments, resilient backs and other qualities of DOM2 horses explain why these horses had replaced all other horses by about 1000 BC, except those ancestral to historic wild horses (Librado et al., 2021), largely through large-scale genetic introgression. Feralization of partly domesticated lineages may have also occurred (Outram et al., 2021), but I disagree with the feral status of the Przewalski’s horse (Gaunitz et al., 2018). The lack of evidence of horse domestication from the Tersek culture sites (Outram et al., 2009) and the evidence of horse hunting at Botai (Olsen, 2006a) imply that genomic samples from the Botai and Tersek culture sites are likely to include samples from both captively bred and hunted horses. Descent from horses at the incipient stage of domestication that were bred in captivity for few hundred years during the 4th millennium BC does not warrant this feral status, especially if free-living wild horses are also included among ancestors. The Przewalski’s horse is the only remaining wild horse and, as the only existing non-DOM2 horse, a living relic. Conservation of this horse is therefore of the utmost importance.

Final remarks and suggestions

81The domestication of horses on the Pontic-Caspian steppe and the subsequent dispersal of DOM2 horses certainly changed the world. Horseback riding probably drove the selection that resulted in these horses, but riding itself was not revolutionary. Because all domestic and semi-domestic species that can be ridden are ridden at least somewhere, horseback riding was "destined" to begin as soon as suitable horses for riding existed, and to become common as soon as it became necessary to herd horses on horseback.

82Riding horses had been experimented with since the second half of the 5th millennium BC, became common and widespread during the early stages of the Yamnaya culture around 3100 BC, and necessary by the middle of the third millennium BC at the very latest, because the daily management and especially moving of large herds of horses over long distances required mounted herding.

83The size and strength of ancestral and early DOM2 horses were not limiting factors for horseback riding. As it became more common, selection began to favour horses with malleable temperaments and resilient backs, which resulted in changes in the ZFPM1 and GSDMC genes respectively. This selection also resulted in the distinct ancestry profile of DOM2 horses, which is estimated here to have emerged in about 2300-2200 BC. Because their initial dispersal is estimated here to have begun at 2300±150 BC, these horses began to disperse elsewhere while they were still emerging.

84Horses may have been used as packhorses for about as long as they have been ridden, but using horses for traction had to wait until they were temperamentally more suitable and the necessary tack was developed. Although chariotry spread together with DOM2 horses during the 2nd millennium BC, common people had little or no access to chariots, which were mostly prestige items of limited practical utility for herding horses, for example.

85Our knowledge on horse domestication and the evolution of equestrianism is rapidly increasing and the "big picture" is emerging, but there will always be gaps in this knowledge due to incomplete evidence over time and space, as well as the preservation bias. Some of these gaps can be filled by reconstructions (subject to constant revisions) that are derived by combining scientific evidence provided by archaeological, biological and behavioural research with knowledge accumulated by equestrian peoples over centuries and by individuals over decades of first-hand equestrian experience. This first-hand experience is necessary to truly understand horses and what it takes to herd, train, ride and drive them, and helps to sift out incorrect assumptions. Additional experience, and therefore knowledge, could also be gained through scientific experimental studies on, for example, taming and training wild equids.

Supplementary information

86The supplementary information is available in .docx from the BMSAP website (https://journals.openedition.org/​bmsap/​11934?file=1; https://journals.openedition.org/​bmsap/​11939?file=1). It includes SI A “Dissemination of horses over the Great Plains and the adjacent regions of North America”; and SI B “Estimation of body size and ability to carry weight” with tables S1-S12 and figures S1-S4.

Haut de page

Bibliographie

Anthony DW (2007) The horse, the wheel, and language: How Bronze Age riders from the Eurasian steppes shaped the modern world. Princeton, NJ: Princeton University Press

Anthony DW (2016) The Samara Valley project and the evolution of pastoral economies in the Western Eurasian steppes. In: Anthony DW, Brown D, Kuznetsov P et al (eds) A Bronze Age Landscape in the Russian Steppes: The Samara Valley Project. Monumenta Archaeologica 37. Cotsen Institute of Archaeology Press at UCLA, Los Angeles, pp 3-36

Anthony DW, Brown DR (2000) Eneolithic horse exploitation in the Eurasian steppes: diet, ritual and riding. Antiquity 74:75-86 [https://doi.org/10.1017/S0003598X00066163]

Anthony DW, Brown DR (2003) Eneolithic horse rituals and riding in the Steppes: new evidence. In: Levine M, Rassamakin Y, Kislenko A et al (eds) Late Prehistoric Exploitation of the Eurasian Steppe. Series: McDonald Institute Monographs. Publisher: McDonald Institute for Archaeological Research, Cambridge, pp 55-68

