23andMe says Bangladeshis are more Bengali than West Bengalis!

As some of you may know 23andMe updated its South Asian ancestry panel. On the whole, I’ll give it a thumbs up, but, you need to be aware of the way they’re framing things. For example, pretty much every Bangladeshi has more “Bengali” ancestry than people from West Bengal.

The profile above on the left is mine. On the right is a friend whose background is West Bengali, of the Kayastha caste. Basically, 23andMe seems to be taking the East Asian enriched ancestry of Bangladeshi Bengalis as more diagnostic of being Bengali.

Now, compare me to a Bengali Brahmin (on the right):

So in all likelihood, Tagore’s ancestry composition would result in not so much “Bengali”….


106 Replies to “23andMe says Bangladeshis are more Bengali than West Bengalis!”

  1. Wasn’t Tagore born in present day Bangladesh? Also half of his family was married to bengali muslims, so perhaps he was at least more bengali than W-Bengal folks, according to this 23 and me.

  2. Steppe proportions, if taken out of context, are a confirmation bias seeking race realist’s wet dream.

    There sure is a lot of circumstantial evidence that if looked at in an unnuanced way, could be used to frame an argument in support of perpetuating the casteist order of S Asia and Bollywood non dalit Punjabi worship.

    Someone once argued on one of those crazy blogs that the soul of a man is through his Y DNA. That R1as are destined leaders. Maybe I should give on life as an HM52….

    S Asia was once a paradise of Sanskritic great builders of civilization. Sadly, they mixed and left off an inferior group of mutts. Thank god though were striated in a feudal order encoded in religious doctrine. It has at least slowed the destined failure….

    1. “S Asia was once a paradise of Sanskritic great builders of civilization. Sadly, they mixed and left off an inferior group of mutts. Thank god though were striated in a feudal order encoded in religious doctrine. It has at least slowed the destined failure….”

      This, I can only hope, are the views you are deriding rather than endorsing.

      But desis being desis, I can see the potential for future Y-dna or mt-dna name dropping, rank pulling based on “steppe” proportion, and similar antics.

      In fact dont be surprised if you soon have dna based astrological advises and matrimonial services 🙂

      1. sarcasm

        But I agree. This shit is next. Anything to worship the fair skinned, “sharp featured,” intellectually, and morally superior overlords.

        Hail Indra. May his light save the noble Kashmiris from both their false abrahamic ideologies but most importantly the invading black apes that only overpower the noble last remaining caucasoids of the North via shear numbers and exploitation of the pathologic altruism of the Sky Father’s people, those who bestowed upon the savages of the subcontinent the opportunity to ascend from their fecal huts and hunter gatherer way to at least serve the ubermensch.

        Jai Shree Ameen

    1. My guess is

      Southern South Asian–> Tam Brahms (interlopers from the North)
      Southern Indian —>Pure Dravidian AKA “the best people” 😛

  3. I’m a Rajasthani Jain and clustered most closely with Southern/Tamil Brahmins in the Harappa Ancestry project. Here is my “new” breakdown per 23andMe:

    Central Asian, Northern Indian and Pakistani: 66.7%
    –> Northern Indian and Pakistani: 48.4%
    –> Bengali and Northeast Indian: 5.9%
    –> Broadly Central Asian, Northern Indian and Pakistani: 12.3%
    Southern South Asian: 16.6%
    –> Southern Indian & Sri Lankan: 15.5%
    –> Broadly Southern South Asian: 1.1%
    Southern Indian subgroup: 6.9%
    Broadly Central & South Asian: 9.8%
    Unassigned: 0.2%

    I’m confused about the difference between Southern South Asian (16.6%) and Southern Indian subgroup (6.9%), and also broadly Central Asian and North Indian, Pakistani (12.3%) vs. Broadly Central & South Asian (9.8%).
    Razib, or someone else, can you help explain?

    1. Amit Shah is the most famous modern day Jain guju vania.

      We hope he too becomes fuhrer one day. Long live the 3rd Rashthra

      Hail Modi

  4. LOL, Indian-american Malayali-christians finally getting some type of recognition:

    From 23&me:
    Malayali subgroup:
    Home to the Malayali (meaning “people of the mountains”), Kerala has been at the heart of a thriving international spice trade for millennia, becoming one of India’s most diverse and prosperous states. This genetic signature reaches high levels among Christian communities who live in Kerala, a large number of whom immigrated to the United States during the last century.

    Much to probably his horror, the Bengali brahmin is 1.5% Malayali… 🙂

    1. “Much to probably his horror, the Bengali brahmin is 1.5% Malayali”

      That could be the reason why both of u elect similar parties. 😛

  5. Just checked my updated 23andMe profile. Why, oh why, did they label South Indian Brahmins the Southern Indian subgroup?

  6. The component is labelled ‘Bengali and NE India’ rather than Bengali with the following populations contributing to the cluster: ‘Assamese, Bangladeshi, Bengali, Khasi, Manipuri, Meghalayan, Nagalandese’.

    In a sense, it would be mo surprise that less East Asian enriched Bengalis would score lower of that component given their geographic location ie West Bengal.

    Bengali (Kulin) Brahmins in particular would be scoring more UP like than generic Bengali given how they plot on PCA and historical migrations.

    What I find surprising is how Bangladeshis can score with that component given they make up a core – my family – Bengali & NE Asian from 55% to 75%, N Indian & Pak 5% to 25%. S Indian 0% to 10%.

    The few Bangladeshi Hindu samples I have seem to score like Bangladeshi Muslims suggesting that in non-Brahmins, it again boils down to a mostly geographic issue. Also, aside from the numerous Sylheti samples, there are a few Dhaka, Barisal and Chittagong samples too.

    This new version seems a bit noisy though – alot of random trace ancestries popping up.

  7. 23andme does a decent job of identifying the major components of ancestry. I am a regular NI Hindu, and they fairly accurately put me in a group of 4 indian provinces I must come from.

    But I must say I have to take their “trace ancestry” components with a pinch of salt. In their earlier analysis, they attributed 0.1% of my ancestry to Japanese, and similar percentage to Finnish. They even boldly predicted that I had a full blooded Japanese ancestor who lived in 1700s.

    That of course left me scratching my head, but even before I could discover any skeletons in the closets of my family, they quietly dropped the Japanese component.

    In their recent updated analysis, they have done away with the Finnish component too. But in its place, now I have two brand new ancestry components. 0.1% of Arab/Levantine, and 0.1% of – of all places – Somalian.

    What is sure is that I am not gonna bug my family elders to tell me the story of our secret Somalian pirate forefather.

    Is it genuine margin of error, or bit of a sensationalism to entice more customers?

    1. Haha my brother and I got those weird Finnish and Somalian traces too with the new beta updates. The Somalian wasn’t too out there but now it’s disappeared anyways. The Finnish is just plain wonky though.

  8. justanotherlurker and thewarlock, will you care to specify what districts your ancestral villages were located in?

    also, is it possible for you to send me your raw data file. needless to say, you can delete all identifying details.

      1. justanotherlurker

        i am just interested in jain genetics. especially jains from rajasthan who have an oral tradition of descent from rajputs. i am just curious and seeking genetic confirmation of this claim.

        1. The tradition of being descended from Rajputs is interesting to me too. My community has the same (and we’re Marwari Baniyas). I wonder if there is any validity too it. I’m almost certain we have Jains in the family too since intermingling between Marwari Jains and Baniyas is frequent.

          1. saurav:

            True, all sorts of claims across many communities about descent. But it seems several are actually proving to be right (like the claim of Jats about more recent Central Asian descent etc..) now through genetic analysis.

