The day of the Dasa

Unless you have been sleeping today you may have noticed two important papers on South Asian historical population genetics have been published. The simple and short paper is An Ancient Harappan Genome Lacks Ancestry from Steppe Pastoralists or Iranian Farmers. The longer paper, which is basically a book if you read the supplements, is The Formation of Human Populations in South and Central Asia (and update on a preprint which came out over a year ago).

So the “Rakhigarhi genome” is finally out. She turns out to be an interesting individual: she has some, but not much, Andamanese-related hunter-gatherer ancestry, a lot of Iranian-farmer-related ancestry, and no steppe ancestry. She is very similar the dozen or so “Indus Periphery” samples found outside of South Asia, in the region’s near-abroad (Khorasan and into Turan). Her mtDNA is U2b2. My mtDNA is U2b. So my mother’s maternal lineage dates back to the IVC period. Not a surprise, but still cool.

The major finding that is of great interest is that the “Iranian-farmer” ancestry of the Indus Valley Civilization population was possibly not “Iranian” at all. That is, it seems unlikely that the West Asian-related ancestry in the IVC people was due to a migration out of the Zagros agricultural hearth. The reasoning here is simple. There was ancient population structure in the Near East at the beginning of the Holocene. There were, roughly, there major groups which expanded, Anatolian farmers, related Levantine farmers, and more distantly related Iranian (Zagros) farmers. These groups intermixed copiously during the Holocene. All the farmers of the Holocene in western Iran and even the hunter-gatherers had some ancestry from the Anatolian lineage.

Anatolian heritage is not present in the IVC people. Because Anatolian ancestry is found in Iranian hunter-gatherers at the beginning of the Holocene, the West Asian-related ancestors of the IVC people must have diverged earlier. One option is that there were a set of hunter-gatherer populations in the territory of modern Iran, Afghanistan, and Pakistan (and possibly northwest India) who were related to each other but differentiated due to distance and separation. Modern Iran is bifurcated by some rather harsh deserts between the west and the east. There is no reason the same could not have applied to the Pleistocene. In particular, during the Last Glacial Maximum.

Related to this, Iosif Lazaridis has a preprint out which argues that the difference between the “Anatolian” and “Iran” clusters lay in differential admixture with “Ancient North Eurasians” (ANE) into the latter. The non-Rakhigarhi paper above highlights the role of Turan in mediated interaction and gene flow between northern Eurasia and Iran-Afghanistan-Central Asia region. The difference between the quasi-Iranian ancestors of the IVC people and those of the Zagros, the Iranians proper, may simply be that the ANE-related admixture was stronger further east. Or not. In some ways, the paper opens up a lot of possibilities as to the landscape of late Pleistocene western Asia. It is a reasonable interpretation in the paper that agriculture was spread not through mass migration (e.g., Bantu expansion, farming in Neolithic Europe, etc.) to northwest South Asia, but through cultural diffusion. But the distribution and origin of the quasi-Iranian population need a lot more ancient DNA.

The origin and distribution of Andamese-related hunter-gatherers (AHG), earlier described as “Ancient Ancestral South Indians” (AASI), also needs more elucidation. It has long been known that the various East Eurasian groups seem to have separated very soon after 40,000 years ago. The AHG clade is only distantly related to the Andamanese themselves, who have more of an affinity with the Hoabinhian people of Southeast Asia. Though the diversity of mtDNA macro-haplogroup M is suggestive of long-term habitation of South Asia by some of the AHG, we cannot reject the possibility that they were intrusive from the east during the Pleistocene or Holocene, at least in part.

The awkward construct proposed by Indian researchers to David Reich to term the ancestral populations “ANI” and “ASI” (Ancestral North Indian and Ancestral South Indian) was to some extent a political move. It left open the possibility of deep geographical indigeneity of most of the ancestry of modern South Asians. I was moderately skeptical because I suspected the ANI was intrusive from West Asia (the Iranian-farmer and steppe migration models). These results do not support that, and it may, in fact, be the case that ANI-like quasi-Iranians occupied northwest South Asia for a long time, and AHG populations hugged the southern and eastern fringes, during the height of the Pleistocene.

What a lot of these questions need are people with detailed paleoclimate knowledge. The human geography would be much easier to infer if we had a sense of the primary carrying capacity. Hunter-gatherers tend to be very thin in desert areas, so those would serve as natural gene flow barriers. The divergence between western and eastern Eurasian populations is rather stark, so one might suppose that the Thar desert region was particularly difficult during the Pleistocene to traverse.

At some point, I have to come back to the “Aryan question.” These papers strongly point to the likelihood that the Aryans were intrusive to the Indian subcontinent.