Anthony D, Brown D (2011) The Secondary Products Revolution, horse riding, and mounted warfare. Journal of World Prehistory 24, 131-160 [https://doi.org/10.1007/s10963-011-9051-9]

Anthony D Brown D (2014) Horseback riding and Bronze Age pastoralism in the Eurasian Steppe. In: Mair VH, Hickman J (eds) Reconfigurating the Silk Road: New Research on East-West Exchange in Antiquity. University of Pennsylvania Museum of Archaeology and Anthropology, Philadelphia, pp 55-71

Anthony DW, Brown DR, George C (2006) Early horseback riding and warfare: The importance of the magpie around the neck. In: Olsen S, Grant S, Choyke A et al (eds) Horses and Humans: The Evolution of the Equine–Human Relationship. British Archaeological Reports International Series 1560:137-156. British Archaeological Reports, Oxford [https://doi.org/10.30861/9781841719900]

Anthony DW, Khokhlov AA, Agapov SA et al (2022) The Eneolithic cemetery at Khvalynsk on the Volga River. Praehistorische Zeitschrift 2022; 97(1):22-67 [https://doi.org/10.1515/pz-2022-2334]

Azzarolli A (1985) An Early History of Horsemanship. Brill, Leiden

Azzarolli A (1998) Outlines of early equitation. In: Anreiter P, Bartosiewicz L, Jerem E (eds) Man and the Animal World: Studies in Archaeozoology, Archaeology, Anthropology and Paleolinguistics in Memoriam Sánder Bökönyi. Archaeolingua, Budaperst, pp 41-53

Barclay HB (1982) Another look at the origins of horse riding. Anthropos 77(1/2):244-249 [www.jstor.org/stable/40460442]

Benecke N (2009) On the beginning of horse husbandry in the southern Balkan Peninsula – the horses from Kirklareli-Kanligecit (Turkish Thrace). In: Mashkour M (ed) Equids in Time and Space. Proceedings of the 9th International Conference of Archaeozoology. Oxbow, Oxford, pp. 92-100

Bennett EA, Weber J, Bendhafer W et al. (2022) The genetic identity of the earliest human-made hybrid animals, the kungas of Syro-Mesopotamia. Science Advances 8(2):eabm0218 [https://doi.org/10.1126/sciadv.abm0218]

Berger J (1986) Wild Horses of the Great Basin: Social Competition and Population Size. University of Chicago Press, Chicago

Bethke B (2017) The archaeology of pastoralist landscapes in the Northwestern Plains. American Antiquity 82(4):798-815 [https://doi.org/10.1017/aaq.2017.44]

Bowling AT, Ruvinsky A (2000) Genetic aspects of domestication breeds and their origin. In: Bowling AT, Ruvinsky A (eds) The Genetics of the Horse, Chapter 2 (CAB International, Wallingford 2000), pp 25-51

Bréhard S, Bălăşescu A (2012) What’s behind the tell phenomenon? An archaeozoological approach of Eneolithic sites in Romania. Journal of Archaeological Science 39(19):3167-3183 [https://doi.org/10.1016/j.jas.2012.04.054]

Brubaker AS, Coss RG (2015) Evolutionary constraints on equid domestication: Comparison of flight initiation distances of wild horses (Equus caballus ferus) and plains zebras (Equus quagga). Journal of Comparative Psychology 129(4):366-376 [http://dx.doi.org/10.1037/a0039677]

Buchanan B (1966) Catalogue of Ancient Near Eastern Seals in the Ashmolean Museum. Vol. 1, Cylinder Seals. Clarendon, Oxford

Bökönyi S (2008) Das Przewalski-Pferd oder Das mongolische Wildpferd. Die Wiederbelebung einer fast ausgestorbenen Tierart. Archaeoligua, Innsbruck

Chechushkov IV, Epimakhov AV (2018) Eurasian steppe chariots and social complexity during the Bronze Age. Journal of World Prehistory 31:435-483 [https://doi.org/10.1007/s10963-018-9124-0]

Christensen JW, Munk R, Hawson L et al. (2021) Rider effects on horses’ conflict behaviour, rein tension, physiological measures and rideability scores. Applied Animal Behavioral Science 234: 105184 [https://doi.org/10.1016/j.applanim.2020.105184]

Clutton-Brock J (1992) Horse Power: A History of the Horse and the Donkey in Human Societies. Harvard University Press, Cambridge, MA

Clutton-Brock J (2012) Animals as Domesticates: A World View Through History. Michigan University Press, Michigan

Crouwel J (2013) Studying the six chariots from the tomb of Tutankhamun – an update. In: Veldmeijer AJ, Ikram S (eds) Chasing Chariots: Proceedings of the First International Chariot Conference (Cairo 2012). Sidestone Press, Leiden, pp 73-93