            I don’t think claim of Rajput descent is that common in Gujarat – Rajputs don’t have the same level of prestige there since they were kicked out of power in the 13th century vs. in Rajasthan where they continued to rule until independence.

            As for Rajasthani Jain claims, they could be real or aporcyphal or partly true…Hard to say at this time. Some supporting points are: a) Gotra/clan names that are even now clearly Rajput – for not all but many Rajasthani Jains. Parmar (of Raja Bhoj of Malwa), Chauhan (of Prithviraj), Solanki (of the last ruling Rajput dynasty of Gujarat with their capital in Northern Gujarat very close to current border with Rajasthan) and Rathod (of the current Jodhpur Royals). b) Numerous recorded documents mostly by Jain monks going back to 9th-10th century CE c) Rituals and traditions especially marriage rituals that are closes to that of Rajputs d) Hierarchy in villages/cities and close collaboration between Rajput elites and Jain elites (Bhamashah is well known in Rajasthan as the benefactor of Rana Pratap)
            Of course many of these things could be through emulation or diffusion given the prestige Rajputs had in Rajasthan..So we will see what genetics says..(there are confounding factors here though: Did Rajputs merge with other groups or accept new members since some of them split off and became Jain; The almost certain mixing of Rajput origin Jains with other Jains of othercaste origin (mostly Vaishya but also some Brahmins – Kashyap is another gotra that I have come across).

            And finally, no one in the community actually takes overt pride in Rajput origins..The few times it has been stated it was mentioned in a matter of fact way, and not to tout superiority..Funnily enough, Rajasthani Jains have a pretty strong superiority complex based on economic success (of the elites but the halo extends to the community) and what they perceive to be the superiority of their religion (on elite markers of India such as Ahimsa, vegetarianism, charity etc.)

          2. I dont know man. Anyway i dont think you need to provide genealogy and such to prove anything nowadays. Like Pakistani rajputs have none of it and still claim to be rajputs. In UP you have folks called lodh-rajputs who are both OBCs and rajputs (apparently) . Go figure

    1. yeah but don’t know your email.

      Father side from Kalol, village near Ahmedabad

      Mother side from near Rajkot

      Half Guju Half Saurast Thanakvasi Jain

      mt U8b’c y H

      Also, I am non-religious. So using this more as proxy to trace ethnic roots and not much else. I hope no one is taking my joking seriously.

      1. U8b’c huh? That is interesting. Based on what I know U8 (xK) is rare as of now. IDK about the U8b’c distribution or its exact origins, but I won’t be surprised if it was brought to south Asia by Iran-like HGs.

      2. thewarlock

        thanks for sharing info. just dabbling in genetics amateurishly.

        you can send the raw file to the temporary gmail address deepa.1974.deepa

        thanks again

    1. I look forward to meet new cousins and to further enlarge my extended family. Pls include the previously mentioned 5 Sacramento rajas into the project while there are still in front of the Taj Mahal.

    1. That’s something about Hilsa fish, correct?

      On a more serious note, our conception of West Bengal as Bengali may have had to do more with the late 18th century.

      It was a transitional region between Bihar and the Bengali heartland, and I doubt that any standardized vernacular existed. The Battle of Plassey, administrative impositions, printing press and education probably led to a language shift in the elites by the 1850s and the masses by the 1900s. Even by the 1905 partition, it seems that the idea of West Bengal being “Bengali” only held with the upper levels.

      1. It was a transitional region between Bihar and the Bengali heartland, and I doubt that any standardized vernacular existed. The Battle of Plassey, administrative impositions, printing press and education probably led to a language shift in the elites by the 1850s and the masses by the 1900s. Even by the 1905 partition, it seems that the idea of West Bengal being “Bengali” only held with the upper levels.

        is this true? i am sure there has/is ia dialect continuum.

        1. Some languages, like Maithili, are remnants of a dialect continuum. However, the reality is that “Hindi” has over time become the language of the middle north, leading the people of Bihar to look westward. The printing press brought Devanagari, and replaced what were otherwise scripts fairly similar to Bengali. The same happened from the East in what is now Bengal.

          Between modern education and mass media, these dialects are dying out. Manbhumi is a remnant of a middle language close to Maithili, but due to pressure to either assimilate to Hindi or Bengali, it is dying out.

  9. How on earth they labelled Bengali and Noetheast India in the same category? Northeast Indians like Khasi have only 15% Western Eurasian vs Bengalis 60% Western Eurasian. 23andme is becoming a joke.

  10. Rajput claims are often a load of horseshit. And what Jats are not central asian lol. A group of Jats, eastern Jats, specifically Rors, is shown to be the most steppe. There is some suggestion of some input from Indus periphery areas of Pashtun like people mixing with some indic people, resulting in Jats, before the group for whatever reason moved into the eastern Indus and Western gangetic plains.

    But again, that is only an input for their ancestries. Their wild claims of being “Kangz and shiet” of Central Asia and being descended from exclusively noble warrior pure caucasoid clans is utter nonsense. They have enough AASI. Most look visibly S Asian. A non insignificant minority doesn’t look conventionally so, especially with heavy grooming, and are this recruited for Bollywood and Bhangra videos, where they carry a plow in one hand and a gun in another, side by side with likes of Sidhu Moosewala, who raps in vulgar Punjabi with a co-opted West Coast Black American style.

    They as a group are historically herders who moved into the areas I described probably literally seeking more fertile pastures. They arrived in an already caste established society and assimilated into this middling role in the rural peasantry but with greater land owning and military opportunity than similar designated shudra groups. But even that I think may be slightly exaggerated. Most of the early power players in for example the Sikh community, known to be now dominated by Jats, were actually in fact Khatris, aka the bhappa Sikhs that Jats mock for their perception of insufficient masculinity.

    Their main luck is greater martial race designation post 1857, British initiated irrigation advances in their region they benefitted massively from as land owners, and also the fact that they resemble West Eurasians the most, out of all Indian populations, thus making it easier for them to enter the entertainment industry, give India’s ancient and historical power dynamics. Tongue twister: Basically they are more Brahmin than the Brahmins in what makes the Brahmins phenotypically Brahmin.

    1. thewarlock:
      dude, calm down. I used Central Asian as a shorthand for relatively more recent (i.e. 1500 year old) Steppe or Central Asian ancestry. This part is indeed true.

      Rajput claims can be horseshit or not – it depends on which community etc. There is enough circumstantial and historical evidence for Rajasthani Jains to claim partial Rajput ancestry, although like I said the commonalities could also be due to emulation or diffusion.
      In any case, unlike OBCs or Jats etc claiming certain ancestries, Rajasthani Jains don’t associate Rajput ancestry with prestige. In all my years, I have only heard it mentioned a few times and that too only by relative more historically aware, educated folks. Also Rajputs don’t phenotypically look *that* different from others in Rajasthan (their higher meat/protein consumption might in the past have resulted in bigger stature )

      This is also similar to West Asian ancestry claims of Muslims. It is now proven that almost all South Asian Muslims have higher West Asian and Central Asian than Hindus of corresponding caste level or class, and upper class/Ashraf types have non-trivial non-Indian ancestry (say 15% to make up a number) though many still are phenotypically as South Asian as anyone else (while many others are South Asian but with a distinct foreignish flavor lol)

      Point is that not all community origin stories are made up; often there is a small grain of truth to them..

      1. that amount of West Asian is on average way way way less than 15% when controlling for caste. But point taken.

        The issue is the oft tiny grain of truth becomes the dominant narrative. Everyone needs a false mythos to prop themselves up. It’s part of human nature, so I shouldn’t be surprised. And point taken about non prestige based connection perception for Rajasthani Jains.