From the Cell paper:

Since language spreads in pre-state societies are often accompanied by large-scale movements of people (Bellwood, 2013), these results argue against the model (Heggarty, 2019) of a trans-Iranian- plateau route for Indo-European language spread into South
Asia. However, a natural route for Indo-European languages to have spread into South Asia is from Eastern Europe via Central Asia in the first half of the 2nd millennium BCE, a chain of transmission now documented in detail with ancient DNA. The fact that the Steppe pastoralist ancestry in South Asia matches that in Bronze Age Eastern Europe (but not Western Europe [de Barros Damgaard et al., 2018; Narasimhan et al., 2019]) provides additional evidence for this theory, as it elegantly explains the shared distinctive features of Balto-Slavic and Indo-Iranian languages (Ringe et al., 2002).

From the Science paper:

Our results not only provide negative evidence against an Iranian plateau origin for Indo-European languages in South Asia, but also positive evidence for the theory that these languages spread from the Steppe. While ancient DNA has documented westward movements of Steppe pastoralist ancestry providing a likely conduit for the spread of many Indo-European languages to Europe (7, 8), the chain-of-transmission into South Asia has been unclear because of a lack of relevant ancient DNA. Our observation of the spread of Central_Steppe_MLBA ancestry into South Asia in the first half of the 2 nd millennium BCE provides this evidence, and is particularly striking as it provides a plausible genetic explanation for the linguistic similarities between the Balto-Slavic and Indo-Iranian sub-families of Indo-European, which despite their vast geographic separation, share the Satem innovation and Ruki sound laws (63). If the spread of people from the Steppe in this period was a conduit for the spread of South Asian Indo-European languages, then it is striking that there are so few material culture similarities between the central Steppe and South Asia in the Middle to Late Bronze Age (i.e. after the middle of the 2nd millennium BCE). Indeed, the material culture differences are so substantial that some archaeologists recognize no evidence of a connection. However, lack of material culture connections does not provide evidence against spread of genes, as has been demonstrated in the case of the Beaker Complex, which originated largely in western Europe, but in Central Europe was associated with skeletons that harbored ~50% ancestry related to Yamnaya Steppe pastoralists (18).

If you look deeper in the paper you see that the authors zeroed in on the period between 2000 and 1000 BCE for a reason. The people of the Eurasian steppe are diverse, and always in flux, and the earlier and later agro-pastoralists were genetically distinct. The Yamnaya culture lacked a “European” element that arrived on the forest-steppe through demographic reflux. The later Indo-European agro-pastoralists, such as the Scythians and Kushans, tended to have East Asian ancestry which is lacking in northwest South Asia. The particular profile found groups such as North Indian Brahmins fits best with the steppe people which were ascendant in the period between 2000 and 1500 BCE.

There is, of course, the assertion by some Indians that Indo-European languages are indigenous to South Asia. If that is the case, then they would have had to expand elsewhere. I won’t address archaeological or linguistic issues. Rather, the problem is that the spread of “steppe” ancestry in the period between 3000 and 1000 BCE across the whole zone of Indo-European speaking languages is so clear that it is the most likely candidate, and the steppe ancestry has origins in the…forest-steppe. Indian counter-arguments are not impossible but tend to be highly complicated.

To me, the more interesting aspect of the story is not the origin of the Indo-Aryans, but how they came into being into what they were as depicted in the Vedas, and later the epics such as the Mahabharata and Ramayana. Let me quote from the Science paper:

Taken together, the poor fits at both extremes of the Indian Cline imply that the Indian Cline does not represent a simple mix of two homogeneous ancestral populations, ANI and ASI. Instead, in the Middle to Late Bronze Age both of these groups were themselves part of metapopulations—relatively well represented by the Steppe Cline and the Indus Periphery Cline—that were not completely homogenized at the time they met and mixed. Most groups in India today can be represented as mixtures of average points along the Steppe Cline (we show below that the ANI fit along the Steppe Cline) and the Indus Periphery Cline (the ASI) but there are deviations from this simple model that contribute to the observed patterns.

Between 1500 and 500 BCE South Asia saw the development of Indian genetics and culture in a way that we understand it today, from the north to the south. One of the striking aspects of the Swat valley samples in the Science paper is that AHG ancestry increases over time (along with steppe ancestry). The Swat people seem to have started out a much higher fraction of IVC sorts, very high on Iranian-related ancestry. But after 1000 BCE they integrated more and more with people to their south and east. Meanwhile, in South India, groups like Nadars from the Tamil country are still about 5% steppe in their heritage, and non-trivial fractions of R1a1a is found among these groups.