Crowell-Davis SL, Weeks J (2005) Maternal behaviour and mare-foal interaction. In: Mills D, McDonnell S (eds) The domestic horse: The origins, development and management of its behaviour. Cambridge University Press, Cambridge, UK, pp 126-138

Davydov VN (2017) Research on Human and Reindeer Relations in Southern Yakutia [Sakha Republic], Anthropology & Archeology of Eurasia 56(1-2):6-31 [https://doi.org/10.1080/10611959.2017.1352313]

De Grossi Mazzorin J, Riedel A, Tagliacozzo A (1998) Horse remains in Italy from the Eneolithic to the Roman period. In: Peretto C, Giunchi C (eds) Proceedings of the XIII International Congress of the UISPP. Forlí, Italy: Abaco, v. 6(1), 87-92

Dietz UL (2003) Horseback riding: man’s access to speed. In: Levine M, Renfrew C, Boyle K (eds) Prehistoric Steppe Adaptation and the Horse. MacDonald Institute for Archaeological Research, Oxford, pp 189-199

Dornburg A, Beaulieu JM, Oliver JC et al (2011) Integrating fossil preservation biases in the selection of calibrations for molecular divergence time estimation conclusions. Systematic Biology 60(4):519-527 [https://doi.org/10.1093/sysbio/syr019]

Drews R (2004) Early Riders: The Beginning of Mounted Warfare in Asia and Europe. Routledge, London

Dyson S, Thomson K (2022) The recognition of pain and learned behaviour in horses with buck. Equine Veterinary Education 34(5):272-280 [https://doi.org/10.1111/eve.13466]

Efkleidou K (2019) Walking your way to death: exploring the relation between the location of Mycenaean chamber tombs and roads in the Argolid. Open Archaeology 5:484-504 [http://doi.org/10.1515/opar-2019-0030]

Eveleth PB, Tanner JM (1976) Worldwide Variation in Human Growth. Cambridge University Press, Cambridge

Ewers JC (2001) The Horse in Blackfoot Indian Culture. Reprinted from the 1985 edition. University Press of the Pacific, Honolulu, Hawaii

Fages A, Hanghøj K, Khan N et al (2019) Tracking five millennia of horse management with extensive ancient genome time series. Cell 177(6):1419-1435.e31 [https://doi.org/10.1016/j.cell.2019.03.049]

Fitzhugh WW (2009) The Mongolian deer stone-khirigsuur complex: dating and organization of a Late Bronze Age menagerie. In: Bemmann J et al. (ed.) Current archaeological research in Mongolia (Bonn contributions to Asian archaeology 4):183-199. VFG-Arch, Bonn

Forbes JD (1959) The appearance of the mounted Indian in Northern Mexico and the Southwest, to 1680. Southwestern Journal of Anthropology 15(2):189-212 [https://doi.org/10.1086/soutjanth.15.2.3628807]

Gaunitz C, Fages A, Hanghøj K et al (2018) Ancient genomes revisit the ancestry of domestic and Przewalski’s horses. Science 360(6384):111-114 [https://doi.org/10.1126/science.aao3297]

Gmelin SG (1770) Reise durch Russlannd zur untersuchung der drey Natur-Reiche. Kayserl. Adademie der Wissenschaften, St. Petersburg

Govedarica B, Manzura I (2019) The Giurgiulesti cemetery in chronological and cultural context of Southeastern and Eastern Europe. Eurasia Antiqua 2016(22):1-39 [ISSN 0949-0434]

Guimaraes S, Arbuckle BS, Peters J et al (2020) Ancient DNA shows domestic horses were introduced in the southern Caucasus and Anatolia during the Bronze Age. Science Advances 6 (38):eabb0030 [https://doi.org/10.1126/sciadv.abb0030]

Hančar F (1956) Das Pferd in prähistorischer and früher historischer Zeit. Wiener Beiträge zur Kulturgeschichte und Linguistik, 11. Institut für Völkerkunde der Universität Wien. Herold Wien

Hayes AM (1903) The Horsewoman: A Practical Guide to Side-Saddle Riding. 2nd edition. Charles Scribner’s Sons, New York

Hayes MH (1889) Illustrated Horse Breaking. W. Thacker and Co., London

Hayes MH (1893) The Points of the Horse. W. Thacker and Co., London

Heptner VG, Nasimovich AA, Bannikov AG (1988) Mammals of the Soviet Union (Vol. 1). Artiodactyla and Perissodactyla. Smithsonian Institution Libraries and National Science Foundation, Washington, DC

Hämäläinen P (2019) Lakota America: A New History of Indigenous Power. Yale University Press, New Haven

Jiang H, Moro A, Liu Y et al. (2020) Two GWAS-identified variants are associated with lumbar spinal stenosis and Gasdermin-C expression in Chinese population. Scientific Reports 10:21069 [https://doi.org/10.1038/s41598-020-78249-7]

