Pages

Sunday, April 16, 2023

A response to “Genetic population structure of the Xiongnu Empire at imperial and local scales” (Lee et al. 2023)

My last blog entry featured data from an upcoming article about the Xiongnu empire. The entry was quite a success with thousands of viewers, in part due to Empress Ashina’s G25 coordinates I suppose. The article is out now, having  been published as part of the Science Advances journal.

Genetic population structure of the Xiongnu Empire at imperial and local scales

JUHYEON LEE1, BRYAN K. MILLER , JAMSRANJAV BAYARSAIKHAN, ERIK JOHANNESSON, ALICIA VENTRESCA MILLER6, CHRISTINA WARINNER, AND CHOONGWON JEONG

SCIENCE ADVANCES

14 Apr 2023

Vol 9, Issue 15

DOI: 10.1126/sciadv.adf390


I have to say that I am a bit disappointed in the article. In my humble opinion there were quite a bunch of odd statements, reaching conclusions and an inclusion of possibly incorrectly dated/attributed samples.


From what I can tell, many people on the internet seem to agree, so I am not just ranting into the void here. What I will do is address some of the points made in the article in typical Musaeum Scythia fashion, and I will formulate some questions at the end.


 I will forward this to the authors involved and hope at least one responds. I am just a little blogger who only gets a few thousand views on a post entry when favoured by the sky gods, so I won't be expecting too much. I hope that my dear readers, many who are very knowledgeable people themselves, also bring up points of contention which the authors might come across as well.


Terrace tomb of Takhyltin Khotgor


To start…

Overall, individuals from the Ulaanzuukh and the Slab Grave cultures present a homogeneous genetic profile that has deep roots in the region and is referred to as Ancient Northeast Asian (ANA) (14). The recent publication of additional genome-wide data for Ulaanzuukh and Slab Grave individuals (15) provided an opportunity to investigate the genetic profile of the Slab Grave individuals across a wider geographical distribution (Fig. 1A) and to refine our genetic modeling of the formation of the Xiongnu more generally.”

The first thing I want to address is that it certainly is not the case that Ulaanzuukh and Slab Grave have a homogenous profile of ANA populations. While this is by far the largest component amongst them, this comes through various sources.


Furthermore, the samples have a heterogenous amount of Yellow River related ancestry. This was not picked up in Jeong 2020, and neither was this in this article. I can quickly show this on G25, but I don’t want to get into a “my model vs their model" and “G25 vs QpAdm” debate, although if you scroll down a bit you can see the G25 anyways.


What I am going to point out is that nearly all Slab Grave samples carried Y-chromosome haplogroup Q-M120. Q-M120 is a lineage that in terms of modern distribution is significantly more present in China, and has shown up in ancient samples from China such as XW-M1R18 from the Xiaowu site of the Yangshao culture, while lacking in MNG_North or MNG_East samples which tend to have clades of haplogroup C and N. Interestingly the members of the Peng clan of the Zhou dynasty were under this lineage as well, and these were likely descendants of northerners affiliated with the Zhou.I think you can make an archaeological case for this partial southern origin too, as both the Ulaanzuukh and Slab Grave culture have strong affinities with Zhukaigou and related cultures of the northern chinese steppe arc. That's not the topic for the day though.



