Botai tamed horses around 3500 BCE — but not the ones we ride today

Before the horse came in: the forest-steppe around 3700 BCE

The country that became the Botai homeland — the rolling forest-steppe of what is now North Kazakhstan Province, drained by the Ishim and its tributary the Iman-Burluk — was, on the eve of horse herding, a thinly peopled landscape of small Neolithic hunting bands. The Atlantic climatic optimum had passed; the steppe was warmer and wetter than today, with belts of birch and pine breaking the open grass. Wild ungulates were abundant. Wild horses (Equus ferus) ranged across the open ground in herds of dozens to low hundreds, alongside saiga antelope, aurochs, red deer, elk, and occasional bears at the woodland margins.¹

The human communities of the late fifth and early fourth millennium BCE in this region — sometimes grouped under the rubric of the Atbasar and the Surtandy cultures — left only thin archaeological traces. They built small, seasonal camps; they hunted with bow and bone-tipped spear; they processed wild ungulate carcasses on the spot, carrying back to camp only the meat-bearing limb bones; they made coarse, sand-tempered pottery and chipped stone tools from the cherts and silicified sandstones of the local outcrops. Their faunal assemblages are mixed — saiga, horse, red deer, elk, aurochs — in the proportions one expects of a forest-steppe hunting economy that takes whatever the landscape produces.²

The wild horse in this world was prey. It was also one of the more difficult prey species in the available repertoire: faster than aurochs over short bursts, more skittish than red deer, capable of sudden mass flight, and prone to migrating across enormous ranges in response to seasonal grass and water. Marsha Levine's Antiquity paper on hippophagy (1998) argued that the long-overlooked food value of horses on grassland — their meat-to-bone yield is excellent, their fat content rises sharply with winter coat, their prehensile lips let them graze through snow that defeats cattle — meant that any human population which solved the problem of horse capture would unlock a calorific surplus the steppe could not match by any other means.³ The problem was solving the capture.

The neighbours: a sketch of the wider Eneolithic

The forest-steppe was not an isolated province. To the west, in the lower Volga and Don basins, the Khvalynsk culture had been burying its dead with copper ornaments and the occasional horse skull since at least 4500 BCE — the horse already mattering as a symbol, even if the relationship was not yet pastoral. To the northwest, the Trypillia (Cucuteni-Tripolye) mega-settlements of present-day Ukraine and Moldova had built towns of two and three thousand people, the largest aggregations of human population anywhere in the world at that date, on a cattle-and-cereal economy that owed nothing to the horse. To the south, in the Caucasus and the Iranian plateau, the early metallurgical traditions that would later become Maikop and Kura-Araxes were under way. Each of these worlds had domestic stock — cattle, sheep, goats, pigs — and each ignored the horse. The forest-steppe band that Botai inhabited was the seam where the available domesticates ran out and where the only large animal worth managing, on the calorific terms Levine described, was the horse.⁴

What categories did not yet exist

By any later standard, the world of the early fourth millennium BCE on this stretch of the steppe was missing several categories. There was no herding of horses. There were no corrals. There was no riding, in any documentable sense. There was no spoke-wheeled vehicle. There were no chariots. There was no koumiss — no fermented mare's milk. There was no concept of the horse as wealth, no horse-burials with grave goods, no chief whose status was recorded in mounts.

What there was, instead, was a settled pattern of opportunistic hunting in which horses were one species among several and the human communities lived in small, mobile camps that left only ephemeral marks. The shift from that world to the world of Botai — a single site with 153 pit-houses, more than 300,000 catalogued bone fragments of which over 99 percent are horse, residential phosphorus signatures inside what were almost certainly horse-corrals abutting the houses themselves — is the change that needs explaining.⁵

