Centum and Satem

I may well be wrong, but there’s no point in waiting until they dig up and sequence every last body in Eurasia. Time to stick my neck out.

Here’s my current best guess concerning the Indo-European expansion:

It all started pretty far to the East. There, some crazy locals first tamed the horse.  I’m thinking that they were something like the Botai culture, riding and hunting horses, but not farmers. They may well have milked those horses – by the way, horse milk is much richer in lactose than human or cow milk.

These early horsemen were genetically similar to the Ancient North Eurasians, or as those who know have dubbed them, Sibermen.

Having horses made them natural raiders, let them expand.  Lactose tolerance might have helped.

A fraction of them conquered some farmers (a mixture of Middle Eastern types and others similar to western hunter-gatherers)  in the eastern Ukraine,  imposed their language, and roared into northern Europe.  This accounts for the Centum languages: they have probably some additions from whatever language those Ukie farmers were speaking. Some people from this group (Tocharians) must have made a wrong turn and eventually ended up in Western China. Even today some Uighurs show the mark of those red-headed strangers.


But that centum expansion  didn’t come first. Other, earlier kinds of Indo-Europeans had already destroyed the old EEF culture in the Balkans – and although they seem to have had some ANE ancestry, they apparently had very little WHG ancestry.  You see mixtures of EEF and ANE in Greece and Albania, but almost no  WHG.  This has to be the result of a separate, early Indo-European expansion.  Looks as if this might have gone on into Anatolia –   the Hitties, Luwian, Palaic.

Nor was it the last. There must have been some peoples in the real Indo-European homeland (farther to the east than the Yamna culture) who had not yet conquered a bunch of motley farmers and had remained mostly ANE.  Considerably later, now charioteers,  some of them moved south, conquered some more Armenian-like farmers (the Bactria-Margiana culture, BMAC, located in the southern part of what used to be Soviet Central Asia, now Trashcanistan) (but no WHG types) and then went on to conquer India and Iran.  Some ended up in odd corners like Nuristan and the Chitral Valley: here’s a young Kalash boy.


Both the European Centum expansion and the later Indo-Aryan expansion carried the same lactose-tolerance mutation., which has never been found in aDNA from the EEF farmers. Of course it could have been around in Old Europe, just still rare, but the simplest explanation for finding the same common mutation in Europe and north India is that it was spread by the Indo-Europeans themselves.

The WHG hunters had blue eyes, the same mutation as today, while the early European farmers had SLC24A5, but as of yet I’ve not seen any ancient examples of the  alleles that give varied hair color in Europeans.  Blondes and redheads are rare in Basques and Sardinians – I’m wondering if the ANE are the source.

This entry was posted in Dietary adaptations, European Prehistory, Genetics, Indo-European, Linguistics and tagged , , , . Bookmark the permalink.

100 Responses to Centum and Satem

  1. Bruce says:

    How much further east than the Yamna culture? West of the Urals, I assume.

  2. Sean says:

    You can’ get away from the fact that those totally different red and blonde colours are not just found more commonly in the population of north west Europe that elsewhere today, they are ten a penny: the diversity is the norm in Europe. And those photos just emphasise that because no-one thinks there is anything noteworthy about red or blonde hair in Europe.

    There were mixtures and conquests but nobody appears to be saying that there was a totally red headed and a wholly blonde popuation, and these mixed with the uniformly black haired populations of the rest of the word. As there are are totally different hair colours like blonde and red in the European populations; from whence the diversity? It seems to me that the only explaination is that novelty and diversity in hair color was what was being selected for in a population. Why? We are not told what this section pressure maybe was and why it only operated on this population, that is apparently thought to be of no interest or relevance .

    OK, here is the essential point where I think someone is sticking their neck out like a giraffe. It seems to be being advanced that the hair diversity originally came from a different population to the population with the diversity of eye color. So it is being said the current appearance of Europeans is due to the mixing of two totally different populations which both have striking diverstity in a visible characterstic. (And not clear which had white skin, maybe that is supposed to have originated in a third originally).

  3. sceptic says:

    A couple of minor changes to your scenario:

    1. Indo-Iranians spread from Sinthashta culture, South East of Yamna I think.

    2. Most likely the expansion in to Iran and India was through elite dominance (like the Mittani) conquest, including things like mercenary bands going in service of a non-Indo-European king, and then taking over the empire. I don’t see a large genetic imprint or population replacement is likely. This would also explain the lack of institutional memory of migration and conquest (it was likely just a few migrants, who wanted to adopt the new lands).

    3. The idea of the east and westward expansions independently conquering farmers is interesting and may solve the issue of shared words for cereal farming in all Indo_European languages while there is no evidence of any such farming in the steppes. They acquired them independently!

    • Matthew M. Robare says:

      Re: 2. Firstly, there was clearly a substantial population replacement in Iran, Pakistan, Afghanistan and India. I understand that Central Asia was pretty well IE before the Turkic peoples moved in. Secondly, how is there a lack of institutional memory of migration and conquest? Isn’t that what the Rig Veda is?

      Re: 3. There was farming on the steppes. Herodotus recorded that the Scythians had a farming caste. Obviously that doesn’t indicate that the farmers were genetically related to the ruling class. Beckwith has other examples of farming, too, as I recall. Also, wouldn’t shared words for cereal farming among dispersed IE languages mean that theyv didn’t acquire them independently?