There is now a good amount of evidence that the Austro-Asiatic Munda expanded into a landscape where unmixed AHG/AASI populations existed. Though the Science paper puts this in the 3rd millennium, I think the period between 2000 and 1000 BCE is more likely, since Austro-Asiatic rice farmers are found in northern Vietnam in 1900 BCE. The existence of unmixed AHG/AASI suggests to me that the expansion and dominance of Dravidian-speaking agricultural societies in much of South India in the form we recognize them today does not predate the arrival of Indo-Aryans by much if at all. Rather than thinking of Indian culture as the application of Indo-Aryan elements atop a Dravidian base, it is more accurate I think to consider them a synthesis that developed simultaneously. Though it is quite likely that the IVC language was related to that of the Dravidians, the impact of the Indo-Aryans shapes most Dravidian-speaking societies both culturally and genetically.

In fact, the Indo-Aryans themselves had changed genetically and culturally by the time they occupied territory within South Asia. They had mixed with people in eastern Iran and Afghanistan, reducing their steppe fraction, and then mixed again with local South Asian populations. The Indo-Iranian soma/homa cult may have been picked up from the culture of Bactria-Margiana.

A major takeaway from these sorts of papers is the uniqueness of humans and the integrative and panmictic power of culture. From a population genetic perspective parameters such as distance and topography matter a lot. Major ecological barriers such as deserts also have an impact. But the spread of Indo-European languages and genes is more than just a matter of diffusion. A powerful cultural organism expanded, assimilated, and in some cases integrated and synthesized, huge swaths of Eurasia. The IVC society was successful for several thousand years. But it is clear that there were plenty of AHG peoples in the Indian subcontinent while they flourished in the northwest. It was the arrival of Indo-Aryans which revolutionized things so that no “pure” AHG community exists in South Asia today.

Ironically, the sons of Indra spread the seed of the Dasa far and wide, from the Himalaya to Kanyakumari.


36 thoughts on “The day of the Dasa

  1. Is it surprising that the Rakhigarhi individual has not much Andamanese-related HG ancestry? I thought I remembered earlier guestimates of up to 40/50%?

  2. Good write-up Razib.

    I especially liked the part on the antiquity of Iranian like ancestry in South Asia, not only because it aligns well with my own views but because you have a lot more flexibility and openness on this issue and have been willing to go where the evidence takes you. Kudos to you for that !

    One thing you missed out on was the all-pervasive genetic influence of IVC people in Eastern Iran & Central Asia in the Bronze Age. It is a story waiting to unfold.

    Lastly, the Dasa were not native people in Rigveda but rather the Iranians with whom the Indo-Aryans had periods of conflict.

  3. It is intriguing how it seems that the original Indo-Aryans were the first people who played a key role in triggering breakdown of the barriers between the old IVC sphere, so to speak, and much of peninsular India (and probably also more easterly parts of Gangetic India at least). Maybe their early “barbarity” led to the breakdown of the “civilised” order of the old subcontinent which seems to have had stricter barriers between Iranian-related peoples and AHG-related peoples (ASI, for example, being considered to form as late as 1500 BC). But then “civilisation” again took over and everyone began arranging themselves into neat endogamous groups by the beginning of the first millennium AD.

  4. I think the Indo-Aryans pushed into South Asia which resulted in pressure on the old IVC elite. Some were eliminated, others arranged with the newcomers and again others preferred to stay independent and moved further down into so far rather uncivilised territory to found new reigns.

    The more the Indo-Aryan unities moved on, the more South the IVC/Dravidian rulers evaded them by moving South,until practically all of SA was civilised but isolated tribal territories.

    Since the principalities integrated the formerly independent SA HG, some might even relied on them, and established networks of trade and ideas, more AASI were integrated and mixed especially with lower social strata.

    I’m not completely sure about the IVC Iranian-related component and how independent it really was. It seems so now, but lets see which connections it had.
    From the cultural and anthropological point of view it was always unlikely that IVC was more on the AASI side imho.

    I wonder how much the characterisation of the IVC Iranian-related component changes estimates for the AASI/Andaman-related component in modern Indian people? Or not at all?

  5. From my rather rudimentary and hurried view which can later on prove so extremely embarrassing to record here, it seems that current Indians might show elevated AASI levels than previous estimates because of the seemingly very high Iranian-related ancestry in the Indus Periphery Cline samples, including the Rakhigarhi individual, according to the current estimates, and Indians might require more AASI than previously suggested if the current Indus Periphery Cline is used as one of their ancestors.

  6. Honestly, I would have thought the other way around, because if the IVC West Asian component is more independent, more of the deviation from West Asians proper could be explained by this without having to explain it by non-West Eurasian AASI ancestry. But probably I look at it the wrong way.

  7. The steppe MLBA enrichment in late historical and modern Indian caste populations may be largely due to a receding indosphere in Central Asia centered around kashmir and similar schools of sanskritic learning established during the Buddhist Era all the way to western china. There were many foreign students and scholars in places like Nalanda.

