No Horse was domesticated in the steppe of Kazakhstan, west Siberia and Ukraine: DNA study

Was the wild steppe horse ever domesticated?

Archaeology and DNA

Sir William Ridgeway had noted on anatomical grounds that the Przewalskii was not the ancestor of the caballus horses (1905:425). Further than this the US Bureau of Animal Husbandry noted in its Report (1910:165) that the modern horse had come from many sources: “But notwithstanding the absence of well preserved skulls it has been possible by making use of the new methods to obtain a considerable amount of evidence that the domestic horses had a multiple origin, that they include amongst their ancestors not only varieties allied to the wild horse which still survives in Mongolia, and varieties adapted for a forest life, but also varieties specialized for ranging over boundless deserts and plateaus, and for living amongst foothills and upland valleys.” (emphasis added). The contribution of the horse genes from many regions in the domestication of horse has been proved, yet any contribution from the steppe-horse has not been supported by these DNA studies of the caballus and the Przewalskii horses (Jansen 2002; Bowling 2003; Wade 2009; Cai 2009:481; Achilli 2011).

The official position in science is that the steppe horse Equus przewalskii has different chromosome number, and that has not been found in any domestic horse and hence it is an entirely different species (Oakenfull 2000; Clark 2006) from which the modern caballus horse could not have descended.

Confirming the non-domesticablility of the Przewalskii horse, Jansen et al (2002:10910) noted, “Modern breeding of the wild Przewalski’s horse initially encountered problems such as pacing, excessive aggression, impotence, and infanticide, leading the Przewalski’s horse to the brink of extinction. The Przewalski’s horse is not ancestral to domestic horses, but if their wild ancestors were similarly intractable, it is unlikely that the technique was mastered many times independently during prehistory. The ease of domestic horse breeding today may be the genetic consequence of selections of particularly amenable beasts some thousands of years ago.”

It can be proposed that the true horse might have lived along with the Przewalskii in the steppe. But that imagination is not allowed. The non-Przewalskii horse could not have lived along with the Przewalskii in the steppe because of the Gause’s Law of Competitive Exclusion, which states: “No two species can equally and successfully occupy the same niche in the same habitat at the same time.” With time one species completely eliminates the other by competition. Hence all the horse bones recovered from the steppe must be considered those of the wild Przewalski. This view is consistent with Levine’s findings too; and the bones with evidence of domestication must be considered imported from outside into the steppe-region.

None of the archaeological claims made so far for the presence of the domestic horse in the steppe have been uncontestable. Outram (2009) found evidence of mare’s milk on pottery at Botai. However the bones of the Botai horses, particularly the vertebral bones showed no damage to them which normally occurs due to riding. After her detailed examination Levine noted, “the material from Botai examined so far most probably was from wild individuals” (Levine 2005:107). Thus presence of milk on a potsherd does not necessarily mean evidence of domestication. It was easy to capture a Przewalskii full-term pregnant mare, keep her captive through delivery, then use her milk for some time before slaughter. For the domestication status of any animal there should be complete package of evidence, not just a stray finding.

More than this, the Bronze Age Botai horses examined by Outram were not indigenous but were imports from outside, as Outram himself noted: “Metrical analysis of horse metacarpals shows that Botai horses resemble Bronze Age domestic horses rather than Palaeolithic wild horses from the same region.” The import of the steppe horses from outside during the Bronze Age is further confirmed by ancient DNA studies (vide infra). Morphological studies too have shown that metrically, the Bronze Age and later domestic horses of Eurasia resembled the European and Indian fossil horses (stenonis, sivalensis etc), but not with the steppe horse.

Ancient Horse DNA

Ancient DNAs of horse recovered from archaeological samples do not support the steppe origin or domestication of horse. Keyser-Tracqui (2005) examined 13 horse DNAs recovered from a third century BC Scythian Era frozen tomb at Berel in Kazakhstan. “It shows that the 15 ancient Asian horse’s sequences obtained did not form a separate cluster.” (p. 206). This means they did not have a common source of origin, and they were not locally domesticated. The study revealed that the DNAs had come from six different clades of Vila.  “No clear geographical affiliation of the specimens studied was thus determined”, they noted.

Intriguingly, none of these third century Scythian horse samples matched with Akhal Teke (Turkmenistan) breed of horse. Keyser-Tracqui noted, “Whereas no matches were found with the Akhal-Teke specimens, some complete matches were observed with Chinese Guanzhong and Tuva horses (the Tuva republic is localized between the Siberia and the Mongolia) as well as with Anatolian horses.” (p. 208)

This finding helps us understand the origin of the Akhal Teke horse. The South Central Asian horse Akhal Teke has no contribution from the hypothetical source of horse located in the steppe. It also means that the Akhal Teke breed is not as old as we think, and it has been formed after the Scythian Era from admixture of Iranian and other breeds.