The transmission: how horses entered the human household at Botai

The site and its discovery

The settlement of Botai sits on a low terrace above the Iman-Burluk river, about three kilometres downstream of the modern village of Nikolskoye in Aiyrtau District. The site was identified in 1980 by Viktor Fyodorovich Zaybert (Зайберт), then a younger archaeologist working out of what became the North Kazakhstan State University in Petropavl. Systematic excavation began under Zaybert's direction in 1981 and continued, with interruptions, into the 2000s; the type-site at Botai itself has been opened to roughly ten thousand square metres. Two further settlements of the same culture — Krasnyi Yar on the Iman-Burluk and Vasilkovka on a tributary further south — have been excavated more recently in collaboration between Kazakh institutions and the Carnegie Museum of Natural History in Pittsburgh, where the American zooarchaeologist Sandra Olsen built the long-running comparative programme.⁶

What Zaybert and Olsen's teams found, when they tallied the bone fragments coming out of the trenches, was an assemblage that has no parallel in the wider Eneolithic. The Botai settlement alone yielded more than three hundred thousand identifiable bone fragments. Over ninety-nine percent of them came from one animal — Equus ferus. The remaining one percent — dogs, the occasional saiga, a few bovids — confirms that the inhabitants were not strictly capable of hunting only horses; they had other species available, and almost entirely ignored them.⁷

The architecture itself was a departure from what Surtandy hunting camps had looked like. The Botai houses were sunken pit-structures, roughly polygonal in plan, three to four metres across, dug a metre into the loess and roofed in turf-on-timber. They were aggregated in tight clusters, often sharing walls, with paths and small open spaces between them — a real village rather than a seasonal camp. Estimates of total population at the type-site run between 200 and 500 people, large enough to need an organised labour pool for hunting, processing, and corralling, and large enough that the local horse population must have been heavily and continuously pressured for centuries. Animals were processed inside or adjacent to the houses; refuse pits, ash deposits, and butchery debris produce the great majority of the catalogued bone count. The site is not a special-purpose hunting station. It is a place people lived, year after year, on horses.

The three evidence lines (Outram 2009)

The case that Botai was a society of horse herders rather than horse hunters was made most influentially in a 2009 paper in Science by Alan Outram of the University of Exeter, working with Natalie Stear, Robin Bendrey, Sandra Olsen, Alexei Kasparov, Victor Zaibert, Nick Thorpe, and Richard Evershed. Outram and colleagues advanced three independent lines of evidence.⁸

The first was bit-wear. Bridled horses, when restrained by a rod through the mouth, develop a distinctive pattern of mechanical damage on the anterior face of the second lower premolar (P2): a chamfered wear bevel, sometimes accompanied by tiny chip-fractures and enamel exposure, that is not produced by ordinary feeding. Robin Bendrey's methodological work on bit-wear had established quantitative thresholds — a bevel of three millimetres or more on the P2's anterior wear surface — that could distinguish the trace of a bit from natural attrition.⁹ Botai horse mandibles, when scored against the threshold, showed bit-wear lesions on a non-trivial fraction of adult animals. Critically, the bits in question need not have been metal; rope, rawhide, or bone bits would all leave the same mechanical signature, and the bone cheekpieces that Bendrey and others have catalogued from Botai assemblages are consistent with a soft-bit-and-bone-cheekpiece bridling system that the later Sintashta bridles would refine into the chariot-control apparatus of the Bronze Age.

The second was metrical. Outram's team measured the metacarpals — the slim foreleg bones whose length-to-width ratio is highly sensitive to selection — of horse skeletons from Botai and compared them against Pleistocene wild horses from Siberia and against later Bronze Age domestic horses from the Sintashta and Andronovo cultures. The Botai metacarpals clustered with the Bronze Age domestic sample, not with the Pleistocene wild sample. The difference is modest and somewhat ambiguous (Pleistocene-versus-Holocene morphometric differences alone could account for some of the shift), but the direction of the signal is consistent.