Kalekna P (2009) The Horse in Human History. Cambridge University Pres, Cambridge

Khazanov AM (1994) Nomads and the Outside World. Second Edition. Cambridge University Press, Cambridge

Kanne K (2018) Becoming equestrian: people and horses in Bronze Age Hungary. PhD dissertation, Northwestern University, Evanston, IL

Kanne K (2022) Riding ruling and resistance: Equestrianism and political authority in the Hungarian Bronze Age. Current Anthropology 63(3):289-329 [https://doi.org/10.1086/720271]

Klecel W, Martyniuk E (2021) From the Eurasian Steppes to the Roman circuses: A review of early development of horse breeding and management. Animals 2021(11):1859 [https://doi.org/10.3390/ani11071859]

Kotova NS (2008) Early Neolithic in the Pontic Steppes. BAR International Series 1735. Oxford University Press, Oxford

Kotova N, Makhortykh S (2010) Human adaptation to past climate changes in the Northern Pontic steppe. Quaternary International 220(1-2):88-94 [https://doi.org/10.1016/j.quaint.2009.09.026]

Kristiansen K (2018) The rise of Bronze Age peripheries and the expansion of international trade 1950-1100 BC. In: Kristiansen K, Lindkvist T, Myrdal J (eds.) Trade and civilization: economic networks and cultural ties, from prehistory to the Early Modern Period. Cambridge University Press [AKO], Cambridge, pp 87-112.

Kuzmina EE (2003) Origins of pastoralism in the Eurasian steppes. In: Levine M, Renfrew C, Boyle K (eds) Prehistoric steppe adaptation and the horse. McDonald Institute, Cambridge, pp 203-232

Kuzmina EE (2008) The Prehistory of the Silk Road. Mair VH (ed). University of Pennsylvania Press, Philadelphia

König Von Borstel U, Glißman C (2014) Alternatives to conventional evaluation of rideability in horse performance tests: suitability of rein tension and behavioural parameters. PLoS One 9: e87285 [https://doi.org/10.1371/journal.pone.0087285]

Lasley BL, Loskutoff NM, Anderson GB (1994) The limitation of conventional breeding programs and the need and promise of assisted reproduction in nondomestic species. Theriogenology 41(1):119-132 [http://dx.doi.org/10.1016/S0093-691X(05)80057-3]

Levine M (1990) Dereivka and the problem of horse domestication. Antiquity 64(245):727-740 [https://doi.org/10.1017/S0003598X00078832]

Levine M (1999) The origins of horse husbandry on the Eurasian steppe. In: Levine M, Rassamakin Y, Kislenko A et al (eds) Late Prehistoric Exploitation of the Eurasian Steppe. Series: McDonald Institute Monographs. Publisher: McDonald Institute for Archaeological Research, Cambridge, pp 5-58

Levine M (2003) Focusing on Central Eurasian archaeology: east meets west. In: Levine M, Renfrew C, Boyle K (ed) Prehistoric Steppe Adaptation and the Horse. MacDonald Institute for Archaeological Research, Oxford, pp 1-7

Levine M (2005) Domestication and early history of the horse. In: Mills DS, McDonnell SM (eds): The Domestic Horse: The Evolution, Development, and Management of Its Behaviour. Cambridge University Press, Cambridge, pp 5-22

Librado P, Khan N, Fages A et al (2021) The origins and spread of domestic horses from the Western Eurasian steppes. Nature 598: 634-640 [https://doi.org/10.1038/s41586-021-04018-9]

Lindner S (2020) Chariots in the Eurasian Steppe: a Bayesian approach to the emergence of horse-drawn transport in the early second millennium BC. Antiquity 94(374):361-380 [https://doi.org/10.15184/aqy.202.37]

Littauer MA, Crouwel JH (1996) The origin of the true chariot. Antiquity 70(270):934-939 [https://doi.org/10.1017/S0003598X00084192]

Littauer MA, Crouwel JH, Raulwing P (2002). Selected writings on chariots and the early vehicles, riding and harness. Brill, Leiden

Mandelbaum DG (1940) The Plains Cree. American Museum of Natural History 37(2):155-316

Maran J (2020) The introduction of the horse-drawn light chariot: divergent responses to a technological innovation in societies between the Carpathian Basin and the East Mediterranean. In: Maran J, Băjenaru R, Ailincăi S-C et al (eds) Contacts between the Balkans, the Aegean and western Anatolia during the Bronze and Early Iron Age. Habelt, Bonn, pp 505-528

Metzner-Nebelsick C (2021) Chariots and horses in the Carpathian lands during the Bronze Age. In: Bragli B, Dietz A Földi Z et al (eds) Distant worlds and beyond. Heidelberg: Propylaeum, Heidelberg, pp 11-131