    It may have nothing to do with Sintashta or arrival of Indo-european languages.

  8. The steppe MLBA enrichment in late historical and modern Indian caste populations may be largely due to a receding indosphere in Central Asia centered around kashmir and similar schools of sanskritic learning established during the Buddhist Era all the way to western china. There were many foreign students and scholars in places like Nalanda.

    no. they tested samples from iron age and historical turan. these individuals are genetically different. see page 260 of the supplements.

    please read the whole paper if you care (also, the admixture decay would be somewhat later)

  9. The only problem with this analysis is that it is based on a single sample in the eastern most periphery of ancient Indus Civilization (Rakhigarhi). Ancient Indus Civilization was most likely racially very diverse based on regional (north/west vs east/south) and class/caste (slaves/migrants of from east/south) variants

  10. True, but the part which matters most is that this Iranian-related, probably long time local population founded and dominated the IVC. The single samples importance lies in the proof for the outliers further North to be with high certainty representative for IVC. Because of that there are higher resolution samples and they are fairly diverse in their ancestry proportions of the two main components.

    Anything just fits in and it was always clear that the influences, cultural and/or genetic on the IVC came from the West but had its unique character. Just not the details.

    Steppe/Aryan had no significant influence on IVC and AASI a secondary one, most likely comparable to WHG influences on early European farmer.

    These aspects being now confirmed.

  11. Razib, would you say this makes the Elamo-Dravidian hypothesis much less likely? Hard to see how it would work unless Elamite speakers came from the east at a very early date.

  12. @Arun, regarding your question, I would guess something’s may be screwed up in Shindre’s Fig2C implementation of the Harappa genome, I6113, to the model from Narasimhan et al, which models individuals as a mix of Indus_Periphery_West (11% AHG, 89% IranN), Steppe_MLBA and residual AHG. To explain at length:

    That graphic then suggests about 19% AHG, in their graphic (about 9% closer to the AHG pole).

    However, the direct (and much simpler) two-way models they implement give: “The only fitting two-way models were mixtures of a group related to herders from the western Zagros mountains of Iran and also to either Andamanese hunter-gatherers (73% ± 6% Iranian-related ancestry; p = 0.103 for overall model fit) or East Siberian hunter-gatherers (63% ± 6% Iranian-related ancestry; p = 0.24)”, e.g. 27% Onge-related to 37% East Siberian HG related (average 32%). A higher level of ENA than presented in Fig2C and probably about 1/3.

    Now, as a check on that, if I plug the measures of red “East Eurasian” ADMIXTURE component and PC positions from Narasimhan’s data explorer (!/
    vizhome/TheFormationofHumanPopulationsinSouthandCentralAsia/AncientDNA) into simply scaling the red component and Eurasian PC1 (which correlate almost perfectly) into a measure using Narasimhan’s estimate of AHG variation in InPe samples (11-50%), then that suggests that I6113 has about 33% AHG. That matches a bit better with the two-way estimates I think.

    Still, it does seem relatively low compared to what I was expecting.

    When Reich was on the lecture circuit in 2018, he presented a clinal graphic modelling Indus Periphery as Turan+Steppe+AHG – (from Alexander Kim’s twitter).

    That shows one Indus Periphery sample close to ASI at about 30:70 Turan:AHG, and the least AHG shifted samples at about 20%. (From this I recently personally extrapolated quite a few reasonable possibilities!). Somehow that seems to have got cut waaay down to 50% AHG at max and 11% minimum.

    So it seems like AHG is generally lowballed in all these samples compared to what Reich was presenting last year. I’d like to see some independent checks on these when the data are all examined by bloggers, to try and understand what went on there.

    I’d add though, these lower AHG numbers seem pretty inconsistent with some qpAdm modelling by Chad Rohlfson from last year – There he models (with a very good fit!) a sample SIS3 as 73-79% Irula, which implies 50-52% AHG, assuming about 66% AHG in Irula. However, this paper only seems to present the same sample, labelled as Indus_Periphery_9, about 40% AHG. So there may be some systematic lowballing of AHG here…

    Likewise sample, I4411, who Rai was touting as a relatively Irula-like Harappan sample, is not in Shinde’s paper…

  13. The Elamo-Dravidian hypothesis was always not generally accepted. But probably we deal with distantly related branches of Iranian-related ancestry, too deeply diverged for an easy to reconstruct root? But I see little chances for Dravidian being anything else but the language of the IVC. Yet other scenarios are still not impossible.

  14. @Matt: Probably thats because the new West Asian component of SA Iranian-related ancestry eats up a lot of the deviation from other West Eurasians in their model which was put to AASI/AHG before?