The third was lipid residue. Working with Richard Evershed, whose Bristol laboratory had pioneered the recovery of degraded animal fats from prehistoric pottery, Outram's team extracted absorbed organic residues from Botai pot sherds. The δ13C and δD values of the recovered fatty acids matched those of horse milk, not horse carcass fat. Several pots had been used to process mare's milk. Whether the milk was drunk fresh or fermented into the alcoholic koumiss that is still the national drink of Kazakhstan, the residue could not say. What it could say is that humans at Botai were milking mares, which requires a tameness and a routine of handling that hunting cannot supply.¹⁰ The dairying signal at Botai pushes the documented use of horse milk back approximately 2,500 years before the next firm evidence and remains the earliest such case anywhere in the world.

The corrals

Outram's three evidence lines were buttressed by independent geochemical work at the related settlement of Krasnyi Yar. There, Sandra Olsen's team had identified a series of post-hole patterns just outside the pit-house cluster that looked architecturally like enclosure walls — not fortification, but pen. Rosemary Capo, a geochemist at the University of Pittsburgh, sampled the soils inside these patterns and ran them against soils elsewhere at the site. The phosphorus levels inside the suspected enclosures were elevated several-fold; the nitrogen levels were low, indicating that the deposition was ancient rather than recent. The signature was consistent with what one would expect from sustained accumulation of horse dung and urine — the geochemical fingerprint of a corral that had stood for years.¹¹ The enclosures were not large by later pastoral standards — each could perhaps have held a few dozen animals — and they sat directly against the houses, with shared walls or short connecting passages. This is the spatial signature of small-herd, household-scale management, not of industrial pastoralism. The Botai relationship with horses was lived inside the home, not at a distance.

Photograph by Fernando Losada Rodríguez. Equus ferus przewalskii (Przewalski's horse) at Woburn Safari Park, England, 2014. CC BY-SA 4.0 via Wikimedia Commons. · CC BY-SA 4.0

The counter-argument (Taylor and Barrón-Ortiz 2021)

The Outram synthesis stood as the textbook account of horse domestication for roughly a decade. In 2021, William Timothy Treal Taylor and Christina Isabelle Barrón-Ortiz published a paper in Scientific Reports titled "Rethinking the evidence for early horse domestication at Botai", which mounted the first serious challenge.¹² Their argument, much briefer than the consensus it attacked, was that the Botai bit-wear scoring was unsound. Naturally-occurring dental abnormalities — periodontal disease, malocclusion, abnormal wear from coarse fodder — could produce P2 wear patterns within the range Bendrey had attributed to bits. The lipid residue case was harder to dispute, but Taylor and Barrón-Ortiz noted that lactation does not require domestication: cooperative milking of a tamed lead-mare in a captive herd is one possible mechanism, but so is targeted harvesting of nursing wild mares.

Outram and colleagues replied in a rebuttal that ran to twice the length of the original challenge, defending the bit-wear scoring and emphasising the convergence of independent evidence lines (bit-wear and metrical and lipid and corral geochemistry).¹³ Most of the field has reached a working compromise. The Botai horses were managed in ways that fall on the spectrum between intensive hunting and full pastoralism: regularly handled, occasionally bridled, milked at least sometimes, kept within corrals, and certainly not entirely wild. Whether one calls this "domestication" in the strict sense depends on the definition one accepts. The current synthesis, articulated by Outram in his 2023 Frontiers in Environmental Archaeology review, treats Botai as the type-case of a "prey-pathway" initial phase of domestication — a centuries-long process of niche construction that was clearly under way at Botai, but had not yet produced a population genetically isolated from the wild stock.¹⁴ The terminological dispute matters less than the underlying picture: people at Botai were doing things to and with horses that no earlier human population had been able to do.

What changed and what was replaced: two domestications, not one

The genetic surprise (Gaunitz 2018)

The most consequential revision of the Botai story came not from archaeology but from ancient DNA. In 2018, a team led by Charleen Gaunitz and Antoine Fages of the Centre for Anthropobiology and Genomics in Toulouse, working under Ludovic Orlando and with Outram and the Kazakh colleagues, published in Science the genomes of forty-two ancient horses — twenty of them from Botai itself, the others from later steppe sites.¹⁵ The expectation, before the data came in, had been that Botai horses would lie genetically near the base of the modern domestic line. They did not.