        1. no worries. sorry my post was a bit strident, but gross generalizations and broad strokes bother me when coming from intelligent folks 🙂

          I quickly searched for rajput jains and there is one tweet here complaining about Jains getting in the way of Rajput (in Rajasthan) rights to animal sacrifice because they “claim common Kuldevis”..:)


          “Interestingly, even Rajputs in Rajasthan are being denied their Right to Sacrifice at the altar of their KULDEVIS by the govts.
          Most are being encroached upon by Jains who claim common Kuldevis.
          We never did it in public but secluded places around the Temple.”

          PS: Opposed to any form of animal sacrifice but also of communities imposing their norms on others. Persuasion should be the way

      2. Interesting that Jains claim Rajput ancestry but don’t associate it with prestige- other vaishya groups from Rajasthan are similar. Merchant groups in that part of India distance themselves from martial and ruling groups. Maybe this is because of ingroup-outgroup dynamics that make ahimsa, vegetarianism, and ritual purity especially important to a vaishya vis-à-vis kshatriya dharma. I read about this in a really interesting book a few years ago called “Alchemies of Violence” by Lawrence A. Babb (Amherst) that explained this well from a social anthropology perspective.

    2. @thewarlock: Some correction for your post:
      -Rors are not Jats, they are fairly similar to eastern Jats in terms of overall autosomal breakdown but they are a distinct community.
      -A Pashtun mixture with a local population wouldn’t make sense since they have a higher steppe ancestry than the Pashtuns and the locals (at least as far as eastern Jats are concerned). Either the mixture was from a Tajik-like population or a hypothetical population more steppe-rich than modern Tajiks (to account for the steppe ancestry)-though that is supposed to be a condition for some kind of a Scythian ancestry, something that I highly doubt if not outright reject.
      Now here are my thoughts related to the post but not directly pertinent to it:
      -A more likely scenario is that eastern Jats (and Rors) are simply a remnant population that didn’t mix as much as the other Indo-Aryans did with the local post-IVC populations. How did this happen? I am not sure. It is just an ad-hoc hypothesis/pet theory of mine for now.
      -The Khatris are probably closer to Jat Sikhs than Hindu Jats are to Jat Sikhs. This might either be because of recent mixture or perhaps because the Jat label is perhaps not a very old one and could have at some point of time acted as a broad category which included several different populations.

      1. well put. Thanks for clarifications. All I can say though is that bollywood and the Punjabi music industry have pumped up the community so much. Every other song mentions the ethnic label.

        Even in America, wherever I have gone from Edison to Atlanta to Houston to SF, I hear the same tropes of their Jatt pride and bragging about how people don’t think they are Indian. My girlfriend is Punjabi with some Jatt lineage on her mom’s side, per her last name, so it never stops lmfao. They are to self promotion what Parsis are to business acumen in the subcontinent. Other communities need to learn how to emulate their self esteem, albeit with less cartoonish levels of confidence and a mythos closer to reality (need a bit of faking for spice ;))

        1. @thewarlocke:
          Well, a recent comment of mine along the with person who I recently responded to seems to have vanished. But I will mention some things from the 2 posts since they were a bit informative. The person in question posted a breakdown which seems to support the mixture theory, ie. Punjabi Jatts are probably a byproduct of the mixture between Khatris and eastern Jats. Though their MLBA ancestry is roughly half-way between Khatris (22.8%) and Rors (36.4%) (Ros being a proxy for eastern Jats), but their AASI ancestry is higher than both, maybe some minor Chamar-like input as well then?

          1. Ah screw it, the comments appeared, nvm, ignore my comment which talks about the 2 supposed missing comments.

          2. As noted in my other comment, Jatts of Central Punjab/Jatt Sikhs have admixture but it is very unlikely to be related to Khatris. Simply put, Jatt Sikhs of Punjab demographically outnumber Hindu/Sikh Khatris and most other Hindu/Sikh Punjabi groups by the millions. Rather, the actual admixture is likely a combination of IVC + BMAC admixture on a Hindu Jaat base due to geography. Within Central/East Punjab, there is no correlation with levels of IVC/BMAC admixture vs. Hindu Jatt admixture as Jatt Sikhs have been mixing with one another across Punjab for hundreds of years.

            As for Jatt Sikh AASI level, we’re talking about 1% differences in Jatt Sikh vs. Khatri averages and approximately 2% difference (closer to 1.5-1.7% in other runs) between Jatt Sikh and Ror averages. The sample sizes on Global 25 are 18 Khatris, 15 Ror and 12 Jatt Sikhs (all unrelated genica members + their relatives/friends). These are very insignificant differences considering you can find 1% AASI differences between siblings or other relatives.

            Rather than Chamar admixture, some of the individual Jatt Sikh samples may just have higher IVC admixture.

            On a separate note, it’s worth noting that, from my experience. Punjabi Khatris and Jatt Sikhs look more “NW” shifted phenotypically relative to Hindu Jaats and Ror (despite them being higher Steppe). This also applies to Kashmiris vs. Khatris/Jatt Sikhs.

      2. Jatt Sikhs and Haryana/West UP Hindu Jaats having many overlapping clan/gotra names (such as Beniwal, Punia, Baidwan, Pannu, Warraich, Gill, Grewal, etc; not all overlap of course). While it is true that Jatt Sikhs cluster closer to Khatris than to Haryana/West UP Jaats on average, there is diversity among Jatt Sikhs. Some are almost identical to Khatris, others are 50 50 and others are almost identical to Hindu Jaats.
        Here is a recent nMonte run from anthrogenica that includes a Mesolithic AASI HG proxy using samples from the new Narasimhan/Shinde papers. The Jatt Sikh average is in between Khatris and Rors (close enough proxy for Haryana/West UP Jaat).

        1. @Paindu
          In that case the mixture theory probably makes sense. Punjab Jats being between Khatris and Rors/eastern Jats would be an indication of just that. Looking over your picture it is interesting to see how the Rors have a comparable amount of both steppe and Iran HG ancestry to Tajiks (in particular the southern ones like Shugnan and Ishkashim). This combo seems forms like ~75% to 80% of their ancestry on average and the only major differences are in the value distribution of the other 25% of their ancestry. No wonder they ended up being close to Tajiks on global25 PCAs. I guess that this would be similar for eastern Jats (with the added difference of less Iran HG and more AASI).

          1. @DaThang

            So, Punjabi Jatts of Central Punjab/Jatt Sikhs have admixture but it is very unlikely to be related to Khatris. Simply put, Jatt Sikhs of Punjab demographically outnumber Hindu/Sikh Khatris by the millions. Rather, the actual admixture is likely a combination of IVC + BMAC admixture due to geography. Y-DNA haplogroup studies (see Mahal 2017) on Jatts seems to suggest Jatt Sikhs and Hindu Jaats are dominated by the same key paternal haplogroups save for higher R1a and much lower Q in Jatt Sikhs.

            Jatt Sikhs represent the space between SPGT Swat types such as Khatris and Hindu Jaats of the Haryana/West UP Gangetic plains. I have only seen 2 Rajasthani Jatts on GEDMatch but they seem to be in between Jatt Sikhs and Hindu Jaats of Haryana/West UP.

            As for Rors, they cluster closely with the 3 Hindu Jaats (2 from Haryana + 1 from West UP) I’ve seen on Global 25. In fact, two of the Hindu Jaats are actually clustering toward the Steppe peak represented by Ror 37. Their AASI levels are identical as well. You seem to have proposed that Hindu Jaats are more AASI than Ror. That doesn’t seem to be the case from the samples I’ve seen (other than 1 potential outlier from Mathura that I’ve come across).

            Note: This discussion is kind of cluttering up the blog so we can discuss this on anthrogenica if you join.