  15. @obs, I’ve had a look at a few more things today, and I’m not as convinced as I was that there’s any lowballing on the Indus_Periphery ranges – probably only by a 3-4% percent and not the larger 5-7% I had guessed. Although I’m not sure if their figures actually reflect their ranges too well, but that’s another story. I’m still fairly convinced that Rakhigarhi either has to have about 34% AHG though. (Otherwise else she is on a very different cline than the Indus_Periphery samples, and that opens a massive can of worms, that would require somehow massive migration of Indus_Periphery like people from somewhere in the Turan into South Asia to resolve, since their models of South Asian population history don’t work without Indus_Periphery like populations).

    I’m not totally sure I buy their argument of a diverging ancestry component that splits from a clade of “Iranian herders” (previously called Zagros Neolithic); but I think since it would have split relatively shallowly as a clade from IranN, I don’t think this *should* greatly affect how AHG would be estimated.

  16. Re; genetics and the spread of languages, I would say the linguistic and archaeological issues do seem pretty paramount, and really insurmountable for South Asian Indo-European from what I know in a limited way.

    Lots of genetic issues seem more resolvable by assuming that language can outrun genetics quite easily when dealing with a cultural innovation complex linked to a particular language which “goes viral” between dense populations, which is not necessarily a questionable assumption (or is it?).

    If A only changes B’s genetics 20%, but it’s language 100%, then B does the same to C, then you’d have an expansion that doesn’t have much correlation to the genetics of A, even if that’s where it originated. And that’s seemly fairly possible if A is not particularly large and a split off moves into a large B, which splits off again bordering a large C, but there is still some compelling drive to change languages. (Something like this was essentially the argument by steppe hypothesis proponents, back when we didn’t have adna that seemed to support that hypothesis, after all?)

    So I don’t think language spreads which are orthogonal to population genetic shifts are necessarily doomed by genetics. The problem for linguists who do propose these sort of models is to show them decisively enough.

  17. Matt,

    There was already a lot of evidence from modern DNA studies which indicated that the Iran N like ancestry in South Asians dates to the end of LGM.

    Unfortunately, people today are too busy into modelling of populations and they don’t pay attention to data that is steadily being churned out on modern DNA.

    I have written a blog on it where I explain in detail how the Iran N ancestry in South Asia is quite old.

    What the Reich team has done now is produce another line of evidence to support what was already evident from previous studies.

    Your opinion counts to little if you keep ignoring this data.

  18. Re; AASI in Austroasiatic groups in India, it does seem somewhat naive that the paper states, of their admixturegraph: the fact that some Austroasiatic-speaking groups in South Asia (e.g., Juang) harbor ancestry from a South Asian group with a higher ratio of AASI-related to Iranian farmer–related ancestry than any groups on the Modern Indian Cline, thus revealing that groups with substantial Iranian farmer– related ancestry were not ubiquitous in peninsular South Asia in the third millennium BCE when Austroasiatic languages likely spread across the subcontinent.

    Or, they just got in some extra AASI (relative to Dai) without Iranian farmer related ancestry in SE Asia before entering into S Asia (wherever we place the boundary). Which we know from SE Asian adna they did. (Sure there may have been AASI ancestry unadmixed with Iran N related ancestry in S Asia until more recently than many would imagine. Not an unreasonable contention. But the Austroasiatic groups relative excess of AASI:IranN related obviously don’t really provide evidence for this as they seem to claim.).

  19. Quote: ‘There is now a good amount of evidence that the Austro-Asiatic Munda expanded into a landscape where unmixed AHG/AASI populations existed.’

    Question to Razib. You mean in 2000BC there exist 3 groups in India in addition to Iranian IVC people.
    AHG who come first from Africa and whose descendants are Andaman tribals today.
    AASI who lived in South India and are majority contributors in todays Adivasis in mainland India.
    AAM (Munda) people who come from southeast Asia with rice crop sometime from 3000-1000BC

    I have heard earliest migrations take place into India 60000 years which must be the AHG peoples who speak Astro Asiatic language.

    1)When do AASI arrive in India? And what language do they speak?
    2)Can the steppe migrants bypass IVC and interact directly with (AASI/AHG +- Munda) before IVC decline?
    3)As we see Vedic religion changes itself significantly in later times by praying to Trinity gods and almost humilating Indra as powerless in front of trinity, how possible it is that the *new* brahmins are actually IVC people with aryans mixture who have hijacked the religion and now dictating the later text after Rgveda after having learnt IE languages? I mostly mean the panch dravida/ south brahmins here.

    Some questions could be amateurish or seem like conspriracy material but I dont have a deep understanding other than casual reading on internet. Sorry for that.


  20. AHG and AASI are different terms for basically the same. These South Asian HG are Andaman-related, which means the Andaman islanders are the closest, largely unmixed group still alive.
    But similar populations were once more widespread as you can see with Negritos and Veddoid rests throughout SEA.

    So the Indian AHG element is just related, not necessarily identical to Andamans. To this might also point the relationship to more Northern Asian population recently hinted to.

    Basicaly I would call this AHG indigenous jungle inhabitants, whereas Iranian-related ancestry came at an unknown date from the West and steppe ancestry from the North-West much later.
    East Asian came from the East obviously and the introduction of rice might have helped West Eurasians a great deal to spread genes and higher culture to the tropics.
    This is true for Iranian and steppe people, Indo-Aryans and Dravidians alike.

    The Indian climate is not perfect for the Near Eastern crops and pure West Eurasians, the latter especially because of the diseases.
    So the combination of mixture with locals and organised mass cultivation of rice might have been crucial for trespassing the biogeographical barrier for West Eurasian culture and genes, at least in an altered form.

    Thats why the Iranian element was concentrated in the North for so long. Only under pressure and with new tools at hand they moved deeper into the more tropical zone.

  21. Another innovation coming in with Indo-Aryans was lactase persistence and raw milk consumption from probably more productive cattle breeds (?) together with more milk products. Probably the cult of the cow was to prevent people from slaughtering animals which were more useful as milk producers? Is there any such theory or explanation around?

  22. Rakhigarhi sample not having much AASI/AHG ancestry means South Asian people are turning to be more western Eurasian shifted than previously thought.

  23. @Kevin

    Makes sense. Reich says max AASI is 55%-60%. If I recall correctly, they were originally thinking 80%? If so, quite the reduction.

    And even now, the AASI estimates might prove bloated still. I mean let’s think about it: the West Eurasian ancestry of Indus_Perhiphery is being defined in relation to Iran_N. But, it’s quite clear that the West Eurasian element unique to South Asia and surrounds is distinctive… the same way CHG is far from being identical to Iran_N. So, even that supposed 10% AASI IN Indus_Perhiphery_West could merely be a function of said distinction (differential affinities in deeper ancestral connections) between IN_P_W and Iran_N. So IN_P_W might possibly end up being wholly West Eurasian (I mean, it already is; 10% isn’t much. But I mean even lower… like 2%).

    On top of this, there is considerable genetic divergence between AHG (Onge, Jarawa, etc.) and AASI. Like serious differentiation. We really have no reason to assume that the AHG percentages will correspond exactly with actual AASI percentages (although the relative levels must be accurate?).

    We really need some Mesolithic (or even earlier) aDNA from Sri Lanka.

  24. Why Sri Lanka and not mainland South Asia? The closer to the Northern population centres the better. Sri Lanka might be more special.

  25. One thing that has not received much attention is the fact that BMAC has been taken off the table as possibly ancestral to the people of the Indian Subcontinent. Prof. Michael Witzel, of Harvard among others, believed that the “Aryans’” in their invasion/migration into India used the BMAC as a stepping stone. Such theories now need reworking.

    The Reich lab people previously believed the Iranian-related ancestry in the Indian Subcontinent was due to a mass migration of Iranian agriculturalists. However, mtDNA data from almost 20 years ago showed that differences in several mtDNA lineages between Iran and India dated to before the LGM. I wrote a comment regarding this in their bioRxiv paper to this effect and am glad to see that now with autosomal data they have pushed back the common origin of Iranian-related ancestry in Iran and India to before 10.000 years ago.

    It is now seen to be possible that in Harappan times there may have been unmixed AASI in the East of the Subcontinent as well as people with unmixed Iranian-related ancestry in the far west. Narasimhan has identified 8 samples in Shahr I Sokhta and 3 in Gonur to have been migrants from the Subcontinent. All of them have detectable amounts of AASI and that is a good criterion to use. But there were probably other migrants but with very low amounts of AASI. It is possible to identify other migrants from the Subcontinent using data they have published. From the 2018 pre-print, we have Sarazm_EN and Gonur1_BA_o. Both of these show high allele sharing with Birhor (Figure S3.10 of the Supplement) and therefore are likely to have been Out-of-India migrants. They have relatively high ANE and relatively low Anatolia_N. From the 2019 Narasimhan paper, we have Aigyrzhal_BA which has high allele sharing with AHG (Figure S 19 of the Supplement). This sample from the IAMC in Krygyzstan also has high ANE and low Anatolia_N. It should be possible to model present-day Indians as a combination of “Indus_Periphery”, “AHG” and “Aigyrzhal_BA” without using Steppe_MLBA.

  26. The steppe-like influence is directly from Eastern CW descendents. There is no way to create this without the European migration.
    Actually the new data made the Indo-Aryans more European because they were not from Yamnaya.
    You can see an almost mass production of chariots and bronze weapons in Sintashta, a highly developed warrior culture. Its also intriguing that the lactase persistence variant is European and the rates are the highest in NW India.
    Distract milk and rice from most predominantly West Asian groups diet living in the more tropical parts of the subcontinent and its obvious how big the impact of the new subsistence was.

    Anyway, you dont get around steppe ancestry.

  27. @wisecracks: the same way CHG is far from being identical to Iran_N

    Although Harvard seem weirdly resistant to acknowledging this – case in point in this paper is a couple samples from Dhzarkutan in the Bactria Margiana Archaeological Complex who pretty much match to the Bronze Age cultures of the Caucasus, or even more so, to unsampled individuals who probably existed at this time and are close present day North Caucasians (i.e. largely of Maykop related ancestry, with some Yamnaya related ancestry). These are UZ-JAR-021 (I4901) and UZ-JAR-022 (I5608).

    They’re dated relatively early in the Dhzarkutan cultural history (2100-1800 BCE), and interestingly have catacomb burials, which may (perhaps naively) invite some comparisons with the Catacomb Culture in the steppes which is roughly 2800–2200 BCE ( They have some obvious shared drift CHG which populations only these Caucasian populations have.

    But as Narasimhan’s paper doesn’t treat CHG as a distinct element (and in this it follows from the odd lapse in Reich’s magnun opus where the CHG aren’t mentioned). So it fails to pick up on this and labels them as steppe-related samples.Which is a shame, because these two females might tell us that you have an interesting kind of long range migration to which the Caspian Sea wasn’t always a hard barrier. Rather than proposing that BMAC interaction was simply a question of only an Iranian related pool with Central Asia interacting with the steppe, Central and West Asia, excluding flow from the Caucasus. (BMAC has a main Iranian related pool really seems resistant to change over time, but outlying people are flowing in from all directions).

    Related point on this paper, but more in praise of it, it’s really interesting how this paper tells us that Central Asian / West Siberian and Western Steppe/NE Europeans populations were interacting.

    You can see that the Western Steppe/NE Europeans related to Corded Ware must have demographically been a preponderant expansion, however also when we see individuals with Central Asian / West Siberian ancestry show up in grave complexes, these are in no way culturally “backward”.

    For instance, the sequence of 5 samples associated to the Potapovka Culture sampled to date, which proceed the Steppe_MLBA culture, are quite heterogenous, and in that context, the sampled individual that is earliest and most clearly associated with a rich burial and with chariot burial has a heavily West Siberian related degree of ancestry (though he is also R1a of Corded Ware derivation). While the other 4 samples are a mix of West Siberian, Yamnaya (and y haplogroups to match, when male), with only one sample being really similar to the later Steppe_MLBA set. Central Asia and West Siberia must never have been “backwards” in this period (see early and advanced Seima-Turbino metallurgy) but probably was demographically overwhelmed over time?

  28. @Matt: Demographic growth is just one factor, but there are many others playing in. That a group of people was in theory technologically more advanced never prevented them from being conquered. Especially not in earlier, prehistoric times.

    If another people were physically stronger, acted more aggressive and like one unit while using superiour tactics, what difference would it make if your spearpoints were of better quality?

  29. Possible, but some would have to look at how you’d evidence a sort of “militaristic culture” thesis. Technology is reasonably hard (e.g. you can say that bronze could have had reasons to be more advanced coming from a Central Asian province as that’s where the tin was, rather than arsenic that tended to kill “smiths”, and so prospectors would’ve gone there early) and it may be possible to estimate population sizes from adna which is again hard science. I suppose you could even say that estimating physical strength is possible (though you would need direct skeletal data adjusted for elite grave bias, I doubt that essentially North Eurasian derived Central Asian people were smaller or anything in this case). While estimating a kind of psychological / espirit de corps etc advantage in warfare seems kind of liable to lots of soft interpretation that falls into line with national myths and ideologies so on.

    That’s even the case when you have literal historical evidence to fall back on, and many of these people have never spoken to us through writing and never will.

  30. Seima-Turbino seems to be more like a today’s corporation than a tribal entity or culture. They are not just skilled bronze artisans, they also feature bone / horn / wood processing artisanship from the Siberian hunter-gatherer toolkit, so it’s but impossible not to conclude that they were a company of the most skilled artisans from different ethnic origins. And their business model seems to have included long-distance winter travel on frozen rivers, so just a few unusually snowy winters could have put them out of business (and indeed they disappeared just as quickly as they emerged).

  31. You can reconstruct how a people were organised and how many were trained in pastoralism, hunting and war to a certain degree. For example, Germanics had much higher rates of free warriors than late Romans, so a fairly small Germanic tribe could muster many more trained warriors than the Romanised civilian population with its rather weak militias, once the legions were gone. Same goes for a lot of clanish and tribal people, or religious fanatic vs. more sedentary and state oriented people, but generally more peaceful people. Even to this day, if you look to Afghanistan for example.

    An example of relevance in prehistory is the overtake of Neolithic populations by Northern European hunter gatherers, like Reich said, from their protected Northern zone, were early farming was less productive. The HGs were healthier, more robust and it seems they took first just some farmer females, abducted them to their clans. Then some of those females started to make pottery in their “new home”, and the deeper they thrusted into Neolithic territory, the more farmer-like they became genetically and culturally. But not like those they conquered, but more specialised on war and husbandry, while the females taught them cultural techniques and field work, planted crops.

    I’m pretty sure they had, early on, zero demographic and even less so technological advantage, but were just more mobile and aggressive, healthier and better trained hunters and warriors.

    Historical cases are known from the Near East, were one wave of pastoralists with zero technological advantage after the other conquered the city states and formed the Semitic states on the base of the more advanced Sumerian culture.

    Obviously there could always have been a minor technological advantage, like a better composite bow or better bred horses and gear, but the point is, general cultural development was not that important or advantageous, but rather esprit the corps and warrior qualities. To a certain point of course. Against machine guns and aircrafts its hard to win with arrows…

    By the way, trade and war were quite often interconnected, because trade made a region, people and their qualities, as well as weaknesses, known. And of course disputes over trading are some of the most common quarrels.
    Its not by chance early Indo-Europeans and Vikings for example were both traders and conquerors, depending on the circumstances and chances they saw.
    Romans allowed only friendly tribals to trade inside of their realm, to visit their urban centres. The hostile ones shouldn’t know too much.

  32. Wasn’t there a downvote feature on this site earlier at some point? I am so itching to put a loving downvote to my wonky and cringeworthy comment about revised AASI estimates for Indians after all these days of partial meltdown following that comment, but it seems the feature is not there now.

  33. “Its also intriguing that the lactase persistence variant is European and the rates are the highest in NW India.” —- The same LCT variant is also present in Toda people of Nilgiris(South India) and as far as i know, they speak a dravidian language different from Tamil.

  34. @td
    The Toda are an interesting case indeed, because they are living further South, are pastoralists phenotype wise on the more Caucasoid side and speak a Dravidian tongue. They breed buffaloes and are heavy milk consumers which hold the animals in high esteem,even sacred.

    Nobody should overestimate the importance of language in India, because of different phases of internal colonisation and intermixture.

    Additionally, Indo-Aryans introduced the lactase persistence variant from Europe and an intensified milk based husbandry and diet.
    But a lot of its spread and selection was happening inside of India. Dont forget the same happened in Europe.

    I cant recall a genetic study done on Toda right now and whether they looked at lactase persistence.
    But I’m confident geneflow from Indo-Aryans took place.

    But of course, in theory it could be older.

  35. I believe “dasa” was the term given by self-identified aryans to all the people that they defeated

    Which means, if a bunch of Aryans defeated other Aryans then the victorious Aryans referred to the defeated Aryans as dasas — and pushed them further east (assam) or south

    It is a case of history being written by the victors

    I am not sure, if there are any documents of battles between Aryans & “natives”. So the original dasas referred to in the vedas might all be defeated aryan kings

    Maybe the incoming aryans were so technologically advanced that their serious fights were only with other Aryans who had come before or after. I am guessing, the native population probably formed alliance with the aryans and took sides in the battles

  36. Well, the battles with the “Dasa” are peculiar in many ways and differentiated from intra-Aryan disputes. Weapons and strategies for example, not just physical appearance.

    It seems to be quite likely that some local elites were assimilated or integrated into the new Aryan power system early on.
    Probably thats how Dravidian survived? As an ally of a powerful Indo-Aryan unit?

    Anyway, I always wondered about the description of the Dasa, which is clearly pointing to a Veddoid phenotype rather than what I expect the average IVC people to look like.

    So probably you are right insofar that they made the most foreign, most extreme phenotype they encountered “the Dasa”, the local hostile people in their myths, even if a lot of the real people they met weren’t like that.

    In tales about the enemies the most extreme, foreign and negative traits being often overemphasized, while the similarities being downplayed. And its quite likely they encountered some extreme Veddoid individuals rather sooner than later when entering South Asia, even if they were a small minority.
    Probably they met the phenotype even before and/or heard stories from BMAC or other people they communicated with before moving in.

    In the propaganda people often create a caricaturesque appearance of the enemy.

    Because Onge-like people will for sure have been rare to non-existent in the North of India, but the description goes in that direction.

Comments are closed.