Target

Distance

CHN_Upper_Yellow_River_LN

MNG_North_N

RUS_Argun_River_Meso

RUS_Tyumen_HG

RUS_Yakutia_Ymyiakhtakh_LN

MNG_Ulaanzuukh_Slab_Grave:BOR001

0.03651264

23.8

34.4

40.6

1.2

0.0

NG_Ulaanzuukh_Slab_Grave:BUL002

0.02488712

42.0

20.6

34.4

3.0

0.0

MNG_Ulaanzuukh_Slab_Grave:MIT001

0.03132711

2.6

64.6

31.8

1.0

0.0

MNG_Ulaanzuukh_Slab_Grave:ULN001

0.02879159

21.0

52.4

26.4

0.2

0.0

MNG_Ulaanzuukh_Slab_Grave:ULN003

0.03269858

10.8

88.4

0.0

0.8

0.0

MNG_Ulaanzuukh_Slab_Grave:ULN006

0.03179151

9.4

79.6

10.0

1.0

0.0

MNG_Ulaanzuukh_Slab_Grave:ULN007

0.02585614

24.4

49.6

26.0

0.0

0.0

MNG_Ulaanzuukh_Slab_Grave:ULN009

0.02535003

1.4

70.0

27.8

0.8

0.0

MNG_Ulaanzuukh_Slab_Grave:ULN010

0.02789474

23.2

75.0

1.6

0.2

0.0

Average

0.02945661

17.6

59.4

22.1

0.9

0.0



That the authors consider this a homogenous Northeast Asian population when it clearly isn’t is of importance, because it  shows how they miss particular details and this influences their later analyses and conclusions. Moving on…


TAK001 - no C14 dating?


The next point is my skepticism around the TAK001 sample. In Jeong 2020 there were a whole bunch of samples that were not C14 dated. Some of these were obviously misdated, such as the medieval Uyghur sample OLN001.B0101, identical to iron age Slab Grave samples with Y-DNA Q-M120. There were some other cases of misdated or mixed up samples too. Three of these which stood in particular to me. These samples have a few things in common:

  • All were found in Northwestern Mongolia near the Altai

  • All were nearly identical to late bronze age pastoralists from Northwestern Mongolia

  • None had radiocarbon dating


Two of these came from the Xiongnu period, SKT007 and TAK001. In addition a third sample of theirs like this was THL001, a mongol period sample from 1400 ad.


Target: RUS_Late_Xiongnu:TAK001

Distance: 2.5789% / 0.02578860

98.4 MNG_Khovsgol_BA

1.4 IRN_Tepe_Hissar_C

0.2 RUS_Sintashta_MLBA

0.0 CHN_Upper_Yellow_River_IA

0.0 MNG_Slab_Grave_EIA_1


Target: MNG_Late_Med:KHL001

Distance: 2.2203% / 0.02220321

95.2 MNG_Khovsgol_BA

2.4 CHN_Upper_Yellow_River_IA

1.8 IRN_Tepe_Hissar_C

0.6 RUS_Sintashta_MLBA

0.0 MNG_Slab_Grave_EIA_1



So colour my surprise when it turned out that the sample from tomb 82 apparently was TAK001.




In this article they did not perform C14 dating on the TAK001 sample either, which is very unfortunate. The authors seem to consider the genome to be genuinely from tomb 82 and the Xiongnu period. Given that tomb 82 had a clear Xiongnu period grave it can hardly be the result of just misdating a grave to a much later period than it actually belonged. But perhaps there was a mix-up of samples in the laboratory?


In the article they also refer to THL001:

It is rather unexpected to observe the presence of Khovsgol_LBA ancestry in a form not associated with SlabGrave ancestry during the Xiongnu period, as it had largely been replaced by Slab Grave in Mongolia by the EIA (14). Khovsgol_LBA ancestry was also reported from the Mongol era site Khalzan Khoshuu, which is located only 95 km away from the TAK site (14).

This is one of the reasons why I suspect something is off with TAK001 because there is no way in hell you have unmixed khovsgol-type populations lasting more than 2000 years in that region when we see several significant population shifts during the metal ages and historical periods. Yet the authors seem to consider this genuine, and it might be that a similar scenario is at hand for the Xiongnu period samples.


I am not outright saying that the sample is misdated or incorrectly attributed to the woman buried in tomb 82, it might be the case that you had Xiongnu elites which were unmixed with any of the iron age nomadic steppe populations or with the people of Han dynasty China and had full autosomal continuity since the late bronze age. However that seems a bit of a hard sell when none of these Khovsgol-continuity samples actually have radiocarbon dates.


If it does turn out to be legit, I’ll eat some crow and start shilling for Vovin’s Yeniseian hypothesis again. Khovsgol Q-L330 armies have gone out, they will catch Bokkok…


Depiction of Shi Le from the Ming Dynasty. Weren’t the Jié supposed to look like Hu with big beards and deep-set eyes? Maybe the painter missed that part…