Mohr E (1971). The Asiatic Wild Horse. JA Allen, London

Moorey PRS (1970) Pictorial evidence for the history of horse-riding in Iraq before the Kassite period. Iraq 32(1):36-50 [https://www.jstor.org/stable/4199890]

Morgunova NL (2015) Pottery from the Volga area in the Samara and South Urals region from Eneolithic to Early Bronze Age. Documenta Praehistorica 42:311-319 [https://doi.org/10.4312/dp.42.22]

Morgunova NL, Khokhlova OS (2013) Chronology and periodization of the Pit-grave culture in the area between the Volga and Ural Rivers based on radiocarbon dating and paleopedological research. Radiocarbon 55(3):1286-1296 [https://doi.org/10.1017/S0033822200048190]

Motuzaitė Matuzevičiūtė G (2013) Neolithic Ukraine: A Review of theoretical and chronological interpretations. Archaeologia Báltica 20:136-149 [https://doi.org/15181/ab.v20i0.812]

Niskanen M (2022) Communication invitée du theme "Les relations entre les humains et les animaux" Body size, conformation, athletic ability and temperament of horses have likely affected the human – horse relationship over time. Bulletins et mémoires de la Société d’Anthropologie de Paris 34(S) | 2021 [https://doi.org/10.4000/bmsap.8132]

Nolan LE (1852) The Training of Cavalry Remount Horses. Parker Son and Bourn, London

Olsen SL (2006a) Early horse domestication: weighing the evidence. In: Olsen SL, Grant LS, Choyke AM et al (ed) Horses and Humans: The Evolution of Human-Equine Relationships. British Archaeological Reports International Series 1560:1-375. British Archaeological Reports, Oxford [https://doi.org/10.30861/9781841719900]

Olsen SL (2006b) Early horse domestication on the Eurasian steppe. In: Zeder MA, Bradley DG, Emshwiller E et al (eds) Documenting Domestication: New Genetic and Archaeological Paradigms. University of California Press, Berkeley, pp 245-269 [https://doi.org/10.1525/9780520932425-020]

Outram A, Bendrey R, Evershed R et al (2021) Rebuttal of Taylor and Barrón-Ortiz 2021 Rethinking the evidence for early horse domestication at Botai [https://doi.org/10.5281/zenodo.5142604]

Outram AK, Kasparov A, Steer N et al (2012) Patterns of pastoralism in later Bronze Age Kazakhstan: new evidence from faunal and lipid residual analysis. Journal of Archaeological Science 39:2424-2435 [https://doi.org/10.1016/j.jas.2012.02.009]

Outram AK, Steer NA, Bendrey R et al (2009) The earliest horse harnessing and milking. Science 323(5919):1332-1335 [https://doi.org/10.1126/science.1168594]

Owen DI (1991) The first equestrian: An Ur III glyptic scene. Acta Sumerologica 13:259-273

Peilieu C (1980) A background perspective: Some notes on the characterization and distribution of domestic animals in China. In: Hoefer JA, Tsuchitani PJ (ed) Animal Agriculture in China. CSCPRC Report No. 11. National Academy Press, Washington, DC, pp 5-15

Raulwing P (2009) The Kikkuli text: Hittite training instructions for chariot horses in the second half of the 2nd millennium BC and their interdisciplinary context [http://www.lrgaf.org/Peter_Raulwing_The_Kikkuli_Text_MasterFile_Dec_2009.pdf]

Rassamakin Y (2012) Eneolithic burial mounds in the Black Sea Steppe: from the first burial symbols to monumental ritual architecture. In: Ancestral Landscape. Burial mounds in the Copper and Bronze Ages (Central and Eastern Europe – Balkans – Adriatic – Aegean, 4th-2nd millennium B.C.) Proceedings of the International Conference held in Udine, May 15th-18th 2008. Maison de l’Orient et de la Méditerranée Jean Pouilloux, 2012, Lyon, pp 293-305 (Travaux de la Maison de l’Orient et de la Méditerranée. Série recherches archéologiques, 58) [http://www.persee.fr/doc/mom_2259-4884_2012_act_58_1_3470]

Rassamakin Y, Menotti F (2011) Chronological development of the Tripolye culture giant-settlement of Taliaki (Ukraine): 14C dating vs. pottery typology. Radiocarbon 53(4):645-657 [https://doi.org/10.1017/S0033822200039102]

Robinson GC (2001) Mustangs and mustanging in southwest Texas. In: Dobie JF, Boatright MC, Ransom HH (eds) Mustangs and Cow horses. University of North Texas Press, Denton, Texas, pp 3-21

Scopa C, Contalbridgo L, Greco A et al (2021) Emotional transfer in human – horse interaction: new perspectives on equine assisted interventions. Animals 2019(9):1030 [https://doi.org/10.3390/ani9121030]