Many of these ancient horses in Kazakhstan had arrived from north Europe: “None of the Scythian haplotypes matched those provided for Przewalski’s horses, which are considered to be a relic population of wild horse of Eurasia (Lister et al. 1998; Oakenfull & Ryder 1998). Interestingly, we noted that BER10-11-13 sequences perfectly matched that of a Viking Age horse bones found in a restricted area in Sweden (Vila` et al. 2001) as well as that of an Icelandic horse, which represents an ancient Norwegian breed.” (p. 208)

This indicates arrival of the European as well as the Chinese horse in Kazakhstan during or before the Scythian period. Older Bronze Age DNAs from the steppe are not reflected in these third century BC samples, implying that the Scythian Era horses were fresh arrivals from other places and the Bronze Age horses had been replaced by the imported horse during the Scythian period. Two of the samples belonged to haplogroup A6, two to hg D and three to C1 (p. 205). Thus the haplogroups ascribed to the horses in this study were A, C and D. The Bhutia (Himalayan, Country Bred Indian) breed of horse has haplogroup composition of A, C, D, F, and it is possible that there was a significant arrival of Himalayan horse from North India to Berel.

The mtDNA haplotypes detected in the ancient horse from Kazakhstan and West Siberia from the periods Copper and Bronze Ages were X3, X3c1, X5, X7a, G3, K2 and D3 (Cieslac 2010:Figure 2). These have not been found in the modern Indian horses (Devi and Ghosh 2013:5862 Table 1). The X5 which was confined to this region during the Bronze Age is found today in the Fulani horse of Africa (Cieslac 2010:3 pdf). Further investigation is needed to confirm that the X5 reached Africa and the steppe both from a common source in Europe. Likewise the X3 too was found in the steppe during the Bronze Age. It appeared later in China in the Iron Age (ibid). This lineage is possibly European in origin and indicates a migration of European horse to the steppe during the Bronze Age. These ancient lineages are absent from the modern Kazakh breed of western China (Tao Zhang 2012:923). This means the steppe horse of the Copper and Bronze Ages neither contributed to the horse population of later India nor to the breeds Kazakh and Akhal Teke.

In the study of Iranian horse breeds by Moridi, it was found that the Sistani breed, native of the Sistan-and-Baluchistan province (adjoining Baluchistan of Pakistan) had the maximum number of haplogroups. The European haplogroup B was present only in the coastally located Arabian-Irani and the Sistani breeds and was not found in the northerly located breeds viz. Kurd, Caspian-Short and Turkoman. In Devi’s study of Indian breeds, only the Marwari and Kathiawari breeds which are nearer to Sistan-and-Baluchistan had haplogroup B, but others including Bhutia, Spiti, Zanskari and Spiti did not have this haplogroup.

The haplogroup B has been found in Guizhou breed of South China and the Mongolian breed of northwest China too (Tao Zhang 2012:923, 924). It is also present in the ancient Chinese DNA (2.9%) and modern Mongolian (6.1%), Korean Cheju (17.6%), Tuva (9.1%, north of Mongolia), Mesenskaya (27.8%) and Orlov (16.7%) breeds (Dawei Cai 2009:838 Table 3).

Thus there is a circumscribed region where the haplogroup B is absent. This region is Tibet, Turkmenistan, Kurdistan and Himalayan India. This fact too indicates that apart from some introgression in the Marwari and Kathiawari breeds, the Indian breeds are pure and have been free from Central Asian and Western influence.

In China there are five regional indigenous horse types viz. Mongolian, Kazakh, Hequ, Tibetan and the Southwest, to each of which several breeds belong (Tao Zhang 2012). The Kazakh breed of the Chinese native horse is the one which was most probably imported from Kazakhstan in the historical periods. Haplogroup B is not found in the Kazakh breed of Chinese horse, and is absent from the Tibetan horse too (Tao Z.:923; Dawei Cai: Table 3). Haplogroup B was not found in Kazakhstan in the Bronze Age, although it appears in the Iron Age (Ceislac 2010:Fig 2) to disappear again from the Kazakh breed. This shows that there were some European horse arrivals to the steppe during the Iron Age.

We can hence conclude that the modern Kazakh breed of horse was imported into Kazakhstan from the Himalayan provinces of India and northeast Iran during the Iron Age, wherein the haplogroup B was absent then. The Bhutia, Spiti, Zanskari and Manipuri Indian horses have remained isolated from the Muslim period influence on the horse breeds in India. The Turkoman, Caspian Short and Tibetan breeds too have remained in isolation. Clearly, instead of being the source of horse to Europe or India or even to Turkmenistan, the steppe was the recipient of the European DNA haplogroup B from Europe, and also several horse lineages from India.