Xiongnu elite ancestry associations

Together, we observe a high degree of genetic heterogeneity and diversity at the elite sites of TAK and SBB, with the highest genetic heterogeneity observed among the lowest-status individuals. In contrast, we find that the highest-status individuals in this study, the aristocratic and local elite females, tended to be less diverse (P = 0.011 for Brown-Forsythe test to compare the PC1 variances of low-status and high-status individuals) with higher levels of eastern Eurasian ancestry. Eastern Eurasian ancestry is represented by the higher PC1 values and the mean of PC1 of the high-status individuals is significantly greater than that of the low-status individuals (P = 0.032 for one-sided Welch’s t test to test the null hypothesis that PC1 mean of high-status individuals is equal to that of low-status individuals). This suggests that elite status and power was disproportionately concentrated among individuals who traced their ancestry back to the preceding EIA Slab Grave groups. Three of the six elite females, one low-status female child, and one low-status male had minor ancestry contributions consistent with Han_2000BP (14, 72), suggesting that interregional connections to groups in Han Dynasty China may have been greater, and more complex, than previously understood.

Before this study, only one other individual from an elite square tomb had been analyzed in a genome-wide manner: DA39 from Tomb 1 at the imperial elite site of Gol Mod 2 in central-north Mongolia (13). This adult male, buried in one of the largest square tomb complexes excavated to date, surrounded by at least 27 satellite burials, and containing rare exotic items such as Roman glass bowls, was likely a chanyu, or ruler of the empire (73, 74). Like the elite women at the western frontier, he also had very high eastern Eurasian ancestry (deriving 39.3 and 51.9% from SlabGrave1 and Han_2000BP, respectively, and the rest from Chandman_IA; data file S2C) and was genetically similar to TAK002 in tomb THL-64 (Fig. 3B). Such patterns of ancestry, stratified by indicators of status and power, provide clues as to the nature of the political formation of the Xiongnu and the relative power dynamics of the empire’s diverse political actors.

Ah the dreaded “elite xiongnu graves” discussion. I’ve said several times how I think this aspect of Xiongnu academia is a red herring, in part due to the fact that all of these elite graves are from late periods of the Xiongnu empire. Imagine a discussion about the genetic origin of elites of the Roman empire based on samples from wealthy graves after the third century crisis, this would be comparable to that. In any case, I found this phrase in particular a bit jarring:

This suggests that elite status and power was disproportionately concentrated among individuals who traced their ancestry back to the preceding EIA Slab Grave groups.

Like I explained before, Q-M120 is by far the most significant paternal lineage in the Ulaanzuukh and Slab Grave populations. This lineage was clearly spread through a patrilineal tradition, at the expense of the lineages in Neolithic Mongolia. We do see some outliers with different lineages such as Q-L330 and N-M2019 in Wang’s article, but these are easily demonstrable to deviate from the core population in their autosomal composition.


During the Xiongnu period, this core Slab Grave lineage had a completely different story. We do have a decent amount of samples from “Genetic evidence suggests a sense of family, parity and conquest in the Xiongnu Iron Age nomads of Mongolia”with Q-M120 as five samples carried it, but this is simply due to the dataset coming from a family grave. In Jeong’s dataset you have IMA002 with Q-M120, and that is about it.


The fact that Slab Grave might not be a singular ancestor is somewhat earlier in the article, but they did a poor job at it:

We note that all qpAdm admixture models equally fit when SlabGrave1 was replaced by “AR_Xianbei_IA” from the :Mogushan archaeological site in Inner Mongolia that belongs to the Iron Age Xianbei context (data file S2C) (27). All but two males (BUL002 and I6365) associated with the Ulaanzuukh and Slab Grave cultures belong to Y-haplogroup Q, all three AR_Xianbei_IA males belong to Y-haplogroup C, and the Xiongnu males harbor both Q and C (data file S1C) (14, 15). Although not conclusive, this suggests that the ANA ancestry source of the Xiongnu-period individuals may not be exclusively traced back to the Slab Grave culture but may also include nearby groups with a similar ANA genetic profile, such as the Xianbei.

May also include? May? C-F1756 is more significantly present in these samples than Q-M120. And then they bring up the Q in Xiongnu samples to link it to Slab Grave, but the Q in Xiongnu samples is primarily not of this kind. I count these Xiongnu samples from Jeong which carry Q:


IMA004: Q-L330

BTO001: Q-L330

SKT008: Q-L330

DEL001: Q-L330

IMA006: Q-M120


You also have SKT007 with Q-L330 but I think this is a misdated sample as well. The separation of Q-L330 and Q-M120 is nearly 30000 years ago! Because of Slab Grave outlier I6359 I can already smell a counter-argument: Slab Grave samples also had Q-L330 so maybe the Xiongnu Q comes from these Slab grave Q-L330.