Serpell JA (1989) Pet-keeping and animal domestication: A reappraisal. In: Clutton-Brock J (ed) The Walking Larder: Patterns of Domestication, Pastoralism, and Predation. Unwyn Hyman, London, pp 10-21

Sherratt A (1997) The secondary exploitation of animals in the Old World (revised from 1983). In: Sherratt A (ed) Economy and Society in Prehistoric Europe: Changing Perspectives. Princeton University Press, Princeton, MA, pp 199-228

Sherratt A (2003) The horse and the wheel: The dialectics of change in the circum-Pontic and adjacent areas, 4500-1500 BC. In: Levine M, Renfrew C, Boyle K (eds) Prehistoric Steppe Adaptation and the Horse. McDonald Institute, Cambridge, pp 233-252

Shishlina NI, Zazovskaya EP, van der Plicht J et al.(2009) Paleoecology, subsistence, and 14C chronology of the Eurasian Caspian steppe Bronze Age. Radiocarbon 51(2):481-499 [https://doi.org/10.1017/S0033822200055879]

Sidnell P (2006) Warhorse: Cavalry in Ancient Warfare. London, Hambledon Continuum

Signor III PW, Lipps JH (1982) Sampling bias, gradual extinction patterns, and catastrophes in the fossil record. In: Silver LT, Schultz PH (eds) Geological Implications of Impacts of Large Asteroids and Comets on the Earth. Special Paper of the Geological Society of America 190:291-296 [https://doi.org/10.1130/SPE190-p291]

Smith CH (1866) The Natural History of Horses, with Memoirs of Gesner. Henry G. Bohn, London

Smith JM Jr (2000) Dietary decadence and dynastic decline in the Mongol Empire. Journal of Asian History 34(1):35-52 [https://www.jstor.org/stable/i40089927]

Stefánsdóttir GJ, Ragnarsson S, Gunnarsson V. et al. (2014) Physiological responses to a breed evaluation field test in Icelandic horses. Animal 11(9):431-439 [https://doi.org/10.1017/S1751731113002309]

Stella E (1518) De Borussiae antiquitatibus. Libri duo. Basileae.

Stépanoff C (2012) Human-animal "joint commitment" in a reindeer herding system. HAU: Journal of Ethnographic Theory 2(2):287-312 [https://doi.org/10.14318/hau2.2.015]

Suri P, Palmer MR, Tsepilov YA et al (2018). Genome-wide meta-analysis of 158,000 individuals of European ancestry identifies three loci associated with chronic back pain. PLoS Genetics 14(9):e1007601 [https://doi.org/10.1371/journal.pgen.1007601]

Taira M, Desforges M (2021) ZFPM1 Necessary for Development of Serotonergic Projections Related to Anxiety and Contextual Fear Learning. The Journal of Neuroscience 41(18):3945-3947 [https://doi.org/10.1523/JNEUROSCI.3125-20.2021]

Taylor WTT, Barrón-Ortiz CI (2021) Rethinking the evidence for early horse domestication at Botai. Scientific Reports 11:7440 [https://doi.org/10.1038/s41598-021-86832-9]

Taylor WTT, Cao J, Fan W et al (2021) Understanding early horse transport in eastern Eurasia through analysis of dentition. Antiquity 95(384):1478-1494 [https://doi.org/10.15184/aqy.2021.146]

Taylor WTT, Clark J, Bayarsaikhan J et al (2020) Early pastoral economies and herding transitions in Eastern Eurasia. Scientific Reports 10:1001 [https://doi.org/10.1038/s41598-020-57735-y]

Taylor WTT, Librado P, Icu MHT et al (2023) Early dispersal of domestic horses into the Great Plains and northern Rockies. Science 379(6639):1316-1323 [https://doi.org/10.1126/science. adc9691]

Tikker L, Casarotto P, Sigh P et al (2020) Inactivation of the GATA cofactor ZFPM1 results in abnormal development of dorsal raphe serotonergic neuron subtypes and increased anxiety-like behavior. The Journal of Neuroscience 40(44):8669-8682 [https://doi.org/10.1523/JNEUROSCI.2252-19.2020]

Trautmann M, Frînculeasa A, Preda-Bălănică B et al (2023) First bioanthropological evidence for Yamnaya horsemanship. Science Advances 9(9):eade2451 [https://doi.org/10.1126/sciadv.ade24]

Turchin P, Currie TE, Turner EAL (2016) Mapping the spread of mounted warfare. Cliodynamics 7:217-227 [https://doi.org/10.21237/C7clio7233509]

Vybornov A, Kulkova M, Kosintsev P et al. (2018) Diet and chronology of Neolithic cultures (from 6500 to 4700 cal BC) in the Lower Volga basin. Radiocarbon 60(5):1597-1610 [https://doi.org/10.1017/RDC.2018.95]

Wilkin S, Miller AV, Fernandes R et al (2021) Dairying enabled Early Bronze Age Yamnaya steppe expansions. Nature 598: 629-633 [https://doi.org/10.1038/s41586-021-03798-4]

Haut de page

Documents annexes

Haut de page

Notes

1 A chariot refers to a type of two-wheeled cart suitable for warfare, hunting and other high-speed horse-powered traction. The term "charioteering" refers to the use of these specific type of carts.