The mtDNA haplogroups A and F were the commonest lineages in the Chinese ancient horse, and were the only ones in the pre-domestication samples (Dawei Cai 2009:837, 840). None of the ancient DNAs recovered from the steppe as reported by Cieslac, belonged to the haplogroup F (Cieslac 2010: Table 1). In the modern horse breeds these two (A and F) in the highest frequency are found in Chinese and Indian Himalayan breeds (34%; Devi and Ghosh 2013:5862 Table 1). F is present in the modern Kazakh breed.

However this combination (high proportion of both A and F) is not present in Russian horses. In fact the minimum frequency of F is found in Vyatskaya (Russia), Mesenskaya (Russia), Orlov (Russia) and European horses (ibid), indicating that the domestication of the Chinese horses was distinct and separate from the steppe and the European horses. The low level of the hg F might have come into the Russian horses as the result of some horse migration from China. McGahern’s study too noted the absence of haplogroup F from the Orlov and Akhal Teke, and its very low frequency in Vyatskaya breeds of steppe and adjoining areas (2006: Fig 1).

Thus haplogroup F belongs to a domestic horse clade which once lived in Himalayan India, Tibet, China and Mongolia, but did not live in the steppe. We can say that the steppe region (Kazakhstan, Ukraine, West Siberia) received migrant domestic horses from China/Mongolia (hg A, F), Tibet (hg E, D, A), Europe (hg B) and India (A, D, F), particularly during the Iron Age and after. Our conclusion is supported by ancient DNA study by Keyser-Tracqui. The ancient DNA study of the domestic horse remains from the Scythian Period recovered from the steppe proved that they all had been imported from outside regions like Anatolia, China etc (Keyser-Tracqui 2005). None of the thirteen such DNAs matched the DNA of today’s horse breeds from Central Asian, steppe and other neighbouring region such as Yakut, Mongolian, Akhal Teke (Turkmenistan) horses, or the Pleistocene horse from northeast Asia, which are generally conjectured as the close relatives of the ancient steppe horse. Matching was not done with the Indian horse.

These DNA findings are consistent with the archaeological finding from the Scythian era that the domestic horse of the south Ural region (Siberia; e.g. Rostovka, Preobrazhenka, Samus’ IV etc) came from Central Asia at about 1300 BCE (Kuz’mina and Mallory 2007:200). In our study, we surmise that these horses had reached Central Asia from further south like Iran and India, and not from north to south. It is consistent with Warmuth’s DNA study which found that the no horse migration from north to south took place in Eurasia (2012b:7).

In a study of horse DNA from ancient horse remains found from China showed 75% of Chinese horses (from 2000 BCE and later) had same DNA as those found from earlier periods, implying their local domestication from the pre-existing wild Chinese horse (Cai 2009:481). The remaining 25% DNAs of post-2000 BCE were new arrivals from outside, but not from the steppe. The conclusion is that the 75% of the late Bronze Age horses of China had been domesticated locally from the wild Chinese horses, and they had contributed to the steppe domestic horse population too; however, none of the Late Bronze Age domestic horses of the steppe had any local ancestry from the wild steppe horse. This type of solid DNA evidence is lacking for the steppe.

Similar DNA study of ancient horse remains from Iberian Peninsula (Spain) revealed many of the modern Spanish lineages had been present throughout the early Neolithic, Copper Age, Bronze Age and Iron Age, indicating a local domestication within the Iberian Peninsula itself from the local wild horse population (Warmuth 2011). Cieslak found that the domestic horse’s DNA lineages B, H1 and J originated from within the Iberian Peninsula and were domesticated probably during the Copper/Bronze Age.

However, some East/Southeast Asian lineages (namely A, D2, X4a) were too present in Iberia at the Bronze Age indicating an early arrival of some domestic horses from outside from the south-east Eurasia possibly through North Africa and India (Cieslak 2010:3). In our view, horse migration from Southeast Asia to the Iberian Peninsula necessarily involved passing through South Asia under a state of domestication. Luis (2006) found that the modern American horses were related to the ancient Iberian (Spanish) horse DNAs from the prehistoric dates.

Cieslak found that all the regions of Eurasia had local Pleistocene horse DNAs preserved in continuity into the modern primitive breeds of the regions. The steppe region, however, was the only exception in this respect which showed no such continuity of lineages in Cieslak’s study. His has been supported from the DNA study of the second millennium BCE remains of horse from the steppe (supra). Thus we conclude that no horse was domesticated in the steppe.