  11. Thewarlock:
    It just struck me that you are so focused on the Jatts because of your particular relationship situation that exposes you a lot to the inner Dynamics of Sikh Jatts. For those of us from Indi ( outside of Punjab) Jatts are not prominent and at best are cartoonish punjabis that some songs mention. Jatts are not really prominent in Bollywood ; Bollywood is a khatri Enterprise who also I think look more Western than Punjabi Jatts.
    PS: how is her family dealing with a Gujju prospective SIL 🙂

    1. lol they cool. enough inter community and even inter racial stuff in both our extended families.

      Bruh you still didn’t answer my question above, if we look alike. I want to see if genotype translated to phenotype. You could be my good twin.

      1. thewarlock:
        Sorry had missed your post earlier.
        Not much phenotypic resemblance, bro (I also have a slightly atypical for the region, look so that plays into it). Though that’s not surprising how much phenotype can diverge even among cousins (and sometime siblings) 🙂 But I agree with you on your pan Indian look. True son of Indus, indeed!

    2. @Justanotherlurker

      Speaking as someone who is intimately familiar with the influence of Jatt Sikhs in Punjab, you are spot on that Jatt Sikhs aren’t very relevant outside the Punjab region. They have some influence in Haryana/Northern Rajasthan and Delhi but are otherwise mostly limited to Punjab.

      On a separate note, while, it is true that Khatris dominate the Bollywood industry, it is Jatt Sikhs who dominate the Punjabi music and film industry/scene in India. Considering the popularity of Punjabi music in India and the South Asian diaspora, they are heavily overrepresented in South Asian music.

  12. Jats being central asian is the biggest con job , they have been pulling over the years. They have been enabled by the second con job group the Gujjars who claim ancestry to Khazars who ruled like on the other side of the Caucuses mountain range. Dont know who and how it all started.

    They are our Pakistani version of “We all came from Iran/Arabia” folks.

    1. Jatts aren’t Central Asian per say but they are clearly Central Asian shifted relative to most South Asian populations (including Kashmiris). Hindu Jaats/Ror are shifted toward Pamiri Tajiks due to their Steppe while Jatt Sikhs are more shifted toward Northern Pashtun groups (lower Steppe and higher IVC/BMAC admixture). In pure genetic distance (see Global 25), these populations are closer to some of these South Central Asian pops than they are most other South Asians outside the NW + some Brahmin groups.
      This may be shocking considering their phenotypes (especially Hindu Jaats + Ror) who resemble an UP Kshatriya more closely vs. a Ishkashimi Pamiri (most South Asian shifted Pamiri group) despite being closer in raw genetic distance to the Ishkashimi than the UP Kshatriya samples from the Metspalu study.

      1. where do these G25 samples come from? is it voluntary ancestry data? If so, is there a possibility of bias?

        anecdotally, I notice more atypically west eurasian shifted people of the groups opt a lot more for testing. Intuitively, this makes sense to me, given the assymetry or atypicality of their phenotype with others of their community, promoting them to explore more, especially if more West shifted looking, given the possibility of seeing exotic ancestry of higher conventional prestige, such as Persian or European.

        I know phenotype doesn’t equal genotype. But they are correlated. I think we need a good bit of random sampling of those who self identify as some of these groups. Then I would trust some of these averages more. Because Narshimha AASI contradicts G25. And phenotype, as you stated, also has inconsistencies with the tested genetic cline. I think it is still possible all of this is accurate. But hopefully more data can verify what is truth.

        1. Global 25 samples come in two ways. Either they are downloaded from publically available academic samples (such as from the Damgaard study) or user provided data. The academic samples include both moderns and ancients.


          I don’t understand what you mean by bias? Are you implying user provided data isn’t representative of the actual ethno-religious groups of South Asia?

          Interesting. So you personally/anecdotally know individuals who are atypically western shifted phenotypically and have observed they are somehow overrepresented in genetic testing? Furthermore, you have actually observed people from the same ethno-religious group scoring differently in correlation with their phenotype?

          Personally, I have to disagree. You can find heavily contrasting phenotypes even among siblings or cousins. However, you can reasonably infer they (relatives) would be rather close autosomal wise.

          I agree there is a correlation between phenotype and genotype. However, it is not always consistent. One of my relatives is noticeably more West Asian shifted in phenotype than another relative yet is actually slightly higher AASI in Global 25.

          Regarding Narasimhan, it doesn’t contrast G25 for most groups from what I’ve noticed. The user provided Khatri and Lohana samples I’ve seen on Global 25 seem to fall in line with the data from the paper. There is just some minor variation depending on the model and individual ancient samples used. Is there any specific group you believe scores noticeably different between Narasimhan and the user provided G25 samples?

          1. @DaThang

            Those Ror and Kalash samples are the same ones in the previous model I posted. These are the individuals instead of the averages. However, the model (different variations of AASI used) is also slightly different.

            Anyways, that model is a bit strange because Ror 37 is actually the most Steppe shifted Ror (in essentially every other run and on Harappa) but it seems like a chunk of his Steppe is going into ANF (Barcin N) or WSHG (West Siberian) as some of the Ror Steppe is not MLBA but “Siberian” like. Anyways, typically, I wouldn’t use a model with both ANF + Sintashta (which is ANF + EHG) + WSHG. I’d use Sintashta + BMAC (Gonur 1 BA, which is Iran N + ANF) + WSHG (if necessary).

            I took a 2nd look at the Narasimhan data.

            The Narasimhan Jatt Sikh samples are much more AASI shifted on “average.” However, we can’t see the individual data so we cannot clarify the potential inclusion of outliers, mixed or mislabeled individuals) that would be visible when looking at extreme differences between components of individual members of the same population.

            I inquired further into the Jatt Sikh samples and they seem to be from a separate private health/disease study (that includes Reich & Moorjani as two of the authors) on South Asians that includes Jatt Sikhs from a hospital in Bathinda.


            They were not collected by Narasimhan himself. I don’t know what kind of rigorous parameters (if any) they applied when collecting their sample data. However, I think it is alarming that the IBD (identity by descent) from the supplementary data in the paper suggests that their Jatt Sikh and scheduled caste samples from Haryana (these would be Chamar, Mazhabi, Balmiki, etc. types) have high shared IBD. That could hint at the possible inclusion of Chamar admixed or Chamars themselves in the Jatt Sikh data considering that Chamars/dalits make up 30-35% of Punjab and commonly use Jatt Sikh names. AFAIK, Jatt Sikhs aren’t related to these groups.

  13. Sir Herbert Risley declared the Jats to be the true representatives of the Vedic Aryans.

    Professor B. S. Dhillon states that Jat people are mainly of Indo-Scythian lineage with composite mixing of Sarmatians, Goths & Jutes in History and study of the Jats. Historian James Tod agreed in considering the Jat people to be of Indo-Scythian Stock.[61] Moreover, Sir Alexander Cunningham, Former Director-General of the Archeological Survey of India, considered the Jat people to be the Xanthii (a Scythian tribe) of Scythian stock who he considered very likely called the Zaths (Jats) of early Arab writers.