      • sceptic says:

        Re 2, there is no archaelogical evidence for population replacement in India/Iran that corresponds to Indo-European timeline. eg. the decline of the Indus Valley and subsequent eastward migration of Neolithic people is dated roughly 2000 – 1000 BC, but skeletally there are no changes. The genetic marker of the “Ancestral North Indian” people is likely to be the large Neolithic farmer population and not Indo-European at all.

        The Rig Veda is almost entirely hymns in praise of the Vedic Gods. There is no mention of migration at all barring a handful of disputed and ambiguous verses. The people who composed it called themselves Aryans and their land the Aryavarta – they never state that they migrated from or to the Aryavarta. There are definitely mentions of wars and victories, but that is not memory of migration. The caricature of the Rig Veda as evidence for bands of conquering nomads is highly overstated. Also the Rig Veda was followed over the next 500-1000 years by thousands of other hymns and expository verses in the subsequent Vedas, and the people who composed those definitely had forgotten any migration from a distant land. (I’m not arguing that there was no migration, just that the memory was lost, as it was lost in Greece for example).

        Re 3: See J P Mallory, “Twenty-first century clouds over Indo-European homelands”: http://www.jolr.ru/files/(112)jlr2013-9(145-154).pdf. The presence of shared farming words is a real puzzle.

        • Karl Zimmerman says:

          There’s no evidence of population replacement. But judging by the data that the Harappa Ancestry Project put together, there is pretty clearly a layer of Northeastern European admixture, which peaks with Jatts (17%), and is at least 5% in many northern Indian groups (particularly high caste).

          And, of course, Aryans wouldn’t have been pure “Northeast Europeans” when they reached India. Admixture with farming groups in modern Turkestan could have resulted in admixture which today shows up as Caucasian, Baloch, and even South Indian. Modern day Central Asians show a lot of admixture from all of these groups after all.

          In a way, the genetic situation in India could be confounded the same way it is in Europe. There’s reason to believe the pre-Iranian population in Central Asia was closely related genetically to the group that brought Near Eastern agriculture into India (e.g., both would count as Ancient North Indian). If the Aryans admixed with them, and then in turn admixed to some degree with the already mixed Harappans, then it would be hard to disentangle the two threads. Just the same way that the mixed Indo-Europeans in Europe make it hard to disentangle the degree to which population replacement versus hybridization occurred in Northern Europe.

          • Greying Wanderer says:

            ” If the Aryans admixed with them, and then in turn admixed to some degree with the already mixed Harappans, then it would be hard to disentangle the two threads.”

            Yes if a migration/invasion pattern has an underlying mechanic (e.g. it takes c. 400 years for a source region to build up an invasion level of population surplus) then you might get a repeated cycle of migrations/invasions from the same source region creating layers of similar genetics in the target region.

          • sceptic says:

            “If the Aryans admixed with them, and then in turn admixed to some degree with the already mixed Harappans, then it would be hard to disentangle the two threads.”

            There is already some linguistic evidence that this might be the case – eg. Indo_Iranian has borrowed words from a language family, and Indo-Aryan has some specific ones from the same family.

  4. Andrew says:

    I have thought about Greg’s suggestion that there is an IE golden family and I think it is possible that it extended into Egypt.

    Some Indo-Euros ruled over the ancient kingdoms in the Fertile Crescent as well. Some of the Kassite rulers of Babylon had Indo-Euro names, the Mitanni were Indic and probably the ones to introduce the chariot to the area. Some of the Pharaohs in the eighteenth dynasty married with Mitanni royals. Pharaoh Akhenaten was a religious heretic who worshipped the Sun disk that seems similar to the Indo-Euro worship of the sky father.


    Akhenaten’s son was King Tut whose DNA was recently extracted and it revealed some of the known sister marriage of the pharaohs. Hawass did not want to publish detailed DNA information that revealed Tut’s ancestry. A leaked source had claimed that Tut was Y chromosome R1b which is a probable Indo-European source.


    Speculative, but plausible. Too bad the data was officially withheld.

    There is a more probable interpretation of the Nebra Sky disk than the current one. The Disk appears to represent the sky as it would be seen in the solar eclipse of April 16, 1699 BCE. The disk is made from copper that came from Austria and the disk was found in Germany. It was found with 3 different types of swords ( a collection perhaps). Interestingly, the eclipse of 1699 BC was not total in Europe, but it was in Egypt. The center of totality was within a few miles of the Upper Nile River Valley per the NASA eclipse database. The disk also had motifs similar to the Indo-European notion of the solar boat.


    OK, yeah. Pots!

    • Paul Conroy says:

      Yeah, I agree. As I mentioned previously, I’ve always though of the Hyksos to be Indo-Europeans. Hence the red-haired pharaohs and the later mention of “red-skinned” peoples of Libya.
      In fact the Sea Peoples may have been early Indo-Europeans…

    • Jim says:

      The Mitanni rulers seem to have spoken an Aryan language but the bulk of the population were Hurrian speakers.

      • Andrew says:

        Yep, there seems to be a pattern that suggest elite dominance where royals speak IE while the majority folk speak another language.

    • Matthew M. Robare says:

      Would a Golden Family really be neccessary? On a time scale of even a few hundred years anyone who has children surviving into adulthood is going to wind up having lots of descendents. Moreover, looking at historical IE cultures, it would seem that they had a great diversity. Despite a close physical proximity and active trading and warfare, Greeks, Indo-Aryans, Indo-Iranians, Italics, Celts, Germanics and other IE speakers zealously maintained their separate identities and traditions. Unlike other peoples who pursued a bigger-is-better strategy, IE speakers were always fracturing into small groups. And even when they did pursue a bigger-is-better strategy, with the Persian, Hellenistic and Roman Empires, it never took very long before fracturing resumed.