The Botai horses, when their genomes were placed in a phylogeny of all available ancient and modern horse genomes, fell on a branch with the Przewalski's horse (Equus przewalskii) of Mongolia — long supposed to be the last truly wild horse and the closest living relative of the ancestral wild stock from which all modern domesticates descend. The relationship was the other way around. Przewalski's horses were not the wild ancestor; they were the feral descendants of horses managed at Botai-like sites in the fourth and third millennia BCE that had escaped human control and reverted to wild behaviour. The modern domestic horse, meanwhile, carried only about 2.7 percent Botai-related ancestry. The lineage that became every horse subsequently ridden in Eurasia descended overwhelmingly from a different ancestral population, on a different stretch of steppe, in a different millennium.¹⁶

The French CNRS press release that accompanied the Science paper put the reversal in plain terms: les chevaux Botaï ne sont pas les aïeux de nos chevaux domestiques, mais ceux des chevaux de Przewalski — the Botai horses are not the ancestors of our domestic horses, but of the Przewalski's horses, the very animals long treated as the last living window into the original wild stock. Two centuries of taxonomic and historical assumption about what a "wild horse" was had been built on a population that turned out to be feral, not wild. The Przewalski's horse seen in zoos and on the Mongolian reserves is, technically, an escaped Eneolithic domesticate.¹⁷

The Pontic-Caspian breakthrough (Librado 2021)

That separate ancestral population was localised three years later by Pablo Librado and Orlando's team in a Nature paper that drew on 273 ancient horse genomes from across Eurasia, dated between 50,000 BCE and 200 CE. The signal was sharp. Modern domestic horses — the lineage now labelled DOM2 — originated on the lower Volga-Don steppes of what is now southern Russia, around 2200 BCE, more than a thousand years after Botai. They then expanded across Eurasia at extraordinary speed, replacing essentially every other horse population in their path, reaching central Europe by ~2000 BCE and East Asia by the mid-second millennium BCE.¹⁸

The genomic signature of the DOM2 horizon is striking on two specific selection targets. The locus GSDMC, associated with vertebral and back-musculature development, shows a hard sweep through the lineage — the modification that gave the horse a back that could carry a rider comfortably and a wagon-team that could pull a load without breaking down. The locus ZFPM1, associated with neurological development and behavioural traits, shows a parallel sweep — the genomic correlate of the docility, the resistance to flight, the trainability that distinguishes a manageable working horse from a tamed wild one. The two changes together describe a horse that was selected, in a few generations of intensive breeding, into a fundamentally new kind of animal.¹⁹ Antoine Fages and colleagues, working from a 278-genome time-series published in Cell in 2019, also identified two now-extinct horse lineages — one in Iberia, one in Siberia, both contributing essentially nothing to modern populations — which together with the Botai-Przewalski branch underline the extent of the genetic turnover: the DOM2 horizon did not merge with existing horse populations across Eurasia, it replaced them.²⁰

The cascade: chariots, Indo-Iranian languages, mounted warfare

The DOM2 horizon coincided with another transformation. The Sintashta culture, which emerged east of the Ural Mountains around 2100 BCE, produced the earliest known spoke-wheeled chariots, paired-horse burials in kurgan graves, and elaborate cheekpieces of bone and antler that show their builders had a sophisticated theory of bit-and-rein control. Librado's genomic data put DOM2 horses ubiquitously in those Sintashta kurgan burials. The chariot, the horse capable of pulling it at speed, and the Indo-Iranian linguistic horizon that carried both eastward into Central Asia and South Asia, are a single bundle that diffuses out of the southern Trans-Ural steppe in the early second millennium BCE.²¹