    Sir Alexander Cunningham, (Former Director-General of the Archaeological Survey of India) wrote: The Xanthii (a Scythian tribe) are very probably the Zaths (Jats) of the early Arab writers. As the Zaths were in Sindh to the west of the Indus, this location agrees very well with what we know of the settlement of the Sakas (Scythians) on the Indian frontier.[67]

    Sir John Marshall, (Former Director-General of the Archaeological Survey of India) wrote: “These Scythian invaders came principally from the three great tribes of Massagetae (great Jats), Sacaraucae, and Dahae (still exists as a Jat clan of Punjab)[68], whose home at the beginning of the second century B.C. was in the country between the Caspian sea (sea) and the Jaxartes river (Central Asia).[69]

    Arthur Edward Barstow wrote: “The Medii, Xanthii, Jatii, Getae and other Scythian races, were gradually working their way from the banks of the Oxus (River valley in Central Asia) into Southern Afghanistan and the pastoral highland about Quetta (a Pakistani city)

    Professor Henry Smith Williams wrote: “The extent of the Scythian invasion has been variously estimated. Some scholars believe that they virtually supplanted the previous population of India (means Punjab), and there seems little doubt that by far the most numerous section of the Punjab population is of Scythian origin.”[73]

    Professor Pritam Singh Gill wrote: “There is a general consensus of opinion that Jats, and with them Rajputs and Gujjars were foreigners who came from their original home, near the Oxus, Central Asia.”[74]

    Professor Tadeusz Sulimirski wrote: “The evidence of both the ancient authors and the archaeological remains point to a massive migration of Sacian (Sakas)/Massagetan (“great” Jat) tribes from the Syr Daria Delta by the middle of the second century B.C. Some of the Syr Darian tribes; they also invaded North India.”[75]

    I. Sara wrote: “Recent excavations in the Ukraine and Crimea. The finds points to the visible links of the Jat and Scythians.”[78]

    Sir Mountstuart Elphinstone Grant Duff wrote: “My conclusion, therefore, is, that the Jats may be of Scythian descent.”[80]

    Ujagir Singh Mahil wrote: “Jat were called Scythians; because they were the inhabitants of the ancient country of Scythia. The Jats who invaded the Punjab and conquered India up to Benares were called Indo-Scythians.”[81]

    James Francis Katherinus Hewitt wrote: “Further evidence both of the early history and origin of the race of Jats, or Getae, is given by the customs and geographical position of another tribe of the same stock, called the Massagetae, or great (massa) Getae.”[82]

    Syed Muhammad Latif wrote: “A considerable portion of the routed army of the Scythians settled in the Punjab. A portion of these settlers, the descendants of Massagetae, were called Getes, from whom sprung the modern Jats.”[84]

    Dr. Gopal Singh wrote: “The Jats of the Panjab, are Scythians in origin and came from Central Asia, whose one branch migrated as far south in Europe as Bulgaria. “[85]

    Satya Shrava wrote: “The Jats are none other than the Massagetae (Great Getae) mentioned in Diodorus as an off-spring of the ancient Saka tribe…. a fact now well-known.”[87]

    B. S. Nijjar wrote: “The Jats are the descendants of Scythians, whose kingdom’s capital was Scythia, in the present Ukraine.

    Steven M. Collins advocates the identification of Massagetaeans as “Great Jits or Jats” of Asia.[89][90]

    Weer Rajendra Rishi advocates that the Jats are none other than the Massagetaeans (Maha/Great Getae).[91]

    Rahul Sankrityayan had identified the Jats as Massagetaeans.[92]

    George Rawlinson has identified the Massagetaeans as “Great Jits or Jats” of Asia.

    Sir Richard Francis Burton wrote: “The Massagetae (greater Jats or Goths) are opposed to the Thyssa (or lesser) Getae, and both used the sagaris.”
    (MT Note – ‘sagaris’ is a two-blades Serbian axe)

  14. Sikhism explains Sikh jatts well possibly? maybe because despite it retaining a lot of tribalism of Hinduism in practice (I am aware it is anti caste in theory) but not same degree of tribalism that Steppe got diluted by khatri mix and extra AASI, even more than khatris, is via some chamar mix.

    1. While possible, it is unlikely that Sikhism has any relation to higher IVC/BMAC admixture in Jatt Sikhs relative to Hindu Jaats. It is just far too recent and Jatt Sikhs are one of the most if not the most endogamous Sikh Punjabi groups. More than likely, Jatt Sikhs picked up their higher IVC/BMAC admixture (relative to Hindu Jaats) due to more “ancient” mixing with the local Punjabi groups NW of the Steppe peak zone in Northeastern Haryana/West UP.

      Also, the 1% difference in AASI levels between the 18 Damgaard Punjabi Khatri samples (from Amritsar I believe; barring 1-2 outliers removed) on Global 25 and 12 Jatt Sikhs on G25 (diaspora samples from genica members) seems a bit small too small to suggest Chamar admixture no? Or are you referring to the Narasimhan “private” Jatt Sikh samples (who score much higher AASI than the samples on Global 25). That might be more related to the social and ethno-cultural dynamics of Punjab. Especially, with respect to the social dominance of Jatts over the Indian Punjab region.

      1. What this seems to suggest then, is that the ‘Jat’ label is probably not a very thorough one with respect to genetic ancestry. Also since you brought up numbers in another post, I suspect that estimates like 82 million are massive overestimates. Based on my own inspections, it looks like there are probably no more than 30 million Jats in Punjab + Haryana + UP.

        1. I think the Jatt label is fine in the context of Punjab, Rajasthan, Haryana and West UP. You have to understand there are millions of Jatts spread across this large geographic region. Not to mention there is not a united “Jatt” identity across religions. The average Jatt Sikh in Punjab or Hindu Jaat in Haryana in India does not see themselves as “ethnic kin” despite overlapping clan names. Religion, politics and language (Punjabi vs. Haryanvi/Hindi/Rajasthani) play a significant role there.

          The Steppe peak is around Eastern Haryana/West UP and while Jatt Sikhs are lower Steppe than them on average (indicating other IVC/BMAC like admixture), they are still higher Steppe on average than other NW groups or Brahmins for that matter. That is what is suggested by the diaspora samples on G25 anyways. The Narasimhan samples are much more AASI shifted (suggesting potential inclusion of outliers, mixed or mislabeled individuals).

          I have no idea where you are getting your numbers but estimates for Jatt Sikhs have put them at approximately 20-25% of the entire Punjab population. They are likely anywhere from 8-10 million spread across Punjab, Haryana, Delhi and the rest of India. Hindu Jaats “should” heavily outnumber them once you include all those in Rajasthan, Haryana, Delhi and West UP. There are also a small number of Muslim Jatts of Punjab/Haryana that didn’t migrate during partition.


          1. the jump to potential inclusion of outliers, mixed, or mislabeled accusations against Narsimha results (which btw included 40+ people, more than for most ethnic groups tested) is what is making me suspect the samples on G25 from diaspora, so commonly cited by amthrogenica users, are skewed by more curious west eurasian shifted types who, in my experience, do more ancestry testing. clearly, sample size is insufficient. but there is massive ego I notice for battle of “most steppe,” so discrediting of that results in accusations of bad data collection seems to be a theme online. you may be correct, but I think more samples are needed from a variety of locations and peoples that self identify as Jats.

            Again, I know phenotype does not equal genotype. But there is positive correlation. The guy wanted to look into Haryana Jats from the central asia paper was stated saying how he wanted to investigate the group, given the men are “6’0 tall ON AVERAGE,” a laughable statement for any major S Asian ethnic group. There definitely was an agenda. Even given that, the data are the data. But I do notice that there is variation and many individuals need to be averaged to minimize error for an already rather diverse appearing group, given asymmetry of results between eastern jatts and western jatts and s central asia vs. narsimha paper for western jatts.

          2. @thewarlock

            I am concerned with your presumption that somehow these “anthrogenica users or their relatives” are all curious West Eurasian shifted types or that West Eurasian shifted types phenotypically in general do more genetic testing. How would you even begin to analyze such a broad statement across diaspora individuals or corroborate their results with their phenotypes (unless you had access to tons of their results and phenotypes)? As I’ve previously stated, I’ve actually observed the opposite within my family and friends. There is no correlation between phenotype and who is a little higher Steppe, Iran N or AASI. At least at the ethno-religious sub-group level. There should be at a more broad level.

            Anyways, I took a 2nd look at the Narasimhan data.

            The Narasimhan Jatt Sikh samples are much more AASI shifted on “average.” However, we can’t see the individual data so we cannot clarify the potential inclusion of outliers, mixed or mislabeled individuals) that would be visible when looking at extreme differences between components of individual members of the same population. I’m not making an absolute statement (or at least not intending to).

            I inquired further into the Jatt Sikh samples and they seem to be from a separate private health/disease study (that includes Reich & Moorjani as two of the authors) on South Asians that includes 41 Jatt Sikhs from a hospital in Bathinda.


            They were not collected by Narasimhan himself. I don’t know what kind of rigorous parameters (if any) they applied when collecting their sample data. However, I think it is alarming that the IBD (identity by descent) from the supplementary data (2nd link above) in the paper suggests that their Jatt Sikh and scheduled caste samples from Haryana (these are Chamar, Mazhabi, Balmiki, etc.) have high shared IBD. That could hint at the possible inclusion of Chamar admixed or Chamars themselves in the Jatt Sikh data considering that Chamars/dalits make up 30-35% of Punjab and commonly use Jatt Sikh names. AFAIK, Chamars, Mazhabis and Balmikis of Haryana are not related to Jatt Sikhs. How would you explain this IBD (suggesting recent shared ancestors) similarity?

            I think with regards to phenotype, we must take a step back. Yes, there is correlation between phenotype and genotype. However, that is more at the group level and there can be exceptions (look at Kashmiris vs. Hindu Jaats/Ror). These exceptions may be related to selection (sexual or climate related). At the individual level, there can be widely contrasting differences in phenotypes and similar autosomal data (you can see this in any South Asian family).

            I’ve already pointed out the issue with Narasimhan samples. I find it astonishing that Jatt Sikhs would share high levels of IBD sharing with Scheduled castes of Haryana.

            Again, differences between “Western” Punjabi Jatts and “Eastern” Haryana/West UP Jatts seems to correlate with geography in terms of additional IVC/BMAC admixture among Punjabi ones. This makes sense considering how non Jatt Punjabis such as Khatris, Arains, Tarkhan, Gujjar, etc. score.

          3. On a side note, thankfully Punjab and Haryana high court ruled out against the bill. an affluent and ultra politically powerful community didn’t need OBC reservation. Patels did the same nonsense riots in Gujarat. These land owning tribes need to check themselves and stop using violence and destruction of property as a way to bully others to get their way. Real dalits need it a lot more than some overnight rich farmer millionaires with their property values ballooning in the last couple decades.


            just btw for goofy 6’0 average claims

          4. @thewarlock
            I’m a fairly rational person. I won’t make ridiculous claims that the average Jatt Sikh or Hindu Jatt/Ror is 6’0. Especially, those in India.

            Anyways, regarding height, that study on Jatt Sikhs is from almost 50 years ago (1972). AFAIK, “max” height is often linked to nutrition and wealth/income (ie access to better nutrition/medicine). A study on diaspora 2nd or 3rd gen Jatt Sikhs in Canada/US/UK would inevitably show differences with the 1972 study and even modern Punjab which straggles behind Western countries in nutrition and GDP per capita but is far ahead of where it was 50 years ago.

            uisashi on anthroscape makes a good point:

            Obviously, Jatt Sikhs are not remotely close to the majority of Punjab’s population (20-25% max) and average height is affected by migration from other states. Regardless, Punjab has one of the highest average male heights in all of India. Based on personal observations, I would think the average 2nd or 3rd gen adult male Jatt Sikh in the western diaspora is around 5’10 (similar to average American white male). I’ve noticed many multi-gen Tamils (especially those in Toronto) happen to be quite tall as well. However, in general, I feel the average multi-gen South Asian in the western diaspora in general tends to be taller than their counterparts back in India, Pakistan, Bangladesh or Nepal. There are exceptions of course.

          5. @thewarlocke
            What was the average height of south Asians in general in the 1970s? The more recent estimates give a grand total of 165 cm for males. I would expect the modern Jat average to be more like 175 cm.
            South Asia hasn’t had a 183cm average male population since the Ganga mesolithic period.

    2. @thewarlock

      I guess, after this extensive correspondence, you are an expert for Jats. Although, I can’t see the meaning of all these listings of specific genes, steppes, grasslands, etc…

      I have two questions. One is related to the list of statements of historians and other people which I presented in my previous comment (they are taken from the site of the World Jat Aryan Foundation). All statements are pretty similar with some variations. But, they have one thing in common.

      Q1 – What is common for all above mentioned statements?

      The second question I already asked but none tried to answer. There is no right or wrong answer, I would expect you (as one of the founding members of SA Serbian Brotherhood) to give your opinion. It is related to the first World Jat Congress in Belgrade, in 2003. This Congress was scheduled for July 2003. but it was stopped by India’s Prime Minister, Atal Bihari Vajpayee. It was, however, held in September 2003.


      Q2 – Why the Congress is scheduled in Belgrade, why India’s PM stopped the Congress and why the Congress was, however, held 2 months later?

      1. I have no idea about why the Congress was stopped of politics behind it. As for false pushing of ancestry materials. Well that has been going on forever. Hell the Kalash were sold as Alexander’s army descendants by some, until recently unequivocally proven wrong. Genetics is clarifying a lot of the biased leaps of faith anthropologists went off of, based on, at least what is recognized now, as partial pseudoscience of phenotypes to make all sorts of wild assumptions. Then bandwagoning happened.

        Also, I am not an expert. I have been around community members, but I am not from the community nor have I formally studied it. But I do notice trends.

        Just for fun, it is tough for me to tell which one is the parody website


        The reputed historian Qanungo writes that the philologists like Dr. Trumpp and Beames [1] very strongly claimed a pure Indo-Aryan descent for Jats both in consideration of their physical type and language, which has been authoritatively pronounced as a pure dialect of Hindi, without the slightest trace of Scythian. [2]

        The Jats have pure Aryan physical features. Their wheatish complexion, oval face with a firm jaw,

        History of the Jats, End of Page-5

        prominent nose, dark eyes, thin lips, well set teeth, long neck, broad shoulders, thin waist and tall stature are unmistakably Aryan. Hardly any Jat will be found with non-Aryan features. They have retained racial purity due to their homogeneity. It can be safely said that if any people have preserved pure Aryan characteristics it is the Jats, Ahirs and Gujars.

        A Jat is fearless and frank in expression.”


        “The Jats are Aryans and like all Aryans originated in india. About 8 million Jats live in the Indian state of Haryana. Jat, pronounced (…er, this is tough…), “Jut” in Bollywood, “Jaat” in Haryana, “Whaat?” in Bihar, “Sikh” or “Humble follower” in Punjab, “Brigadier/colonel/lieutenant/armywala” in Pakistan, “Yahya Khan” in Bangladesh (they realized later on that he was actually Pashtun but it didn’t make a difference), “Pakistani” in Abaad Kashmir, “Indian” in Barbaad Kashmir, “Gadha” in Jataka tales, and “Crazy foreigner” everywhere else, is a word with a complex etymology but entirely obvious meanings.

        They are an extremely humble, docile, group of people, known for their intellectual prowess and quiet nature. However, despite this, every great person who has ever lived has turned out to be a Jat. Unbelievable as such a claim as this seems, you only need to ask any Jat, or read any book written by a Jat to see that this is true.

        Some famous Jats include Alexander the Great, William the Conqueror, Genghis Khan, Charlemagne, Shakespeare, Einstein, anyone whose ever been in a fight and won, Napoleon, Superman, Ramesses II, Optimus Prime, Ozymandias, King Arthur, Yoda, Indiana Jones and Morpheus, amongst others.

        Whilst their origin is disputed by many “reputable” sources such as western scholars and modern factual evidence, they are actually descended from a race of super-powerful, warlike demi-gods who destroyed anything that crossed their path without even breaking a sweat, and who could travel through time and divide by zero, and who created the universe itself. Again, read any book written by a Jat to see that these claims are indeed, true.”


  15. “I’m confused about the difference between Southern South Asian (16.6%) and Southern Indian subgroup (6.9%), and also broadly Central Asian and North Indian, Pakistani (12.3%) vs. Broadly Central & South Asian (9.8%).
    Razib, or someone else, can you help explain?”

    Southern Indian Subgroup basically mean South Indian Brahmins. Quoting from 23andme website..

    “Of these many groups, we were able to identify a genetic signature that reaches high levels among people with ancestry from southern India who say they are Brahmin. This group, labeled “Southern Indian Subgroup,” was probably identified in our Ancestry Composition analysis because members of these communities migrated to the United States at higher rates than others and are therefore more genetically represented in the 23andMe customer database.”

    Broadly Central & South Asian means all we know is that this component comes from Central and SA, but beyond that it is not possible to identify any subgroup in this region. Quoting from their website..
    “Some Central & South Asian DNA is difficult to assign confidently to one population and receives a “Broadly” designation.”

    Same goes for all other “broadly” classifications.

    Hope this helps.

  16. @Paindu

    Fair arguments. I just want to see more published data from a larger sample of Jatt Sikhs that has more rigorous collection than GED match uploads. You could very well be totally correct. Let’s see how this AASI pans out with more sampling under controlled conditions. Future should be interesting

    1. @thewarlock

      I honestly don’t think larger sample sizes will show any significant differences. While possible, I personally doubt it. The bigger issue is whether the samples are available publicly or not. This way the individuals in an average can be analyzed to determine if there are extreme outliers that are completely outside the cluster on a PCA or on different ends of the South Asian cline.

      Because, the Damgaard study had approximately 15-20 Punjabi Khatris and Gujarati Brahmins and even they had a few clear outliers or potentially mislabeled individuals. Even the Pathak study that included Ror, Kamboj, Rajasthan Gujjars, etc. had 2 Kamboj samples clustering with mid-caste South Indians while the rest were firmly in the Khatri to Northern Pashtun cline. They were likely mislabeled Chamars or some lower caste population that identified as Kamboj considering they clustered on a totally different end of the South Asian cline. The point being that you can’t even take academic samples to be fully accurate.

      I also wanted to point out that I have the pictures of some individuals with whom I share on 23andMe and their GEDMatch results. Some might be quite surprised to see how extreme the phenotype-genotype discrepancy can be. As an example, I mean some individuals who look more clearly South Asian than the pictures you posted of yourself yet score very low SI (relatively to other South Asians) and very high NE Euro + Med.

      1. Do you know if the Ror samples are available in Harappa Gedmatch format? It would be interesting to see how they compare.

        1. @INDTHINGS

          Most of the individual Ror and Kamboj Harappa/admixture results as well as those of ancients (Indus Periphery, BMAC and Steppe) have been posted on the anthrogenica forum.

  17. @DaThang
    No. Not accurate, from what I have seen sources say.. I am tired of old tropes of N Indians being giants compared to S Indians. This is simply not true. Indians are just short overall. You can argue the Punjab average is brought down by the 30% chamars but Jats are the biggest group in Punjab and the average of 168.4 is no where near 175. To me, when I visited India, the chamars, who I assumed to be the street laborers and servants, were shorter but not tremendously so. Honestly, a lot of that could be chalked up to bad nutrition compared to the other groups. The only trend seems to be that where nutrition is good, height is good. These statements about massive men are nonsensical. I am not saying you made them. But I have heard enough guys say, “In my village in Punjab, all the men are 6’0 tall. Jatts are 6’0 tall with good body and are warriors.” In reality, maybe an average of like 171cm for Jatts and maybe 165cm for Chamars, if you want to put chamars as around the same height of some of the S Indian state averages. But even that is a massive assumption based on preconceived notions of what groups we think are taller.
    Here is a paper for average height of 20-29 year old males from India by state (2011)
    Table 1:
    Jammu and Kashmir 168.0
    Himachal Pradesh 166.4
    Punjab 168.4
    Uttaranchal 165.2
    Haryana 168.1
    Delhi 165.9
    Rajasthan 167.2
    Chhattigargh 163.7
    Madya Pradesh 165.8
    Uttar Pradesh 164.5
    Sikkim 160.0
    Arunachal Prad1610esh
    Nagaland 163.1
    Manipur 163.4
    Mizoram 162.9
    Tripura 161.6
    Meghalaya 157.5
    Assam 163.3
    West Bengal 163.8
    Jharkand 162.5
    Orissa 162.9
    Bihar 163.6
    Goa 165.3
    Gujarat 166.2
    Maharasthra 165.7
    Andhra Pradesh 164.6
    Karnataka 165.7
    Kerala: 167.8
    Tamil Nadu: 165.5
    India Average: 165.2

    Btw, just by religious break down
    Average heights
    Hindu 165.1
    Muslim 165.3
    Christian 163.8
    Sikh 170.0
    Buddhist 164.3
    Jain 170.0

    Now, go tell a Sikh that Jains have a similar average height. You will get laughed at. In reality, height in S Asia is largely about nutrition. Those groups that eat well and have money are taller. Period.

    1. @Thewarlocke
      Unfortunately none of them are broken down by caste. If there was a caste based data from 1970 then perhaps we could have made a ratio of any overall increase or decrease since then and apply it to current populations in order to roughly estimate the sub-population height averages as of 2011. Considering that Jats tend to form a minority population of the states that they inhabit (20% to 25% of Punjab, 30% in Haryana and 6% in UP), I think that their impact would add up overall proportionally to their state height (probably the reason why Punjab and Haryana are taller than UP, Himachal, Uttarakhand and Rajasthan on average). If I had to guess, then I’d say that they are ~5 cm taller than the state averages where they form a visible minority (20% or more of the population). Probably an average height of ~173 cm then

      1. I disagree. Even dalits and banias of North are taller than Southern and Central counterparts. I think Jatt height is highly exaggerated due to false mythos. like I said, self promotion seems to be speciality. I stick to a 170-171cm average. No one has data regardless. Also, jats are pretty damn well fed compared to many others. A 5’8 average is nothing special. Again, like even Paindu said, enough Jatts are even more AASI shifted looking than my Guju Vania self.

        I am a 176cm and 170lbs. Of course there are many people taller than me in India. But when I visited, even in rich areas, most men were shorter than me. Also, bone structure wise, pretty much everyone except for a small minority had smaller wrists and ankles (good surrogate marker for bone structure) compared to most N Americans.

        Blacks and Whites have same average height in USA. But blacks more on extremes. Jatts strike me that way. Their extremes are wild. But mean is only slightly bigger. Humans tend to compartmentalize and streotypes for the ease of approximations because going into slight nuances for everything would clutter the mind into a state of inefficiency. The widely exaggerated stereotypes are an example of this.