Target: MNG_Slab_Grave_EIA_1:I6359

Distance: 2.5405% / 0.02540503

78.2 MNG_Khovsgol_BA

21.6 MNG_Slab_Grave_EIA_1

0.2 CHN_Upper_Yellow_River_IA


Note that this sample, a genetic outlier (so not genetically Slab-Grave) and the only of its kind uncovered so far had the subclade Q-L330 > Q-YP771 > Q-BZ180 according to Haplotreeinfo. Q-BZ180 being about 6000 years old.


The Xiongnu Q-L330 carriers had these subclades:

IMA004: Q-L330 > No further designation

BTO001: Q-L330 >  Q-YP771 > Q-BZ180

SKT008: Q-L330 > Q-YP771 > Q-BZ433 > Q- L332

DEL001: Q-L330 > Q-YP771 > Q-BZ433 > Q- L332


As you can see, a variety in Q-L330 going back thousands of years prior to the formation of these iron age cultures, with one Xiongnu having a match with the Q-L330 clade found in the Slab Grave outlier. If we extend Xiongnu to broader Hunno-Sarmatian period samples we also have Q-M25 such as Q-L715 present amongst samples from the Bulan-Koba culture, whose samples seem like Eastern Scythian remnants. 


All these clades come from Khovsgol-related populations and such clades were common in Scytho-Siberian populations, the Chandman samples from Jeong even carry this. These do not have Slab Grave origins. The only Xiongnu sample among the list with Q-M120 is IMA006.


So by en large the Q they are referring to in Xiongnu has little to do with Slab Grave. The C lineage they refer to, C-F1756 was not found amongst the Slab Grave samples, and clearly come from a population that was different from the currently sampled Slab Grave population.


Funnily enough DA39 was mentioned, the only male sample from these aristocratic elite burials we have coming from the Arkhangai tomb. What was not mentioned in the article is that this sample was R1a-Z93, a lineage not found in any of the Slab Grave samples but was ubiquitous amongst Scytho-Siberian populations as it was the founder lineage of Indo-Iranian peoples.



From the perspective of Y-DNA haplogroups, there is no large Slab Grave continuity or contribution to the Xiongnu, nor is there an association with Slab Grave haplogroups to be found in elite samples. The authors do not make a strong case for such a Slab Grave continuity when there is this much of a discrepancy between Slab Grave haplogroup lineages and Xiongnu lineages, nor is their case strengthened when the samples do not showcase a closer affinity with Slab Grave than with other Northeast Asian populations and could equally be modeled with different Northeast Asian source populations such as the Xianbei samples from the Hulunbuir region.


What about a correlation between elite status and Slab Grave autosomal ancestry then? I decided to divide the samples into three groups: aristocratic elites from the terrace tombs, local elites, and commoners in the same trend the authors did, with source populations similar to ones the authors used.


Aristoricratic elites:

Target

Distance

CHN_Upper_Yellow_River_IA

MNG_Slab_Grave_EIA_1

MNG_Chandman_IA

MNG_Khovsgol_BA

MNG_Xiongnu_East_Asian:DA39

0.02404569

48.0

45.2

6.8

0.0

TAK002:TAK002.A0101

0.01626133

44.6

52.0

1.2

2.2

RUS_Late_Xiongnu:TAK001

0.02594314

0.0

0.8

3.0

96.2

Average

0.02208339

30.9

32.7

3.7

32.8


If we take TAK001 to be legit as the authors do,  the amount of Slab grave ancestry in these three comes to about 33%. If we don’t, then it is around 50% ish. Also funnily enough TAK001 has more west eurasian ancestry than a bunch of the Shombuuzyn Belchir samples due to its Ancient North Eurasian ancestry but was described as eastern.


The “local” elites of Shombuuzyn Belchir:

Target

Distance

MNG_Slab_Grave_EIA_1

MNG_Chandman_IA

MNG_Khovsgol_BA

CHN_Upper_Yellow_River_IA

TKM_Gonur1_BA

SBB002:SBB002.A0101

0.04150491

98.4

0.0

0.0

0.0

1.6

SBB003:SBB003.A0101

0.03261868

43.0

39.6

0.0

11.0

6.4

SBB007:SBB007.A0101

0.02748112

97.2

0.0

0.0

0.0

2.8

SBB008:SBB008.A0101

0.02034559

55.2

9.0

21.8

9.8

4.2

Average

0.03048757

73.5

12.2

5.5

5.2

3.8


Or simply said the women of Shombuuzyn Belchir, as all of these were women. As you can see these samples are mostly of ANA extraction highlighted by the high amount of Slab Grave ancestry in this model. However, SBB003 has significant input from Scytho-Siberian populations too and some southern Central Asian input in addition, probably coming from the same source.


The “low status” commoners of SBB and TAK:


Target

Distance

MNG_Chandman_IA

MNG_Slab_Grave_EIA_1

TKM_Gonur1_BA

MNG_Khovsgol_BA

CHN_Upper_Yellow_River_IA

SBSBB001:SBB001.A0101

0.02999014

55.4

41.8

2.8

0.0

0.0

SBB010:SBB010.A0101

0.03013561

49.6

40.4

8.4

0.0

1.6

TAK006:TAK006.A0101

0.05803741

0.0

100.0

0.0

0.0

0.0

TAK008:TAK008.A0101

0.03663809

76.0

0.0

10.8

13.0

0.2

TAK009:TAK009.A0101

0.03378645

61.6

0.0

15.4

22.0

1.0

Average

0.03771754

48.5

36.4

7.5

7.0

0.6


On the flip side, all of these samples are male. The three from Takhiltyn Khotgor came from satellite burials and the two from Shombuuzyn Belchir had “poor” burial inventory.


So if we’d order these three groups by the degree of “Slab Grave” related ancestry, then the first place goes to group 2, the  “Local elite” of Shombuuzyn Belchir. The third place would actually be the aristocratic elites if you count TAK001 as the authors do, but if you don’t they would be second place with the commoners having the lowest amount of Slab Grave related ancestry.


What is clear is that the strongest correlation is with Han-related ancestry, clearly due to Heqin policies. 2/3 or if you ask me 2/2 of the aristocratic elite samples had significant Han-related ancestry.  Finding small amounts of Han ancestry in “lower status” samples does not negate this, just like having low status samples model 100% as Slab Grave is not an argument against the authors’ claims that Slab Grave is disproportionately associated with elite status.


The most important point however is that these inferences are made based on data retrieved from very small sample sizes. We are discussing the genetic dynamics of elite and commoners of an empire lasting nearly three centuries based on a handul of samples from two archaeological sites covering a short window of the imperial period of the Xiongnu.


Low status at Takhiltyn Khotgor

All this talk of high status Xiongnu samples brings me to another point, and that is that the smaller graves of Takhiltyn Khotgor were referred to as “low status” by the authors. You can wonder how low status these burials would have been if they were included in a cemetery dedicated to someone clearly of very high standing within the Xiongnu empire. This lower status seems based on burial size and grave goods compared to the large terrace tombs, as well as the physical condition of the people.


These three skeletons all had significant wear on their skeletons and were described as having robust muscle insertions, a sign of significant physical labour. This could be interpreted as having a lower status, but this is troubled by something I have seen quite consistently in physical descriptions of iron age nomadic elites in soviet academic circles, which is that the elite males are commonly described as being very well-developed in terms of muscles. These elite males in these societies were not doing hard labour, but they certainly didnt sit on their saddle all day doing nothing.


One such example was the eite chieftain from Tugozvonovo, a Hunno-Sarmatian period site of the Altai, described as a “heroically built” man of tall stature (187cm) with “unusually developed arm and leg muscles”. Another example of muscular elite males are the Pazyryk culture males from Badashar and Tuekta, described here.


You can also be certain that TAK008, TAK009 and TAK006 were not native to the Altai region given that their autosomal profiles point to Eastern Xinjiang for the first two and The Khingan range for the latter one. Although without isotopes we cannot be certain where they grew up, this could be in the area they were buried in. Let's say these people were buried here without a family nearby, who is constructing their graves? Who is providing their graves with burial goods?


This isn’t to say that I think these guys were the secret elites or something, but that I think the “low status” designation itself is not exactly strongly demonstrated. I like Miller’s presentation on commoners of the Xiongnu elite, Miller was also one of the co-authors of the article. Here he paints a more nuanced picture, also asking the same question whether how much of a “commoner” the people interred in these satellite burials actually were:



Shombuuzyn Belchir

Focusing on Shombuuzyn Belchir in particular:

At the SBB site, we found lower overall genetic diversity, and specifically, no individuals with very high levels of western Eurasian ancestry (Fig. 3E). However, the individuals with the highest levels of western Eurasian ancestry were both adult males (SBB001 and SBB010), although they derived their western ancestry from slightly different sources (Chandman_IA for SBB001 and Chandman_IA and Gonur1_BA for SBB010). As at TAK, the highest-status graves belonged to females (SBB002, SBB003, SBB007, and SBB008), whose modeled ancestries all included SlabGrave1, with other minor ancestry contributions. The genetic determination of SBB007 as female was particularly noteworthy because the grave goods included horse-riding equipment, a gilded iron belt clasp, and a Han-painted lacquer cup, which have been assumed in other contexts to be accouterments associated with male horse-mounted warriors. Similarly, SBB010, an adult male, was buried with a bone tube case containing an iron needle, indicating that sewing implements were not exclusively associated with women.

SBB003 on G25 could be modeled as being 40% Chandman derived or 50% derived from a Sagly-like source. How is that minor? It is nearly half of her ancestry.


In regards to the data of this site being extrapolated to “local elites” across the Xiongnu empire as a whole, I have my gripes with this as we clearly see a closely related population featured. The rate of homozygosity between some of the samples attests to this. Sure there is internal genetic variation, but the two y-chromosomes are of the same C-F1756 lineage which connects to the core ancestry of the Shombuulyn Belchir populations. Ironically the samples with this eastern Y-chromosome haplogroup have significant Scytho-Siberian ancestry, and indication of patrilocality of the community perhaps.


You know what this pattern reminds me of? The Pannonian Avars. You had a population where both the elites and commoners carried the same Y-chromosome haplogroup and while there was genetic variation due to genetic intermixing, these belonged to the same population based on paternal descent. Pretty much one large tribe, or a population which grew out of one. It seems to be the case that the Shombuuzyn Belchir site represents a similar case.


If you look at the graves themselves, they largely seem to be conducted in the same manner. I didn’t see anything particularly pointing towards elites when I first came across the site. The distinction in class, commoner versus local elite,  is based on the amount of burial goods present in the graves.



The “commoner” samples at Shombuuzyn Belchir are also the only males from Shombuuzyn Belchir, and the wealthy burials are also just all the women of Shombuuzyn Belchir. A pattern?


Xiongnu Gynocracy

Last, the concentration of wealth and elite status specifically among women in these frontier communities warrants further attention. It has been previously noted that Xiongnu-era female graves in peripheral regions of the empire tend to be especially wealthy and high status (75, 76). The elite women’s graves at TAK and SBB conform to fashions befitting higher-ranking persons in Xiongnu society. The association between the social rank and biological sex is statistically significant with P value of 0.002 for the Fisher’s exact test. The only females not elaborately buried in wood-plank coffins were children. The prominent status of Xiongnu women at TAK and SBB speaks to the powerful role of women in the empire and their likely prominent place in strategies for expansion and the integration of new realms and territories.

This was also a rather odd interpretation of the data. Think of the process:

  • Burials with high amount of grave goods are uncovered

  • These burials are then proclaimed to be local elite tombs of the Xiongnu

  • These burials then only contain women


Conclusion? Xiongnu women had a powerful role in the empire and their expansion.


I want to point out that these areas had been under Xiongnu control since about 200 BC, these 50 BC burials do not represent a Xiongnu expansion. What they might represent is the consolidation of Xiongnu power in these regions in the first century BC as you had rebellions and self-proclaimed Chanyus in this era during the period of the first Xiongnu civil war, which slightly predates these tombs.


Back to the burials. Maybe these elite burials as extravagant as they are, might not be a all that representative of the elites within the Xiongnu empire. Not only are they very late but apparently these are disproportionate towards women.  Does this attest the power of women within the Xiongnu? How can we be sure that their wealth and splendour attests to their power and prominence, rather than those of their fathers and husbands? A queen does not make a matriarchy after all.


Now we know that female lineages do play somewhat of a role here as within the social order of the Xiongnu elite. You had clans who were in the elite circle because their women were married to the men of a particular clan. The maternal dynastic tribes so to say. The Luandi clan of Touman and Modu had such a bond with the Huyan clan. The Huyan clan was then later replaced with the Xubu clan.


However the power in the Xiongnu realm clearly lay with the men. Wise king on the left, wise king on the right, the Chanyu being a patrilineal dynasty and most of the Xiongnu populations having patrilneal social structures.


It begs the question, in a society of kings and chieftains,why is there such a weak representation of elite males in these two Xiongnu burial sites? Were their burials hidden to prevent looting? Probably not since DA39 comes from a quite extravagant burial. Were they buried somewhere else then? This might be a possibility.


In any case, what is clear to me is that we have an incomplete understanding of these elites based on the current excavated graves and the ancient samples from these graves. All the Xiongnu aristocratic  elite tombs come from the latter part of the first century BC or later, which is the final part of the Xiongnu empire. Burials identified as local elites due to a higher amount of burial goods are incredibly shifted towards women, which cannot be representative of the actual elites. Most of the common burials belong to men, all the males in this article belong to so-called commoner or low-status burials.


These are the main points of the article which I contend. My questions for the authors on the genetic front would be:

  • What guarantee is there that TAK001 is not the genome of a misdated sample or a mis-attributed sample when no radiocarbon dating had been done on the sample since published and in this article?

  • If genuine, how do the authors explain a near-complete autosomal continuity with samples more than a thousand years earlier when Scytho-Siberian populations were significantly present throughout the iron age in Western Mongolia, and Central-Eastern Mongolia housed Slab Grave and pre-Xiongnu populations? How does THL001, a supposed medieval mongol period sample of the same autosomal profile fit in this?

  • Had the authors been aware of the variety of Yellow river related ancestry in the Ulaanzuukh and Slab Grave populations as well as the discrepancy between the Slab Grave and some Hunnic samples on this front, would they still have proclaimed Slab Grave as the primary ancestral population of the Xiongnu and their elite?

  • Were the authors aware that most Xiongnu Q were of entirely different branches than the Slab Grave Q, also coming from different populations? If not, how does this affect their conclusions that Xiongnu came from Slab Grav based on the presence of Q in both populations?

  • If the AR_Xianbei samples worked equally well as Slab Grave and had a stronger y-dna match with several previously published and three (out of five) of the new published samples, why was Slab Grave used as a primary source?


I have some questions for the archaeologists involved too:

  • Considering that not all elite terrace tombs belonged to women (DA39), what does the gender ratio of other excavated terrace tombs look like?

  • Taking Shombuuzyn Belchir as a case site, how do the archaeologists interpret the finding that all wealthy burials belonged to women and that commoner burials belonged to men? Might these findings highlight issues in the methodology of determining elite status based on burial inventory?

  • Could there be gender-specific rites relating to burial goods, creating a false sense of social strata based on burial goods? Or were male “local elites” buried elsewhere? 

  • How do authors explain the lack of elite burials uncovered from the early phase of the Xiongnu empire in general?

  • Will there be future archaeological projects in China to bridge the gap between Xiongnu history and XIongnu archaeology, the latter almost entirely focused on Mongolia while China plays a large role in the former?


I hope at least one of the authors responds to these questions. While an interesting topic, one dear to my heart, I find that the authors have come to conclusions which I find a bit far-fetched and sometimes even at odds with their own data. More troubling is how one of the key samples from the article, TAK001, has a dubious genetic profile for its period and is not C14 dated.


If anyone has some other points of contention please post them here as well. If you want to help me out you could share this with other academics dealing with the Xiongnu, or steppe nomad history in general. In the meantime I’ll continue writing on my next blog entries, if all goes well there should be a very interesting post about the various ancient peoples of Xinjiang coming out in a short while.


See also:

1 comment:

  1. I created a Blogger account just to post this comment that I appreciate your efforts in bringing us these detailed and very interesting articles. I am one of those silent readers, but thought now was finally the time to show my appreciation for your blog. Keep up the good work!

    ReplyDelete