Modern DNA Studies

Yinghui Ling (2010: Fig 1) examined the Y chromosomes of the existing Chinese horse breeds. They found that there were two male lineages in the Chinese horse, designated haplogroup A and haplogroup B. The haplogroup A was distributed all over China, but the haplogroup B was found only in the southern regions of China adjoining the Indian boarders and Southeast Asia. The study was not extended to include India. However from the data generated it can be said that there the male horse lineage (hg B) was domesticated in the Himalayan region, and that this lineage extended into Tibet and Southeast Asia, but it has not spread far north into China.

The genome-wide SNP data study of the horse breeds showed that the modern Iberian and three Middle Eastern breeds (viz. Caspian, Arabian and Akhal Teke breeds) formed one single genetic cluster (Petersen 2013:6 pdf; 7pdf Fig2). The Middle Eastern breeds clustered with the Iberian breeds, indicating ancient migrations of domestic horse between Iberia and the respective regions of Middle East. Such migrations involved North Africa, Arabian Peninsula and Iran when the climate was better during the Bronze Age or before. From Iran horses were taken to Central Asia during the Copper and Bronze Ages, when there was large scale human migration from Iran to Central Asia. Yet the migrations seem to have taken place from East Iran during the Scythian Era because the Akhal Teke DNA does not match any ancient Scythian DNAs. Or, the Akhal Teke was formed after this time. Thus horse came to the steppe from south, west and east. It is also notable from the study of Petersen that the steppe, or Central Asia or Ukrainian breeds of horses do not form any cluster (see dendrogram Petersen: Figure 2), indicating that they did not originate from any distinct or demarcatable source.

The mitochondrial DNA haplogroups A to G are widely distributed in world. However there is some pattern. McGahern’s study demonstrated that the domestic horse breeds of the world belong to two distinct genetic cohorts, eastern and western, on the basis of their mtDNA variation. However the steppe countries like Kazakhstan and Ukraine did not come within any of the two cohorts indicating mixed origins from both east and west (McGahern 206:495).

The haplogroup F, very common in China, Mongolia, Tibet, Kazakhstan and India is absent from the Kurd, Turkoman and Sistani breeds of Iran (Dawei Cai 2009:Table 3; Devi and Ghosh 2013; Moridi:4 pdf, Fig1). However the hg F is present in the Arabian and the Caspian breeds, in which it might have reached from India and Tibet during the historical Arabian Empire period. In the steppe region consisting of Kazakhstan and West and South Siberia, the hg F occurs in the Bronze Age but not in the Iron Age (Cieslac 2010:Fig 2; Keyser-Tracqui 2005). That means the Iron Age import into the steppe was mainly from south (Iran) and west (Europe) where haplogroup F did not exist then.

Similarly, the haplogroup B is considered European (Iberian) in origin and it is not found in the Indian country-bred or the mountain breeds of India and Kazakhstan. Some of the Iranian breeds have haplogroup B, others do not (Moridi: Fig 1). In the ancient samples, B is absent from China, Korea and Mongolia (Cieslac 2010:Fig 2). It is absent from West and South Siberia and Kazakhstan during the Bronze Age but appears there in the Iron Age (ibid). Today it is absent from the Kazakh breed of China yet is found in the western steppe breeds. This indicates migration of some horse from Europe to the steppe during the Iron Age.

Lira and colleagues (2010) demonstrated that horse was domesticated in Iberia from the wild horses present there since Late Pleistocene period. They also found that haplogroup D1 had not been present in Bronze Age Iberia and arrived there during the medieval period. Warmuth and colleagues (2011:4 pdf, Fig 1B) too concluded that horse was domesticated in Iberia from native wild horses.

Warmuth (2011) noted that apart from Iberia, there was another likely source for the European domestic horse–the region of Iran south of the Caspian Sea. The archaic Caspian breed lives here. “Our investigation of genetic diversity in traditional European horse breeds reveals two hotspots of genetic diversity, one in the Caspian region of western Asia and one in the Iberian Peninsula.” (Warmuth 2011:2 pdf). However the steppe was not found to be a genetic hotspot in their study. This study was not extended to the further east and had not included India and China.

Study of the autosomal genes by Warmuth (2012b:7pdf) demonstrated that there was a general migration of the true horse from east to west before domestication. However, horses of Lithuania and Kazakhstan showed evidence of recent arrival from the East, implying import after domestication. This fact about Kazakhstan militates against Central Asia having been a place of domestication of horse.

The study (2012b:7) ruled out any migration of horse from north to south: “There was no significant correlation between genetic diversity and latitude, despite written accounts documenting a continued flux of horses from the steppe lands in the north into both India and much of China (Gommans 1994)” (Warmuth 2012b:7pdf). This finding rules out the common conjecture that the horse arrived into Iran and India from north.

DNA study of living Greek horses revealed that many of them (Crete, Pindos and Pinias breeds) had arrived there from the Middle East route, and not from Ukraine/East Europe route (Bömcke 2010:7/9pdf). This too goes against the domestication of the horse in the steppe, and is consistent with the arrival of the horse from Iran and India to Middle East and from there to Southeast Europe.

