We know that the Caucasus Mountains formed a persistent prehistoric barrier to cultural and population movements. Nevertheless, an even more persistent frontier to population movements in Europe, especially since the Neolithic, is the Pontic-Caspian steppe – forest-steppe ecotone.
Like the Caucasus, this barrier could certainly be crossed, and peoples and cultures could permeate in both directions, but there have been no massive migrations through it. The main connection between both regions (steppe vs. forest-steppe/forest zone) was probably through its eastern part, through the Samara region in the Middle Volga.
The chances of population expansions crossing this natural barrier anywhere else seem quite limited, with a much less porous crossing region in the west, through the Dnieper-Dniester corridor.
A Persistent ecological and cultural frontier
It is very difficult to think about any culture that transgressed this persistent ecological and cultural frontier: many prehistoric and historical steppe pastoralists did appear eventually in the neighbouring forest-steppe areas during their expansions (e.g. Yamna, Scythians, or Turks), as did forest groups who permeated to the south (e.g. Comb Ware, GAC, or Abashevo), but their respective hold in foreign biomes was mostly temporary, because their cultures had to adapt to the new ecological environment. Most if not all groups originally from a different ecological niche eventually disappeared, subjected to renewed demographic pressure from neighbouring steppe or forest populations…
Before the emergence of pastoralism, the cultural contacts of the Pontic region (i.e. forest-steppes) with the Baltic were intense. In fact, the connection of the north Pontic area with the Baltic through the Dnieper-Dniester corridor and the Podolian-Volhynian region is essential to understand the spread of peoples of post-Maglemosian and post-Swiderian cultures (to the south), hunter-gatherer pottery (to the north), TRB (to the south), Late Trypillian groups (north), GAC (south), or Comb Ware (south) (see here for Eneolithic movements), and finally steppe ancestry and R1a-Z645 with Corded Ware (north). After the complex interaction of TRB, Trypillia, GAC, and CWC during the expansion of late Repin, this traditional long-range connection is lost and only emerges sporadically, such as with the expansion of East Germanic tribes.
A barrier to steppe migrations into northern Europe
One may think that this barrier was more permeable, then, in the past. However, the frontier is between steppe and forest-steppe ecological niches, and this barrier evolved during prehistory due to climate changes. The problem is, before the drought that began ca. 4000 BC and increased until the Yamna expansion, the steppe territory in the north Pontic region was much smaller, merely a strip of coastal land, compared to its greater size ca. 3300 BC and later.
This – apart from the cultural and technological changes associated with nomadic pastoralism – justifies the traditional connection of the north Pontic forest-steppes to the north, broken precisely after the expansion of Khvalynsk, as the north Pontic area became gradually a steppe region. The strips of north Pontic and Azov steppes and Crimea seem to have had stronger connections to the Northern Caucasus and Northern Caspian steppes than with the neighbouring forest-steppe areas during the Upper Palaeolithic, Mesolithic, and Neolithic.
NOTE. We still don’t know the genetic nature of Mikhailovka or Ezero, steppe-related groups possibly derived from Novodanilovka and Suvorovo close to the Black Sea (which possibly include groups from the Pannonian plains), and how they compare to neighbouring typically forest-steppe cultures of the so-called late Sredni Stog groups, like Dereivka or partly Kvityana.
Despite the Pontic-Caspian steppes and forest-steppes neighbouring each other for ca. 2,000 km, peoples from forested and steppe areas had an obvious advantage in their own regions, most likely due to the specialization of their subsistence economy. While this is visible already in Palaeolithic and Mesolithic hunter-gatherers, the arrival of the Neolithic package in the Pontic-Caspian region incremented the difference between groups, by spreading specialized animal domestication. The appearance of nomadic pastoralism adapted to the steppe, eventually including the use of horses and carts, made the cultural barrier based on the economic know-how even stronger.
Even though groups could still adapt and permeate a different territory (from steppe to forest-steppe/forest and vice-versa), this required an important cultural change, to the extent that it is eventually complicated to distinguish these groups from neighbouring ones (like north-west Pontic Mesolithic or Neolithic groups and their interaction with the steppes, Trypillia-Usatovo, Scythians-Thracians, etc.). In fact, this steppe – forest-steppe barrier is also seen to the east of the Urals, with the distinct expansion of Andronovo and Seima-Turbino/Andronovo-like horizons, which seem to represent completely different ethnolinguistic groups.
As a result of this cultural and genetic barrier, like that formed by the Northern Caucasus:
1) No steppe pastoralist culture (which after the emergence of Khvalynsk means almost invariably horse-riding, chariot-using nomadic herders who could easily pasture their cows in the huge grasslands without direct access to water) has ever been successful in spreading to the north or north-west into northern Europe, until the Mongols. No forest culture has ever been successful in expanding to the steppes, either (except for the infiltration of Abashevo into Sintashta-Potapovka).
2) Corded Ware was not an exception: like hunter-gatherer pottery before it (and like previous population movements of TRB, late Trypillia, GAC, Comb Ware or Lublin-Volhynia settlers) their movements between the north Pontic area and central Europe happened through forest-steppe ecological niches due to their adaptation to them. There is no reason to support a direct connection of CWC with true steppe cultures.
3) The so-called “Steppe ancestry” permeated the steppe – forest-steppe ecotone for hundreds of years during the 5th and early 4th millennium BC, due to the complex interaction of different groups, and probably to the aridization trend that expanded steppe (and probably forest-steppe) to the north. Language, culture, and paternal lineages did not cross that frontier, though.
EDIT (4 FEB 2019): Wang et al. is out in Nature Communications. They deleted the Yamna Hungary samples and related analyses, but it’s interesting to see where exactly they think the trajectory of admixture of Yamna with European MN cultures fits best. This path could also be inferred long ago from the steppe connections shown by the Yamna Hungary -> Bell Beaker evolution and by early Balkan samples:
Let me begin this final post on the Corded Ware—Uralic connection with an assertion that should be obvious to everyone involved in ethnolinguistic identification of prehistoric populations but, for one reason or another, is usually forgotten. In the words of David Reich, in Who We Are and How We Got Here (2018):
Human history is full of dead ends, and we should not expect the people who lived in any one place in the past to be the direct ancestors of those who live there today.
Another recurrent argument – apart from “Siberian ancestry” – for the location of the Uralic homeland is “haplogroup N”. This is as serious as saying “haplogroup R1” to refer to Indo-European migrations, but let’s explore this possibility anyway:
We have now a better idea of how many ancient migrations (previously hypothesized to be associated with westward Uralic migrations) look like in genetic terms. From Damgaard et al. (Science 2018):
These serial changes in the Baikal populations are reflected in Y-chromosome lineages (Fig. SA; figs. S24 to S27, and tables S13 and SI4). MAI carries the R haplogroup, whereas the majority of Baikal_EN males belong to N lineages, which were widely distributed across Northern Eurasia (29), and the Baikal_LNBA males all carry Q haplogroups, as do most of the Okunevo_EMBA as well as some present-day Central Asians and Siberians.
The only N1c1 sample comes from Ust’Ida Late Neolithic, 180km to the north of Lake Baikal, which – together with the Bronze Age sample from the Kola peninsula, and the medieval sample from Ust’Ida – gives a good idea of the overall expansion of N subclades and Siberian ancestry among the Circum-Arctic peoples of Eurasia, speakers of Palaeo-Siberian languages.
What we should expect from Uralic peoples expanding with haplogroup N – seeing how Yamna expands with R1b-L23, and Corded Ware expands with R1a-Z645 – is to find a common subclade spreading with Uralic populations. Let’s see if it works like that for any N-X subclade, in data from Ilumäe et al. (2016):
Within the Eurasian circum-Arctic spread zone, N3 and N2a reveal a well-structured spread pattern where individual sub-clades show very different distributions:
N1a1-M46 (or N-TAT), formed ca. 13900 BC, TMRCA 9800 BC
N1a1a2-B187, formed ca. 9800 BC, TMRCA 1050 AD:
The sub-clade N3b-B187 is specific to southern Siberia and Mongolia, whereas N3a-L708 is spread widely in other regions of northern Eurasia.
N1a1a1a-L708, formed ca. 6800 BC, TMRCA 5400 BC.
N1a1a1a2-B211/Y9022, formed ca. 5400 BC, TMRCA 1900 BC:
The deepest clade within N3a is N3a1-B211, mostly present in the Volga-Uralic region and western Siberian Khanty and Mansi populations.
N1a1a1a1a-L392/L1026), formed ca. 4400 BC, TMRCA 2800 BC:
The neighbor clade, N3a3’6-CTS6967, spreads from eastern Siberia to the eastern part of Fennoscandia and the Baltic States
N1a1a1a1a1a-CTS2929/VL29, formed ca. 2100 BC, TMRCA 1600 BC:
In Europe, the clade N3a3-VL29 encompasses over a third of the present-day male Estonians, Latvians, and Lithuanians but is also present among Saami, Karelians, and Finns (Table S2 and Figure 3). Among the Slavic-speaking Belarusians, Ukrainians, and Russians, about three-fourths of their hg N3 Y chromosomes belong to hg N3a3.
In the post on Finno-Permic expansions, I depicted what seems to me the most likely way of infiltration of N1c-L392 lineages with Akozino warrior-traders into the western Finno-Ugric populations, with an origin around the Barents sea.
This includes the potential spread of (a minority of) N1c-B211 subclades due to contacts with Anonino on both sides of the Urals, through a northern route of forest and forest-steppe regions (equivalent to the distribution of Cherkaskul compared to Andronovo), given the spread of certain subclades in Ugric populations.
NOTE. An alternative possibility is the association of certain B211 subclades with a southern route of expansion with Pre-Scythian and Scythian populations, under whose influence the Ananino culture emerged -which would imply a very quick infiltration of certain groups of haplogroup N everywhere among Finno-Ugrics on both sides of the Urals – , and also the expansion of some subclades with Turkic-speaking peoples, who apparently expanded with alliances of different peoples. Both (Scythian and Turkic) populations expanded from East Asia, where haplogroup N (including N1c) was present since the Neolithic. I find this a worse model of expansion for upper clades, but – given the YFull estimates and the presence of this haplogroup among Turkic peoples – it is a possibility for many subclades.
N1a1a1a1a2-Z1936, formed ca. 2800 BC, TMRCA 2400 BC:
The only notable exception from the pattern are Russians from northern regions of European Russia, where, in turn, about two-thirds of the hg N3 Y chromosomes belong to the hg N3a4-Z1936—the second west Eurasian clade. Thus, according to the frequency distribution of this clade, these Northern Russians fit better among other non-Slavic populations from northeastern Europe. N3a4 tends to increase in frequency toward the northeastern European regions but is also somewhat unexpectedly a dominant hg N3 lineage among most Turcic-speaking Volga Tatars and South-Ural Bashkirs.
N1a1a1a1a4-M2019 (previously N3a2), formed ca. 4400 BC, TMRCA 1700 BC:
Sub-hg N3a2-M2118 is one of the two main bifurcating branches in the nested cladistic structure of N3a2’6-M2110. It is predominantly found in populations inhabiting present-day Yakutia (Republic of Sakha) in central Siberia and at lower frequencies in the Khanty and Mansi populations, which exhibit a distinct Y-STR pattern (Table S7) potentially intrinsic to an additional clade inside the sub-hg N3a2
The second widespread sub-clade of hg N is N2a. (…):
N1a2b-P43 (B523/FGC10846/Y3184), formed ca. 6800 BC, TMRCA ca. 2700 BC:
The absolute majority of N2a individuals belong to the second sub-clade, N2a1-B523, which diversified about 4.7 kya (95% CI = 4.0–5.5 kya). Its distribution covers the western and southern parts of Siberia, the Taimyr Peninsula, and the Volga-Uralic region with frequencies ranging from from 10% to 30% and does not extend to eastern Siberia (…)
The “European” branch suggested earlier from Y-STR patterns turned out to consist of two clades
N1a2b2a-Y3185/FGC10847, formed ca. 2200 BC, TMRCA 800 BC:
N2a1-L1419, spread mainly in the northern part of that region.
N1a2b2b1-B528/Y24382, formed ca. 900 BC, TMRCA ca. 900 BC:
N2a1-B528, spread in the southern Volga-Uralic region.