        1. >Even dalits and banias of North are taller than Southern and Central counterparts.
          Do you have anything to back that up?
          >Also, jats are pretty damn well fed compared to many others. A 5’8 average is nothing special.
          I don’t ever recall saying that 173 cm is special. It just seemed to fit the proportions decently (Assuming that the average north Indian is 166 cm and so 166 x 0.7 + 173 x 0.3 = 168.1- this is obviously not the only solution to the combination question).
          As far as the heights of modern south Asian populations are concerned, I suspect that both the steppe and the AASI (at least the Ganga mesoliithic type of AASI) heavy populations are over-represented among tall south Asians (considering how tall the steppe people and mesolithic south Asian people were). Something could explain why modern south Asians are so much shorter compared to their mesolithic south Asians counterparts- they have a lot of the divergent Iran-like HG (which I currently suspect might be responsible for genetic impact leading to lower heights) ancestry along with continued exposure to farming (which likely contributes the environmental + long term selection impact leading to lower heights).

          1. that’s fair calculation combo. I didn’t know mesolithic had tall heights too. I am curious about it. Do you have source link for that? Thanks for info.

        2. When I referred to Jatts more South Asian looking than yourself, I was mostly referring to Hindu Jaats/Ror. Jatt Sikhs will have those types too but they’re a smaller proportion among Punjabi Jatts than Haryanvi/Rajasthan/West UP ones. At least based on my observations as well as those of my Punjabi friends actually from India.
          As for height, the average Jatt Sikh male in India is probably around 5’7 to 5’8. That’s pretty realistic. It’s in the western diaspora in the UK, US and Canada among diaspora folks where you can find individuals closer to 5’10 on average and a strong minority well over 6’0. However, I’d argue multi-gens are likely closer to their “nutrition” maxed heights for their respective ethno-religious groups. All else being equal in terms of nutrition, I still think some South Asian ethno-religious groups will end up taller than others on average similar to how some European countries are taller than others. IMO, Punjabis and Tamils tend to some of the taller South Asian pops on average.

          1. Yeah anecdotally, in the non first gen US S Asians, Punjabis and S indians seem to be the tallest. Bengalis and Gujaratis are shorter. Probably close to 5’10 for Punjabis, maybe 5’9+ for South Indians (with Keralites 5’10, tamils 5’9, and Andhara 5’8), and 5’8 for others. Although, Jains seem a bit taller to me than other Gujus. In that study I linked, by religion, Sikhs and Jains had the highest average by religious group at 170cm. I have enough 6’0+ male relatives, even in India. But that again could largely be nutrition.

            But empirical collection is needed to confirm of course.

            And yeah Paindu, I agree from what I have seen with Sikh vs. Hindu Jats. For whatever, reason the former have a slight afghan bent that I can’t quite put my eye on that really diverges from my very classic indic look. Anyway, I facially look like what I am the most. Just a regular guju dude.

          2. >When I referred to Jatts more South Asian looking than yourself, I was mostly referring to Hindu Jaats/Ror

            Which is interesting considering that they have lower AASI than Punjabi Jats do. IDK maybe the AASI estimates are wrong (since no actual AASI sample is available, it is all “simulated” as of now) and Rors/eastern Jats have both a higher steppe and a higher AASI ancestry than Punjabi Jats do, meanwhile Punjabi Jats definitely seem to have a higher Iran HG ancestry (given the current breakdowns and the global25 pca for south Asia).

          3. @DaThang
            So, I don’t think a 1.5-2% difference in AASI (comparing those on G25) is going to make much of a difference in phenotype. At that point, there are other factors at play such as selection (including climate related) or inbreeding (in the case of groups such as the Kalash or Chitpavan Brahmins).
            The only thing I could think (outside of selection) that could be causing the phenotype discrepancy on average between the two Jatt groups is that Jatt Sikhs have more recent Turan/BMAC ancestry than Hindu Jatts? This based on the presumption that Hindu Jatts/Ror represent a mix between a Steppe heavy pop (on the paternal side) with a Chamar like population (on the maternal side) with little further admixture afterwards. Meanwhile, Jatt Sikhs have that base with additional IVC + BMAC/Turan ancestry on top. A mt-DNA study on Jatt Sikhs vs. Hindu Jatts/Ror would be quite interesting.

          4. @Paindu:
            I suspect that the Hindu Jats might be quite different in terms of paternal and maternal haplogroups when compared to Rors. Didn’t you mention that they have less R1a than Punjab Jats?
            >Jatt Sikhs and Hindu Jaats are dominated by the same key paternal haplogroups save for higher R1a and much lower Q in Jatt Sikhs
            Or was this a typing error? The Jat average overall for all R (including R2) is 28.5%, so if it is less than Punjab Jats then they might end up being around ~20% R1a, which is comparable to the R1a frequency in Gujars. Rors are around 35% R1a and definitely have more steppe paternal than maternal lineages, meanwhile Gujars probably derived the steppe ancestry roughly equally on both sides.
            That being said, it would be interesting to see a large mtDNA study on eastern Jats (with like a sample of 1000 people), with the result being compared to the mtDNA distribution in Rors and Punjab Jats.

          5. @DaThang
            So, Hindu Jatts do see to have much higher Q (probably closer to 15%) than both Jatt Sikhs (Punjabi) and Ror. They also have lower R1a than both Jatt Sikhs and Ror at around 25% R1a if you were to use the Y-STR predictions (which can separate R2 from R1a as well) from the 2017 Mahal Study on about 200 total Jatt Sikhs and Hindu Jatts. In contrast, based on the non-academic Y-DNA data of about 300 Jatt Sikhs I’ve seen collected by anthrogenica members, Jatt Sikhs are closer to 40% R1a and per the Pathak study that included Ror, they are at about 35% R1a.

            Anyways, well the Q difference is noticeable, I’m very curious if there any significant differences on the maternal lineages of Jatt Sikhs vs. Ror/Hindu Jatts. Especially, with regards to proportions of M lines vs. Steppe vs. BMAC/”West Eurasian” IVC lines.

          6. @paindu
            I definitely share your interest regarding the mtDNA lineages. It was a surprise to find out that Haryana Jats have more Q than Punjab Jats. What kind of Q is this? Is it the one that is thought to have been in south Asia for over 10,000 years or is it a more recent subclade (Hunnic/any later steppe input)?
            BTW do you happen to know the difference in L-M20 frequencies in Sikh Jats vs Hindu Jats?

            Oh and are you a member on anthrogenica? We could continue this conversation via PMs if you are.

      1. This is odd since the rest of India only thinks of the word “Jatt” in close proximity to some sort of a joke. When I hear the word ‘Jatt’ I am reminded of the guy Amir Khan beats up in the beginning of ‘Dangal’.

        (All this is probably quite unfair to Jatts.)

      2. Well, Aamir character itself is a Jatt.

        Also i think Jatt are seen as country bumpkins by outsiders, but still they also have positive stereotypes as brave etc. So all in all, it cancels one another.

        1. “Well, Aamir character itself is a Jatt.”

          Yes, but that is not the image that pops into my head when I think ‘Jatt’. Not that Aamir Khan is particularly cerebral looking but he lacks the tall physique and that certain attitude.

  18. @Thewarlocke:

    There is a lower end estimation as well but I can’t find it. The lower end had the male average at 176 cm (forgot the female average) so even the lower end estimation was taller than EHG and Yamnaya average by a bit. The high end estimation of 180 cm+ completely overshadows any modern day south Asian average. They also had correspondingly large skulls. Based on what I know these guys were about as large as and as big brained as the EEMH hunter-gatherers. You could probably google what their crania looked like. Mahadaha, Damadana and Sarai Nahar Rai are of particular importance (just as a hint).

    Meanwhile here is an upper paleolithic height estimation for Sri Lanka on this page:
    174 cm for males and 166 cm for females.

    South Asia took a huge hit in terms of average height since the mesolithic.
    (test, comment isn’t getting posted somehow)


Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.