Hence the steppe region of today was not the source of the horse populations for Central Asia, China, India, Iran, Greece and Spain–we can say from the above studies. The study by Warmuth (2012b:5,Table 1) also showed that Ukraine, Kalmykia (Russia) and Kyzilorda (Kazakhstan) of the steppe had no private or unique alleles, while other regions like Jammu (India), Yunnan (Southwest China adjoining northeast India) and Naryn (Kyrgyzstan not far from the northern reaches of India) had many unique alleles each indicating local evolution in these areas.[1] Thus Kyrgyzstan, located immediately to the north of Pamir, was the only place north of the Indo-Iranian plains where we find any evidence of in-situ horse domestication. No place to the further north, including Ukraine, Kazakhstan, Mongolia and Siberia has any DNA evidence of local domestication of horse.

This reflects that Ukraine and the other steppe regions possibly had received all their domestic horse DNAs from outside, while Jammu, Kyrgyzstan and Yunnan—the regions adjoining the Himalayas where sivalensis, the Indian wild horse once prevailed–had local domestication of horse. Considering facts from every angle it becomes obvious that the steppe origin of horse domestication cannot be sustained any more, and it was an academic hoax.

The Multiple Primary Domestication events of Horse:

Necessarily associating the horse domestication with the Aryans was the most unfortunate event of historiography. The DNA studies revealed that horse had not been domesticated only at one or a few places, but at seventy seven places throughout Eurasia, and can be grouped into seventeen DNA types (Jansen, 2002; Vila, 2001; Lippold 2011a; Cieslak 2010). Lippold (2011b) noted that the DNA remains of the wild horse from ancient Siberia, Alaska and Yukon proved that they all belonged to the Przewalskii, and it is enough to prove that the caballus horse had not lived there in the wild then. Lippold also found that all the ancient remains of horse with features of domestication (dated 800 BCE) belong to the non-Przewalskii type of DNA, clearly indicating the arrival of the domestic horse from outside the range of the Przewalskii horse.

All the regions of Eurasia other than Mongolia, Siberia and the steppe at the Bronze Age had domestic horses which had been locally captured (Kavar 2008). Clearly Siberia, Mongolia (and North America and steppe) were the regions where ancient domestic horses had been imported from outside, and not locally domesticated.

Lindgren (2004) and Lau (2009) found (DNA study) that although wild mares had been recruited from all over Eurasia, yet on the male side there were only one or just a few stallions. It was also confirmed that none of the male progenitors was the steppe or Central Asian Przewalskii stallion (Lindgren:336). In our opinion this progenitor stallion could have been from the sivalensis stock (vide infra).

DNA studies found that some northwest European domestic ponies namely the Fjord, Icelandic and Shetland ponies have a single cluster of DNA, which originated very early just after the Late Glacial period (about 10,000 BCE; Jansen:10908). Achilli (2011) found one lineage of European horse (haplogroup L) was domesticated in Europe from where it seems to have spread to Middle East and Asia.

Study of DNA recovered from the Neolithic and Bronze Age horse bones showed that many of the extant European lineages had already been there at those times, and some even during the late Pleistocene (Achilli 2011:3-4 pdf; Cieslak 2010:3; Lira 2012). In fact at least one lineage of modern domestic horse Lusitani Group C had been domesticated quite early, possibly during the Neolithic period itself (Lira 2009). The haplogroup D lineage, which is the most prominent lineage of Iberia, arrived here only during the Bronze Age. Another study found that at least one breed of horse had been domesticated in Spain much before Indo-European linguistic arrival to the area (Achilli et al 2011:4 pdf). Solis (2005) showed many horse breeds of the Iberian Peninsula are autochthonous and have been domesticated locally in Europe itself (Solis: 677). The study by Royo noted common DNA motifs in Iberian and Barb (North African) horses, which is consistent with our view that the late Bronze Age arrivals of the domestic horse to Iberia took place through North Africa. These studies rule out the association of the European horse with the Indo-European culture.

A recent study by Devi (2013) revealed that the 59 Indian horse samples studied carried 35 haplotypes, which gives one of the highest index of genetic diversity in the word. Such figure is consistent with the oldest horse domestication event having taken place in India. A total of seven major mtDNA haplogroups (A–G) was identified in the Indian horse breeds that indicated the abundance of mtDNA diversity. The haplogroup D constituted 33% of Indian horse breeds. The Manipuri breed of horse comes from the eastern end of India and it has remained segregated from the other horse breeds of India and outside. It is presumed to be the local descendant of the Wild Asiatic Horse. However, it was noted that the Manipuri horse showed closest affinity with the various Indian horse breeds as well as the Thoroughbred horse, and not with the Chinese or Central Asian breeds (ibid).

