VIII.21. Siberia

VIII.21.1. West Siberia

The Middle Bronze Age Pre-Andronovo cultural horizon of the beginning of the 2nd millennium BC in south-western Siberia was constituted by all of the forest-steppe and southern forests from the Trans-Urals to the Yenisei River, which were occupied by several cultures and groups: Tashkovo (Middle Tobol), Loginovo (Middle Ishim), Odino, Krotovo, and Elunino (the Middle Irtysh, Upper Ob’, and Altai areas), as well as the Samus’ culture (Tom–Chulym rivers) (Koryakova and Epimakhov 2007).

They display clear evidence of the diversified economy contributed by the productive and non-productive branches. Their similar pottery morphology and decoration shows alternating holed and combed motifs that cover the entire pot’s surface and reach back to an earlier epoch when they were dominant. These cultures remained rather distinctive and were only slightly touched by steppe influence (Koryakova and Epimakhov 2007).

The Tashkovo and Krotovo are more significant, contributing to later cultural formations in western Siberia, with the latter continuing into the Late Bronze Age. The Tashkovo culture is represented by villages with a circular or semi-circular layout, of ten to twelve houses on the banks of small rivers or lakes, whereas Krotovo shows open dispersed settlements with one- or two-chamber houses of the semisubterranean type and flat burial grounds (Koryakova and Epimakhov 2007).

The eastern Krotovo–Elunino territory (Middle Irtysh and Upper Ob’) shows flat burial grounds situated on elevated riverbanks or terraces, with sprinkled ochre, burials of separated skulls or skeletons without skulls, collective multi-layered burials, and secondary fractured burials. Some graves show rich metal objects, among which are some of Seima–Turbino type (Koryakova and Epimakhov 2007).

The arrival of Andronovo-like cultural horizon in the forest-steppe and southern forest regions replaced these cultures or displaced them to the north and east into the taiga.

viii.21.1. Yukaghirs

Based on proposed Indo-Uralic community (Kortlandt 2002; Kloekhorst 2008; Hyllested 2009), a macro-family formed by Indo-European, Uralic, and Yukaghir is quite likely to have migrated back from the Trans-Urals region with hunter-gatherer pottery (see §ii.3. Indo-Uralians). Remnant populations from this migration probably include two Sintashta outliers of Khvalynsk Eneolithic-like ancestry (ca. 2000–1650 BC), one R1b1a1-P297, the other R1b1a1a-M73 (Narasimhan et al. 2018), both lineages probably related to westward migrations through the Urals. Similarly, there is a possible sample of hg. R1b1a1-P297 in Darra-e Kūr, Afghanistan (ca. 2700 BC), with fully Iran Neolithic-like ancestry. The presence of hg. R1b1a1a-M73 lineages among diverse central Asian populations, in particular among Turkic-speaking groups (see below), also suggest potential remnant groups of hg. R1b1a1-P297 in West Siberia during the Neolithic.

Nevertheless, the continuous population expansions and replacements in Siberia have obscured the potential migration routes for Yukaghirs, whose language has been recorded only recently. Ancient DNA sampled from historical central Siberian peoples show certain discordances with a simplistic model of a macro-Yeniseian community in the taiga zone and southern Siberia up to the Altai–Sayan zone (Kim et al. 2018), and this complexity is in turn compatible with ancient Eurasian and Indo-Uralic population movements through Siberia . Modern Yukaghirs show among ten sampled individualsin a similar distribution to Tungusic Evenks and Evens—four hg. C2-M217, three N1a-L279, two R1a-M420, as well as (not present in Tungusic peoples) one I2a-L460 and one O-M175 (Fedorova et al. 2013).

Four samples of the Bronze Age Glazkovo culture, from the Lake Baikal (ca. 18th–13th c. BC), one of hg R1-M173, show ancestry compatible with modern north-east Siberian populations, compatible with their later described spread to the north (de Barros Damgaard, Marchi, et al. 2018).