  5. a very knowing American says:

    Donald Ringe’s cladistic reconstructions of Indo-European (cited in David Anthony’s book, but you can track down more recent stuff online), which seem like the best bet to give more-or-less correct branching orders (but not dates), have Indo-Iranian nested deep inside the tree, with Balto-Slavic as a close neighbor. This would seem to fit with a history in which eastern Battle Axe types migrate further east, through forested lands north of the steppe, and eventually contribute to founding the Sintashta culture in the Urals, the likely Indo-Iranian urheimat. David Anthony’s archeological reconstruction (pp 371-389) also fits this scenario. This would be instead of Proto-Indo-Iranians being guys who stayed behind on the steppes. Of course linguistic and cultural ancestry doesn’t have to match genetic ancestry all that well.

    Also of note, in Ringe’s work Germanic is uniquely unruly. It keeps jumping between close-to-Italo-Celtic and close-to-Balto-Slavic, suggesting Proto-Germanic might be some kind of hybrid.

    David Anthony. 2007. The Horse, the Wheel and Language: How Bronze Age Riders from the Eurasian Steppes Shaped the Modern World. Princeton University Press

  6. Samuel Ohanian says:

    Finns have high blonde rates. How high is their ANE?

    • High. One of the ironies of people now talking about ANE as a sign of Indo-Europeans is that in Europe it tends to peak in Finno-Ugric speakers. But then, blonde and ginger hair seem to peak in Finno-Ugric speakers as well.

      This isn’t any contradiction to the ANE-IE theory, though, since linguistic reconstrunctions put Finno-Ugric and Indo-European homelands in contact and Finno-Ugric is not like the other non-IE languages of Europe/Anatolia in that it does share some basic vocabulary, the pronouns and some grammar with IE while eg Basque seems to have nothing to do with IE. It’s very plausible that the original FU speakers and the original IE speakers were genetically similar or even the same.

      • Jim says:

        A genetic relationship between Indo-European and Uralic-Yukagir languages seem to be accepted by many linguists.

      • Samuel Ohanian says:

        True, but I think linguists think F-U and I-E have similarities because of contact not because of a genetic relationship. And Finno-Ugric expanded recently and fast as well.

        • subcomputer says:

          Both hypotheses come up, but the genetic relationship is usually considered the probable if unprovable. Most texts discussing suggested genetic relations between families usually contain a line such as “If another family were proven to be connected to Indo-European it would by far most likely be Finno-Ugric; if it was another family then one of the prime hurdles would be explaining the lack of connection to Finno-Ugric.”

    • Sean says:

      Finland was the last place in Europe to abandon hunter gathering I believe. And it is not simply uniform blonde hair in North Europe which is the problem.The diversity and proliferation of novelty is the unusual thing in Europe. Where one should expect the the highest amount of diversity need not concern us here although ‘A Decreasing Gradient of 374F Allele Frequencies in the Skin Pigmentation Gene SLC45A2, from the North of West Europe to North Africa.’ says “The highest allele frequency is observed in Denmark”.

      SLC45A2 solute carrier family 45, member 2 [ Homo sapiens (human) ] “Mutations in this gene are a cause of oculocutaneous albinism type 4, and polymorphisms in this gene are associated with variations in skin and hair color. Multiple transcript variants encoding different isoforms have been found for this gene”

      The Interplay between Natural Selection and Susceptibility to Melanoma on Allele 374F of SLC45A2 Gene in a South European Population “In order to assess also the association of L374F to hair or eye color, we genotyped a subset of 344 individuals from which we had paired information for these traits. We observed that the ancestral allele G (374L) was associated with black (OR = 2.14; p = 0.0018) and dark brown hair (OR = 2.24; p = 0.0189), and the darkest eye color […] we observed that the polymorphism Leu374Phe (L374F, rs16891982) was statistically associated with skin color variability within this sample. In particular, allele 374F was significantly more frequent among the individuals with lighter skin.[…] The age of the expansion of the allele in this case was estimated to be of 16,480 years ”

      To me the above implies that although the hair, eye colour diversity and skin color presently associated with west Europeans may have been present in some invaders, it originated in Europe.

      • Greying Wanderer says:

        “To me the above implies that although the hair, eye colour diversity and skin color presently associated with west Europeans may have been present in some invaders, it originated in Europe.”

        Or originated in a broad range across northern Eurasia for one reason but was more strongly selected for in one particular region for additional reasons.

    • JayMan says:

      For reference in case anyone needs it. Disparate sources so can’t vouch for accuracy:

      • Kate says:

        Yes. I notice redheads are in difficult terrain and awkward shaped states whilst yellow-heads are on the plain in nice square states. Do you think the redhead density at the juncture of 10 southern states is a place of choosing (good camouflage in autumn) or do you think redheads were driven into mountainous (and/or cold) areas?

        All the patterns regarding nwEuropeans are so striking that were it another group of people one feels these things would be axiomatic.

        The blond regions in nwEurope, rather like LP stats, seem to radiate from a point location. Red hair seems more as if it is pushed to marginal locations, Wales, Denmark. Unless, is it possible? red hair actually came via the Atlantic coastal route from the Levant, having dropped down to the Levant from the Tarim Basin via east Iran.

        • Greying Wanderer says:

          “Red hair seems more as if it is pushed to marginal locations”

          Or the red layer extends all across the north but where it overlaps with a second layer radiating out of the Baltic it turns blond. That to me is the key question – does the distribution represent a red layer that is submerged in the mid part of its range or are the red areas islands?