Horse statue (three feet tall) dating 7000 BCE was found recently from Asir (Arabia Felix, near Abha, southwest Arabia; Science News, BBC News, Reuters etc). Arabia was not inhabitable by horse during the last glacial owing to cold desert like conditions. Any early Holocene presence of the horse must have been from India, where the sivalensis horse had survived the Last Glacial maximum. There is plenty of evidence of Indian migration to the East Arabian coast (human DNA, Underhill 2009:2 and 3; shrew, mice, Duplantier) during late Pleistocene and early Holocene. Evidence from the other regions indicates that the Indian sivalensis was domesticated as early as 8,000 BCE. Hence it could have been taken to the Arabian coastal region by the early Holocene migrants.


Climate change and the arrival of the Domestic Horse in Central Asia:

The archaeology of the steppe has confused the archaeologists because the steppe archaeological sites have horse bones which had been hunted for meat, or captured live then sacrificed ritually. They belong to the Przewalskii horse which was never domesticated, but lived in the wild in the steppe. It is extremely difficult, if not impossible, to make distinction between the Przewalskii and the true-horse bones. There is no evidence to support that any non-Przewalskii type wild horse lived in the steppe about 2000 BCE or earlier.

In spite of the several horse hoaxes raised about the steppe, the recent archaeological studies serious challenge the long held view that steppe was the home of the true horse which we have today as the domestic species. The archaeological study done at Begash (Kazakhstan) confirmed in a recent study: “While pastoral herding of sheep and goats is evident from the Early Bronze Age, the horse appears only in small numbers before the end of the first millennium BC” (Frachetti and Benecke 2009: Abstract).

The paper adds the “horse use seems to commence gradually and is not highly associated with early and middle Bronze Age pastoralists.” (ibid:1025). The authors find, the “percentages of horse remains at Begash remain below 6 per cent until approximately AD 50 (Phase 3b)”, and “The domestic horse is documented at Begash by the start of the second millennium BC, but its impact on pastoralism is not clear.” In our view, such stray domestic horses as the ones documented from the second millennium BC Begash were Bronze Age imports from the further south i.e. Iran and northwest India.  Challenging the whole hypothesis, Frachetti and Benecke note: “Thus the data from Begash draw into question the general view that Eurasian pastoralism diffused eastward as a result of mounted horsemen in the Bronze Age”.

A study at Tentek-sor (northern Caspian, Kazakhstan) revealed that horse bones do not increase between 4000 BCE and 2000 BCE (Koryakova), and the samples did not contain any domestic sheep or cattle, even wild aurochs was only 5% (Frachetti 2012:7; Koryakova 2007). However, after 2000 BCE, we get a very large number of domestic cattle (60 to 90%) in the steppe and Central Asia, indicating the arrival of pastoralists with cattle into this region only after 2000 BC (Koryakova:88, 65, 146-147). This indicates that probably all the horse bones from earlier than 2000 BC dates are of the wild horses.

Koryakova (p.54) noted “but horse bones are extremely rare in the Kurgans”, and “a larger group of specialists share the idea that classical steppe nomadism appeared in the first millennium BCE” (ibid:55). This is consistent with the palaeo-climatic studies of Asia too. Earlier than this period there were forests in Central Asia and South Siberia although breaking and fast changing into open grass-lands due to human interference after 2000 BC. Yet forests dominated the region up to 1000 BC which were cleared by the farmers for agro-pastoral purposes. However, it is the total conversion into grassland and desert ecosystem which would have forced man to adopt the nomad existence. Thus nomadism was not the product of the mood or temperament of the particular nationality, but a geographic-ecological compulsion. It was a specialized niche for man.

The fossil pollen studies from Central Asia, South Siberia, Northern and Western China and other steppe zones have shown that the steppes converted into forests in the early Holocene (8000-6000 BC). Jiang (2006) found that the inner Mongolian steppe changed into birch-pine (Betula/Pinus) forest at 10,500 BC-7,200 BC period, and evolved into woodland with these trees dominating at 7,200-4,700 BC period, but reverted back to the steppe after 4,700 BC. It is at this very time, i.e. at 8,500 BC that the North American horses became extinct, a latest report based on ancient DNA says (Haile 2009). Earlier such extinction was stated to have taken place about 12,000 BC on the basis of fossil studies (Buck and Bard 2007). Clearly extinction of the North American horse occurred owing to the same climatic reasons which were operative in Central Asia and India. In Asia too horse became extinct from the region which we call the steppe.