We also have a good idea of the distribution of haplogroup R1a-Z645 in ancient samples. Its subclades were associated with the Corded Ware expansion, and some of them fit quite well the early expansion of Finno-Permic, Ugric, and Samoyedic peoples to the east.
This is how the modern distribution of R1a among Uralians looks like, from the latest report in Tambets et al. (2018):
Among Fennic populations, Estonians and Karelians (ca. 1.1 million) have not suffered the greatest bottleneck of Finns (ca. 6-7 million), and show thus a greater proportion of R1a-Z280 than N1c subclades, which points to the original situation of Fennic peoples before their expansion. To trust Finnish Y-DNA to derive conclusions about the Uralic populations is as useful as relying on the Basque Y-DNA for the language spread by R1b-P312…
Among Volga-Finnic populations, Mordovians (the closest to the original Uralic cluster, see above) show a majority of R1a lineages (27%).
Hungarians (ca. 13-15 million) represent the majority of Ugric (and Finno-Ugric) peoples. They are mainly R1a-Z280, also R1a-Z2123, have little N1c, and lack Siberian ancestry, and represent thus the most likely original situation of Ugric peoples in 4th century AD (read more on Avars and Hungarians).
Among Samoyedic peoples, the Selkup, the southernmost ones and latest to expand – that is, those not heavily admixed with Siberian populations – , also have a majority of R1a-Z2123 lineages (see also here for the original Samoyedic haplogroups to the south).
To understand the relevance of Hungarians for Ugric peoples, as well as Estonians, Karelians, and Mordovians (and northern Russians, Finno-Ugric peoples recently Russified) for Finno-Permic peoples, as opposed to the Circum-Arctic and East Siberian populations, one has to put demographics in perspective. Even a modern map can show the relevance of certain territories in the past:
Summary of ancestry + haplogroups
Fennic and Samic populations seem to be clearly influenced by Palaeo-Laplandic peoples, whereas Volga-Finnic and especially Permic populations may have received gene flow from both, but essentially Palaeo-Siberian influence from the north and east.
The fact that modern Mansis and Khantys offer the highest variation in N1a subclades, and some of the highest “Siberian ancestry” among non-Nganasans, should have raised a red flag long ago. The fact that Hungarians – supposedly stemming from a source population similar to Mansis – do not offer the same amount of N subclades or Siberian ancestry (not even close), and offer instead more R1a, in common with Estonians (among Finno-Samic peoples) and Mordvins (among Volga-Finnic peoples) should have raised a still bigger red flag. The fact that Nganasans – the model for Siberian ancestry – show completely different N1a2b-P43 lineages should have been a huge genetic red line (on top of the anthropological one) to regard them as the Uralian-type population.
It is not hard to model the stepped arrival, infiltration, and/or resurge of N subclades and “Siberian ancestries”, as well as their gradual expansion in certain regions, associated with certain migrations first – such as the expansions to the Circum-Arctic region, and later the Scythian- and Turkic-related movements – , as well as limited regional developments, like the known bottleneck in Finns, or the clear late expansion of Ugric and Samoyedic languages to the north among nomadic Palaeo-Siberians due to traditions of exogamy and multilingualism. This fits quite well with the different arrival of N (N1c and xN1c) lineages to the different Uralic-speaking groups, and to the stepped appearance of “Siberian ancestry” in the different regions.
It is evident that a lot of people were too attached to the idea of Palaeolithic R1b lineages ‘native’ to western Europe speaking Basque languages; of R1a lineages speaking Indo-European and spreading with Yamna; and N lineages ‘native’ to north-eastern Europe and speaking Uralic, and this is causing widespread weeping and gnashing of teeth (instead of the joy of discovering where one’s true patrilineal ancestors come from, and what language they spoke in each given period, which is the supposed objective of genetic genealogy…)
As far as I know – and there might be many other similar pet theories out there – there have been proposals of “modern Balto-Slavic-like” populations (in an obvious circular reasoning based on modern populations) in some Scythian clusters of the Iron Age.
NOTE. I will not enter into “Balto-Slavic-like R1a” of the Late Bronze Age or earlier because no one can seriously believe at this point of development of Population Genetics that autosomal similarity predating 1,500+ years the appearance of Slavs equates to their (ethnolinguistic) ancestral population, without a clear intermediate cultural and genetic trail – something we lack today in the Slavic case even for the late Roman period…
We also know of R1a-Z280 lineages in Srubna, probably expanding to the west. With that in mind, and knowing that Palaeo-Germanic was in close contact with Finno-Samic while both were already separated but still in contact, and that Palaeo-Germanic was also in contact and closely related to a ‘Temematic’ distinct from Balto-Slavic (and also that early Proto-Baltic and Proto-Slavic from the Roman Iron Age and later were in contact with western Uralic) this will be the linguistic map of the Iron Age if R1a is considered to expand Indo-European from some kind of “patron-client” relationship with west Yamna:
My problem with this proposal is that it is obviously beholden to the notion of the uninterrupted cultural, historic and ethnic continuity in certain territories. This bias is common in historiography (von Falkenhausen 1993), but it extends even more easily into the lesser known prehistory of any territory, and now more than ever some people feel the need to corrupt (pre)history based on their own haplogroups (or the majority haplogroups of their modern countries). However, more than on philosophical grounds, my rejection is based on facts: this picture is not what the combination of linguistic, archaeological, and genetic data shows. Period.
Nevertheless, if Yamna + Corded Ware represented the “big and early expansion” of Germanic and Italo-Celtic peoples proper of the dream Nazi’s Lebensraum and Fascist’s spazio vitale proposals; Uralians were Siberian hunter-gatherers that controlled the whole eastern and northern Russia, and miraculously managed to push (ethnolinguistically) Neolithic agropastoralists to the west during and after the Iron Age, with gradual (and often minimal) genetic impact; and Balto-Slavic peoples were represented by horse riders from Pokrovka/Srubna, hiding then somewhere around the forest-steppe until after the Scythian expansion, and then spreading their language (without much genetic impact) during the early Middle Ages…so be it.
Even though proposals of an Eastern Uralic (or Ugro-Samoyedic) group are in the minority – and those who support it tend to search for an origin of Uralic in Central Asia – , there is nothing wrong in supporting this from the point of view of a western homeland, because the eastward migration of both Proto-Ugric and Pre-Samoyedic peoples may have been coupled with each other at an early stage. It’s like Indo-Slavonic: it just doesn’t fit the linguistic data as well as the alternative, i.e. the expansion of Samoyedic first, different from a Finno-Ugric trunk. But, in case you are wondering about this possibility, here is Häkkinen’s (2012) phonological argument:
The case of Samoyedic is quite similar to that of Hungarian, although the earliest Palaeo-Siberian contact languages have been lost. There were contacts at least with Tocharian (Kallio 2004), Yukaghir (Rédei 1999) and Turkic (Janhunen 1998). Samoyedic also:
a) has moved far from the related languages and has been exposed to strong foreign influence
b) shares a small number of common words with other branches (from Sammallahti 1988: only 123 ‘Uralic’ words, versus 390 ‘Uralic’ + ‘Finno-Ugric’ words found in other branches than Samoyedic = 31,5 %)
c) derives phonologically from the East Uralic dialect.
The phonological level is taxonomically more reliable, since it lacks the distortion caused by invisible convergence and false divergence at the lexical level. Thus we can conclude that the traditional taxonomic model, according to which Samoyedic was the first branch to split off from the Proto-Uralic unity, is just as incorrect as the view that Hungarian was the first branch to split off.
Late Uralic can be traced back to metallurgical cultures thanks to terms like PU *wäśka ‘copper/bronze’ (borrowed from Proto-Samoyedic *wesä into Tocharian); PU *äsa and *olna/*olni, ‘lead’ or ‘tin’, found in *äsa-wäśka ‘tin-bronze’; and e.g. *weŋći ‘knife’, borrowed into Indo-Iranian (through the stage of vocalization of nasals), appearing later as Proto-Indo-Aryan *wāćī ‘knife, awl, axe’.
It is known that the southern regions of the Abashevo culture developed Proto-Indo-Iranian-speaking Sintashta-Petrovka and Pokrovka (Early Srubna). To the north, however, Abashevo kept its Uralic nature, with continuous contacts allowing for the spread of lexicon – mainly into Finno-Ugric – , and phonetic influence – mainly Uralisms into Proto-Indo-Iranian phonology (read more here).
The northern part of Abashevo (just like the south) was mainly a metallurgical society, with Abashevo metal prospectors found also side by side with Sintashta pioneers in the Zeravshan Valley, near BMAC, in search of metal ores. About the Seima-Turbino phenomenon, from Parpola (2013):
From the Urals to the east, the chain of cultures associated with this network consisted principally of the following: the Abashevo culture (extending from the Upper Don to the Mid- and South Trans-Urals, including the important cemeteries of Sejma and Turbino), the Sintashta culture (in the southeast Urals), the Petrovka culture (in the Tobol-Ishim steppe), the Taskovo-Loginovo cultures (on the Mid- and Lower Tobol and the Mid-Irtysh), the Samus’ culture (on the Upper Ob, with the important cemetery of Rostovka), the Krotovo culture (from the forest steppe of the Mid-Irtysh to the Baraba steppe on the Upper Ob, with the important cemetery of Sopka 2), the Elunino culture (on the Upper Ob just west of the Altai mountains) and the Okunevo culture (on the Mid-Yenissei, in the Minusinsk plain, Khakassia and northern Tuva). The Okunevo culture belongs wholly to the Early Bronze Age (c. 2250–1900 BCE), but most of the other cultures apparently to its latter part, being currently dated to the pre-Andronovo horizon of c. 2100–1800 BCE (cf. Parzinger 2006: 244–312 and 336; Koryakova & Epimakhov 2007: 104–105).
The majority of the Sejma-Turbino objects are of the better quality tin-bronze, and while tin is absent in the Urals, the Altai and Sayan mountains are an important source of both copper and tin. Tin is also available in southern Central Asia. Chernykh & Kuz’minykh have accordingly suggested an eastern origin for the Sejma-Turbino network, backing this hypothesis also by the depiction on the Sejma-Turbino knives of mountain sheep and horses characteristic of that area. However, Christian Carpelan has emphasized that the local Afanas’evo and Okunevo metallurgy of the Sayan-Altai area was initially rather primitive, and could not possibly have achieved the advanced and difficult technology of casting socketed spearheads as one piece around a blank. Carpelan points out that the first spearheads of this type appear in the Middle Bronze Age Caucasia c. 2000 BCE, diffusing early on to the Mid-Volga-Kama-southern Urals area, where “it was the experienced Abashevo craftsmen who were able to take up the new techniques and develop and distribute new types of spearheads” (Carpelan & Parpola 2001: 106, cf. 99–106, 110). The animal argument is countered by reference to a dagger from Sejma on the Oka river depicting an elk’s head, with earlier north European prototypes (Carpelan & Parpola 2001: 106–109). Also the metal analysis speaks for the Abashevo origin of the Sejma-Turbino network. Out of 353 artefacts analyzed, 47% were of tin-bronze, 36% of arsenical bronze, and 8.5% of pure copper. Both the arsenical bronze and pure copper are very clearly associated with the Abashevo metallurgy.
The Abashevo metal production was based on the Volga-Kama-Belaya area sandstone ores of pure copper and on the more easterly Urals deposits of arsenical copper (Figure 9). The Abashevo people, expanding from the Don and Mid-Volga to the Urals, first reached the westerly sandstone deposits of pure copper in the Volga and Kama basins, and started developing their metallurgy in this area, before moving on to the eastern side of the Urals to produce harder weapons and tools of arsenical copper. Eventually they moved even further south, to the area richest in copper in the whole Urals region, founding there the very strong and innovative Sintashta culture.
Regarding the most likely expansion of Eastern Uralic peoples:
Nataliya L’vovna Chlenova (1929–2009; cf. Korenyako & Ku’zminykh 2011) published in 1981 a detailed study of the Cherkaskul’ pottery. In her carefully prepared maps of 1981 and 1984 (Figure 10), she plotted Cherkaskul’ monuments not only in Bashkiria and the Trans-Urals, but also in thick concentrations on the Upper Irtysh, Upper Ob and Upper Yenissei, close to the Altai and Sayan mountains, precisely where the best experts suppose the homeland of Proto-Samoyed to be.