          ” is it possible? red hair actually came via the Atlantic coastal route from the Levant”

          I wouldn’t be at all surprised if it originally came out of the middle east, spread widely and then declined except in some refuge regions.

          • Paul Conroy says:

            I commented elsewhere on the relative strangeness of the Aran Islanders, off the coast of Galway, in the West of Ireland – they have also the supposedly highest concentration of Red Hair in the world.

            I remember reading that Neolithic Germans were much more Red haired than modern Germans a few years ago somewhere.

          • Kate says:

            Also, (somebody once told me) there are redheads on the islands off the coast of Portugal. So were redheads a beach-combing population or were they driven to the coast? I found a ref to a Medieval quote that was distinctly unPC regarding redhairedness so maybe redheads have always been marginalised (there seems to be a prevalence of robustness amongst some redheads too). But redhair seems set to survive well. Redheads have become the acceptable face of …plus, red does very well on mixing, the whole world could end up with red undertones.

          • Bruce says:

            Tacitus noted the red hair and stout limbs of the Picts and concluded that they must have come from Gaul. I assume the Iron age Gauls were red haired.

          • Paul Conroy says:

            I thought that was said about the Caledonians, not the Picts…

            The Caledonians being the older layer, and the Picts being a newer layer, possibly from Pictou in SW France.

        • JayMan says:


          The redheads in North America trace where various Celts landed, primarily Scots (Ulster Scots in Appalachia and the West, Highland Scots in New England):

          More Maps of the American Nations | JayMan’s Blog

          Likewise, it seems the blonds in North America correspond to where there’s extensive Scandinavian (and to a lesser extent, German) settlement.

          • Kate says:

            Yeah, I’ve read American Nations, studied it; it’s the work you’ve done that most interests me. One small comment if I may, being not from US I could really have done with the maps being the same scale as I don’t have instant State-recognition so I often have to trace the boundaries to compare the patterns you describe. I only mention it in case there’s an easy way to do that but I’m not tekky so maybe it requires special software.

            Interesting that folks from the north European plain ended up on the north American plain. Must be the result of what land was available when people turned up but, I wonder about people at that time deciding where to go and what factors influenced their decisions. And I wonder if the SWPL satire on camping is actually a throw back to pastoralism. It seems uncanny the way settlers took to the west. Maybe nwEuros just love a good campfire. It’s Grey’s comments that get me thinking this way, the way he tells the story of ancient migrations in a very graphic and human way. 🙂

          • Paul Conroy says:


            You may not be aware of this, but possibly 1/2 of Scots-Irish or Ulster-Scots are not in fact of Scottish descent… they’re just repackaged Native Irish… and in some cases English.

            How do I know this? Because I study Y-DNA, especially R1b-M222 and have matches in the South who are astounded to find that they are patrilineally of Native Irish descent, despite being Presbyterian…
            What you have is various people who became Presbyterian, later calling themselves Scots-Irish to draw a line between themselves and Catholic Irish, especially Famine Irish.

            My Southern relatives migrated to the US South in the period 1620 – 1750 mostly, or before the main Scots-Irish migration. I would suggest that the earliest settlers in the South were largely drawn from the Irish Midlands, both Native Irish and English settlers. The Irish Midlanders speak with a drawl, and use certain idiom that today is part and parcel of the Southern accent. It is in no way related to the Lallans accent of Scotland or Northern Ireland.

        • Sean says:

          The Ulster settlers were from where the valleys were boggy, so they tended to choose the high ground, wrongly believing it to be better farmland. Nobody drove those people off, more the other way about.

  7. a very knowing American says:

    Going back to the very early part of the expansion, and expanding on Cochran, it does look like there might have been a horse stage and a cow stage. In the horse stage, horse-riding and horse-milk-drinking warriors from the East (Khvalynsk?) conquer much of the western steppe, and set themselves up as rulers (Sredni-Stog culture), and do the same for a big chunk of the Balkans and Anatolia (which get pretty thoroughly trashed). At this stage, only ruling elites descended from the invaders are lactose tolerant. In the Western steppe, you get lactose tolerance increasing in frequency among the conquered folk, who eventually make the shift to intensive cattle pastoralism and milk-drinking (Yamnaya), and then the new tri-sectional society sends migrants to points west and north.

  8. Karl Zimmerman says:

    The bifurcated distribution of red hair strikes me as unusual. We all know it peaks in Northwest Europe (particularly in Scots and Irish) but it also peaks strongly apparently among certain groups in Northeast Europe – most notably among the Udmurts, but also to a lesser extent among the Komi and Mari.

    Unless there were two independent mutations, this seems a unlikely distribution. Of course, what’s between the two groups is the Baltic and Scandinavia – where the blondest people on Earth live. The logical conclusion thus might be that red hair was an earlier mutation, and that to some degree blond hair displaced it. This seems logical, given we know there are evolutionary disadvantages to red hair (e.g., cannot suntan, so you’ll burn to a crisp everywhere that’s not dreary, overcast, and high latitude). Indeed, according to some maps I’ve seen, the hours per year of sunshine are significantly higher in the Baltic than they are either in the British Isles or Uralic Russia. And even before the new finds with archeogenetics, there were indications that the blond hair mutation was Neolithic (or more recent) while red hair may have been significantly older.

    The question is when did these mutations arise? The presence of red hair among Uralics suggests that it existed in at least low frequencies among hunter-gatherer groups which had some ANE admixture. The blondness of Estonians and Finns might suggest the same about blond hair, but we have fairly good evidence that both groups actually had language shift away from Indo-European languages in late prehistory. So it may well be that blondism was a trait which either originated within or near the root of Indo-Europeans.

