The Tarim Mummies were the last of the Paleo-Siberians


The paper that reported on the data from the famous “Tarim Mummies” is out. If you don’t have a good grasp of the alphabet soup of ancient early Holocene populations, the results are going to be hard to parse. So I’ll make it simple for you: it now looks that the Tarim populations from 4,000 years ago are among the last people who were mostly “Ancestral North Eurasian” (ANE), and, they had no connection to populations in Europe. The second part is important because Victor Mair and others who have examined the mummies are wont to proclaim that the “Loulan Beauty” and her peers were “Caucasoid” due to their physical features. This may still be technically true, but the inference that this has to do with migration from the west of a European-origin population turns out to be false.

Being wrong is not a big deal. My own suspicion and assumption that these were part of the early movement of Iranian people eastward also turned out to have been wrong too. This is clear because the Tarim population from 4,000 years ago didn’t have any gene flow from the eastern Yamnaya, the Afansievo, let alone the larger Corded Ware reflux to the steppe (Iranian people did move into the Tarim zone later; the languages of the southern rim of the basin in historic times were East Iranian, and Iranians seem to have arrived in the north in Mongolia late in the Bronze Age).

I’ve added some stuff to the plot below to make it clear:

The PCA above is consistent with the Tarim mummies being mostly descended from ANE, though they have a minority of northern East Asian ancestry.

Interestingly, the earlier remains from Dzungaria are mostly descended from Afanasievo populations with a minority of ANE ancestry. The authors conclude, correctly I think, that this points to the likely origins of the Tocharian languages from the Afanasievo, and the possibility (I bet) that the ancient Yamnaya language was similar to that of the Tocharians. The fact that the Tarim people seem to have been mostly very distant branches of R1b illustrates the origin of the R lineage deep in Siberia during the Pleistocene. R and Q are clearly from the Paleo-Siberians.

Finally, let’s talk about the famous “European” or “white” Buddhist monks depicted at Tufan:

Tajik Girl

The Chinese describe some of these people as having light hair and eyes. In other words, they looked like Europeans. Previous work had argued that this was due to the Tocharians being descended from European-like people. But we now have enough evidence from the Yamnaya to know that very few were pale-eyed or light-haired. Rather, they were a dark-haired and dark-eyed population with olive skin. Unless there was later natural selection for these characteristics, this isn’t due to the Afanasievo ancestry of the Tocharians (who by the period 500-1000 AD were mostly localized to the northeast of the modern Tarim basin, around Turfan). Rather, the Tocharians were themselves a mix of people, and I believe those with Europoid physical appearance had those because they were heavily Iranianized in ancestry.

Today most people associate “Iranian” with Iran, and Persia, but Persians emerged on the southwestern frontier of the Iranian world, as heirs of Anshan and Elam (these were non-Indo-European societies). For much of history, Iranian-speaking people spanned the zone between Hungary and Mongolia and were much more physically and culturally diverse. To this day many Tajiks could pass for European, and these people have some of the highest fractions of “steppe herder” ancestry in the world. A minority of the Sintashta likely had blue eyes going by their genomes, so I think the origin of these “Europoid” people has to be interactions with the expanding Andronovo-horizon in the latter period of the Bronze Age.

These results in this paper show that the core population 4,000 years ago in the region of the Tarim that was later home to the Tocharians was inhabited by an ANE/Paleo-Siberian population, with a minority component of ancestry derived from northern East Asians. This ancestry dates to the early Holocene, 10,000 years ago. The later Tocharians probably absorbed these people, but I believe they were a mix of post-Afanasievo populations and Iranians. The former gave the Tocharians their unique and very basal Indo-European language, and the latter were responsible for “European” physical features so noted by the Han Chinese chroniclers in the 1st millennium A.D.

Note: the ANE are closer to West Eurasians than East Eurasians, but they are very distantly related to the former. Their ancestors seem to have diverged from European and West Asian hunter-gatherers 35,000 to 40,000 years ago.

The Yamnaya origins of the Tocharians?

Eurogenes points me to a new paper in Current Biology, Ancient Genomes Reveal Yamnaya-Related Ancestry and a Potential Source of Indo-European Speakers in Iron Age Tianshan. The conclusion:

Combining both the genetic and archaeological evidence, we here provide the first direct evidence of an early stage of population admixture around 2,100 BP in Xinjiang in Western China. Our study supports the “Steppe hypothesis” over the “Bactrian Oasis hypothesis” for the peopling of the Xinjiang region. The high amount of Yamnaya or Afanasievo-related ancestry in the Iron Age Xinjiang individuals indirectly supports the introduction of Indo-European languages into the region that survived in the form of Tocharian until the late first millennium CE. We note that we need more individuals from different sites and time periods to shed more light on the genetic history of the Tarim basin and the whole Xinjiang region.

The ethnolinguistic pattern of the Tarim Basin in early historic times was complicated. In addition to very distinct “Tocharian” languages, some of the cities were dominated by Iranian peoples. The modern-day Uyghurs are almost certainly descended from some of these populations. Uyghur men carry both R1b and R1a.

These results from ~2,000 years ago of ten individuals (five of them well dated) suggest that the northern Tianshan was populated by peoples that were direct descendants of the post-Yamnaya Afanasievo, who mixed to varying degrees with trans-Siberian populations (the mixture seems to have had a wide range of fractions). Two of the males carried R1b, the haplogroup dominant among the Yamnaya, but not latter cultures of the steppe (Andronovo). These individuals lacked European farmer ancestry, again indicative of their isolation from dynamics on the western forest-steppe that resulted in genetic differences between the successors of the Yamnaya and the Yamnaya proper.

The argument here is somewhat by elimination. Historical records indicate that some of the cities of the Tarim, particular those of the southern fringe of the basin, were Iranian speaking. Additionally, Iranian cultures are associated with haplogroup R1a, and the Sintashta-Andronovo cultures all had European farmer ancestry. In contrast, R1b is rare outside of Europe (though it is found in Kalash and Yaghnobi), but is found among Uyghurs and among these samples. Tocharians are the most likely descendants of these people, who arrived in the region almost 5,000 years ago.

This explains how the Tocharian languages were so distinct, and, their deep separation from other Indo-Europeans. The Tocharians were isolated and diverged very early. Later they were joined by Iranian groups. Eventually both these were absorbed by Turkic populations, first the Uyghurs, and later the Karluk Turks (the modern Uyghurs revived an ancient ethnonym).