The downstream impact of that bundle on settled Bronze Age civilizations was profound and rapid. The Hyksos, who took Lower Egypt around 1650 BCE, came with chariots and horse-pairs of recognisably steppe-derived breed. The Hittite empire built its core military arm around chariotry, and the surviving Hittite training texts — most famously the Kikkuli text, composed in the fourteenth century BCE by a Mitannian horse-trainer for the Hittite king Suppiluliuma I — describe in extraordinary detail a months-long conditioning regimen for chariot horses that assumes horse-handling knowledge centuries deep at the moment of writing. The Mycenaean palace economies of Late Bronze Age Greece, the Shang dynasty's chariot corps in north China, the chariot-using armies of Mitanni and Assyria — all rest on the DOM2 expansion, and all began collecting chariots and the specialist personnel required to operate them within a few centuries of the Sintashta horizon.²²

Figure from Lindner, S. 'Chariots in the Eurasian Steppe: a Bayesian approach to the emergence of horse-drawn transport in the early second millennium BC.' Antiquity 94 (374), 2020, pp. 361–380. doi:10.15184/aqy.2020.37. CC BY-SA 4.0 via Wikimedia Commons. · CC BY-SA 4.0

It is important here not to overread. Librado and colleagues explicitly rejected the earlier consensus that linked the DOM2 horse to the massive Yamnaya pastoralist expansion into Bronze Age Europe around 3000 BCE, the expansion that brought Indo-European languages west. The Yamnaya, on the new genetic timing, did not ride DOM2 horses; their expansion is now ascribed to wagons and walking, not riding. The horse-borne wave is later, smaller, and aimed eastward — into the Tarim Basin, the Indus, eventually onto the chariot-using courts of Mitanni and Shang. What it carried with it linguistically was Indo-Iranian — Sanskrit, Avestan — not the western branches of Indo-European.²³ David Anthony's 2007 synthesis The Horse, the Wheel, and Language, which articulated the older link between horse-borne Yamnaya and Indo-European spread, accordingly needs to be read with the post-2018 genetic timing in mind: the wagon part of the argument is intact, the horseback-riding part has had to be substantially rebuilt.

What Botai's contribution actually was

The Botai contribution to the eventual horse civilization of Eurasia, then, was real but indirect. Botai was where humans first solved the problem of routine, handled coexistence with the horse — bridling, herding, milking. The skills, the techniques, and quite probably some of the early stock were known across the steppe by the time the lower Volga breeders began their work two millennia later. When the new selection pressure was applied — bigger animals, calmer temperament, stronger backs, longer-lasting traction — it was applied to a species the steppe peoples already knew how to handle. The genetic line that came out of the Volga-Don sweep replaced nearly everything that had come before. The handling tradition, the dairying tradition, the corralling tradition, the very concept of the horse as a partner rather than a quarry — these were the legacy that survived the genetic turnover.

The point bears restating because the recent press coverage of the 2018 reversal has sometimes implied that the Botai story has been "overturned" or "disproved". It has not. What the ancient DNA has done is sharpen the picture into two domestications rather than one. The first, on the Iman-Burluk in the late fourth millennium BCE, produced a managed but not yet genetically distinct population, and a body of practical knowledge about how to live with horses. The second, on the lower Volga-Don fifteen centuries later, produced the genetically distinct population that became every horse the world rides today. Both events are real; both were necessary; neither replaces the other.

Koumiss and the secondary-products dimension

A specific institutional inheritance is worth singling out. The fermented mare's milk drink — koumiss in the Russian-derived English form, qımız in modern Kazakh — that Outram and Evershed traced in Botai pottery in 2009 is still a national beverage of Kazakhstan and Kyrgyzstan, drunk daily by millions of people in those countries and across the broader Central Asian steppe. The continuity is not direct (modern dairy-mares are DOM2 horses, not Botai horses), but the practice of milking mares, the rituals around fermentation, and the integration of mare's milk into the calorific economy of pastoral life trace back, through an unbroken chain of practice, to the very first humans known to have processed it. The 5,500-year-old residue in a Botai sherd is the same drink, on the same stretch of steppe, in the same handed-down practice. Few foodways have that kind of continuity in the archaeological record.²⁴

Mare's milk also turns out to have been the entry point into a much broader category of pastoral nutrition. Andrew Sherratt's "secondary products revolution" framework, formulated in the 1980s and now substantially confirmed by the archaeobotanical and lipid-residue record, treated the systematic exploitation of living animal products — milk, wool, traction, dung — as a distinct phase that followed the initial domestication of cattle, sheep, and goats by several thousand years in most parts of the world. The Botai mare-milk residue is, on present evidence, the earliest secondary-products signal anywhere for horses. The shift from horse-as-meat to horse-as-living-resource happened first here, on the forest-steppe, before it happened with the descendant DOM2 lineage anywhere else.