    • “The blondness of Estonians and Finns might suggest the same about blond hair, but we have fairly good evidence that both groups actually had language shift away from Indo-European languages in late prehistory”

      Eh, what? We have zero evidence of this and we have some very good evidence against it, eg. we have evidence of some non-Finno-Ugric and non-Indo-European substrate language in areas that now speak Finnish. Eg. all large lakes in Finland seem to have non-Finno-Ugric and non-Indo-European names while you can find Finnic names for lakes in Russia in areas that now speak Russian, indicating two replacements, some paleo-European language replaced by what’s now Finnish and Russian replacing various Finnic languages.

      There is no trace of Indo-European languages at all in historically Finnish-speaking areas of Finland in nature place names and all known language shifts have gone the other way, Finnic to Germanic or Balto-Slavic. Of course there is admixture from neighbors but no evidence of some language shift and plenty of evidence against some across-the-board IE substrate (why is there evidence of some other language then?).

      As for redheads, it’s a single gene (and it is the same gene) so you just need to take that out to make a ginger blonde and drift and founder effects can alter frequencies of a single trait easily. The likelihood of founder effects / drift at work is very evident in how it seems to follow the branching of the Finno-Ugric languages, red hair is notably elevated from the “average” in the Permic branch and notably lowered in the Finno-Saami-Mordvinic branch. It is more or less guaranteed that you’d see such patterns in physical traits even if in a hypothetical scenario where population splits completely followed language splits with no admixtures etc.

      • Karl Zimmerman says:

        Finland and Estonia have very high rates of lactose tolerance. In the case of Finland, among the highest in the world. This suggests that Indo-Europeans were pretty clearly part of the early population mix of both nationalities.

        Razib has posted his thoughts regarding this before. If there’s anything which seems off to you about the supposition, I’d like to hear it.

        I think it is pretty widely understood that the present Finnish population descends from a relatively small population and is highly “inbred” by European standards. The group could have been at the extreme southern edge of Finland, and encompassed a mixture of (proto-Baltic?) farmers and Finnic hunter-gatherers.

        As far as I know, the genetics of Estonians don’t even display the Siberian weirdness of the Finns, and look pretty much identical to Balts and Poles.

        • I did point out that Razib Khan was wrong on his blog. In this post actually thought the high ANE ancestry (ie “Paleolithic Siberian”) of Finns is a sign of recent East Asian origins of Finns and that the other parts of Finnish ancestry are Indo-European. This is just so wrong that you can’t possibly be more wrong and he must know it now as his latest posts are indeed about ANE as a sign of Indo-Europeans.

          As for you, what reason do we have to suppose that lactose tolerance was

          a) spread by Indo-Europeans? (if the original spread of Indo-European was connected to pastoralist inventions you wouldn’t expect them to have a high tolerance since then it would be new and not yet selected)
          b) limited to Indo-Europeans? (after all Europe presumably had plenty of languages and cultures before, maybe Indo-Europeans even picked it up from elsewhere)
          c) not under recent selection, destroying connections to languages (ie you would only need some admixture as a seed and you’d end up with another tolerant population)
          d) not already transmitted from proto-Indo-Europeans to proto-Finno-Ugrians or vice versa? or some other earlier interaction?

          (proto-Finno-Ugric has some early Indo-European loanwords for domesticated animals which might indicate some transmission of pastoralism. That might even be the original spark of the FU spread considering that eg reindeer herding seems to have been invented as an analogy with pastoralism.)

          “As far as I know, the genetics of Estonians don’t even display the Siberian weirdness of the Finns, and look pretty much identical to Balts and Poles.”

          Estonians have a high ANE component, the peak in Europe in the last study I checked, and no East Asian like component has showed up in studies so far. Finns and Saamis have some East Asian like component but it’s a small contribution and we have no proof yet of whether it’s Siberian in origin and from what time (there are after all several migrations from Asia in the archaeological record and they all could have left some genes behind).

          Words like “Siberian weirdness” have especially little meaning here when we’re talking about the big discovery that ancient Siberians and today’s Siberians are very different and that the ancient Siberians or an ancient-Siberian-like component is a big component in Europeans while a modern Siberian-like component is not.

          • Karl Zimmerman says:

            I’m not going to sit here and re-argue what Razib said, but he pretty clearly distinguished between the Ancient North Eurasian versus recent Siberian admixture within the Finnish population. His main point is since many Comb Ceramic/Pitted Ware areas now show no evidence of recent Siberian admixture (e.g., the kind that shows up on ADMIXTURE runs, not ANE), that the Siberian admixture must be post Comb Ceramic – and even post Corded Ware. Hence, it entered Finland only after Indo-Europeans did.

            Now, you may be right that Uralic peoples in general originally had no recent Siberian admixture, and that came later through groups like the Nenets (if your post is the one I think it is). But this is requires an extra migration, and I think Occam’s Razor leans towards Razib’s side without more information (we know there is DNA from Karelian hunters soon to come, so some of this will be solved ASAP).

            As to your wider points about lactose tolerance, some have merit. I do think the adaptation was such a game changer though that if it originated in the Uralic peoples we wouldn’t see Indo-European being so dominant today. The two groups did, however, have very close contact in the formative period. It may be that given the Uralic peoples were right next to Indo-Europeans as they were developing their impressive cultural toolkit they managed to adopt enough bits (both in genetics and artifacts) to remain competitive while more distant groups didn’t have a chance. Sort of similar to why more African megafauna survived, albeit on a much shorter scale.