The Cherkaskul’ culture was transformed into the genetically related Mezhovka culture (c. 1500–1000 BCE), which occupied approximately the same area from the Mid-Kama and Belaya rivers to the Tobol river in western Siberia (cf. Parzinger 2006: 444–448; Koryakova & Epimakhov 2007: 170–175). The Mezhovka culture was in close contact with the neighbouring and probably Proto-Iranian speaking Alekseevka alias Sargary culture (c. 1500–900 BCE) of northern Kazakhstan (Figure 4 no. 8) that had a Fëdorovo and Cherkaskul’ substratum and a roller pottery superstratum (cf. Parzinger 2006: 443–448; Koryakova & Epimakhov 2007: 161–170). Both the Cherkaskul’ and the Mezhovka cultures are thought to have been Proto-Ugric linguistically, on the basis of the agreement of their area with that of Mansi and Khanty speakers, who moreover in their Fëdorovo-like ornamentation have preserved evidence of continuity in material culture (cf. Chlenova 1984; Koryakova & Epimakhov 2007: 159, 175).
The Mezhovka culture was succeeded by the genetically related Gamayun culture (c. 1000–700 BCE) (cf. Parzinger 2006: 446; 542–545).
From the Gamayun culture descend Trans-Urals cultures in close contact with Finno-Permic populations of the Cis-Ural region:
[Proto-Mansi] Itkul’ culture (c. 700–200 BCE) distributed along the eastern slope of the Ural Mountains (cf. Parzinger 2006: 552–556). Known from its walled forts, it constituted the principal Trans-Uralian centre of metallurgy in the Iron Age, and was in contact with both the Anan’ino and Akhmylovo cultures (the metallurgical centres of the Mid-Volga and Kama-Belaya region) and the neighbouring Gorokhovo culture.
[Proto-Hungarian] via the Vorob’evo Group (c. 700–550 BCE) (cf. Parzinger 2006: 546–549), to the Gorokhovo culture (c. 550–400 BCE) of the Trans-Uralian forest steppe (cf. Parzinger 2006: 549–552). For various reasons the local Gorokhovo people started mobile pastoral herding and became part of the multicomponent pastoralist Sargat culture (c. 500 BCE to 300 CE), which in a broader sense comprized all cultural groups between the Tobol and Irtysh rivers, succeeding here the Sargary culture. The Sargat intercommunity was dominated by steppe nomads belonging to the Iranian-speaking Saka confederation, who in the summer migrated northwards to the forest steppe
[Proto-Khanty] Late Bronze Age and Early Iron Age cultures related to the Gamayunskoe and Itkul’ cultures that extended up to the Ob: the Nosilovo, Baitovo, Late Irmen’, and Krasnoozero cultures (c. 900–500 BCE). Some were in contact with the Akhmylovo on the Mid-Volga.
Parpola (2012) connects the expansion of Samoyedic with the Cherkaskul variant of Andronovo. As we know, Andronovo was genetically diverse, which speaks in favour of different groups developing similar material cultures in Central Asia.
Juha Janhunen, author of the etymological dictionary of the Samoyed languages (1977), places the homeland of Proto-Samoyedic in the Minusinsk basin on the Upper Yenissei (cf. Janhunen 2009: 72). Mainly on the basis of Bulghar Turkic loanwords, Janhunen (2007: 224; 2009: 63) dates Proto-Samoyedic to the last centuries BCE. Janhunen thinks that the language of the Tagar culture (c. 800–100 BCE) ought to have been Proto-Samoyedic (cf. Janhunen 1983: 117– 118; 2009: 72; Parzinger 2001: 80 and 2006: 619–631 dates the Tagar culture c. 1000–200 BCE; Svyatko et al. 2009: 256, based on human bone samples, c. 900 BCE to 50 CE). The Tagar culture largely continues the traditions of the Karasuk culture (c. 1400–900 BCE), (…)
The use of a map of “Siberian ancestry” peaking in the arctic to show a supposedly late Uralic population movement (starting in the Iron Age!) seems to be the latest trend in population genomics:
I guess that would make this map of Neolithic farmer ancestry represent an expansion of Indo-European from the south, because Anatolia, Greece, Italy, southern France, and Iberia – where this ancestry peaks in modern populations – are among the oldest territories where Indo-European languages were recorded:
Probably not the right interpretation of this kind of simplistic data about modern populations, though…
Overall, and specifically at lower values of K, the genetic makeup of Uralic speakers resembles that of their geographic neighbours. The Saami and (a subset of) the Mansi serve as exceptions to that pattern being more similar to geographically more distant populations (Fig. 3a, Additional file 3: S3). However, starting from K = 9, ADMIXTURE identifies a genetic component (k9, magenta in Fig. 3a, Additional file 3: S3), which is predominantly, although not exclusively, found in Uralic speakers. This component is also well visible on K = 10, which has the best cross-validation index among all tests (Additional file 3: S3B). The spatial distribution of this component (Fig. 3b) shows a frequency peak among Ob-Ugric and Samoyed speakers as well as among neighbouring Kets (Fig. 3a). The proportion of k9 decreases rapidly from West Siberia towards east, south and west, constituting on average 40% of the genetic ancestry of FU speakers in Volga-Ural region (VUR) and 20% in their Turkic-speaking neighbours (Bashkirs, Tatars, Chuvashes; Fig. 3a).
However, this ‘something’ that some people occasionally find in some Uralic populations is also common to other modern and ancient groups, and not so common in some other Uralic peoples. Simply put:
I already said this in the recent publication of Siberian samples, where a renamed and radiocarbon dated Finnish_IA clearly shows that Late Iron Age Saami (ca. 400 AD) had little “Siberian ancestry”, if any at all, representing the most likely Fennic (and Samic) ancestral components before their expansion into central and northern Finland, where they admixed with circum-polar peoples of asbestos ware cultures.
I will say that again and again, any time they report the so-called “Siberian ancestry” in Uralic samples, no matter how it is defined each time: it does not seem to be that special something people are looking for, but rather (at least in a great part) a quite old ancestral component forming an evident cline with EHG, whose best proximate source are Baikal_EN (and/or Devil’s Gate) at this moment, and thus also East European hunter-gatherers for Western Uralic peoples:
So either Samara_HG, Karelia_HG, and many other groups from eastern Europe all spoke Uralic according to this ADMIXTURE graphic (and the formation of steppe ancestry in the Volga-Ural region brought the Proto-Indo-European language to the steppes through the CHG/ANE expansion), or a great part of this “Siberian ancestry” found in modern Uralic-speaking populations is not what some people would like to think it is…
PCA clines can be looked for to represent expansions of ancient populations. Most recently, Flegontov et al. (2018) are attempting to do this with Asian populations:
For some Turkic groups in the Urals and the Altai regions and in the Volga basin, a different admixture model fits the data: the same West Eurasian source + Uralic- or Yeniseian-speaking Siberians. Thus, we have revealed an admixture cline between Scythians and the Iranian farmer genetic cluster, and two further clines connecting the former cline to distinct ancestry sources in Siberia. Interestingly, few Wusun-period individuals harbor substantial Uralic/Yeniseian-related Siberian ancestry, in contrast to preceding Scythians and later Turkic groups characterized by the Tungusic/Mongolic-related ancestry. It remains to be elucidated whether this genetic influx reflects contacts with the Xiongnu confederacy. We are currently assembling a collection of samples across the Eurasian steppe for a detailed genetic investigation of the Hunnic confederacies.
There are potential errors with this approach:
The main one is practical – does a modern cline represent an ancestral language? The answer is: sometimes. It depends on the anthropological context that we have, and especially on the precision of the PCA:
The ‘Europe’, ‘Middle East’, etc. clines of the above PCA do not represent one language, but many. For starters, the PCA includes too many (and modern) populations, its precision is useless for ethnolinguistic groups. Which is the right level? Again, it depends.
The other error is one of detail of the clines drawn (which, in turn, depends on the precision of the PCA). For example, we can draw two paralell lines (or even one line, as in Flegontov et al. above) in one PCA graphic, but we still don’t have the direction of expansion. How do we know if this supposed “Uralic-speaking cline” goes from one region to the other? For that level of detail, we should examine closely modern Uralic-speaking peoples and Circum-Arctic populations:
The real ancient Uralic cluster (drawn above in blue) is thus probably from a North-East European source (probably formed by Battle Axe / Fatyanovo-Balanovo / Abashevo) to the east into Siberian populations, and to the north into Laplandic populations (see below also on Mezhovska ancestry for the drawn ‘European cline’, which some may a priori wrongly assume to be quite late).
The fact that the three formed clines point to an admixture of CWC-related populations from North-Eastern Europe, and that variation is greater at the Palaeo-Laplandic and Palaeo-Siberian extremities compared to the CWC-related one, also supports this as the correct interpretation.
However, judging by the two main clines formed, one could be alternatively inclined to interpret that Palaeo-Laplandic and Palaeo-Siberian populations formed a huge ancestral “Uralic” ghost cluster in Siberia (spanning from the Palaeo-Laplandic to the Palaeo-Siberian one), and from there expanded Finno-Samic on one hand, and “Volga-Ugro-Samoyed” on the other. That poses different problems: an obvious linguistic and archaeological one – which I assume a lot of people do not really care about – , and a not-so-obvious genetic one (see below for ancient samples and for the expansion of haplogroup N).
Unlike this PCA with ancient samples, where Bell Beaker clines could be a rough approximation to the real sources for each population, and where a cluster spanning all three depicted Early Bronze Age clusters could give a rough proximate source of European Bell Beakers in Hungary (and where one can even distinguish the Y-DNA bottlenecks in the L23 trunk created by each cline) the PCA of modern Uralic populations is probably not suitable for a good estimate of the ancient situation, which may be found shifted up or down of the drawn “Uralic” cluster along East European groups.
After all, we already know that the Siberian cline shows probably as much an ancient admixture event – from the original Uralic expansion to the east with Corded Ware ancestry – as another more recent one – a westward migration of Siberian ancestry (or even more than one). While we know with more or less exactitude what happened with the Palaeo-Laplandic admixture by expanding Proto-Finno-Samic populations (see here), the Proto-Ugric and Pre-Samoyedic populations formed probably more than one cline during the different ancient migrations through central Asia.
Apparently, the Corded Ware expansion to the east was not marked by a huge change in ancestry. While the final version of Narasimhan et al. (2018) may show a little more detail about other forest-steppe Seima-Turbino/Andronovo-related migrations (and thus also Eastern Uralic peoples), we have already had enough information for quite some time to get a good idea.
Mezhovska‘s position is similar to the later Pre-Scythian and Scythian populations. There are some interesting details: apart from haplogroup R1a-Z280 (CTS1211+), there is one R1b-M269 (PF6494+), probably Z2103, and an outlier (out of three) in a similar position to the recently described central/southern Scythian clusters.
NOTE. The finding of R1b-M269 in the forest-steppe is probably either 1) from an Afanasevo-Okunevo origin, or 2) from an admixture with neighbouring Andronovo-related populations, such as Sargary. A third, maybe less likely option is that this haplogroup admixed with Abashevo directly (as it happened in Sintashta, Potapovka, or Pokrovka) and formed part of early Uralic migrations. In any case, since Mezhovska is a Bronze Age society from the Urals region, its association with R1b-Z2103 – like the association of R1b-Z2103 in Scythian clusters – cannot be attributed to “Thracian peoples”, a link which is (as I already said) too simplistic.
The drawn “European cline” of Hungarians (see above), leading from ‘west-like’ Mansi to Hungarian populations – and hosting also Finnic and Estonian samples – , cannot therefore be attributed simply to late “Slavic/Balkan-like” admixture.
Karasuk – located further to the east – is basically also Corded Ware peoples showing clearly a recent admixture with local ANE / Baikal_EN-like populations. In terms of haplogroups it shows haplogroup Q, R1a-Z2124, and R1a-Z2123, later found among early Hungarians, and present also in ancient Samoyedic populations now acculturated.
The most interesting aspect of both Mezhovska and Karasuk is that they seem to diverge from a point close to Ukraine_Eneolithic, which is the supposed ancestral source of Corded Ware peoples (read more about the formation of “steppe ancestry”). This means that Eastern Uralians derive from a source closer to Middle Dnieper/Abashevo populations, rather than Battle Axe (shifted to Latvian Neolithic), which is more likely the source prevalent in Finno-Permic peoples.