What the cost was: the wild relatives and the downstream bill

Przewalski's horse: a story of near-extinction

The cost of the Botai transmission, on the strict definition of cost-of-the-transmission-itself, is small. No city was sacked at the moment a corral was built. No population was conquered. No autonomy was surrendered. The act of bringing horses into the household, on its own terms, was peaceful.

Where the cost shows up is in what happened to the Botai horse line over the subsequent five millennia. After the DOM2 horizon swept through the steppe in the early second millennium BCE, the Botai-lineage horses lost their managed status. The settlements were abandoned (the Botai culture itself ended around 3100 BCE for reasons that remain debated — climatic change, soil exhaustion under the heavy local trampling pressure, competition from later cultures). The horses that had been kept in the corrals dispersed, reverted to wild behaviour, and survived as a relict population on the Dzungarian Gobi at the Mongolia-China border, where the eponymous Russian explorer Nikolai Przhevalsky encountered them in the 1870s and shipped specimens back to St Petersburg. By the early twentieth century, the wild herd was being heavily hunted — for trophies, for zoo specimens, and by Mongolian and Soviet military expeditions during and after the Second World War.²⁵

The species was declared extinct in the wild in 1969, when the last confirmed sighting of a wild herd was made in southwestern Mongolia. For roughly two decades the entire global population of Equus przewalskii — at one point as low as twelve effective founders — existed only in zoo enclosures across Europe, North America, and Mongolia, where the species had become the project of a small conservation network coordinated through the Prague Zoo's international studbook. The captive population grew slowly. Inbreeding depression, hybridization risk with domestic horses, and the question of where to put a recovering herd were unresolved through most of the 1970s and 1980s.²⁶ The genetic bottleneck the species passed through is one of the narrowest documented for any large mammal in the twentieth century — a near-total reduction of evolutionary heritage to twelve breeding individuals — and the recovered Przewalski's population today shows the fingerprints of that bottleneck in its drastically reduced heterozygosity compared with both modern domestic horses and ancient genome samples.

The reintroduction

In 1992, after a decade of diplomatic and logistical preparation, the first sixteen Przewalski's horses were released into Hustai National Park in the Mongolian Khentii. The project was a partnership between the Foundation for the Preservation and Protection of the Przewalski Horse (Netherlands) and Mongolian conservation authorities. Eighty-four animals were eventually returned across the 1990s and 2000s. The Hustai herd reached 260 by 2009. Parallel reintroductions to the Great Gobi B Strictly Protected Area in southwestern Mongolia and to Khomiin Tal in the west have, by the mid-2020s, brought the wild Mongolian population to roughly 850 animals across three sites — still a small total, and still vulnerable to severe winters, disease outbreaks, and the persistent risk of hybridization where Mongolian domestic mares enter the reserves during snowy conditions. The horse is no longer extinct in the wild. It survives because a coalition of Mongolian rangers, Czech and Dutch zoologists, and Russian-era studbook keepers managed not to let the last twelve effective founders go.²⁷

That recovery story has an unusual provenance: the founders' breeding records have been kept continuously in a single international studbook held at Prague Zoo since 1959, and the descendants of every individual horse can be traced through the studbook to the original wild-caught animals of the late nineteenth century. The Przewalski's herd is thus, in a literal record-keeping sense, the best-documented large-mammal population in the world. That documentation made the reintroduction possible — genetic matching of founders to release sites depended on it — and gives the population a kind of bureaucratic continuity that the Botai horses themselves, of course, never had.