          • On IE in Finland says:


            You are right that Finns cannot be modeled as a mixed Siberian and Indo-European peoples. The Siberian component is clearly minor and peaks in the Arctic, far from the Uralic Urheimat. Finns descend from several Mesolithic and Neolithic European migrations and some Siberian migration. It is not known to what extent the Siberian component is connected to Uralic languages.

            However, there are several proponents for a strong Indo-European presence in Finland, which was later Uralisized. Most notable this recent doctoral dissertation:
            Your opinion that there are no signs of IE in Finland is a bit radical, and I dont think it represents mainstream views.

    • Greying Wanderer says:

      “The logical conclusion thus might be that red hair was an earlier mutation, and that to some degree blond hair displaced it.”

      Or maybe added to it?

      if red hair is a partial depigmentation couldn’t one variant of blondism simply be more depigmentation added on top of the original so red hair / green eyes -> blond hair / blue eyes?

      If so you’d have a buried layer of red hair under the blond layer implying a population expansion leaving red islands (Ireland. Mari etc) perhaps in areas that were either unsuitable for a horse based pastoralism or were the last to be conquered?

  9. jamesd127 says:

    ANE redheads, brown eyes
    WHG blue eyes.,

    Pretty sure that when you mix red heads and blue eyes, you are combining different genes for fairer skin, so that the after a short period of selection for vitamin D production, the descendents of the mix will be substantially whiter than either ancestral group.

    Separate groups, separate adaptions to vitamin D shortage. Combine them. If continued selection for vitamin D, much whiter descendents.

    Upon conquering peoples less heavily selected for vitamin D, and mingling, then we get diversity of hair and eye color.

    • Sean says:

      But one gene has an effect on skin plus eyes and hair. Why the polymorphisms in this gene in relation to hair color? Moreover, “Our estimates for the age of 374F placed the origin of this allele in Europe within the last 10,690–36,070 years (with a selection coefficient of 0.0243) using the individuals from the Hapmap, and within the last 25,270–35,290 years (with a selection coefficient of 0.0127) using the individuals from the 1KGP. These estimations support the work by Beleza et al. [42], who reported that the age of the allele ranged between 8,260–31,780 years under a dominant model and between 6,188–26,964 years under an additive model,”

  10. Toddy Cat says:

    Quick, someone mention how this confirms the work of Joseph Greenberg (which is sort of does) and watch the pissed-off linguists charge into the comments, insisting that language and DNA have nothing, Nothing, NOTHING to do with one another…

    • subcomputer says:

      Greenberg’s great, his work has been attacked on several levels, and to be sure, the way he did mass comparison meant you had to be insanely careful about data pollution, but some of his work did predict some later discoveries pretty well.

      As far as linguists and DNA having NOTHING to do with language, sure current opinions are a bit of an over correction, but it’s an interesting comparison to one of my archaeology profs who would drill “Pots Aren’t People”, “I have chopsticks, am from China?” and similar rote drills about DNA vs physical culture into our heads.

      • Toddy Cat says:

        Yeah, I’m sure that Greenberg made some mistakes, but I’m always amused at how much any mention of his work drives a certain kind of linguist up the wall. Some of them seem to take the position that, sure, he was more or less right about Africa, and he was more or less right about Eurasia, and it’s looking more and more like he was right about the Americas, but he shouldn’t have been, because his method wasn’t the approved one.

        As for your old Prof, I had similar professors back in the 1970’s. I’ll bet they just hate DNA testing…

  11. Richard Sharpe says:

    I used to think the ‘c’ in centum was pronounced like the first letter of satem. I guess century confused me.

    • Greg Pandatshang says:

      The “c” in century underwent a sound change very similar to what happened in the satem languages, but much later on. The change of /k/ to a sibilant affricate or fricative is crosslinguistically very, very common. However, it is typically motivated by proximity to a front vowel such as /i/ or /e/, or some similar sound, as in “century”. What makes the satem development unusual is that it happened across the board. I wonder if this had something to do with near-ubiquity of /e/ in Proto-Indo-European: although not every word with /k/ had it followed by /e/, there would have almost always been some closely related word where it was followed by /e/.

  12. Steve Sailer says:

    It would be nice to have maps embedded in the text. Inland Eurasia is a bit of a blur to me.

  13. teageegeepea says:

    Fermented horse milk is rather sour, but the sweetness helps balance that out.

  14. Greying Wanderer says:

    I think this all makes sense except i’d quibble about

    1) “It all started pretty far to the East. There, some crazy locals first tamed the horse”

    I think it’s much more likely this occurred at the furthest edge of the farmers. As farming extends further into more marginal terrain the balance between crops and animals shifts until farming eventually tapers off and there are real word examples where ranchers on the edge of the farming zone recruit local HGs to be herders for them e.g. aborigines in Australia, Gauchos in South America. I think it’s more likely the farmers themselves were responsible for the HG -> herder transition. This would put the ground zero in East Ukraine.

    Once the transition has occurred those first steppe pastoralists could expand dramatically to the east and as there would also be a lot of room for that expansion to the east so the center of gravity of the new pastoralist population would shift east also.)


    2) “These early horsemen were genetically similar to the Ancient North Eurasians, or as those who know have dubbed them, Sibermen.”

    If the ANE as a whole were descended from the interior HGs of the mammoth steppe then they would have existed over a very wide range in the far north from the edges of western Europe all the way to Siberia with a center of gravity close to Siberia.