Their initial admixture with (Palaeo-)Siberian populations is thus seen already starting by this time in Mezhovska and especially in Karasuk, but this process (compared to modern populations) is incomplete:
We know now that Samic peoples expanded during the Late Iron Age into Palaeo-Laplandic populations, admixing with them and creating this modern cline. Finns expanded later to the north (in one of their known genetic bottlenecks), admixing with (and displacing) the Saami in Finland, especially replacing their male lines.
So how did Ugric and Samoyedic peoples admix with Palaeo-Siberian populations further, to obtain their modern cline? The answer is, logically, with East Asian migrations related to forest-steppe populations of Central Asia after the Mezhovska and Karasuk periods, i.e. during the Iron Age and later. Other groups from the forest-steppe in Central Asia show similar East Asian (“Siberian”) admixture. We know this from Narasimhan et al. (2018):
(…) we observe samples from multiple sites dated to 1700-1500 BCE (Maitan, Kairan, Oy_Dzhaylau and Zevakinsikiy) that derive up to ~25% of their ancestry from a source related to present-day East Asians and the remainder from Steppe_MLBA. A similar ancestry profile became widespread in the region by the Late Bronze Age, as documented by our time transect from Zevakinsikiy and samples from many sites dating to 1500-1000 BCE, and was ubiquitous by the Scytho-Sarmatian period in the Iron Age.
Flegontov: Present day Turkic speakers fall into two clusters of admixture patterns (Uralic/Yenisean and Tungussic/Mngolic) based on genomic data with ancient Turks belonging almost exclusively to the first cluster. #ISBA8
The Ugric-speaking Sargat culture in Western Siberia shows the expected mixture of haplogroups (ca. 500 BC – 500 AD), with 5 samples of hg N and 2 of hg R1a1, in Pilipenko et al. (2017). Although radiocarbon dates and subclades are lacking, N lineages probably spread late, because of the late and gradual admixture of Siberian cultures into the Sargat melting pot.
The observed reduction in the genetic distance between the Middle Tagar population and other Scythian like populations of Southern Siberia(Fig 5; S4 Table), in our opinion, is primarily associated with an increase in the role of East Eurasian mtDNA lineages in the gene pool (up to nearly half of the gene pool) and a substantial increase in the joint frequency of haplogroups C and D (from 8.7% in the Early Tagar series to 37.5% in the Middle Tagar series). These features are characteristic of many ancient and modern populations of Southern Siberia and adjacent regions of Central Asia, including the Pazyryk population of the Altai Mountains.
Before the Iron Age, the Karasuk and Mezhovska population were probably already somehow ‘to the north’ within the ancient Steppe-Altai cline (see image below9 created by expanding Seima-Turbino- and Andronovo-related populations. During the Iron Age, further Siberian contributions with Iranian expansions must have placed Uralians of the Central Asian forest-steppe areas much closer to today’s Palaeo-Siberian cline.
However, the modern genetic picture was probably fully developed only in historic times, when Samoyedic and Ugric languages expanded to the north, only in part admixing further with Palaeo-Siberian-speaking nomads from the Circum-Arctic region (see here for a recent history of Samoyedic Enets), which justifies their more recent radical ‘northern shift’.
This late acquisition of the language by Palaeo-Siberian nomads (without much population replacement) also justifies the wide PCA clusters of very small Siberian populations. See for example in the PCA from Tambets et al. (2018):
For their relationship with modern Mansi, we have information on Hungarian conqueror populations from Neparáczki et al. (2018):
Moreover, Y, B and N1a1a1a1a Hg-s have not been detected in Finno-Ugric populations [80–84], implying that the east Eurasian component of the Conquerors and Finno-Ugric people are probably not directly related. The same inference can be drawn from phylogenetic data, as only two Mansi samples appeared in our phylogenetic trees on the side branches (S1 Fig, Networks; 1, 4) suggesting that ancestors of the Mansis separated from Asian ancestors of the Conquerors a long time ago. This inference is also supported by genomic Admixture analysis of Siberian and Northeastern European populations , which revealed that Mansis received their eastern Siberian genetic component approximately 5–7 thousand years ago from ancestors of modern Even and Evenki people. Most likely the same explanation applies to the Y-chromosome N-Tat marker which originated from China [86,87] and its subclades are now widespread between various language groups of North Asia and Eastern Europe .
The genetic picture of Hungarians (their formed cline with Mansi and their haplogroups) may be quite useful for the true admixture found originally in Mansi peoples at the beginning of the Iron Age. By now it is clear even from modern populations that Steppe_MLBA ancestry accompanied the Uralic expansion to the east (roughly approximated in the graphic with Afanasievo_EBA + Bichon_LP EasternHG_M):
An interesting aspect of the paper, hidden among so many relevant details, is a clearer picture of how the so-called Yamnaya or steppe ancestry evolved from Samara hunter-gatherers to Yamna nomadic pastoralists, and how this ancestry appeared among Proto-Corded Ware populations.
Please note: arrows of “ancestry movement” in the following PCAs do not necessarily represent physical population movements, or even ethnolinguistic change. To avoid misinterpretations, I have depicted arrows with Y-DNA haplogroup migrations to represent the most likely true ethnolinguistic movements. Admixture graphics shown are from Wang et al. (2018), and also (the K12) from Mathieson et al. (2018).
1. Samara to Early Khvalynsk
The so-called steppe ancestry was born during the Khvalynsk expansion through the steppes, probably through exogamy of expanding elite clans (eventually all R1b-M269 lineages) originally of Samara_HG ancestry. The nearest group to the ANE-like ghost population with which Samara hunter-gatherers admixed is represented by the Steppe_Eneolithic / Steppe_Maykop cluster (from the Northern Caucasus Piedmont).
Steppe_Eneolithic samples, of R1b1 lineages, are probably expanded Khvalynsk peoples, showing thus a proximate ancestry of an Early Eneolithic ghost population of the Northern Caucasus. Steppe_Maykop samples represent a later replacement of this Steppe_Eneolithic population – and/or a similar population with further contribution of ANE-like ancestry – in the area some 1,000 years later.
This is what Steppe_Maykop looks like, different from Steppe_Eneolithic:
NOTE. This admixture shows how different Steppe_Maykop is from Steppe_Eneolithic, but in the different supervised ADMIXTURE graphics below Maykop_Eneolithic is roughly equivalent to Eneolithic_Steppe (see orange arrow in ADMIXTURE graphic above). This is useful for a simplified analysis, but actual differences between Khvalynsk, Sredni Stog, Afanasevo, Yamna and Corded Ware are probably underestimated in the analyses below, and will become clearer in the future when more ancestral hunter-gatherer populations are added to the analysis.
2. Early Khvalynsk expansion
We have direct data of Khvalynsk-Novodanilovka-like populations thanks to Khvalynsk and Steppe_Eneolithic samples (although I’ve used the latter above to represent the ghost Caucasus population with which Samara_HG admixed).
We also have indirect data. First, there is the PCA with outliers:
Second, we have data from north Pontic Ukraine_Eneolithic samples (see next section).
Third, there is the continuity of late Repin / Afanasevo with Steppe_Eneolithic (see below).
3. Proto-Corded Ware expansion
It is unclear if R1a-M459 subclades were continuously in the steppe and resurged after the Khvalynsk expansion, or (the most likely option) they came from the forested region of the Upper Dnieper area, possibly from previous expansions there with hunter-gatherer pottery.
Supporting the latter is the millennia-long continuity of R1b-V88 and I2a2 subclades in the north Pontic Mesolithic, Neolithic, and Early Eneolithic Sredni Stog culture, until ca. 4500 BC (and even later, during the second half).
Only at the end of the Early Eneolithic with the disappearance of Novodanilovka (and beginning of the steppe ‘hiatus’ of Rassamakin) is R1a to be found in Ukraine again (after disappearing from the record some 2,000 years earlier), related to complex population movements in the north Pontic area.
NOTE. In the PCA, a tentative position of Novodanilovka closer to Anatolia_Neolithic / Dzudzuana ancestry is selected, based on the apparent cline formed by Ukraine_Eneolithic samples, and on the position and ancestry of Sredni Stog, Yamna, and Corded Ware later. A good alternative would be to place Novodanilovka still closer to the Balkan outliers (i.e. Suvorovo), and a source closer to EHG as the ancestry driven by the migration of R1a-M417.
The first sample with steppe ancestry appears only after 4250 BC in the forest-steppe, centuries after the samples with steppe ancestry from the Northern Caucasus and the Balkans, which points to exogamy of expanding R1a-M417 lineages with the remnants of the Novodanilovka population.
4. Repin / Early Yamna expansion
We don’t have direct data on early Repin settlers. But we do have a very close representative: Afanasevo, a population we know comes directly from the Repin/late Khvalynsk expansion ca. 3500/3300 BC (just before the emergence of Early Yamna), and which shows fully Steppe_Eneolithic-like ancestry.
Compared to this eastern Repin expansion that gave Afanasevo, the late Repin expansion to the west ca. 3300 BC that gave rise to the Yamna culture was one of colonization, evidenced by the admixture with north Pontic (Sredni Stog-like) populations, no doubt through exogamy:
This admixture is also found (in lesser proportion) in east Yamna groups, which supports the high mobility and exogamy practices among western and eastern Yamna clans, not only with locals:
We don’t have a comparison with Ukraine_Eneolithic or Corded Ware samples in Wang et al. (2018), but we do have proximate sources for Abashevo, when compared to the Poltavka population (with which it admixed in the Volga-Ural steppes): Sintashta, Potapovka, Srubna (with further Abashevo contribution), and Andronovo:
The two CWC outliers from the Baltic show what I thought was an admixture with Yamna. However, given the previous mixture of Eneolithic_Steppe in north Pontic steppe-forest populations, this elevated “steppe ancestry” found in Baltic_LN (similar to west Yamna) seems rather an admixture of Baltic sub-Neolithic peoples with a north Pontic Eneolithic_Steppe-like population. Late Repin settlers also admixed with a similar population during its colonization of the north Pontic area, hence the Baltic_LN – west Yamna similarities.
NOTE. A direct admixture with west Yamna populations through exogamy by the ancestors of this Baltic population cannot be ruled out yet (without direct access to more samples), though, because of the contacts of Corded Ware with west Yamna settlers in the forest-steppe regions.
A similar case is found in the Yamna outlier from Mednikarovo south of the Danube. It would be absurd to think that Yamna from the Balkans comes from Corded Ware (or vice versa), just because the former is closer in the PCA to the latter than other Yamna samples. The same error is also found e.g. in the Corded Ware → Bell Beaker theory, because of their proximity in the PCA and their shared “steppe ancestry”. All those theories have been proven already wrong.
NOTE. A similar fallacy is found in potential Sintashta→Mycenaean connections, where we should distinguish statistically that result from an East/West Yamna + Balkans_BA admixture. In fact, genetic links of Mycenaeans with west Yamna settlers prove this (there are some related analyses in Anthrogenica, but the site is down at this moment). To try to relate these two populations (separated more than 1,000 years before Sintashta) is like comparing ancient populations to modern ones, without the intermediate samples to trace the real anthropological trail of what is found…Pure numbers and wishful thinking.
Most mtDNA lineages found are characteristic of the early Neolithic farmers in south-eastern and central Europe of the Starčevo-Kőrös-Criş and LBK cultures. Haplogroups N1a, T2, J, K, and V, which are found in the Neolithic BKG, TRB, GAC and Early Bronze Age samples, are part of the mitochondrial ‘Neolithic package’ (which also includes haplogroups HV, V, and W) that was introduced to Europe with farmers migrating from Anatolia at the onset of the Neolithic17,31.
A noteworthy proportion of Mesolithic haplogroup U5 is also found among the individuals of the current study. The proportion of haplogroup U5 already present in the earliest of the analysed Neolithic groups from the examined area differs from the expected pattern of diversity of mtDNA lineages based on a previous archaeological view and on the aDNA findings from the neighbouring regions which were settled by post-Linear farmers similar to BKG at that time. A large proportion of Mesolithic haplogroups in late-Danubian farmers in Kuyavia was also shown in previous studies concerning BKG samples based on mtDNA only, although these frequencies were derived on the basis of very small sample sizes.