That is the cost in the direct sense: the line that came out of Botai survived as a wild population only by a margin so narrow that it required a four-decade global captive-breeding programme to keep alive at all. The Botai transmission's most direct biological inheritor was, by the middle of the twentieth century, an animal that no longer existed outside cages.

The downstream bills

The cost in the indirect sense — the cost downstream of the transmission, paid by populations Botai never met — is larger but is properly the bill of later technologies built on the horse, not of the bridling itself.

When DOM2 horses pulled spoke-wheeled chariots out of the Sintashta steppe in the early second millennium BCE, the technology cascade that followed reorganised the military economy of every Bronze Age civilization within its eventual reach. The Hyksos, who took Lower Egypt around 1650 BCE, came with chariots and horses derived from this lineage. The Hittites, the Mitanni, the Mycenaeans, the Shang dynasty's chariot corps — all are downstream of the same DOM2 breakthrough. The Late Bronze Age palace economies maintained chariot fleets at enormous cost: a single chariot pair required years of training, specialist drivers, careful breeding programmes, and continual veterinary attention.²⁸

When mounted horsemanship matured in the early first millennium BCE — the Scythians from around 800 BCE, the Sarmatians, the Xiongnu — the cost framework shifted again. Mounted nomadic populations, fielding cavalry forces capable of operating across the open steppe at speeds settled infantry could not match, became the single most persistent military problem of every settled civilization within range of the grass. The Han Chinese under Wudi spent decades and tens of thousands of conscripts on the Xiongnu wars; the Romans spent the entire Imperial period managing the Sarmatians and the later Hunnic incursions; the Sasanians and Byzantines faced wave after wave of Turkic mounted attack; the great Mongol conquests of the thirteenth century CE killed, by modest demographic estimates, tens of millions of people across China, Central Asia, Iran, and Eastern Europe.

None of those costs are the cost of Botai. They are the cost of mounted warfare on settled civilization, which is downstream of DOM2, which is downstream of the Pontic-Caspian breakthrough of ~2200 BCE, which is itself only loosely descended from the Botai handling tradition. To assign the cost of the Mongol invasions to a fourth-millennium Kazakh village that bridled the wrong horse line would be historiographic confusion of a kind the atlas exists to refuse. The Botai transmission's bill is small: a wild line nearly lost, then narrowly saved.

The same care is owed to the question of the secondary-products bill. Mounted pastoralism — the Eurasian way of life that the koumiss-drinking, horse-milking Kazakh, Kyrgyz, Mongol, and Turkic peoples have run more or less continuously for three millennia — is a working tradition with a deep ecological footprint. Heavy steppe grazing reshapes grassland community structure, locks open vast areas against forest succession, and depends on regular cycles of mobility between summer and winter pasture. The cultural patterns it sustains are remarkable; the demographic and environmental costs of overstocking, when they materialise, are paid by the herding communities themselves, not by outsiders. The Soviet collectivisation of Kazakh pastoralism in the early 1930s, which killed approximately a million and a half people in the Asharshylyk famine and slaughtered most of the country's herds, is a real and recent demographic catastrophe — but it is the cost of collectivisation, not of horse pastoralism, and the right historical address for it is Moscow in the 1930s, not Botai in the 4th millennium BCE.

What endures

What endures from Botai is the fact of having been first. Five and a half thousand years after a population on the Iman-Burluk river first put a bit in a horse's mouth and milked a mare into a pottery jar, humans across half the planet drink fermented mare's milk in direct continuity of practice; ride animals descended through a parallel but related lineage; and depend on a working partnership with one species, Equus, that was the first large animal humans learned to manage at scale without killing. The horse and human relationship is one of the most consequential interspecies bonds in the species' history, and its first documented chapter was written, in horse-bone and pottery sherds and corral geochemistry, on a stretch of Kazakh steppe that has been quiet since the houses were abandoned in the late fourth millennium BCE.²⁹