    If so the segment that developed a horse culture and dramatically expanded everywhere a horse based pastoralism was suitable would be only one segment of ANE. If so there ought to be surviving populations also descended from ANE but distinct from the IE ANE in regions close to where the mammoth steppe zone used to be but which were not suitable for horse-based pastoralism e.g. mountains and marshes i.e. places like the Caucasus, bits of Scandinavia, that big marsh in eastern Europe I’ve forgotten the name of etc.

    In other words I think it’s more likely the IE developed out of one segment of ANE and then expanded dramatically over the top of various other non-IE segments of ANE except in regions where a horse-based pastoralism wasn’t suitable so the source of the expansion isn’t necessarily related to the center of gravity of ANE and might be (likely was imo) from one of the edges.

    (So I still think Hyperboreans is a better label for ANE than Sibermen.)


    3) “Having horses made them natural raiders, let them expand. Lactose tolerance might have helped.”

    An interesting quote here from a Russian visiting the steppe in 1870


    “The food of the Mongols also consists of milk prepared in various ways … When it comes to “white foods” (anything made from milk), almost everything is heated due to the brucellosis problem within the country. The only thing that they commonly drink raw is mare’s milk just taken from the mare when it is still warm.”

    So maybe the benefit of lactose tolerance is proportional to the brucellosis problem in a particular region which is maybe proportional to the mixture of animal types in that region i.e. for example is the proportion of mares and cattle to sheep and goats ~ proportion of milk that can be drunk raw?

    Hence (maybe) c. 30% LP on the steppe -> sweeping to c. 90% LP in regions where milk from mares and goats/sheep was replaced with milk from cattle?


    4) Also the R1b / R1a distribution looks like two waves so i think either

    a) the Cucuteni of west Ukraine were R1b and (the IE variant of) R1a were the PIE of east Ukraine and R1a both spread unopposed to the east and at the same time gradually pushed R1b west eventually causing the demise of LBK in the process


    b) Cucuteni were standard G farmers and R1b were the PIE around the Black Sea and R1a were originally further east (around the Caspian for example) and the Cucuteni farmers turned the R1b into pastoralist raiders (and got squished by them as a result) but at the same time pastoralism spread east changing the center of gravity in R1a’s favor eventually leading to them out-numbering R1b (on the steppe) and pushing them to the west.

  15. Greying Wanderer says:

    If Vitamin D3 is needed for good eyesight (?)


    “New research from the Institute of Ophthalmology at University College London revealed striking eye benefits from vitamin D3 supplementation in older mice.”

    do you also get vitamin D3 via the eyes as well as the skin?

    If so then lighter eyes might be selected for in northern latitudes and changes in hair color could be a side effect of some genes selected for depigmentation of eye color?

    (Or alternatively do lighter eyes partially compensate somehow for eyesight problems caused by lack of vitamin D3?)

    • Bruce says:

      Interesting. When I think of dog breeds with blue eyes, I usually think of northern breeds e.g. spitz-type dogs. I wonder if blue eyes are adaptive in the far north.

    • Sean says:

      “do you also get vitamin D3 via the eyes as well as the skin?”

      Pesky facts, yes there is evidence of a certain amount of direct synthesis from indirect light bouncing UVB photons on parts of the eye. Of course, NEVER deliberately expose your eye to an unusual level of UV by looking at the sun or glare, (even in eye protection). As I understand it eye colour is not due to simple depigmentation, the colours are produced by complex Tyndall effect scattering at different levels of the reflecting surface of the eye.

      But let me ask you this, why have Europeans got brown eyes green eyes, gray eyes and violet eyes? If it was to do with vitamin D there would be just the genetic adaptation that WORKED. Why are there multiple transcript variants encoding different isoforms on SLC45A2?

      Once last time, red hair confers noticeably lighter skin while blonde hair does not, so why are blondes not less common than redheads, eh? but such fair skinned people synthesis no more less vitamin D for a given amount of UVB exposure ‘no significant correlations were found with constitutive or facultative skin pigmentation’. A large British study found actual lower vitamin D levels in the fair skinned at least partly because it seems ‘FAIR-skinned individuals also appear to be less able to make and process vitamin D in the body, regardless of how long they sit in the sun for’.

      • Karl Zimmerman says:

        My understanding is there are three things which are thought change the color of eyes.

        1. Melanin content. Can of course range from so heavy that your iris cannot be seen as separate from your pupil, to so light that the underlying structural color is evident).
        2. Lipochrome content, which is a separate yellow-brown pigment. In eyes with a good deal of melanin, some Lipochrome results in the rare amber eye color. In eyes with very little melanin, it results in green eyes, as the yellow pigment combines with the blue structural color.

        3. Underlying structure of the Iris. This is more controversial, but it’s thought the difference between blue and gray eyes is due to greater collagen deposits in the iris, so that instead of Rayleigh scattering the iris has Mie scattering. Presumably darker-eyed people have natural variation here as well, but the structural differences are hidden by pigments.

        I’ve always wondered why pheomelanin, which redheads have in their hair and skin (freckles, etc) doesn’t also show up in iris color. It may be because the iris develops from the neuroectoderm, meaning developmentally speaking it’s closer to the brain than the skin, and thus covered by a different set of controlling genes.

      • Sean says:

        Correction to last link http://www.bbc.co.uk/news/health-15151930

        Karl Zimmerman, If you can get brown eyed blondes with tawny skin then blonde hair can’t be a side effect of depigmentation for greater vitamin D synthesis. And as blonde hair is more common than red hair, vitamin D synthesis couldn’t explain red hair even if redheads’ fair skin gave them more vitamin D synthesis, which it doesn’t.

        • Greying Wanderer says:

          “If you can get brown eyed blondes with tawny skin then blonde hair can’t be a side effect of depigmentation for greater vitamin D synthesis. And as blonde hair is more common than red hair, vitamin D synthesis couldn’t explain red hair even if redheads’ fair skin gave them more vitamin D synthesis, which it doesn’t.”

          Multiple responses to the same problem?

          1) Fish -> no problem. Are brown eyes in Europe more common close to the sea?
          2) An interior HG solution to the vitamin D problem
          3) A farmer solution to the vitamin D problem
          4) side effect of combining solution (2) and solution (3)
          5) unintended consequence of hair and eye color diversity on mate selection

          (nb if (5) existed and was mostly based on novelty then it would inherently balance out i.e. as blond went over 50% the novelty effect would reverse).

          • Sean says:

            Canadian Amerindians that live on caribou which contains hardy any vitanin D are dark.

            What is your evidence for saying that light skin makes more vitamin D? It doesn’t and there are two links above one a lab study and another a large scale study of British people that prove fair skin synthesises no more and probably less vitamin D for any given amount of sunlight.

            Red hair goes with pale skin, possibly because it came from furry-as-a -bear Neandertals. But there has to be a shortage od one sex otherwise mate selection balances out

  16. Greying Wanderer says:

    After reading the comments a third possibility occurs to me. What if a transition to pastoralism and the development of a horse culture were two separate events?

    i.e. HGs on the edge of the farming zone get turned into pastoralists by the farmers. Steppe pastoralism (foot based) then spreads east until it reaches a population who mostly hunt horses and that population is the one which combines the two.

    So in effect you’d have PIE north of the Black Sea adjacent to the farmers and the actual full IE further east?

  17. IC says:

    Lots of people with strong ideology will not like your post and photos. But truth is stubborn and can not be explained away.

  18. Matt says:

    Thoughts on this:

    • Centum-Satem split effectively so Centum is Celtic, Germanic, Italic, Greek, Albanian and (weirdly) Tocharian, while Satem is Baltic, Slavic, Armenian, Iranian and Indo Aryan.

    Greg seems to have presented this as Europe, other than Southeast Europe (i.e. the children of Yamnaya / Corded Ware, I guess), as centum, other groups are satem. Seems not really the case? Instead centum binds together West Europeans and Southeast Europeans (including Armenians and Albanians), while satem everyone from Lithuania down to India.

    If anything, centum seems kind of R1b / farmer while satem seems R1a. And centum is Bell Beaker territory, more or less. Another fit would be that centum territory is EEF territory proper, while satem territory is on a gaggle of other groups. But even those don’t fit the Tocharians.

    • If the idea is that the Botai-like culture contributed to the Yamnaya, that sounds plausible. Of course, its the bit of seeing the Tocharians as then coming from Yamnaya that sounds odd. Do the Tocharians go back out East from Yamnaya past the Botai area into Xinjiang? Not impossible, but a weird reflux. And surely, the Tocharian people were a vestige of a wider family, not just a little group who lost their way. If Tocharians do this, then does that alter our assumptions more pure Botai-like people being left to contribute to India at a later date?
    • On lactase, there’s some in Iberia by the Copper Age (25% in a sample on Pontikos’ blog), but I’m not sure whether these are Beaker influenced or what. Copper Age samples in South and West Europe tend to be Oetzi type EEF so far, but who knows if these particular samples were.

    It seems like you have intensive dairying for sure among Neolithic farmer cultures, which is why the idea of the invention of pastoralism (as in, heavy use of dairying and domestic animals) made not that much sense, particularly the mixed pastoralism that is optimal for Europe. The invention of horse riding and raiding would perhaps make somewhat more sense.

    Botai culture didn’t have any cows or sheep or goats, a pure horse culture, while Neolithic farmers did.

    Apparently this also meant the Botai weren’t that nomadic, as nomadism is apparently more of an adaptation to protect animals that can’t stand harsh climates, more than driven by grazing (so the interwebs have told me), and was practiced by later herders who adopted a cattle, sheep and goat package in climates too harsh for these animals to survive (of essentially Near Eastern Neolithic origin). (http://www.carnegiemnh.org/science/default.aspx?id=16611 – Sedentary Horse Pastoralism)

    So for the Botai themselves, horse riders. But perhaps not yet with enough social changes in the direction of mobility to be much cop as raiders (horse riding raiders might not work so well if you don’t live on the move and you can ride your horse fine, and raid long distances… but then the slow moving farmers come and burn down your even slower moving village).

    • That said, apparently the early Corded Ware Indo-Europeans out of the steppe IIRC weren’t big into horse riding by this stage, more oxen and wagons. So I wonder if they switched out a horse heavy steppe package for something cattle heavy and more Europe ready, by mixing with these ancient Ukrainians, and those that didn’t would’ve lost. Nick some innovations, then get a raiding culture, but with more of a heavy infantry and cattle raiding accent (how Irish? but also rather core Indo-European). Not sure about in India.
  19. On Ancient “North” Eurasian, I think we’ll need more samples in West-Central Asia to determine whether this is really a “North” Eurasian clade or a “central” Eurasian clade that bleeds into the more populous at the edges, without really much of a exchange between them.

  20. If it’s already in west-central Eurasia, no need to putz about with separate extra dispersals from way East. Just Yamnaya plus a cline of Early Neolithic plus extra “ANE” plus ASI (whatever that proves to be) that’s already in place.