Conversion back to the steppe and desert ecosystems took place at different times in different areas. Many areas remained forest as late as 2000 BC. At Yolin Am steppe (Southern Mongolia), it was found that it was a forest between c. 3600 BC and 2000 BC, and Betula (birch) and Salix (willow) trees dominated (Miehe 2007:156, Table 1). In general, however, after 3000 BC, more and more of Central Asia converted into steppe and desert (Zhao 2009; Zhao 2008: cited in F. Chen Editorial 2009:1). There was an abrupt change to arid climate at 2500 BC in many regions of China (Zhao 2009: Abstract; Chen, W. 2009). Miehe et al noted in their study of the succession of ecologies in the Gobi desert, that birch and willow pollens and charcoal were present in the soil layers up to 3000 BC (calibrated radiocarbon date), however birch became extinct from that site after that time (Miehe 2007:163; 156 Table 1).

Dense forest and desert are the places where horses die. In the dense forest, that harbours tiger, panther etc, horse can be easily predated because it cannot run fast enough in there to escape from the carnivores. Hence in all likelihood, hardly any horse may have lived before 4000 BC in the regions which we know today as the Eurasian steppe.

On the other hand, northwest India became a semi-arid region in about 33,000 BC and continued to be so until the end of the Glacial period (Petraglia 2009). Semi-arid ecosystems consist of deserts and grasslands like savannah, steppe, Sahel etc, but no dense forests. The Sivalensis horse must have found the northwest Indian grasslands as the ideal habitat. This situation lasted up to 6,000 BC, after which the region became moist leading to the growth of dense forests in Northwest India, which was not a friendly ecosystem for the horse. But then, the former Thar Desert evolved into grasslands (Deotare 2004a:Abstract), which stayed so until 2200 BC. Thus horses could have lived conveniently in the Thar between 6,000 BC and 2200 BC.

On the basis of the recent archaeological and palaeo-environmental studies, we are in a position to say that the horse domestication could have been possible in Central Asia and the north Pontic-Caspian steppe only after 2000 BC, if at all it took place in any of the two regions. The Bronze Age economy of this region remained mainly cattle and goat dominated and the classical horse based nomadism appeared only in the first millennium BCE after the aridity of the region increased enough to eliminate the possibility of farming-pastoralism.

[1] Number of private alleles: Whole of Europe 3; Whole of Kazakhstan 2; Kyrgyzstan 5; Xinjiang 5; India (only Jammu district) 4; Altai 2; Mongolia 2; Yunnan (Southeast Asia, politically part of Republic of China) 3.

And

Michael Witzel and the Indian Willow

Michael Witzel’s Concoction of Absence of Willow from India

Premendra Priyadarshi

The way he distorted, misrepresented and concocted history will not be ignored by the posterity. It will be difficult for the people of the future to believe that he did not know the truth, particularly when the fact will be known to them that people like me had pointed out the errors of facts in his works. They will know that he had the option of retracting from or correcting his earlier stands about the Aryan migration after knowing the loopholes in his theories. Many in future will condemn him of telling the untruth, and he will be accused of racial bigotry.

So what are the facts which expose Witzel to the extent of condemnation?

They are many, many! Today we will see the example of Indian ‘willow’ only. Other examples will be presented in future in my blogs.


Witzel claimed that willow is not found in India, and that it was not found in India when the Aryans arrived here, and that they (the Aryans) thrust the name vetasa on to the ‘reeds’, after finding no willow tree in northwest India (Witzel 2005:374; also 2009 Fulltext:5 n32).

[Witzel (2005:373) wrote, “Some of them therefore exhibit a slight change in meaning; a few others possibly are applications of old, temperate zone names to newly encountered plants, such as ‘willow’> ‘reed, cane’. Again, this change in meaning indicates the path of the migration, from the temperate zone into India.”]

He did not check the facts before writing. At least 40 species of willow (Salix) are native to northwest India (Pakistan), Nepal, Kashmir and other high altitude regions of north India, even if we ignore the species present in Afghanistan.

Witzel also did not know the palynological study of archaeological remains that this particular plant willow’s pollens have been found in abundance from Mehrgarh locality from tenth millennium BC up to the fourth millennium BC (Jarrige 2008:151). That means the Rig Vedic plant vetasa was willow, if the Rig Veda had been composed before the fourth millennium BCE, in the Indus-Saraswati Valley around Mehrgarh. The possibility of such an early date of the Rig-Veda is gaining hold as history is unfolding itself through the results of DNA studies, archaeogenetics, archaeology and by revision of the linguistic studies is pushing back the date of composition of the Rig Veda (Kazanas 2009).

Even after that till date, willow is found in many northern parts of India and Nepal. Some of the finest willows of furniture and basket grade are found in India under the species names Salix types deticulata, elegans etc. At least 30 species of Salix are found in Nepal and 28 in Pakistan itself. Hence Witzel’s wild guess that willow might not be found in India proves wrong and collapses his entire credibility as a reliable author.