A significant genetic influence of HG populations persisted in this region at least until the Eneolithic/Early Bronze Age period, when steppe migrants arrived to central Europe. The presence of two outliers from the middle and late phases of the BKG in Kuyavia associated with typical Neolithic burial contexts provides evidence that hunter-farmer contacts were not restricted to the final period of this culture and were marked by various episodes of interaction between two societies with distinct cultural and subsistence differences.
The identification of both mitochondrial and Y-chromosome haplogroup lineages of Mesolithic provenance (U5 and I, respectively) in the BKG support the theory that both male and female hunter-gatherers became part of these Neolithic agricultural societies, as has been reported for similar cases from the Carpathian Basin, and the Balkans. The identification of an individual with WHG affinity, dated to ca. 4300 BCE, in a Middle Neolithic context within a BKG settlement, provides direct evidence for the regional existence of HG enclaves that persisted and coexisted at least for over 1000 years, from the arrival of the LBK farmers ca. 5400 BCE until ca. 4300 BCE, in proximity with Neolithic settlements, but without admixing with their inhabitants.
The analysis of two Late Neolithic cultures, the GAC and CWC, shows that steppe ancestry was present only among the CWC individuals analysed, and that the single GAC individual had more WHG ancestry than previous local Neolithic individuals. (…) The CWC’s affinity to WHG, however, contrasts with results from published CWC individuals that identified steppe ancestry related to Yamnaya as the major contributor to the CWC genomes, while here we report also substantial contributions from WHG that could relate to the late persistence of pockets of WHG populations, as supported by the admixture results of N42 and the finding of the 4300-year-old N22 HG individual. These results agree with archaeological theories that suggest that the CWC interaction with incoming steppe cultures was complex and that it varied by region.
About the analyzed CWC samples, it is remarkable that, even though they are somehow related to each other, they do not form a tight cluster. Also, their Y-DNA (I2a), and this:
When compared to previously published CWC data, our CWC group (not individuals) is genetically significantly closer to WHG than to steppe individuals (Z = −4.898), a result which is in contrast with those for CWC from Germany (Z = 2.336), Estonia (Z = 0.555), and Latvia (Z = 1.553).
Włodarczak (2017) talks about the CWC period in Poland after ca. 2600 BC as a time of emergence of an allochthnous population, marked by the rare graves of this area, showing infiltrations initially mainly from Lesser Poland, and later (after 2500 BC) from the western Baltic zone.
Since forest sub-Neolithic populations would have probably given more EHG to the typical CWC population, these samples support the resurge of ‘local’ pockets of GAC- or TRB-like groups with more WHG (and also Levant_Neolithic) ancestry.
The known presence of I2a2a1b lineages in GAC groups in Poland also supports this interpretation, and the subsistence of such pockets of pre-steppe-like populations is also seen with the same or similar lineages appearing in comparable ‘resurge’ events in Central Europe, e.g. in samples from the Únětice and Tumulus culture.
About the Bronze Age sample, we have at last official confirmation of haplogroup R1a1a (sadly no subclade*) at the very beginning of the Trzciniec period – in a region between western (Iwno) and eastern (Strzyżów) groups related to Mierzanowice – , which has to be put in relation with the samples from the final Trzciniec period in the Baltic published in Mittnik et al. (2018).
EDIT (8 OCT 2018): More specific subclades have been published, including a R1a-Z280 lineage for the Bronze Age sample (see spreadsheet).
This confirms the early resurge of R1a-Z645 (probably R1a-Z282) lineages at the core of the developing East European Bronze Age, a province of the European Bronze Age that emerged from evolving Bell Beaker groups in Poland.
I don’t have any hope that the Balto-Slavic evolution through BBC Poland → Mierzanowice/Iwno → Trzciniec → Lusatian cultures is going to be confirmed any time soon, until we have a complete trail of samples to follow all the way to historic Slavs of the Prague culture. However, I do think that the current data on central-east Europe – and the recent data we are receiving from north-east Europe and the Iranian steppes, at odds with the Indo-Slavonic alternative – supports this model.
I guess that, in the end, similar to how the Yamna vs. Corded Ware question is being solved, the real route of expansion of Proto-Balto-Slavic (supposedly spoken ca. 1500-1000 BC) is probably going to be decided by the expansion of either R1a-M458 (from the west) or R1a-Z280 lineages (from the east), because the limited precision of genetic data and analyses available today are going to show ‘modern Slavic’-like populations from the whole eastern half of Europe for the past 4,000 years…
I was reading The Bronze Age Landscape in the Russian Steppes: The Samara Valley Project (2016), and I was really surprised to find the following excerpt by David W. Anthony:
The Samara Valley links the central steppes with the western steppes and is a north-south ecotone between the pastoral steppes to the south and the forest-steppe zone to the north [see figure below]. The economic contrast between pastoral steppe subsistence, with its associated social organizations, and forest-zone hunting and fishing economies probably explains the shifting but persistent linguistic border between forest-zone Uralic languages to the north (today largely displaced by Russian) and a sequence of steppe languages to the south, recently Turkic, before that Iranian, and before that probably an eastern dialect of Proto-Indo-European (Anthony 2007). The Samara Valley represents several kinds of borders, linguistic, cultural, and ecological, and it is centrally located in the Eurasian steppes, making it a critical place to examine the development of Eurasian steppe pastoralism.
Khokhlov (translated by Anthony) further insists on the racial and ethnic divide between both populations, Abashevo to the north, and Poltavka to the south, during the formation of the Abashevo – Sintashta-Potapovka community that gave rise to Proto-Indo-Iranians:
Among all cranial series in the Volga-Ural region, the Potapovka population represents the clearest example of race mixing and probably ethnic mixing as well. The cultural advancements seen in this period might perhaps have been the result of the mixing of heterogeneous groups. Such a craniometric observation is to some extent consistent with the view of some archaeologists that the Sintashta monuments represent a combination of various cultures (principally Abashevo and Poltavka, but with other influences) and therefore do not correspond to the basic concept of an archaeological culture (Kuzmina 2003:76). Under this option, the Potapovka-Sintashta burial rite may be considered, first, a combination of traits to guarantee the afterlife of a selected part of a heterogeneous population. Second, it reflected a kind of social “caste” rather than a single population. In our view, the decisive element in shaping the ethnic structure of the Potapovka-Sintashta monuments was their extensive mobility over a fairly large geographic area. They obtained knowledge of various cultures from the populations with whom they interacted.
Interesting is also this excerpt about the predominant population in the Abashevo – Sintashta-Potapovka admixture (which supports what Chetan said recently, although this does not seemed backed by Y-DNA haplogroups found in the richest burials), coupled with the sign of incoming “Uraloid” peoples from the east, found in both Sintashta and eastern Abashevo:
The socially dominant anthropological component was Europeoid, possibly the descendants of Yamnaya. The association of craniofacial types with archaeological cultures in this period is difficult, primarily because of the small amount of published anthropological material of the cultures of steppe and forest belt (Balanbash, Vol’sko-Lbishche) and the eastern and southern steppes (Botai-Tersek). The crania associated with late MBA western Abashevo groups in the Don-Volga forest zone were different from eastern Abashevo in the Urals, where the expression of the Old Uraloid craniological complex was increased. Old Uraloid is found also on a single skull of Vol’sko-Lbishche culture (Tamar Utkul VII, Kurgan 4). Potentially related variants, including Mongoloid features, could be found among the Seima-Turbino tribes of the forest-steppe zone, who mixed with Sintashta and Abashevo. In the Sintashta Bulanova cemetery from the western Urals, some individuals were buried with implements of Seima-Turbino type (Khalyapin 2001; Khokhlov 2009; Khokhlov and Kitov 2009). Previously, similarities were noted between some individual skulls from Potapovka I and burials of the much older Botai culture in northern Kazakhstan (Khokhlov 2000a). Botai-Tersek is, in fact, a growing contender for the source of some “eastern” cranial features.
The wave of peoples associated with “eastern” features can be seen in genetics in the Sintashta outliers from Narasimhan et al. (2018), and it probably will be eventually seen in Abashevo, too. These may be related to the Seima-Turbino international network – but most likely it is directly connected to Sintashta through the starting Andronovo and Seima-Turbino horizons, by admixing of prospective groups and small-scale back-migrations.
Corded Ware – Yamna similarities?
So, if peoples of north-eastern Europe have been assumed for a long time to be Uralic speakers, what is happening with the Corded Ware = IE obsession? Is it Gimbutas’ ghost possessing old archaeologists? Probably not.
It is about certain cultural similarities evident at first sight, which have been traditionally interpreted as a sign of cultural diffusion or migration. Not dissimilar to the many Bell Beaker models available, where each archaeologist is pushing certain differences, mixing what seemed reasonable, what still might seem reasonable, and what certainly isn’t anymore after the latest ancient DNA data.
The initial models of Gimbutas, Kristiansen, or Anthony – which are known to many today – were enunciated in the infancy of archaeological studies in the regions, during and just after the fall of the USSR, and before many radiocarbon dates that we have today were published (with radiocarbon dating being still today in need of refinement), so it is only logical that gross mistakes were made.
We have similar gross mistakes related to the origins of Bell Beakers, and studying them was certainly easier than studying eastern data.
Gimbutas believed – based mainly on Kurgan-like burials – that Bell Beaker formed from a combination of Yamna settlers with the Vučedol culture, so she was not that far from the truth.
The expansion of Corded Ware from peoples of the North Pontic forest-steppe area, proposed by Gimbutas and later supported also by Kristiansen (1989) as the main Indo-European expansion – , is probably also right about the approximate origins of the culture. Only its ‘Indo-European’ nature is in question, given the differences with Khvalynsk and Yamna evolution.
Anthony only claimed that Yamna migrants settled in the Balkans and along the Danube into the Hungarian steppes. He never said that Corded Ware was a Yamna offshoot until after the first genetic papers of 2015 (read about his newest proposal). He initially claimed that only certain neighbouring Corded Ware groups “adopted” Indo-European (through cultural diffusion) because of ‘patron-client’ relationships, and was never preoccupied with the fate of Corded Ware and related cultures in the east European forest zone and Finland.
So none of them was really that far from the true picture; we might say a lot people are more way off the real picture today than the picture these three researchers helped create in the 1990s and 2000s. Genetics is just putting the last nail in the coffin of Corded Ware as a Yamna offshoot, instead of – as we believed in the 2000s – to Vučedol and Bell Beaker.
So let’s revise some of these traditional links between Corded Ware and Yamna with today’s data:
Even more than genetics – at least until we have an adequate regional and temporary sampling – , archaeological findings lead what we have to know about both cultures.
It is essential to remember that Corded Ware, starting ca. 3000/2900 BC in east-central Europe, has been proposed to be derived from Early Yamna, which appeared suddenly in the Pontic-Caspian steppes ca. 3300 BC (probably from the late Repin expansion), and expanded to the west ca. 3000.
The question at hand, therefore, is if Corded Ware can be considered an offshoot of the Late PIE community, and thus whether the CWC ethnolinguistic community – proven in genetics to be quite homogeneous – spoke a Late PIE dialect, or if – alternatively – it is derived from other neighbouring cultures of the North Pontic region.
NOTE. The interpretation of an Indo-Slavonic group represented by a previous branching off of the group is untenable with today’s data, since Indo-Slavonic – for those who support it – would itself be a branch of Graeco-Aryan, and Palaeo-Balkan languages expanded most likely with West Yamna (i.e. R1b-L23, mainly R1b-Z2103) to the south.
The convoluted alternative explanation would be that Corded Ware represents an earlier, Middle PIE branch (somehow carrying R1a??) which influences expanding Late PIE dialects; this has been recently supported by Kortlandt, although this simplistic picture also fails to explain the Uralic problem.
❔ Kurgans: The Yamna tradition was inherited from late Repin, in turn inherited from Khvalynsk-Novodanilovka proto-Kurgans. As for the CWC tradition, it is unclear if the tumuli were built as a tradition inherited from North and West Pontic cultures (in turn inherited or copied from Khvalynsk-Novodanilovka), such as late Trypillia, late Kvityana, late Dereivka, late Sredni Stog; or if they were built because of the spread of the ‘Transformation of Europe’, set in motion by the Early Yamna expansion ca. 3300-3000 BC (as found in east-central European cultures like Coţofeni, Lizevile, Șoimuș, or the Adriatic Vučedol). My guess is that it inherits an older tradition than Yamna, with an origin in east-central Europe, because of the mound-building distribution in the North Pontic area before the Yamna expansion, but we may never really know.