A medieval individual near the Yana river (ca. AD 1350) shows hg. N1a1a1a1a4a1-M1993 (formed ca. 1700 BC, TMRCA ca. AD 450), and falls within the widespread Neosiberian cline evidenced by a recent sample from Ust’Belaya (ca. AD 1300) near Lake Baikal (Sikora et al. 2018). The prevalent presence of N1a1a1a1a4-M2019 (formed ca. 4400 BC, TMRCA 1700 BC) in central Siberia and Yakutia, and at lower frequencies in Khants and Mansis (Ilumae et al. 2016), is most likely the result of the expansion of Yukaghir-and Altaic-related languages with acculturated Palaeosiberian clans.

The complex evolution of north-east Asia can be seen in the replacement of lineages from Ekven Iron Age samples (ca. 400 BC – AD 400), predominantly of hg. Q-M242, at least one within the Q1b1a1a-M3 tree (formed ca. 13200 BC, TMRCA ca. 11400 BC), prevalent in the Americas, and two samples of hg. C2a-L1373, at least one within the C2a1a1a2-F3918 tree (formed ca. 12400 BC, TMRCA ca. 10800 BC), prevalent today in Eurasian populations likely related to Altaic expansions (Sikora et al. 2018).

The expansion of Yukaghir was probably then coincident with Bronze Age migrations, likely continued to the east with N1a1a1a1a4a-M1993 during the Iron Age–Early Middle Ages, as Ugrians and Samoyeds expanded to the north (see §viii.17. Ugrians and Samoyeds). This later expansion probably displaced populations of mainly N1a1a1a1a3b-B202 lineages (formed ca. 2800, TMRCA ca. 600 BC) to the extreme north-eastern Siberia, where they retained their Chukotko-Kamchatkan languages, supported by their prevalent N1a1a1a-L708 lineages (ca. 92%).

Turkic-speaking Evenks cluster together and overlap with Yukaghirs, and both in turn cluster closely to Nganasans from the Taymyr Peninsula and to Chukchi-speaking Koryaks, revealing a complex acculturation of different East Asian peoples in recent times (Karafet et al. 2018). The wide cluster formed by modern Yukaghirs, including southern Samoyedic speakers in the west and Chukotko-Kamchatkan speakers in the east (see Suppl. Graph. 15) further supports the relatively recent expansion of Yukaghir into the Circum-Arctic region.

viii.21.2. Turkic peoples and Mongols

Cultures succeeding Afanasevo in the Altai region show different lineages and the partial resurgence of WSHG ancestry coupled with different Afanasevo-, Steppe MLBA-, AP-, and AEA-related contributions, suggesting the emergence of different local West Siberian populations (Hollard et al. 2018): the Chemurchek culture (ca. 2300–1800 BC), with one hg. C-M130; the Okunevo culture (ca. 2300–1800 BC), with one hg. R1b1a1b1-L23, three likely N-Z4813, and three likely Q1b1a-L54; the Elunino culture (ca. 2300–1700 BC), with one hg. Q-M242, and one Q1b-M346; the Munkh-Khairkhan culture (ca. 1700–1400 BC), with two N-M231 lineages; and the Sagsai culture (ca. 1400–800 BC), with four hg. Q1b1a-L54, four R1a1a1b2-Z93, and one C-M130.

MBA sites Takhilgat Uzuur and Tsagaan Asga in the Mongolian Altai Mountains also show apparent succeeding replacements from hg. Q1b1a-L54 in the 3rd millennium BC to R1a1a1b2-Z93 in the late 2nd millennium BC, to C-M130 in the early 1st millennium BC  (Hollard et al. 2014).

To the east, near Lake Baikal, a resurgence of AP ancestry (up to ca. 50%) coupled with Afanasevo-related ancestry (ca. 10%) found in EBA individuals is continued during the LBA. Samples of the Deer Stone-Khirigsuur Complex from Khövsgöl in northern Mongolia (ca. 1200–800 BC) show a slightly higher contribution of AP ancestry evidenced by a ‘northern’ shift in the PCA, similar to Karasuk or Okunevo samples (Jeong, Wilkin, et al. 2018). In terms of haplogroups, there are ten probably Q1b-L56, possibly all of Q1b1a-L54 subclade Q1b1a3-L330 (formed ca. 16000 BC, TMRCA ca. 5900 BC), with an estimated successful dispersal of these subclades starting in the Mesolithic ca. 6300 BC from central Asia (Grugni et al. 2019), possibly initially accompanying Dene-Yeniseian-related languages (see §v.8. Palaeosiberians).

A small contribution of Steppe MLBA ancestry (ca. 4-7%) suggests close contacts with Andronovo-related peoples. In particular, the presence of one outlier (ca. 1130–900 BC), of hg. R1a1a1b2a2a1-Z2123, is consistent with the appearance of admixed forest-steppe populations of Eastern Steppe MLBA ancestry like Karasuk in the Altai (ca. 1200–800 BC). Supporting these contacts of Karasuk with eastern Asian steppes is also the Karasuk outlier and the sample of hg. Q1a2a-L712. This points to the spread of pastoralism in the region mediated by acculturation and exogamy more than population replacement (Jeong, Wilkin, et al. 2018). Another sample (ca. 1420-1130 BC) of hg. N1a1a-M178, mtDNA U5a2d1, suggestslike the different groups succeeding Afanasevo in the Altai regionthe acculturation of northern Eurasian communities with post-Neolithic expansions.

To the south-east of Lake Baikal, in the Houtaomuga site from Manchuria, there is genetic continuity from ca. 10000 BC until the Iron Age, but haplogroup N1b1-CTS582, found in the Early Neolithic (ca. 5430–5320 BC), is replaced in the Bronze Age by C2b-L1373, which continues in the Early Iron Age (Ning 2018). Twelve Donghu individuals from the Jinggouzi site (ca. 770–476 BC) show C2a1a-F4032 lineages (formed ca. 12700 BC, TMRCA ca. 12300 BC), as do four samples from Xianbei (ca. AD 4th–10th c.) (Zhang, Wu, et al. 2018; Li et al. 2018). The expansion of the Donghu seems to have caused the expansion of C2-M217 lineages in the Trans-Baikal area.

The first Turkic-speaking community is usually identified with the Xiongnu confederation, with ancient Y-chromosomal data indicating a heterogeneous multi-ethnic cultural organisation, likely emerging initially from local East Asian groups to the east of the Tian Shan Mountains, who showed admixture with central steppe nomads. A more recent West Eurasian ancestry is found among western Xiongnu groups, with Central Sakas being the closest source for their admixture (de Barros Damgaard, Marchi, et al. 2018).

Sampled Xiongnu individuals (ca. 300 BC – AD 200) include hg. R1a1a-M198 and C2-M217 in Duurlig Nars, and hg. N1a1-Tat and Q-M52 and in Egyin Gol Valley (Kim et al. 2010; Petkovski 2006); two likely hg. O2a2b-P164 from Omnogobi, one early sample from the West Xiongnu in Khövsgöl (ca. 330 BC), of hg. R1b2-PH155, and one from an aristocratic burial in Arkhangai (ca. AD 1), likely of hg. R1b2b-PH200+ (de Barros Damgaard, Marchi, et al. 2018).

The homeland of Turkic peoples is difficult to pinpoint based on scarce samples through wide temporal transects, due to the multiple population replacements in the central and eastern Asian steppes, forest-steppes, and forests, and to the linguistic data and complex ethnogenesis legends pointing to a composite grouping of diverse elements since the reconstructible stage of the language (Golden 1992).

Based on Indo-European (Iranian and Tocharian) and Uralic influences, Proto-Turkic is supposed to have been spoken ca. 1000 BC in some area from the Trans-Urals area to the Altai, with the forest zones of West Siberia being the most likely candidate based on the Indo-Iranian expansions through the steppes, as well as the Uralic expansions through the forest-steppe and forest regions. From there, the ancestors of the Turks migrated east into the Baikal area, where the Xiongnu confederation eventually emerged, and Huns later migrated to West Eurasia (Golden 1992).

The two Sintashta outliers of Khvalynsk Eneolithic-like ancestry, of reported haplogroups R1b1a1-P297 and R1b1a1a-M73, probably correspond to ancient populations of R1b1a1a-M73 lineages widespread from the west in the southern Urals up to the Lake Baikal in the east since the Neolithic, following a WSHG ancestry cline. This is compatible with the Mesolithic expansion of Eurasian through Inner Asia (see §ii.1. Eurasians), and the isolated development of Altaic in the Neolithic (see §v.8. Palaeosiberians).

Haplogroup R1b1a1a-M73 is reported with low frequencies among modern Siberian populations, such as Ugric and Samoyedic peoples, especially southern Selkups (Tambets et al. 2018), and with increasing frequencies among south Siberian populations, in particular among Turkic-speaking Bashkirs (Jeong, Balanovsky, et al. 2018) and Teleuts near the Altai (Karafet et al. 2018), which suggests its presence among Turkic-speaking peoples before their expansion into the Trans-Baikal area and the creation of a community integrating diverse local populations.

The earlier emergence of Turkic-speaking peoples from the forming multi-ethnic groups in the Trans-Baikal area further supports that Turko–Mongolic, and not only Turkic, expanded from the west. Similarly, the link of different Altaic-related expansions since the early first millennium BC associated with bottlenecks of West Siberian N1a1a1a1a-L1026/L392 lineages (see §v.8. Palaeosiberians) further supports the connection of Altaic with the West Siberian forest-steppes. The adaptation of horseback riding for more stressful and difficult activities such as warfare—characteristic of these nomadic groups—in the eastern steppes, started probably at the end of the Deer Stone–Khirigsuur Complex or slightly later, evidenced by findings from the Altai to the Baikal region, including the appearance of the bronze snaffle bit and innovations in equine dentistry (Taylor, Bayarsaikhan, et al. 2018).

The Huns, likely representatives of the earliest Turkic-speaking groups in Eurasia, emerged following minor male-driven East Asian gene flow into the preceding Sakas that they invaded, ca. 2,000 years ago. They displaced Iranian-speaking groups (ancestors of the Wusun and Kangju) to the south-east of the Tian Shan mountains, where they became isolated (see §viii.19. Iranians). Sampled nomads from the Kargaly in the Tian Shan region include two early ones (ca. 800–700 BC), one of hg. Q-M242, clustering with Siberian peoples (close to Yeniseians), one intermediate (ca. 425 BC), of hg. R1a1a1b2a2-Z2124, and a later one (ca. 35 BC), of hg. R1b1a1a1b-Y20750 (formed ca. 5300 BC, TMRCA ca. 3300 BC), a subclade of R1b1a1a-M73 (de Barros Damgaard, Marchi, et al. 2018).

This variability is also found later among Huns from Tian Shan (ca. AD 60–600), with three hg. R1a1a1b2a2-Z2124, two R1-M173, one hg. R1b2b-PH200, one N1a1a-M178, and one Q1a2a1-L715+, in contrast to Tian Shan Sakas, who were all of R1-M173 subclades. Nomads from the central steppe (ca. AD 360), either Huns or Sarmatians, show one sample of hg. R1a1a1b2-Z93. Huns sampled from the Carpathian Basin (AD 5th c.) show hg. Q1a2-M25, one R1b1a1b1a1a1-U106, and one R1a1a1b2a2-Z2124 (Neparáczki et al. 2019). All Huns show an increased shared drift with West Eurasians compared to Xiongnu (de Barros Damgaard, Marchi, et al. 2018).

A sample of the Mongolian Rouran Khaganate from Khermen Tal shows the continuation of haplogroup C2a1a-F4032 in the region (Li et al. 2018), probably expanding with Mongolian-speaking peoples in the eastern steppes. Samples from Baiyin Huangwan Han dynasty tombs from a north-western Chinese farming area near the Xiongnu states, spanning from the Western Han Dynasty (202 BC – AD 8) to the Eastern Han Dynasty (AD 25–220), show significant genetic contribution from the northern Eurasian populations, as well as an accommodation to the nomadic lifestyle, which supports the acculturation of the Xiongnu population to the Han culture (Li, Ma, and Wen 2018).

After the defeat of the Xiongnu (ca. AD 552), a part of their population migrated to Pannonia, where they became known as the Avars and allied with the Longobards to defeat the Gepids, creating the Avar Khaganate (AD 567-805). Two early female Avars from Szólád (ca. AD 540–640) show an ancestry similar to Eastern Europeans, with contributions of East Asian ancestry, clustering close to modern West Slavs, which justifies the Central Asian admixture found in a Gepid and a medieval Bavarian individual of the region (Amorim et al. 2018). Twenty-three individuals from a group of elite burials in Hungary (AD 7th–8th c.) show a majority of Inner Asian origin (up to 64%) of their mtDNA (Csáky et al. 2018), although there is high intergroup variation (Šebest et al. 2018).