Philology of ‘willow’ (Latin salix and Sanskrit vetasa)

Moreover, Witzel relied on the biased Eurocentric philology of vetasa given by others, and did not check whether any modern Indo-Aryan language has a cognate word of vetasa meaning ‘willow’. He as well as Monier-Williams gave the meaning ‘reed’ for Sanskrit vetasa, which was wrong because the later Indo-Aryan derivatives of vetasa like bet, bed etc certainly meant ‘willow’ in languages like Prakrit, Nepali, Kashmiri and Dardic etc. He is also silent about the origin or etymology of the unique Latin word salix (willow). Thus we can see:

Salix, sallow

 Lat. salix (willow) is a loanword from Germanic (Valpi:415). The cognates are found only in the Celtic and Germanic branches of IE, and that cannot warrant its inclusion as an Indo-European word. Cognates are: M. Irish sail, sa(i)lech, Welsh helyg-en, O. Brit. name Salico-dūnon, Gaul. name Salicilla; O.H.G. sal(a)ha, M.H.G. salhe, Ger. Salweide; O.E. sealh, O.Ice. selja (willow, from *salhjōn).

However most of the cognates in these two branches usually mean something flexible, ‘rope’, ‘twisted’ etc. However the Saxon root suggested for all these i.e. *sal means ‘black’ (Valpi:415) and the PIE root *sal2 means ‘salt’, ‘grey’, ‘saliva’ etc, having no specific semantic feature which could be associated with the willow tree.

Hence in all probability, the words represent an older linguistic substratum of Europe. The Altaic words like Tungus-Manchu *ǯalikta and Uralic like Finnish salava, jalava and Hungarian szilfa meaning ‘elm’ are enough evidence to suggest this fact (see Starostin’s Database).


The other word is Sanskrit ‘vetasa’. Its cognates mean willow in Indo-Aryan, Iranian, Germanic and Greek branches:

Av. vaēiti (willow, willow-stick); Gk. ἰτέα; O.Ice. vīðir, O.E. wīðig; M.L.G. wīde, O.H.G. wīda all meaning ‘willow’ (Pokorny: 1120-1122).

However Latin vitis does not mean ‘willow’ at all but but vine. Contrasting this, Sanskrit vēta-, vētasá, vētra etc all are cognates to this group of words, and mean ‘willow’ in northern Indo-Aryan languages.

Sanskrit u, vīĮu (strong) seems to be related with Proto-Indo-Aryan vēu (bamboo, willow) and its derivative Kashmiri vīr, vīrü (white willow; CDIAL 12091). Turner gives other cognate words from the Indo-Aryan branch having the meaning ‘willow’ as follows:

Proto-Indo-Aryan vēta (CDIAL 12097), Pashai-Dardic vei, wēu (willow), Dardic bīk (willow), Shina-Dard bĕu, bĕvĕ (willow); Proto-Indo-Aryan veta-daṇḍa (willow-stem, CDIAL 12098); From Sanskrit vetasa (CDIAL 12099), Prakrit vēdasa, vēasa (willow), Ashkun-Kaffiri  wis (willow), Kashmiri bisa (willow), Lahnda bīs, Nepali baĩs (willow), Dameli-Kafiri-Dardic bigyē (willow), Proto-Indo-Aryan *vēu–, vētrá–. *vētuka—(willow).

Other Indo-Aryan cognates meaning ‘willow’ and listed by some other authors are: Assamese bheha (salix), Punjabi bed (willow, Salix types, Singh:110); Nepali beu (Turner Nepali:456), bais, biu (Turner Nepali:458).

Persian cognates meaning willow are: bada, bīd, bed, bīdī, bīde (Steingass:165, 217-8).

The examination of the cognate words meaning ‘willow’ as provided by Turner from the modern and extinct Indo-Aryan languages reveals that the real meaning of the Old Indian or Vedic vetasa must have been ‘willow’ too. Prakrit, Northwest Indo-Aryan (Dardic), Kashmiri, Lahnda, Nepali and Assamese have the cognates meaning ‘willow’. Hence the Eurocentric stand taken by these scholars is not correct.

Jarrige, J-F, 2008, Pragdhara 18:135-154.

Kazanas, N., 2009, Indo-Aryan Origins and Other Vedic Issues, Aditya Prakashan, New Delhi.

Witzel, Michael, 2005, “Indocentrism: autochthonous visions of ancient India”, in Bryant, Edwin and Patton,L.L. (Eds.), The Indo-Aryan Controversy: Evidence and Inference in Indian History, Routledge, pp 341-404.

Witzel, Michael 2009, Fulltext, The linguistic history of some Indian domestic plants, J Biosciences 34(6): 829-833. “Fulltext” of this article is available at