❌ Burial rite: Yamna features (with regional differences) single burials with body on its back, flexed upright knees, poor grave goods, common orientation east-west (heads to the west) inherited from Repin, in turn inherited from Khvalynsk-Novodanilovka. CWC tradition – partially connected to Złota and surrounding east-central European territories (in turn from the Khvalynsk-Novodanilovka expansion) – features single graves, body in fetal position, strict gender differentiation – men on the right, women on the left -, looking to the south, graves with standardized assemblages (objects representing affirmation of battle, hunting, and feasting). The burial rites clearly represent different ideologies.
❌ Corded decoration: Corded ware decoration appears in the Balkans during the 5th millennium, and represents a simple technique whereby a cord is twisted, or wrapped around a stick, and then pressed directly onto the fresh surface of a vessel leaving a characteristic decoration. It appears in many groups of the 5th and 4th millennium BC, but it was Globular Amphorae the culture which popularized the drinking vessels and their corded ornamentation. It appears thus in some regional groups of Yamna, but it becomes the standard pottery only in Corded Ware (especially with the A-horizon), which shows continuity with GAC pottery.
❌ Economy: Yamna expands from Repin (and Repin from Khvalynsk-Novodanilovka) as a nomadic or semi-nomadic purely pastoralist society (with occasional gathering of wild seeds), which naturally thrives in the grasslands of the Pontic-Caspian, lower Danube and Hungarian steppes. Corded Ware shows agropastoralism (as late Eneolithic forest-steppe and steppe groups of eastern Europe, such as late Trypillian, TRB, and GAC groups), inhabits territories north of the loess line, with heavy reliance of hunter-gathering depending on the specific region.
❌ Cattle herding: Interestingly, both west Yamna and Corded Ware show more reliance on cattle herding than other pastoralist groups, which – contrasted with the previous Eneolithic herding traditions of the Pontic-Caspian steppe, where sheep-goats predominate – make them look alike. However, the cattle-herding economy of Yamna is essential for its development from late Repin and its expansion through the steppes (over western territories practising more hunter-gathering and sheep-goat herding economy), and it does not reach equally the Volga-Ural region, whose groups keep some of the old subsistence economy (read more about the late Repin expansion). Corded Ware, on the other hand, inherits its economic strategy from east European groups like TRB, GAC, and especially late Trypillian communities, showing a predominance of cattle herding within an agropastoral community in the forest-steppe and forest zones of Volhynia, Podolia, and surrounding forest-steppe and forest regions.
❔ Horse riding: Horse riding and horse transport is proven in Yamna (and succeeding Bell Beaker and Sintashta), assumed for late Repin (essential for cattle herding in the seas of grasslands that are the steppes, without nearby water sources), quite likely during the Khvalynsk expansion (read more here), and potentially also for Samara, where the predominant horse symbolism of early Khvalynsk starts. Corded Ware – like the north Pontic forest-steppe and forest areas during the Eneolithic – , on the other hand, does not show a strong reliance on horse riding. The high mobility and short-term settlements characteristic of Corded Ware, that are often associated with horse riding by association with Yamna, may or may not be correct, but there is no need for horses to explain their herding economy or their mobility, and the north-eastern European areas – the one which survived after Bell Beaker expansion – did certainly not rely on horses as an essential part of their economy.
NOTE: I cannot think of more supposed similarities right now. If you have more ideas, please share in the comments and I will add them here.
✅ EHG: This is the clearest link between both communities. We thought it was related to the expansion of ANE-related ancestry to the west into WHG territory, but now it seems that it will be rather WHG expanding into ANE territory from the Pontic-Caspian region to the east (read more on recent Caucasus Neolithic, on , and on Caucasus HG).
NOTE. Given how much each paper changes what we know about the Palaeolithic, the origin and expansion of the (always developing) known ancestral components and specific subclades (see below) is not clear at all.
❔ CHG: This is the key link between both cultures, which will delimit their interaction in terms of time and space. CHG is intermediate between EHG and Iran N (ca. 8000 BC). The ancestry is thus linked to the Caucasus south of the steppe before the emergence of North Pontic (western) and Don-Volga-Ural (eastern) communities during the Mesolithic. The real question is: when we have more samples from the steppe and the Caucasus during the Neolithic, how many CHG groups are we going to find? Will the new specific ancestral components (say CHG1, CHG2, CHG3, etc.) found in Yamna (from Khvalynsk, in the east) and Corded Ware (probably from the North Pontic forest-steppe) be the same? My guess is, most likely not, unless they are mediated by the Khvalynsk-Novodanilovka expansion (read more on CHG in the Caucasus).
❌ WHG/EEF: This is the obvious major difference – known today – in the formation of both communities in the steppe, and shows the different contacts that both groups had at least since the Eneolithic, i.e. since the expansion of Repin with its renewed Y-DNA bottleneck, and probably since before the early Khvalynsk expansion (read more on Yamna-Corded Ware differences contrasting with Yamna-Afanasevo, Yamna-Bell Beaker, and Yamna-Sintashta similarities).
NOTE 1. Some similarities between groups can be seen depending on the sampled region; e.g. Baltic groups show more similarities with southern Pontic-Caspian steppe populations, probably due to exogamy.
NOTE 2. We have this information on the differences in “steppe ancestry” between Yamna and Corded Ware, compared to previous studies, because now we have more samples of neighbouring, roughly contemporaneous Eneolithic groups, to analyse the real admixture processes. This kind of fine scale studies is what is going to show more and more differences between Khvalynsk-Yamna and Sredni Stog-Corded Ware as more data pours in. The evolution of both communities in archaeology and in PCA (see below) is probably witness to those differences yet to be published.
❌ R1: Even though some people try very hard to think in terms of “R1” vs. (Caucasus) J or G or any other upper clade, this is plainly wrong. It is possible, given what we know now, that Q1a2-M242 expanded ANE ancestry to the west ca. 13000 BC, while R1b-P279 expanded WHG ancestry to the east with the expansion of post-Swiderian cultures, creating EHG as a WHG:ANE cline. The role of R1a-M459 is unknown, but it might be related to any of these migrations, or others (plural) along northern Eurasia (read more on the expansion of R1b-P279, on Palaeolithic Q1a2, and on R1a-M417).
NOTE. I am inclined to believe in a speculative Mesolithic-Early Neolithic community involving Eurasiatic movements accross North Eurasia, and Indo-Uralic movements in its western part, with the last intense early Uralic-PIE contacts represented by the forming west (Mariupol culture) and east (Don-Volga-Ural cultures, including Samara) communities developing side by side. Before their known Eneolithic expansions, no large-scale Y-DNA bottleneck is going to be seen in the Pontic-Caspian steppe, with different (especially R1a and R1b subclades) mixed among them, as shown in North Pontic Neolithic, Samara HG, and Khvalynsk samples.
Corded Ware and ‘steppe ancestry’
If we take a look at the evolution of Corded Ware cultures, the expansion of Bell Beakers – dominated over most previous European cultures from west to east Europe – influenced the development of the whole European Bronze Age, up to Mierzanowice and Trzciniec in the east.
The only relevant unscathed CWC-derived groups, after the expansion of Sintashta-Potapovka as the Srubna-Andronovo horizon in the Eurasian steppes, were those of the north-eastern European forest zone: between Belarus to the west, Finland to the north, the Urals to the east, and the forest-steppe region to the south. That is, precisely the region supposed to represent Uralic speakers during the Bronze Age.
This inconsistency of steppe ancestry and its relation with Uralic (and Balto-Slavic) peoples was observed shortly after the publication of the first famous 2015 papers by Paul Heggarty, of the Max-Planck Institute for Evolutionary Anthropology (read more):
Haak et al. (2015) make much of the high Yamnaya ancestry scores for (only some!) Indo-European languages. What they do not mention is that those same results also include speakers of other languages among those with the highest of all scores for Yamnaya ancestry. Only these are languages of the Uralic family, not Indo-European at all; and their Yamnaya-ancestry signals are far higher than in many branches of Indo-European in (southern) Europe. Estonian ranks very high, while speakers of the very closely related Finnish are curiously not shown, and nor are the Saami. Hungarian is relevant less directly since this language arrived only c. 900 AD, but also high.
These data imply that Uralic-speakers too would have been part of the Yamnaya > Corded Ware movement, which was thus not exclusively Indo-European in any case. And as well as the genetics, the geography, chronology and language contact evidence also all fit with a Yamnaya > Corded Ware movement including Uralic as well as Balto-Slavic.
Both papers fail to address properly the question of the Uralic languages. And this despite — or because? — the only Uralic speakers they report rank so high among modern populations with Yamnaya ancestry. Their linguistic ancestors also have a good claim to have been involved in the Corded Ware and Yamnaya cultures, and of course the other members of the Uralic family are scattered across European Russia up to the Urals.
NOTE. Although the author was trying to support the Anatolian hypothesis – proper of glottochronological studies often published from the Max Planck Institute – , the question remains equally valid: “if Proto-Indo-European expands with Corded Ware and steppe ancestry, what is happening with Uralic peoples?”
For my part, I claimed in my draft that ancestral components were not the only relevant data to take into account, and that Y-DNA haplogroups R1a and R1b (appearing separately in CWC and Yamna-Bell Beaker-Afanasevo), together with their calculated timeframes of formation – and therefore likely expansion – did not fit with the archaeological and linguistic description of the spread of Proto-Indo-European and its dialects.
In fact, it seemed that only one haplogroup (R1b-M269) was constantly and consistenly associated with the proposed routes of Late PIE dialectal expansions – like Anthony’s second (Afanasevo) and third (Lower Danube, Balkan) waves. What genetics shows fits seamlessly with Mallory’s association of the North-West Indo-European expansion with Bell Beakers (read here how archaeologists were right).
More precise inconsistencies were observed after the publication of Olalde et al. (2017) and Mathieson et al. (2017), by Volker Heyd in Kossinna’s smile (2017). Letting aside the many details enumerated (you can read a summary in my latest draft), this interesting excerpt is from the conclusion:
Simple solutions to complex problems are never the best choice, even when favoured by politicians and the media. Kossinna also offered a simple solution to a complex prehistoric problem, and failed therein. Prehistoric archaeology has been aware of this for a century, and has responded by becoming more differentiated and nuanced, working anthropologically, scientifically and across disciplines (cf. Müller 2013; Kristiansen 2014), and rejecting monocausal explanations. The two aDNA papers in Nature, powerful and promising as they are for our future understanding, also offer rather straightforward messages, heavily pulled by culture-history and the equation of people with culture. This admittedly is due partly to the restrictions of the medium that conveys them (and despite the often relevant additional detail given as supplementary information, which is unfortunately not always given full consideration).
While I have no doubt that both papers are essentially right, they do not reflect the complexity of the past. It is here that archaeology and archaeologists contributing to aDNA studies find their role; rather than simply handing over samples and advising on chronology, and instead of letting the geneticists determine the agenda and set the messages, we should teach them about complexity in past human actions and interactions. If accepted, this could be the beginning of a marriage made in heaven, with the blessing smile of Gustaf Kossinna, and no doubt Vere Gordon Childe, were they still alive, in a reconciliation of twentieth- and twenty-first-century approaches. For us as archaeologists, it could also be the starting point for the next level of a new archaeology.
The question was made painfully clear with the publication of Olalde et al. (2018) & Mathieson et al. (2018), where the real route of Yamna expansion into Europe was now clearly set through the steppes into the Carpathian basin, later expanded as Bell Beakers.
A total of 286 samples of Uralic-speaking individuals, of those 121 genotyped in this study, were analysed in the context of 1514 Eurasian samples (including 14 samples published for the first time) based on whole genome single nucleotide polymorphisms (SNPs) (Additional file 1: Table S1). All these samples, together with the larger sample set of Uralic speakers, were characterized for mtDNA and chrY markers.