In terms of Y-chromosome haplogroups, early Avars (ca. AD 570–660) show the intrusion of N1a1a-M178, found among six out of eleven males, with at least four of them subclades of N1a1a1a1a3-Y16323 (formed ca. 2900 BC, TMRCA ca. 2900 BC), in turn a subclade of N1a1a1a1a-L392 (formed ca. 4300 BC, TMRCA ca. 2900 BC). At least one of them estimated to be N1a1a1a1a3a-F4205+ (TMRCA ca. 500 BC), a haplogroup confirmed in all seventeen sampled males from the Avar Khaganate in the Danube–Tisza Interfluve (ca. 600–775 BC). Haplotypes confirm that these samples share their closes relatives today among Siberian populations, including Buryats, Mongolians, Yakuts, Mansis and Khants, in line with the presence of hg. N1a1a-M178 in Bronze Age Inner Mongolia and late medieval Yakuts (Csáky et al. 2019).

Two individuals from a secondary power centre east of the Tisza (ca. 650-700 BC) show hg. Q1a-F1096 and Q1b-M346, the latter of likely Altaian or South Siberian paternal origin (Csáky et al. 2019). Other early haplogroups include one C2-M217 and one R1a1a1b2a2-Z2124. All these samples support an original Siberian expansion of this haplogroup from a region close to Lake Baikal. Samples probably related originally to European paternal lines include one early Avar of hg. G2a-P15, and another one of hg. I1-M253 (Neparáczki et al. 2019).

Middle or Late Avar samples (ca. 650-710 BC) show one hg. C2-M217 and one N1a1a1a1a3-Y16323, showing continuity with the previous period, but also one E1b1b1a1b1a-V13. Given the nature of the Avar polity as a Slavic-speaking territory during the last century of its existence (ca. 700–800 BC), and the finding of E1b1b1a1b1a-V13 lineages later during the Hungarian Conqueror period (see §viii.17.1. Ugrians) and among early Slavs, it is highly likely that the expansion of hg. E1b1b1a1b1a-V13 from the Carpathian Basin is related to the Slavonic expansion to neighbouring regions (Neparáczki et al. 2019).

The Turkic Khaganate assumed military and political organisation of the steppes as the Hunnic Empire broke up and dispersed (AD 6th c.), following the emergence of the Turks, the blacksmiths of the Rourans (Suppl. Fig. 17). Their elite soldiers are genetically closer to East Asians than the preceding Huns of the Tian Shan mountains, with one sample from Berygavoya (ca. 690 BC), of hg. R1-M173. A genetic outlier of the central steppe (ca. AD 270), of hg. R1-M173, shows pronounced European ancestry, and thus ongoing contacts with Europe (de Barros Damgaard, Marchi, et al. 2018).  

The Turkic Khaganate was eventually replaced by short-lived steppe cultures, such as the Kipchak and the Tungusik Kimak poulations, which spread southwards towards the Tian Shan mountains and westward towards the Ural Mountains to form the Kimak Khaganate in the central steppe (ca. 8th–11th c.). One sample from Kimak nomads of the Central Steppes (ca. AD 665), of hg. R1b1a1a1a-Y14051, does not show elevated East Asian ancestry (de Barros Damgaard, Marchi, et al. 2018).

The Kima Khaganate was replaced by local Kipchak groups allied with the Cuman of West Eurasia, hence probably originating near the area of Tuva. Two individuals dating to the Cuman–Kipchak alliance (ca. AD 1050) show one hg. C-M130 and increased East Asian ancestry, and the other one pronounced European ancestry, which is compatible with the incorporation of western and eastern steppe populations. The Karakhanid Khaganate from Turan incorporated some of these groups, with three samples (ca. AD 950–1250) showing further East Asian influx compared to earlier Turks (de Barros Damgaard, Marchi, et al. 2018).

Other unassigned early medieval Turkic samples show different proportions of East Asian ancestry, including nine from Tian Shan, among them one early (ca. AD 800–1000), of hg. J2a-M410, and one later (ca. AD 1170) of hg. C2a1a1b1b1-Y12825+ (formed ca. 1000 BC, TMRCA ca. 100 BC); one from the Central Steppe (ca. AD 735) of hg. R1b1a1a1a-Y14051+; and one from the Caspian Steppe (ca. AD 700) of hg. R1a1a1b2a2-Z2124+. Among two samples (ca. AD 1250) of the medieval Jochi Khan’s Golden Horde in the central steppes, there is one of clearly East Asian ancestry and corresponding PCA cluster, of hg. C2-M217, and one of West Eurasian descent, of hg. R1a1a1b1a2a-Z280, which is further proof of the assimilation of different groups into succeeding Turkic organisations (de Barros Damgaard, Marchi, et al. 2018).

Among modern Tatars, descendants from elite clans of the Golden Horde belong to haplogroup R1b1a1a-M73 (Akchurin et al. 2018). There seems to be a general trend during the Iron Age and medieval times to a distribution of R1a1a1b2a2-Z2124 lineages in the Pontic–Caspian steppes, of R1b1a1a-M73 lineages in the central steppes, and of R1-M173 (likely R1b2-PH155) and C2a1a1b1b1-Y12825 in the eastern steppes, which may reflect to some extent the different alliances formed by multi-ethnic groups since the time of the formation of the Xiongnu confederacy and the Hunnic expansion, although it may also reflect the initial contacts between peoples of the eastern steppe before the formation of the Xiongnu community. Modern peoples from investigated Xinjiang sites show hg. R1b-M343 including R1b1a1-P297, R1b1a1a-M73 (up to 9%), and R1b2-PH155 subclades, among a majority of typically central Asian lineages, including hg. R1a1a1b2-Z93, especially R1a1a1b2a2-Z2124 (Liu et al. 2018), whose origin cannot be properly interpreted without specific subclades.

Eventually, these khaganates were conquered by the Mongol Empire, which emerged through the unification of East Mongolian and Trans-Baikal tribes, expanding under the rule of Genghis Khan (ca. AD 13th c.). Modern Mongolian tribes show a mixture of East Asian lineages, mainly C2 subclades (ca. 42%), including C2b-F1067 (ca. 29%) and C2a-L1373 (ca. 13%), but also O-M175 (ca. 24%) and N-M231 (ca. 18%). In the PCA, Mongolians cluster in close genetic proximity to a group of North Asian Siberians, including Altaians, Tuvinians, Evenki, and Yakut, with eastern tribes Abaga, Khalkha, Oirat, and Sonid showing the least differentiation, with close interaction between northern Eurasian populations (Bai et al. 2018). Common Mongols likely expanded mainly with a Y-chromosome bottleneck of haplogroup C2a1a1c1-F3796 (TMRCA ca. 500 BC), whose expansion pattern is consistent with the diffusion of most Mongolic-speaking populations (Wei et al. 2018).

Manchu-Tungusic is proposed to have spread either from the Trans-Baikal area or from the Amur River region. Ancient samples from the West Liao River region shows high dynamism, similar to Trans-Baikal areas[39], which may support the emergence of Tungusic-speaking peoples from previous population movements through the eastern steppes. The presence of shared isoglosses with Turkic and Mongolic to the west, forming a likely Altaic family, and with Koreanic (and Japonic) to the east, with less clear links to Altaic, makes the identification of Tungusic still more complicated. The Proto-Tungusic society has been associated either with the Donghu or with an ancestral group from the Amur River region sharing links with eastern peoples. The potential attribution of vowel harmony in Proto-Tungusic to contacts with Mongolic languages (Ko, Joseph, and Whitman 2014) makes the identification of the language with one or the other group still harder.

Present-day Tungusic-speaking peoples, of varied lineages and ancestry, share a similar history to that found among Palaeosiberian peoples integrated among Finnic, Samic, Samoyedic, or Yukaghir-speaking populations, of acculturated Palaeosiberians adopting languages in recent times (Fedorova et al. 2013). This tradition of exogamy notwithstanding, it seems that the recent expansion of C2a1a1b1-M86 lineages among some southern Tungusic groups (as well as other C2-M217 lineages) may have been associated with their expansion from the south-east (Balanovska et al. 2018).