2 Horses can reach grass through snow cover of up to 50-60 cm by pawing with their hooves (Khazanov, 1994:50; Ewers, 2001:42).

3 The ZFPM1 gene is associated with docility and stress tolerance (Tikker et al., 2020; Taira and Desforges, 2021) and the GSDMC gene with a lower risk of developing chronic back problems (Suri et al., 2018; Jiang et al., 2020).

4 This effect is the reverse of the Signor-Lipps effect (Spllil-Rongis is Signor-Lipps spelled backwards), which results in the latest evidence predating the latest occurrence due to the reduced frequency of occurrence towards the end (Signor and Lipps, 1982; Dornburg et al., 2011).

5 Gaunitz et al. (2018: Supplementary Material) thought that the Dereivka sample dates to the Iron Age presumably based on one intrusive horse burial (see Anthony and Brown, 2000). I consider this material to be Eneolithic because this site and the material recovered (except from the one intrusive burial) is soundly dated to the Eneolithic (see Anthony, 2007: 247-249, Table 11.1). I excluded the Iron Age horse and four juvenile specimens.

6 Khalkha Mongol men average 66.5 kg in body weight (Eveleth and Tanner, 1976: Table S1), to which clothing plus riding tack probably add about 12-15 kg. An average Icelandic rider weighs 83 kg (I have assumed fully clothed) and the combined weight of saddle and saddle blanket is 8.7 kg. The weight of bridle, reins, etc. may not have been added to this figure (Stefánsdóttir et al., 2014).

Haut de page

Table des illustrations

Titre Figure 1
Légende A drawing of a carved bone amulet depicting a horse from the Syezzheye cemetery |Dessin d’une amulette en os sculpté représentant un cheval du cimetière de Syezzheye
Crédits after Anthony, 2007: Fig. 9.8 / d’après Anthony, 2007: Fig. 9.8
URL http://journals.openedition.org/bmsap/docannexe/image/11881/img-1.png
Fichier image/png, 70k
Titre Figure 2
Légende Eneolithic mace heads from Novoorsk (A) and Suvorovo (B) |Têtes de masse énéolithiques de Novoorsk (A) et Suvorovo (B)
Crédits redrawn after Anthony (2007: Fig. 11.5) / redessinées d’après Anthony (2007 : Fig. 11.5)
URL http://journals.openedition.org/bmsap/docannexe/image/11881/img-2.png
Fichier image/png, 194k
Titre Figure 3
Légende A) Akkadian (2334-2154 BC) seal impression from Kish; B) Ur III (2112-2004 BC) seal impression from Abbakalla dated to 2037-2029 BC |A) Impression de sceau de la période akkadienne (2334-2154 av. J.-C.) de Kish; B) Impression de sceau de la période Ur III (2112-2004 av. J.-C.) d’Abbakalla datée de 2037-2029 av. J.-C.
Crédits A: redrawn from Buchanan, 1966 / redessinée d’après Buchanan, 1966; B: redrawn from Owen, 1991 / redessiné d’après Owen, 1991
URL http://journals.openedition.org/bmsap/docannexe/image/11881/img-3.png
Fichier image/png, 118k
Titre Figure 4
Légende Changes over time in the ancestry component (k=6) maximized in DOM2 horses from the earliest NEO-NCAS horses to DOM2 horses dated no earlier than 1000 BC (based on Librado et al., 2021: Extended Data Fig. 4). The blue horizontal bars represent time frames for the earliest milking of horses and their initial dispersal. The numbers refer to sites: 1) Varfolomeevka, 2) Algay, 3) Turganik, 4) Oroshaeomoe I, 5) Semenovka 1, 6) Aygurskiy 2, 7) Repin, 8) Sosnovka. The dashed line is the Loess curve through NEO-NCAS, C-PONT and DOM2 horses |Les changements temporels de la composante d’ascendance (k=6) sont maximisés chez les chevaux DOM2, des premiers chevaux NEO-NCAS aux chevaux DOM2 datés au plus tôt de 1000 av. J.-C. (basé sur Librado et al., 2021 : Données étendues Fig. 4). Les barres horizontales bleues représentent les délais pour la première traite des chevaux et la dispersion initiale. Les numéros renvoient aux sites : 1) Varfolomeevka, 2) Algay, 3) Turganik, 4) Oroshaeomoe I, 5) Semenovka 1, 6) Aygurskiy 2, 7) Repin, 8) Sosnovka. La ligne pointillée est la courbe de lœss passant par les chevaux NEO-NCAS, C-PONT et DOM2
URL http://journals.openedition.org/bmsap/docannexe/image/11881/img-4.png
Fichier image/png, 136k
Titre Figure 5
Légende Years when horses were first acquired by selected groups of Native Americans (A). Distances and rates of dissemination of horses (B). The distances are walking distances |Années d’acquisition des chevaux de groupes sélectionnés d’Amérindiens (A) Distances de diffusion et vitesses de diffusion des chevaux (B). Les distances sont des distances à pied
URL http://journals.openedition.org/bmsap/docannexe/image/11881/img-5.png
Fichier image/png, 233k
Titre Figure 6
Légende Dispersal of DOM2 horses from the Pontic-Caspian steppe. Dispersal times based on distances between adjacent points on the map (see figure 5) are given in parentheses to differentiate them from those based on total distances |Dispersion des chevaux DOM2 de la steppe pontique-caspienne. Les temps de dispersion basés sur les distances entre points adjacents sur la carte (voir figure 5) sont donnés entre parenthèses pour les différencier de ceux basés sur les distances totales
URL http://journals.openedition.org/bmsap/docannexe/image/11881/img-6.png
Fichier image/png, 203k
Titre Figure 7
Légende A Przewalski’s stallion named "Vaska" photographed in 1904. This image is in the public domain|L’étalon de Przewalski nommé "Vaska" photographié en 1904. Cette image est dans le domaine public
URL http://journals.openedition.org/bmsap/docannexe/image/11881/img-7.png
Fichier image/png, 363k
Titre Figure 8
Légende Average ancestral DOM2 horse (about 135 cm tall) and average Eneolithic / Bronze Age man (about 175 cm tall) from the Pontic-Caspian steppe |Cheval ancestral moyen DOM2 (environ 135 cm) et homme moyen de l’Énéolithique / de l’âge du bronze (environ 175 cm) de la steppe pontique-caspienne
URL http://journals.openedition.org/bmsap/docannexe/image/11881/img-8.png
Fichier image/png, 188k
Titre Table 1
Légende Mean, standard deviation, minimum and maximum values of estimated withers heights (WH), body masses (BM), and maximum rider weights (XRW). Sources are as follows: 1) Table S1; 2) Kosintsev (1995: 6 referenced by Kuzmina 2008: 44); 3) Kanne (2018: Table 10.5 and references therein); 4) De Grossi Mazzorin et al. (1998: Table 1) |Moyenne, écart type, valeurs minimales et maximales des hauteurs au garrot estimées (WH), des masses corporelles (BM), et des poids maximaux du cavalier (XRW). Les sources sont les suivantes : 1) Tableau S1 ; 2) Kosintsev (1995 : 6 référencé par Kuzmina 2008 : 44) ; 3) Kanne (2018 : Tableau 10.5 et références y figurant) ; 4) De Grossi Mazzorin et al. (1998 : Tableau 1)
URL http://journals.openedition.org/bmsap/docannexe/image/11881/img-9.png
Fichier image/png, 388k
Titre Figure 9
Légende Part of a seal impression from the Bactria-Margiana Archaeological Complex (BMAC), Afghanistan, dated to 2100-1800 BC|Partie d’une empreinte de sceau du complexe archéologique Bactria-Margiana (BMAC), Afghanistan, datée de 2100-1800 av. J.-C.
URL http://journals.openedition.org/bmsap/docannexe/image/11881/img-10.png
Fichier image/png, 85k
Haut de page

Pour citer cet article

Référence électronique

Markku Niskanen, « The prehistoric origins of the domestic horse and horseback riding »Bulletins et mémoires de la Société d’Anthropologie de Paris [En ligne], 35 (1) | 2023, mis en ligne le 22 avril 2023, consulté le 02 juin 2023. URL : http://journals.openedition.org/bmsap/11881 ; DOI : https://doi.org/10.4000/bmsap.11881

Haut de page

Auteur

Markku Niskanen

Research Unit for History, Culture and Communications, University of Oulu, POB 8000, 90014 Oulu, Finland ; markku.niskanen[at]oulu.fi

Articles du même auteur

Haut de page

Droits d’auteur

CC-BY-NC-ND-4.0

Creative Commons - Attribution - Pas d'Utilisation Commerciale - Pas de Modification 4.0 International - CC BY-NC-ND 4.0

https://creativecommons.org/licenses/by-nc-nd/4.0/

Haut de page
  • Logo Société d'Anthropologie de Paris
  • Logo Fonds National pour la Science Ouverte
  • Revue soutenue par l’Institut des sciences humaines et sociales du CNRS
    CNRS - Institut national des sciences humaines et sociales
  • OpenEdition Journals
Search OpenEdition Search

You will be redirected to OpenEdition Search