The question as which material cultures may have co-spread together with proto-Uralic and Uralic languages depends on the time estimates of the splits in the Uralic language tree. Deeper age estimates (6,000 BP) of the Uralic language tree suggest a connection between the spread of FU languages from the Volga River basin towards the Baltic Sea either with the expansion of the Neolithic culture of Combed Ware, e.g. [6, 7, 17, 26] or with the Neolithic Volosovo culture . Younger age estimates support a link between the westward dispersion of Proto-Finno-Saamic and eastward dispersion of Proto-Samoyedic with a BA Sejma-Turbino (ST) cultural complex [14, 18, 27, 28] that mediated the diffusion of specific metal tools and weapons from the Altai Mountains over the Urals to Northern Europe or with the Netted Ware culture , which succeeded Volosovo culture in the west. It has been suggested that Proto-Uralic may have even served as the lingua franca of the merchants involved in the ST phenomenon . All these scenarios imply that material culture of the Baltic Sea area in Europe was influenced by cultures spreading westward from the periphery of Europe and/or Siberia. Whether these dispersals involved the spread of both languages and people remains so far largely unknown.
The population structure of Uralic speakers
To contextualize the autosomal genetic diversity of Uralic speakers among other Eurasian populations (Additional file 1: Table S1), we first ran the principal component (PC) analysis (Fig. 2a, Additional file 3: Figure S1). The first two PCs (Fig. 2a, Additional file 3: Figure S1A) sketch the geography of the Eurasian populations along the East-West and North-South axes, respectively. The Uralic speakers, along with other populations speaking Slavic and Turkic languages, are scattered along the first PC axis in agreement with their geographic distribution (Figs. 1 and 2a) suggesting that geography is the main predictor of genetic affinity among the groups in the given area. Secondly, in support of this, we find that FST-distances between populations (Additional file 3: Figure S2) decay in correlation with geographical distance (Pearson’s r = 0.77, p < 0.0001). On the UPGMA tree based on these FST-distances (Fig. 2b), the Uralic speakers cluster into several different groups close to their geographic neighbours.
We next used ADMIXTURE , which presents the individuals as composed of inferred genetic components in proportions that maximize Hardy-Weinberg and linkage equilibrium in the overall sample (see the ‘Methods’ section for choice of presented K). Overall, and specifically at lower values of K, the genetic makeup of Uralic speakers resembles that of their geographic neighbours. The Saami and (a subset of) the Mansi serve as exceptions to that pattern being more similar to geographically more distant populations (Fig. 3a, Additional file 3: S3). However, starting from K = 9, ADMIXTURE identifies a genetic component (k9, magenta in Fig. 3a, Additional file 3: S3), which is predominantly, although not exclusively, found in Uralic speakers. This component is also well visible on K = 10, which has the best cross-validation index among all tests (Additional file 3: S3B). The spatial distribution of this component (Fig. 3b) shows a frequency peak among Ob-Ugric and Samoyed speakers as well as among neighbouring Kets (Fig. 3a). The proportion of k9 decreases rapidly from West Siberia towards east, south and west, constituting on average 40% of the genetic ancestry of FU speakers in Volga-Ural region (VUR) and 20% in their Turkic-speaking neighbours (Bashkirs, Tatars, Chuvashes; Fig. 3a). The proportion of this component among the Saami in Northern Scandinavia is again similar to that of the VUR FU speakers, which is exceptional in the geographic context. It is also notable that North Russians, sampled from near the White Sea, differ from other Russians by sporting higher proportions of k9 (10–15%), which is similar to the values we observe in their Finnic-speaking neighbours. Notably, Estonians and Hungarians, who are geographically the westernmost Uralic speakers, virtually lack the k9 cluster membership.
We also tested the different demographic histories of female and male lineages by comparing outgroup f3 results for autosomal and X chromosome (chrX) data for pairs of populations (Estonians, Udmurts or Khanty vs others) with high versus low probability to share their patrilineal ancestry in chrY hg N (see the ‘Methods’ section, Additional file 3: Figure S13). We found a minor but significant excess of autosomal affinity relative to chrX for pairs of populations that showed a higher than 10% chance of two randomly sampled males across the two groups sharing their chrY ancestry in hg N3-M178, compared to pairs of populations where such probability is lower than 5% (Additional file 3: Figure S13).
In sum, these results suggest that most of the Uralic speakers may indeed share some level of genetic continuity via k9, which, however, also extends to the geographically close Turkic speakers.
We found that it is the admixture with the Siberians that makes the Western Uralic speakers different from the tested European populations (Additional file 3: Figure S4A-F, H, J, L). Differentiating between Estonians and Finns, the Siberians share more derived alleles with Finns, while the geographic neighbours of Estonians (and Finns) share more alleles with Estonians (Additional file 3: Figure S4M). Importantly, Estonians do not share more derived alleles with other Finnic, Saami, VUR FU or Ob-Ugric-speaking populations than Latvians (Additional file 3: Figure S4O). The difference between Estonians and Latvians is instead manifested through significantly higher levels of shared drift between Estonians and Siberians on the one hand and Latvians and their immediate geographic neighbours on the other hand. None of the Uralic speakers, including linguistically close Khanty and Mansi, show significantly closer affinities to the Hungarians than any non-FU population from NE Europe (Additional file 3: Figure S4R).
Time of Siberian admixture
The time depth of the Globetrotter (Fig. 5b) inferred admixture events is relatively recent—500–1900 AD (see also complementary ALDER results, in Additional file 13: Table S12 and Additional file 3: Figure S7)—and agrees broadly with the results reported in Busby et al. . A more detailed examination of the ALDER dates, however, reveals an interesting pattern. The admixture events detected in the Baltic Sea region and VUR Uralic speakers are the oldest (800–900 AD or older) followed by those in VUR Turkic speakers (∼1200–1300 AD), while the admixture dates for most of the Siberian populations (>1500 AD) are the most recent (Additional file 3: Figure S7). The West Eurasian influx into West Siberia seen in modern genomes was thus very recent, while the East Eurasian influx into NE Europe seems to have taken place within the first millennium AD (Fig. 5b, Additional file 3: Figure S7).
Affinities of the Uralic speakers with ancient Eurasians
We next calculated outgroup f3-statistics  to estimate the extent of shared genetic drift between modern and ancient Eurasians (Additional file 14: Table S13, Additional file 3: Figures S8-S9). Consistent with previous reports [45, 50], we find that the NE European populations including the Uralic speakers share more drift with any European Mesolithic hunter-gatherer group than Central or Western Europeans (Additional file 3: Figure S9A-C). Contrasting the genetic contribution of western hunter-gatherers (WHG) and eastern hunter-gatherers (EHG), we find that VUR Uralic speakers and the Saami share more drift with EHG. Conversely, WHG shares more drift with the Finnic and West European populations (Additional file 3: Figure S9A). Interestingly, we see a similar pattern of excess of shared drift between VUR and EHG if we substitute WHG with the aDNA sample from the Yamnaya culture (Additional file 3: Figure S9D). As reported before [2, 45], the genetic contribution of European early farmers decreases along an axis from Southern Europe towards the Ural Mountains (Fig. 6, Additional file 3: Figure S9E-F).
We then used the qpGraph software  to test alternative demographic scenarios by trying to fit the genetic diversity observed in a range of the extant Finno-Ugric populations through a model involving the four basic European ancestral components: WHG, EHG, early farmers (LBK), steppe people of Yamnaya/Corded Ware culture (CWC) and a Siberian component (Fig. 6, Additional file 3: Figure S10). We chose the modern Nganasans to serve as a proxy for the latter component because we see least evidence for Western Eurasian admixture (Additional file 3: Figure S3) among them. We also tested the Khantys for that proxy but the model did not fit (yielding f2-statistics, Z-score > 3). The only Uralic-speaking population that did not fit into the tested model with five ancestral components were Hungarians. The qpGraph estimates of the contributions from the Siberian component show that it is the main ancestry component in the West Siberian Uralic speakers and constitutes up to one third of the genomes of modern VUR and the Saami (Fig. 6). It drops, however, to less than 10% in most of NE Europe, to 5% in Estonians and close to zero in Latvians and Lithuanians.
One of the notable observations that stands out in the fineSTRUCTURE analysis is that neither Hungarians nor Estonians or Mordovians form genetic clusters with other Uralic speakers but instead do so with a broad spectrum of geographically adjacent samples. Despite the documented history of the migration of Magyars  and their linguistic affinity to Khantys and Mansis, who today live east of the Ural Mountains, there is nothing in the present-day gene pool of the sampled Hungarians that we could tie specifically to other Uralic speakers.
Perhaps even more surprisingly, we found that Estonians, who show close affinities in IBD analysis to neighbouring Finnic speakers and Saami, do not share an excess of IBD segments with the VUR or Siberian Uralic speakers. This is eIn this context, it is important to remind that the limited (5%, Fig. 6) East Eurasian impact in the autosomal gene pool of modern Estonians contrasts with the fact that more than 30% of Estonian (but not Hungarian) men carry chrY N3 that has an East Eurasian origin and is very frequent among NE European Uralic speakers . However, the spread of chrY hg N3 is not language group specific as it shows similar frequencies in Baltic-speaking Latvians and Lithuanians, and in North Russians, who in all our analyses are very similar to Finnic-speakers. The latter, however, are believed to have either significantly admixed with their Uralic-speaking neighbours or have undergone a language shift from Uralic to Indo-European .ven more striking considering that the immediate neighbours—Finns, Vepsians and Karelians—do.
With some exceptions such as Estonians, Hungarians and Mordovians, both IBD sharing and Globetrotter results suggest that there are detectable inter-regional haplotype sharing ties between Uralic speakers from West Siberia and VUR, and between NE European Uralic speakers and VUR. In other words, there is a fragmented pattern of haplotype sharing between populations but no unifying signal of sharing that unite all the studied Uralic speakers.
The paper is obviously trying to find a “N1c/Siberian ancestry = Uralic” link, but it shows (as previous papers using ancient DNA) that this identification is impossible, because it is not possible to identify “N1c=Siberian ancestry”, “N1c=Uralic”, or “Siberian ancestry = Uralic”. In fact, the arrival of N subclades and Siberian ancestry are late, both events (probably multiple stepped events) are unrelated to each other, and represent east-west demic diffusion waves (as well as founder effects) that probably coincide in part with the Scythian and Turkic (or associated) expansions, i.e. too late for any model of Proto-Uralic or Proto-Finno-Ugric expansion.
On the other hand, it shows interesting data regarding ancestry of populations that show increased Siberian influence, such as those easternmost groups admixed with Yeniseian-like populations (Samoyedic), those showing strong founder effects (Finnic), or those isolated in the Circum-Artic region with neighbouring Siberian peoples in Kola (Saami). All in all, Hungarians, Estonians and Mordovians seem to show the original situation better than the other groups, which is also reflected in part in Y-DNA, conserved as a majority of R1a lineages precisely in these groups. Just another reminder that CWC-related ancestry is found in every single Uralic group, and that it represents the main ancestral component in all non-Samoyedic groups.
The qpGraph shows the ancestor of Yamna (likely Khvalynsk) and Corded Ware stemming as different populations from a common (likely Neolithic) node – whose difference is based on the proportion of Anatolian-related ancestry – , that is, probably before the Indo-Hittite expansion; and ends with CWC groups forming the base for all Uralic peoples. Below is a detail of the qpGraph on the left, and my old guess (2017) on the right, for comparison:
#EDIT (22 sep 2018): I enjoyed re-reading it, and found this particular paragraph funny:
Despite the documented history of the migration of Magyars  and their linguistic affinity to Khantys and Mansis, who today live east of the Ural Mountains, there is nothing in the present-day gene pool of the sampled Hungarians that we could tie specifically to other Uralic speakers.
Experienced researchers, particularly those interested in population structure and historical inference, typically present STRUCTURE results alongside other methods that make different modelling assumptions. These include TreeMix, ADMIXTUREGRAPH, fineSTRUCTURE, GLOBETROTTER, f3 and D statistics, amongst many others. These models can be used both to probe whether assumptions of the model are likely to hold and to validate specific features of the results. Each also comes with its own pitfalls and difficulties of interpretation. It is not obvious that any single approach represents a direct replacement as a data summary tool. Here we build more directly on the results of STRUCTURE/ADMIXTURE by developing a new approach, badMIXTURE, to examine which features of the data are poorly fit by the model. Rather than intending to replace more specific or sophisticated analyses, we hope to encourage their use by making the limitations of the initial analysis clearer.
The default interpretation protocol
Most researchers are cautious but literal in their interpretation of STRUCTURE and ADMIXTURE results, as caricatured in Fig. 1, as it is difficult to interpret the results at all without making several of these assumptions. Here we use simulated and real data to illustrate how following this protocol can lead to inference of false histories, and how badMIXTURE can be used to examine model fit and avoid common pitfalls.
STRUCTURE and ADMIXTURE are popular because they give the user a broad-brush view of variation in genetic data, while allowing the possibility of zooming down on details about specific individuals or labelled groups. Unfortunately it is rarely the case that sampled data follows a simple history comprising a differentiation phase followed by a mixture phase, as assumed in an ADMIXTURE model and highlighted by case study 1. Naïve inferences based on this model (the Protocol of Fig. 1) can be misleading if sampling strategy or the inferred value of the number of populations K is inappropriate, or if recent bottlenecks or unobserved ancient structure appear in the data. It is therefore useful when interpreting the results obtained from real data to think of STRUCTURE and ADMIXTURE as algorithms that parsimoniously explain variation between individuals rather than as parametric models of divergence and admixture.
For example, if admixture events or genetic drift affect all members of the sample equally, then there is no variation between individuals for the model to explain. Non-African humans have a few percent Neanderthal ancestry, but this is invisible to STRUCTURE or ADMIXTURE since it does not result in differences in ancestry profiles between individuals. The same reasoning helps to explain why for most data sets—even in species such as humans where mixing is commonplace—each of the K populations is inferred by STRUCTURE/ADMIXTURE to have non-admixed representatives in the sample. If every individual in a group is in fact admixed, then (with some exceptions) the model simply shifts the allele frequencies of the inferred ancestral population to reflect the fraction of admixture that is shared by all individuals.
Several methods have been developed to estimate K, but for real data, the assumption that there is a true value is always incorrect; the question rather being whether the model is a good enough approximation to be practically useful. First, there may be close relatives in the sample which violates model assumptions. Second, there might be “isolation by distance”, meaning that there are no discrete populations at all. Third, population structure may be hierarchical, with subtle subdivisions nested within diverged groups. This kind of structure can be hard for the algorithms to detect and can lead to underestimation of K. Fourth, population structure may be fluid between historical epochs, with multiple events and structures leaving signals in the data. Many users examine the results of multiple K simultaneously but this makes interpretation more complex, especially because it makes it easier for users to find support for preconceptions about the data somewhere in the results.
In practice, the best that can be expected is that the algorithms choose the smallest number of ancestral populations that can explain the most salient variation in the data. Unless the demographic history of the sample is particularly simple, the value of K inferred according to any statistically sensible criterion is likely to be smaller than the number of distinct drift events that have practically impacted the sample. The algorithm uses variation in admixture proportions between individuals to approximately mimic the effect of more than K distinct drift events without estimating ancestral populations corresponding to each one. In other words, an admixture model is almost always “wrong” (Assumption 2 of the Core protocol, Fig. 1) and should not be interpreted without examining whether this lack of fit matters for a given question.
Because STRUCTURE/ADMIXTURE accounts for the most salient variation, results are greatly affected by sample size in common with other methods. Specifically, groups that contain fewer samples or have undergone little population-specific drift of their own are likely to be fit as mixes of multiple drifted groups, rather than assigned to their own ancestral population. Indeed, if an ancient sample is put into a data set of modern individuals, the ancient sample is typically represented as an admixture of the modern populations (e.g., ref. 28,29), which can happen even if the individual sample is older than the split date of the modern populations and thus cannot be admixed.
This paper was already available as a preprint in bioRxiv (first published in 2016) and it is incredible that it needed to wait all this time to be published. I found it weird how reviewers focused on the “tone” of the paper. I think it is great to see files from the peer review process published, but we need to know who these reviewers were, to understand their whiny remarks… A lot of geneticists out there need to develop a thick skin, or else we are going to see more and more delays based on a perceived incorrect tone towards the field, which seems a rather subjective reason to force researchers to correct a paper.
A potential hindrance to our advice to upgrade from PCA graphs to PCA biplots is that the SNPs are often so numerous that they would obscure the Items if both were graphed together. One way to reduce clutter, which is used in several figures in this article, is to present a biplot in two side-by-side panels, one for Items and one for SNPs. Another stratagem is to focus on a manageable subset of SNPs of particular interest and show only them in a biplot in order to avoid obscuring the Items. A later section on causal exploration by current methods mentions several procedures for identifying particularly relevant SNPs.
One of several data transformations is ordinarily applied to SNP data prior to PCA computations, such as centering by SNPs. These transformations make a huge difference in the appearance of PCA graphs or biplots. A SNPs-by-Items data matrix constitutes a two-way factorial design, so analysis of variance (ANOVA) recognizes three sources of variation: SNP main effects, Item main effects, and SNP-by-Item (S×I) interaction effects. Double-Centered PCA (DC-PCA) removes both main effects in order to focus on the remaining S×I interaction effects. The resulting PCs are called interaction principal components (IPCs), and are denoted by IPC1, IPC2, and so on. By way of preview, a later section on PCA variants argues that DC-PCA is best for SNP data. Surprisingly, our literature survey did not encounter even a single analysis identified as DC-PCA.
The axes in PCA graphs or biplots are often scaled to obtain a convenient shape, but actually the axes should have the same scale for many reasons emphasized recently by Malik and Piepho . However, our literature survey found a correct ratio of 1 in only 10% of the articles, a slightly faulty ratio of the larger scale over the shorter scale within 1.1 in 12%, and a substantially faulty ratio above 2 in 16% with the worst cases being ratios of 31 and 44. Especially when the scale along one PCA axis is stretched by a factor of 2 or more relative to the other axis, the relationships among various points or clusters of points are distorted and easily misinterpreted. Also, 7% of the articles failed to show the scale on one or both PCA axes, which leaves readers with an impressionistic graph that cannot be reproduced without effort. The contemporary literature on PCA of SNP data mostly violates the prohibition against stretching axes.
The percentage of variation captured by each PC is often included in the axis labels of PCA graphs or biplots. In general this information is worth including, but there are two qualifications. First, these percentages need to be interpreted relative to the size of the data matrix because large datasets can capture a small percentage and yet still be effective. For example, for a large dataset with over 107,000 SNPs for over 6,000 persons, the first two components capture only 0.3693% and 0.117% of the variation, and yet the PCA graph shows clear structure (Fig 1A in ). Contrariwise, a PCA graph could capture a large percentage of the total variation, even 50% or more, but that would not guarantee that it will show evident structure in the data. Second, the interpretation of these percentages depends on exactly how the PCA analysis was conducted, as explained in a later section on PCA variants. Readers cannot meaningfully interpret the percentages of variation captured by PCA axes when authors fail to communicate which variant of PCA was used.
Five simple recommendations for effective PCA analysis of SNP data emerge from this investigation.
Use the SNP coding 1 for the rare or minor allele and 0 for the common or major allele.
Use DC-PCA; for any other PCA variant, examine its augmented ANOVA table.
Report which SNP coding and PCA variant were selected, as required by contemporary standards in science for transparency and reproducibility, so that readers can interpret PCA results properly and reproduce PCA analyses reliably.
Produce PCA biplots of both Items and SNPs, rather than merely PCA graphs of only Items, in order to display the joint structure of Items and SNPs and thereby to facilitate causal explanations. Be aware of the arch distortion when interpreting PCA graphs or biplots.
Produce PCA biplots and graphs that have the same scale on every axis.
I read the referenced paper Biplots: Do Not Stretch Them!, by Malik and Piepho (2018), and even though it is not directly applicable to the most commonly available PCA graphs out there, it is a good reminder of the distorting effects of stretching. So for example quite recently in Krause-Kyora et al. (2018), where you can see Corded Ware and BBC samples from Central Europe clustering with samples from Yamna:
NOTE. This is related to a vertical distorsion (i.e. horizontal stretching), but possibly also to the addition of some distant outlier sample/s.
The so-called ‘Yamnaya’ ancestry
Every time I read papers like these, I remember commenters who kept swearing that genetics was the ultimate science that would solve anthropological problems, where unscientific archaeology and linguistics could not. Well, it seems that, like radiocarbon analysis, these promising developing methods need still a lot of refinement to achieve something meaningful, and that they mean nothing without traditional linguistics and archaeology… But we already knew that.
Also, if this is happening in most peer-reviewed publications, made by professional geneticists, in journals of high impact factor, you can only wonder how many more errors and misinterpretations can be found in the obscure market of so many amateur geneticists out there. Because amateur geneticist is a commonly used misnomer for people who are not geneticists (since they don’t have the most basic education in genetics), and some of them are not even ‘amateurs’ (because they are selling the outputs of bioinformatic tools)… It’s like calling healers ‘amateur doctors’.
NOTE. While everyone involved in population genetics is interested in knowing the truth, and we all have our confirmation (and other kinds of) biases, for those who get paid to tell people what they want to hear, and who have sold lots of wrong interpretations already, the incentives of ‘being right’ – and thus getting involved in crooked and paranoid behaviour regarding different interpretations – are as strong as the money they can win or loose by promoting themselves and selling more ‘product’.
As a reminder of how badly these wrong interpretations of genetic results – and the influence of the so-called ‘amateurs’ – can reflect on research groups, yet another turn of the screw by the Copenhagen group, in the oral presentations at Languages and migrations in pre-historic Europe (7-12 Aug 2018), organized by the Copenhagen University. The common theme seems to be that Bell Beaker and thus R1b-L23 subclades do represent a direct expansion from Yamna now, as opposed to being derived from Corded Ware migrants, as they supported before.
NOTE. Yes, the “Yamna → Corded Ware → Únětice / Bell Beaker” migration model is still commonplace in the Copenhagen workgroup. Yes, in 2018. Guus Kroonen had already admitted they were wrong, and it was already changed in the graphic representation accompanying a recent interview to Willerslev. However, since there is still no official retraction by anyone, it seems that each member has to reject the previous model in their own way, and at their own pace. I don’t think we can expect anyone at this point to accept responsibility for their wrong statements.
I love the newly invented arrows of migration from Yamna to the north to distinguish among dialects attributed by them to CWC groups, and the intensive use of materials from Heyd’s publications in the presentation, which means they understand he was right – except for the fact that they are used to support a completely different theory, radically opposed to those defended in Heyd’s model…
Now added to the Copenhagen’s unending proposals of language expansions, some pearls from the oral presentation:
Corded Ware north of the Carpathians of R1a lineages developed Germanic;
R1b borugh [?] Italo-Celtic;
the increase in steppe ancestry on north European Bell Beakers mean that they “were a continuation of the Yamnaya/Corded Ware expansion”;
“Corded Ware groups  stopped their expansion and took over the Bell Beaker package before migrating to England” [yep, it literally says that];
Italo-Celtic expanded to the UK and Iberia with Bell Beakers [I guess that included Lusitanian in Iberia, but not Messapian in Italy; or the opposite; or nothing like that, who knows];
2nd millennium BC Bronze Age Atlantic trade systems expanded Proto-Celtic [yep, trade systems expanded the language]
1st millennium BC expanded Gaulish with La Tène, including a “Gaulish version of Celtic to Ireland/UK” [hmmm, datBritish Gaulish indeed].
You know, because, why the hell not? A logical, stable, consequential, no-nonsense approach to Indo-European migrations, as always.
Also, compare still more invented arrows of migrations, from Mikkel Nørtoft’s Introducing the Homeland Timeline Map, going against Kristiansen’s multiple arrows, and even against the own recent fantasy map series in showing Bell Beakers stem from Yamna instead of CWC (or not, you never truly know what arrows actually mean):
I really, really loved that perennial arrow of migration from Volosovo, ca. 4000-800 BC (3000+ years, no less!), representing Uralic?, like that, without specifics – which is like saying, “somebody from the eastern forest zone, somehow, at some time, expanded something that was not Indo-European to Finland, and we couldn’t care less, except for the fact that they were certainly not R1a“.
This and Kristiansen’s arrows are the most comical invented migration routes of 2018; and that is saying something, given the dozens of similar maps that people publish in forums and blogs each week.
It’s hard to accept that this is a series of presentations made by professional linguists, archaeologists, and geneticists, as stated by the official website, and still harder to imagine that they collaborate within the same professional workgroup, which includes experienced geneticists and academics.
I propose the following video to close future presentations introducing innovative ideas like those above, to help the audience find the appropriate mood: