Saturday, December 30, 2023

An individual with or without Sarmatian-related ancestry in Roman Britain?

Just over a week ago this article came out. I live under a rock so I have no idea what kind of coverage this has in mainstream media but knowing how british media has reported on ancient DNA in general - I don’t want to find out.

An individual with Sarmatian-related ancestry in Roman Britain Marina Silva, Thomas Booth Joanna Moore, David Bowsher, Janet Montgomery, Pontus Skoglund Published: December 19, 2023 DOI: 

In the second century CE the Roman Empire had increasing contact with Sarmatians, nomadic Iranian speakers occupying an area stretching from the Pontic-Caspian steppe to the Carpathian mountains, both in the Caucasus and in the Danubian borders of the empire.1,2,3 In 175 CE, following their defeat in the Marcomannic Wars, emperor Marcus Aurelius drafted Sarmatian cavalry into Roman legions and deployed 5,500 Sarmatian soldiers to Britain, as recorded by contemporary historian Cassius Dio.4,5 Little is known about where the Sarmatian cavalry were stationed, and no individuals connected with this historically attested event have been identified to date, leaving its impact on Britain largely unknown. Here we document Caucasus- and Sarmatian-related ancestry in the whole genome of a Roman-period individual (126–228 calibrated [cal.] CE)—an outlier without traceable ancestry related to local populations in Britain—recovered from a farmstead site in present-day Cambridgeshire, UK. Stable isotopes support a life history of mobility during childhood. Although several scenarios are possible, the historical deployment of Sarmatians to Britain provides a parsimonious explanation for this individual’s extraordinary life history. Regardless of the factors behind his migrations, these results highlight how long-range mobility facilitated by the Roman Empire impacted provincial locations outside of urban centers.

To put it bluntly, this article has a lot of issues. The focal point of the article is that Offord Cluny 203645l was of Sarmatian descent, Sarmatian being mentioned like 35 times in a 2-minute article and features in the title. Yet this individual does not seem to be very Samatian at all. In fact, none of the analyses provided actually seem to be in favour of this individual being a Sarmatian.

Even stranger seems to be this hammering on the potential military context of Offord Cluny 203645, a narrative the authors pulled out of thin air as there is no evidence pointing towards this. You would think it is as if they found him in a Centurion’s grave or buried in complete steppe nomad body armour, but no he is from some unmarked burial on a farm.

This article highlights a growing concern of mine regarding the quality of some academic articles covering ancient populations by way of ancient DNA. I might be wrong here, but it is my impression that over the last few years there have been quite an amount of articles that either have poor methodologies, incorrect information or dubious interpretations of data, and I feel that the amount of these has increased when compared to 2017 or so. It might simply be that there is a larger amount of articles being published compared to five years ago, but I also have the feeling ancient DNA articles do not get the rigorous reviews or pushbacks they require. This article for instance had three anonymous reviewers and it got published in Current Biology, which is supposed to be a reputable journal.

You might be surprised by my grievances in regards to this paper because on first glance it probably looks alright to you. I mean, people did move across the Roman empire so what is the issue with a Sarmatian individual in Britain?

A few months ago I came across the abstract posted below, and after I had read it I already had some concerns about this article:

My suspicions arose from the obfuscation of the Steppes and Caucasus as if it is “some distant land over there” when these are two highly distinctive regions, the mention of C4 plant heavy diet as well as the lack of WHG ancestry, a component carried in small figures by sarmatians.

Although the published article seems a bit toned down from this abstract based on the title, it suffices to say that my suspicions were warranted. In this article they used three channels of data to come to the conclusion that this individual was a Sarmatian; genomic data, isotopic data and historical records. 


The main point of evidence in this article comes from the genetic side. As it is only pertained to one sample you can briefly read through their methodology.

From the PCA in the article you can see that Offord Cluny 203645 sits close to the more steppe-shifted Late bronze age samples from Armenia, as well as the less steppe-shifted individuals from the early medieval Alanic period, both of these placed somewhere in between Sarmatians and Armenian samples from antiquity.

In the eastern Caucasus you have populations with significant Steppe_EMBA ancestry and the people have genetic makeups related to those of late bronze age Armenians. In the western Caucasus there probably was gene flow from Srubnaya-related populations in the late bronze age,  as evidenced by the Himera outlier from the Caucasus with R1a-Z93, steppe_MLBA ancestry yet lacking east asian ancestry typical of iron age nomads.

Given the overlap in other ancestries such as steppe, anatolian neolithic farmer, caucasus hunter-gatherers, neolithic iranians etc. this might mean that the affinity towards Sarmatians is superficial and simply the result of significant Steppe_EBA or Steppe_MLBA ancestry.

The aforementioned scenario might be relevant for this individual as a single source model containing just Armenia_LBA managed to pass quite successfully:

We first tested different single-source qpWave models rotating through different populations (STAR Methods), with a focus primarily on populations from the Caucasus and the Pontic-Caspian steppe, in addition to other populations from south and northern Europe (Figure 2A). The only accepted single source is Armenia_LBA (p values = 0.345 and 0.560), whereas Armenia_Antiquity, Sarmatian groups, and populations from Britain (England_Roman or England_IA) are rejected as single sources (Figure 2B and Data S2B).

The author, being under the impression that Armenia is representative of the Caucasus as a whole, then postulated that since Armenia_Antiquity samples carried different genetic profiles compared to Armenia_LBA, Armenia_LBA could not be a proper source population and thus requiring input from populations such as Sarmatians:

“However, Armenia_LBA dates to ∼1200–850 BCE and thus predates Offord Cluny 203645 by up to approximately one millennium. Recent studies revealed ancestry changes in Armenia during the first millennium BCE, which resulted in different ancestry patterns in the region by the time of Offord Cluny 203645.12,13 Therefore, Armenia_LBA is likely not a good representative of the ancestry observed in the Caucasus in the first millennium CE (Figure 1C).”

Had the authors looked more carefully at their own PCA they would have seen a bunch of modern Caucasian ethnic groups clustering fairly closely to Armenia_LBA anyways, showcasing that populations with similar genetic profiles must’ve persisted from the iron age to our current day. In any case imagining that the armenian samples from bronze age to antiquity are a good representative of the Caucasus as whole shows a serious lack of awareness when it comes to the region, as it is known for many ethno-linguistically distinct populations and sharp genetic distinction over short geographic distances. And it always has been that way for a long time if this quote from Strabo is anything to go by:

“All speak different languages because of the fact that, by reason of their obstinacy and ferocity, they live in scattered groups and without intercourse with one another.”

- Strabo, Geography, Book XI, Chapter 2 

The description of this sample based on genetics reminds me of that individual from Himera, as it also passed with a single-source model containing Armenia_MBA [1]:

Sicily_Himera_480BCE_5 This individual could be modeled as a 4-way model (p=1.73E-2; Dataset S5), as a 4-way model of Turkey_N_Barcin (33.4 ± 3.7%), WHG (1.3 ± 3.5%), Russia_Samara_EBA_Yamnaya (20.1 ± 6.4%) and CHG (45.2 ± 5.8%). This is consistent with no WHG contribution; however, the corresponding 3-way model is rejected (p=4.02E-3). Using more proximal sources of ancestry, we tested 1-, 2-, and 3-way models of Sicily_IA, Greece_Crete_BA, Greece_LBA, Turkmenistan_IA.SG, Armenia_EBA, Armenia_MBA.SG, Armenia_MBA, Armenia_LBA.SG, Iran_Hasanlu_IA.SG, CentralSteppe_IA_Nomadic_Steppe, TianShan_IA_Nomadic, TianShan_IA_Nomadic_o3, CentralSteppe_IA_Nomadic_o, Turkey_IA_o3.SG, WesternSteppe_IA_Nomadic_intermediate, Tajikistan_Ksirov_Kushan and Iran_IA_HajjiFiruz. A single 1-way model with Armenia_MBA as the source was valid (p=0.293).

Given that this individual is from the antiquity period and lacks Scythian/Sarmatians input, yet is still similar to Armenia_MBA/LBA populations shows that the methodology of the authors does not hold up.

The proximate models of the samples which passed were two-way models with Armenia_antiquity and two groups of Sarmatian samples.

I pasted values from the two proximate models seen above in this graph here:




source 1

source 2




Russia_Sarmatian_PonticSteppe, Armenia_Antiquity







Russia_Sarmatian_SouthernUrals, Armenia_Antiquity






The P-values on this individual are not exactly stellar, and I know that we are not supposed to chase p-values but the one-way model on the other hand had a significantly higher probability value (0.3451717) than the pair of two-way models.

You can see the other models which the authors tried below, as you can see it is not exactly a rigorous set of models which they tested, especially as most are completely redundant given what we know of the ancestry of the individual.

The exclusion of Armenia_LBA ancestry in this set of rotating models is an odd choice as well given that it worked quite well as a single source reference population for this individual. At the very least they should have run some models combining Armenia_LBA and Sarmatian, or perhaps attempted some three-way reference models in addition to their one and two-way models.

If we cannot look at affinities towards steppe ancestry as an indication of Sarmatian ancestry, what about other components? Strangely enough the authors never managed to do the most basic sanity check when it comes to modelling any potential Scythian-derived population - did these populations carry East Asian ancestry?

Sarmatians were definitely on the low end of East Asian ancestry in Scytho-Siberians but on average should carry about 9-10% or so. This indicates that if you have 10% Sarmatian ancestry, about 1% of your ancestry should be East Asian in origin. 

Unfortunately their PCA was not really designed for this purpose, although populations with higher shares of East Eurasian ancestry have a more northwards position on this PCA. When looking at the position of this sample he plots within the upper ranges of Arm_LBA populations and very closely to the Alanic period samples furthest removed from Sarmatians, which does not point in favour of there being an excess affinity towards East Asians relative to Armenia_LBA.

Given the methodology of the authors, it is basically impossible to tell how much Sarmatian ancestry this individual would have had. Considering that a one-way model with Armenia_LBA passes and there is no clear information on affinities towards East Asian ancestry it might be the case that this individual lacks Sarmatian ancestry completely. 

Perhaps haplogroups can tell us more?

Analysis of the Y chromosome and mitochondrial DNA (mtDNA) of Offord Cluny 203645, tracing paternal and maternal lineages, respectively, also point to ancestry from outside of western Europe, in particular his paternal lineage: R1b-Y13369 (a sub-branch of R1b1a1b1b/R1b-Z2103) (Data S1B). This lineage has been previously identified in skeletal remains ranging from the Late Bronze Age to the Urartian period recovered from present-day Armenia,13 whereas its present-day phylogeny is dominated by samples from the Caucasus, Anatolia, and Near East (Yfull tree v.11.01.00). Offord Cluny carried mtDNA haplogroup K1a (Data S1C), found in Pre-Pottery Neolithic Anatolia and the Levant, and in Europe since the Neolithic.14,15 Although subclades of haplogroup K1a, found at frequencies of ∼5% across all regions in the UK Biobank dataset,16 have been previously identified in ancient individuals from Britain ranging from the Neolithic to the early Medieval period, these all belong to different sublineages than the one observed in Offord Cluny.

So R1b-Z2103, the Yamnaya clade, is an interesting finding as it is a haplogroup shared between Caucasian populations and iron age steppe populations, although not frequently in the latter. The subclade Y13369 identified by the authors is upstream of the L584 clade which as you can see is clearly related to the migration of steppe populations across the Caucasus in the bronze age - evidenced in both ancient samples and modern distributions.

At Ftdna they had further investigated into the clade and uncovered this:

"Offord Cluny 203645" forms a new branch together with old and new Big Y testers under R-L584. The customers' ancestry is reported to be from around Georgia. Two new branches in total: R-M207>M173>M343>L754>L761>L389>P297>M269>L23>Z2103>M12149>Y13369>L584>FTA62508>FT147157 (new)>FTF5689 (new)

When it comes to Sarmatians one example with Z2103 is I0575, carrying the subclade R-Y23099. This clade has a completely different distribution and history compared to the subclade seen in the Offord Cluny sample, which clearly derived from the pool of Z2103 clades which had been present in the Trans-Caucasus since the bronze age. 

On the maternal side, Offord Cluny had mtdna haplogroup K1a, a haplogroup which was not common amongst Sarmatians at all. In my current list of ~60 Sarmatian samples there are none with Mtdna K1a actually.


Overall, our results suggest that there may have been substantial diversity among groups identified as Sarmatians, some of which could have had ancestry that in our data is most closely represented by Armenia_Antiquity.

This just simply is not the case. The only “diversity” is seen in two particular batches of samples; the samples from the Alanic period in the North Caucasus, and samples from the Sarmatian period in Hungary. In both these regions Sarmatians migrants were not the only populations inhabiting the area. We cannot assume that genetic samples from such a mixed context were Sarmatians or identified as one, nor should genomes of individuals who had migrated to other areas and intermixed with the preceding inhabitants be interpreted as genetic diversity within the source population.

If we look towards the Sarmatians from proper Sarmatian cultures we actually see a fairly tight genetic cluster across hundreds of kilometres of geographical distance. Sure over different time periods and in different areas the ancestries shift a bit, but if you take a random Sarmatian from Moldova, and one from Kazakhstan, they would be highly similar.

As it currently stands, the analyses of autosomal ancestry done by the authors insufficiently demonstrates that Offord Cluny 203645 had Sarmatian ancestry. Neither paternal or maternal haplogroups provide a direct link to Sarmatians, making it unlikely that any of his recent ancestors (parents, grandparents) were Sarmatians.

It is possible this individual would have carried some Sarmatian ancestry though. After all, several ethnic groups in the Caucasus nowadays have some Sarmatian ancestry and it would be around this period that this started flowing in. It is just the case that the author’s efforts fail to properly demonstrate to. But even if this was the case, the authors should have called this article a Caucasian-related individual in Britain, not a Sarmatian-related individual, as such gene flow was becoming part of the Caucasian genetic makeup (in some areas).


Beyond genetics, the authors also invoked isotopic data to further their claims that Offord Cluny 203645 was of Sarmatian origins. Unfortunately my knowledge in this specific aspect is a bit sub-par compared to ancient DNA data, and thus I will have to assume that the values and meaning of these values provided by the authors are correct.

The main part here is the data which reveals information about the diet of the Offord Cluny individual, giving insight into the lifestyle of the individual in his earlier years in life.

Offord Cluny 203645 had high δ13C values combined with low δ15N values, indicating a childhood diet rich in non-native C4 crops with little input from marine resources. Incremental dentine analysis (Figure 3B and Data S3B) revealed that his diet underwent a substantial change around the age of 5 years, when δ13C values drop from ∼−12‰ to ∼ −16‰, reflecting a clear shift from eating predominantly C4 plant protein to eating a mixed C3/C4 diet with a possible increase in meat protein indicated by a concomitant rise in δ15N. A second change in diet occurred after the age of 9, when the δ13C profile started falling, reaching ∼ −19‰ around the age of 13, which is approaching an entirely C3 based diet. As there is no clear evidence of wide consumption of C4 crops during the Roman occupation of Britain (despite some sporadic findings of millet)30 and they were not common components of diet in western provinces of the Roman Empire, these two shifts in diet could represent a relocation around the age of 5 years old and again, after the age of 9 years old, which could reflect at least two periods of movement across Europe within the first ∼14 years of his life. It is not possible to distinguish a gradual one-way transition in diet over several years of life from a fairly rapid change, due to increased overlapping in the orientation of the dentine incremental layers.31,32 Nevertheless, the gradual drop in δ13C values observed after the age of 9 could reflect either a sustained increased consumption of C3 crops over several years or possibly a multi-year migration, e.g., westward across Europe to Britain, through regions of gradually diminishing availability of C4 foods such as millet.

The first thing that stood out was the inference of a low contribution from marine resources to the early childhood diet of Offord Cluny. This is a bit of a contrast with the lifestyles of Sarmatians. Sarmatians were pastoralists and most of their food came by way of their domesticated animals, but given the proximity to water bodies such as the black sea, caspian sea, volga, don, kuban etc. marine resources played a noticeable part in the diets of Sarmatians.

Evidence supporting this can be found in the presence of the diphyllobothrium tapeworm in Sarmatian sites in the Volga-Ural region,likely acquired through the consumption of undercooked fish [2].. 

In the North Caucasus steppes, Sarmatian individuals from Klin-Yar consumed a significant amount of fish, leading to a reservoir effect that distorted their radiocarbon dates and made them appear much older than their actual age [3].

Sarmatian male from the Klin-yar cemetery.

The most significant insight in the diet relates to the low figures of  δ13C in the early parts of childhood, consistent with a diet significantly composed of C4 plants such as millet. As you can see in the graph there is a very sudden drop around age 5, with some recovery before another drop commencing around age 9, declining further.

The issue with this that in the regions inhabited by Sarmatians, millet was a secondary grain compared to wheat [4]:

Intensification of wheat and barley consumption in the Trans-Urals occurred during the Iron Age when Sauro-Sarmatian and Sargat interaction spheres spread across large swaths of the Eurasian steppe, evident in shared prestige good assemblages, warrior equipment, and mortuary rituals. This region was deeply involved in exchange with areas in central and southeastern Kazakhstan that consumed millet, yet Trans-Ural groups opted instead to grow wheat and/or barley. Early evidence for a dietary focus on wheat and/or barley, rather than millet, suggests that Trans-Urals populations may have been trading these cultivars to other areas of the steppe that lacked the water resources to cultivate these crops. Investments in different farming technique or a higher value placed on wheat and/or barley may also have been factors affecting the decision to focus on the production of these grains.

More specifically within the Caucasus area, the data from Kichmalka II and Klin-Yar  to the Sarmatian and Alan periods also reveal information on the plant consumption within the region [5]:

There was a significant similarity between the groups of the Sarmatian and Alanian periods (p > 0.05): among the main part of the people, values δ13C and δ15N varied in less than 1 ‰. One sample among the human samples of the Sarmatian period was distinguished by higher values of δ13C and δ15N. Unfortunately, only two isotope signals of domestic animals (sheep/goat and horse) of that period are available. As in the case of the Koban people, ∆δ13 C human-animal was over 2 ‰, and ∆δ15 N human-animal was less than 5 ‰ (Fig. 3). The data obtained indicate that people of the Sarmatian period consumed mainly C3-plants and a sufficient amount of meat and/or milk from animals that fed on this type of plant. The diet of some individuals could have contained a certain share of C4-plants (burial 31).

As opposed to the two periods mentioned above, the difference between the average δ13 C values in bone collagen of humans and animals of the Alanian period did not exceed the trophic step (Fig. 4). The δ13 C values of almost all the samples (with the exception of one, from burial 28 (b)) indicated a diet based on C3-plants. The data obtained do not contradict the results of archaeobotanical studies in the Alanian settlements of the 5th–8th centuries AD located in the areas of the Kislovodsk Basin adjacent to the Kichmalka II site. Wheat dominated in the regional archaeobotanical complex of the time (65 %), while the share of millet was only 9 % (Sergeev, 2021). One may conclude the dominance of C3-plants in the diet of the Alanian population of the 5th–8th centuries AD.

Thus amongst Sarmatians or people living a Sarmatian lifestyle, we should expect signals that indicate a small contribution of millet to the diet, as well as an increased reliance on animal protein. In the mountainous regions of the Caucasus proper however, things look quite different, as millet was intensively cultivated since the later bronze age [6]:

The intensification in millet consumption in the northern Caucasus by c. 1200 cal BC (Fig. S4) coincided with the emergence of an extensive exchange system centered in the Caucasus that facilitated movement of raw materials required for tin-bronze metal working and finished metal goods. While millet has been identified in the southern Caucasus as early as c. 1600 cal BC, the introduction of this crop to mountainous zones only occurred as network interactions became externally focused. This only occurred after 1200 cal BC when imported goods such as Assyrian helmets were recovered from local burial contexts and locally made weapons and armor were exported as prestige goods.

Ironically, the brief dietary shift between age 5-7 is actually what is most consistent with a Sarmatian diet as δ13C drops really low and δ15N increases substantially, the figures seemingly fairly in line with those from the Sarmatian and Alan periods from Klin-Yar:

This could point towards the migration of this route briefly having gone through the steppes, and the 1-2 year period the diet of this individual shifted could fit a brief presence in such areas, migrating from the Caucasus towards Europe across this region. However there are many factors at play here and different migration routes could have provided the same values seen here. Factors such as the precipitation or even the δ15N/δ13C values present in plants consumed by animals humans consume will also affect their δ15N/δ13C values. Also, in terms of δ15N the iron age Koban samples showcase similar values.

A final point of information retrieved from isotopes gives insight into the environment the individual first grew up in:

Stable isotopes support long-distance mobility

The results of the carbon (C), nitrogen (N), oxygen (O), and strontium (Sr) isotope analyses are presented in Figure 3. The 87Sr/86Sr value from Offord Cluny 203645’s second mandibular molar (reflecting the first 5 to 6 years of his childhood22) was 0.709037 ± 0.000012 (2 SE), and strontium concentration from the same tooth was 104.2 parts per million (ppm), both of which are within the range expected for Britain23,24 (Figure 3A and Data S3A). However, this is a common 87Sr/86Sr ratio that can be produced by a wide range of geological terrains, and humans with similar values can be found in a variety of places. On the other hand, δ18O values were lower than what would be expected if he had spent the first years of his childhood in Britain (Figure 3A and Data S3A) and are instead indicative of regions with a colder or more continental climate, being consistent with levels of precipitation recorded today in regions at high altitude.25 Similar combinations of Sr and O isotope ratios have been observed in Roman-period populations in continental Europe.

Although it is not fully conclusive, the precipitation levels suggest that Offord Cluny 203645 may have initially grown up at higher latitudes, which fits the Caucasus mountains perfectly. However, the areas of the Caucasus region the Sarmatians lived in were not the actual Caucasus mountains, but the plains adjacent to them.

In short, the isotopic data is consistent with an individual that grew up on a diet with significant C4 plants such as millet, a lack of marine resources and living at altitude. This is a poor match for attested Sarmatian lifestyles, including the Sarmatians of the Caucasus region. 


The final point I want to bring up is history, because the authors use it to validate their claims - when once again, it does the opposite…

One possibility, given the radiocarbon date obtained (126–228 cal. CE; median 176 cal. CE), would be the historically attested deployment of Sarmatian cavalry in 175 CE, following Roman emperor Marcus Aurelius’s victory in the Marcomannic Wars, as described by the Roman historian Cassius Dio.4,5 In this scenario, the dietary shifts we see in Offord Cluny 203645 would be explicable if he was associated with groups of Sarmatians who moved into central Europe before or during the Marcomannic wars,3 although the plausibility of this interpretation depends on whether children were likely to have been part of movements of Sarmatians across Europe. Little is known about where the 5,500 Sarmatians were stationed in Britain.

Although little is known of the 5500 sarmatians, anyone familiar with either Sarmatian or Roman history should be able to tell you that the Sarmatians referred to here were the Iazyges, who at the time lived in Pannonia and fought alongside the Marcomanni. The Iazyges had moved from the western Pontic steppe towards Hungary in the first century AD. They did not arrive from the Caucasus, and had been attested living hundreds of kilometres away from the Caucasus prior to their migration to Hungary by Strabo.

But if history isn’t their expertise, maybe  I should not blame them right? Well, we have ancient DNA from the exact area where the Iazyges originally hailed from as well as ancient samples from the Sarmatian period in Hungary, which is named after the Iazyges which were one of the various populations inhabiting Hungary at the time, the majority of them not being Sarmatian. We also have Sarmatian-derived samples in Roman context which undoubtedly are linked to the Iazyges as well, and a recent sample (PCA0492) from the Prusz Gdanski site in Poland found in a Wielbark culture context also fits this bill based on genomic profile [7].

None of these samples look remotely similar to Offord Cluny 203645 in any sense, as the Iazyges-related genomic samples seem to have a hodgepodge of Pannonian, Graeco-Roman, Daco-Thracian, Germanic or Balto-Slavic ancestry, with a reasonable (15-50%) amount of input from proper steppe Sarmatians.

If the authors wanted to speculate on this individual having been part of these illustrious 5500 Sarmatians in Britain, perhaps they should have compared the Offord Cluny individual to samples from the very population these Sarmatians were derived from - the Iazyges.

So what about the Alans then? We know of Alan samples in the Caucasus with Caucasus-related ancestry, so perhaps this individual was an early Alan? Probably not.

The Alans contemporary to Offord Cluny would have very recently gone on their westwards spread from the Caspian towards the Western steppes, taking territories from other Sarmatian groups such as the Siraces and Aorsi in the second century AD. The Alans were primarily inhabitants of steppes and at this time were only beginning to settle the steppes adjacent to the Caucasus, after having caused a brief population hiatus in these areas by defeating the Siraces.

Since Offord Cluny 203645  predates the period of the Caucasian Alans, an Alanic population which formed after Alanic tribes consolidating into the steppes adjacent to the Caucasus due to Hunnic pressure, we cannot take the genetic affinity towards some later Alanic period samples as an indication of the individual being an Alan. Alans at the time of Offord Cluny 203645 in general should still genetically look like Sarmatians, with some individuals perhaps having minor Caucasian input.  Especially when considering we have Hunnic samples in Hungary two centuries later that look like regular Sarmatians - and you can guess which Sarmatian group was significantly present within the Hunnic confederation.

So to summarise;

  • We have an individual that for certain derived most of his ancestry from non-Sarmatian peoples. Based on the limited genetic models provided it is difficult to state how much Sarmatian ancestry this person had, if it had any at all as that is also a viable possibility based on their genetic analysis. Neither maternal or paternal haplogroup of Offord Cluny provides a direct link to Sarmatians.

  • The inferred geographical area and diet in childhood are inconsistent with a cultural Sarmatian origin. A childhood with a millet-rich diet, at altitude and without fish contradicts the typical Sarmatian lifestyle of nomadic pastoralism in plain regions, often in the vicinity of water bodies.

  • Since we have no burial context for this individual we have no clues that point toward this individual even identifying as a Sarmatian or even being a member of their culture so to say.

  • The evidence for this individual having a military origin is not present, the suggestions are guesswork based on a somewhat contemporary date to a Sarmatian movement into Britain (not from the Caucasus) and the individual being around 18-25 years old at death.

  • The suggested historical explanation invoking the Iazyges of the Marcomannic war fails for obvious reasons since these Sarmatians arrived there from Central Europe, inconsistent with the genomic origins of the individual

Then it begs the question, why this heavy focus on the Sarmatian aspect?

I genuinely think it has to do with that theory of linking Arthurian tales to Sarmatians due to a surprising amount of similarities and parallels you can find between the Arthurian legend and the Nart Sagas.Some historians have proposed that this Arthurian-Nart Saga connection came about with the movement of these Sarmatians into Britain. Most people proposing this seem to consider that given how most of the King Arthur story developed in France, during the middle ages, if anything Alans in Gaul should be considered as a point of influence. 

In any case I have a feeling this theory, fairly popular some decades ago and even featured in the movie King Arthur from 2004 may have influenced the heavy emphasis  on Sarmatians - when everything points towards the various populations of the Caucasus mountains instead. 

I also wonder if the authors when finalising had a slight change of tune in regards to this individual. If you recall the abstract specifically said “Sarmatian individual” whereas this article’s title has Sarmatian-related ancestry instead. Another point is how the abstract mentioned how a  C4-rich diet was consistent with a Pontic-Caspian steppe location, a claim I have not been able to find in the final article. For good reason as it would have been incorrect.

All-in-all, I find this a rather bizarre article. The authors took what would be an interesting discovery of an individual of Caucasian origins being in Roman Britain, right at the time period when there was an interesting history between Rome and the kingdoms of Pontus, Iberia and Albania, and wrote an article how the Offord cluny individual was a Sarmatian that served in Rome’s military and was part of the “5500 Sarmatians” dubiously linked to the Arthurian legends in popular pop-history theories. The data provided by the authors seems to only serve the story they had imagined, judging by the way how this evidence rationally would go against the individual being  of Sarmatian origin yet they argue a Sarmatian origin for it.


  1. Reitsema et al. (2022). The diverse genetic origins of a Classical period Greek army. Proceedings of the National Academy of Sciences.

  2. Slepchenko, Sergey, Evgeny Pererva, Ivanov Sergey, and V.M. Klepikov (2019). Archaeoparasitological analysis of soil samples from Sarmatian Burial Ground Kovalevka I, 2nd–1st centuries BCE, Russia. Journal of Archaeological Science: Reports. 26. 101874.

  3. Belinskij, A.B. & Härke, Heinrich (2018). Ritual, society, and population at Klin-Yar (North Caucasus): Excavations 1994-1996 in the Iron Age to early medieval cemetery (Archäologie in Eurasien, Bd. 36) [XVIII + 416 pages].

  4. Ventresca Miller, A.R., Makarewicz, C.A. Intensification in pastoralist cereal use coincides with the expansion of trans-regional networks in the Eurasian Steppe. Sci Rep 9, 8363 (2019).

  5. Babenko A.N., Dobrovolskaya M.V., Vasilyeva E.E., Korobov D.S. Reconstructed Paleodiets and Subsistence Strategies of the Central Ciscaucasian Population (1000 BC to 1000 AD), Based on Collagen Isotope Analysis of Bone Samples from the Kichmalka II Burial Ground. Archaeology, Ethnology & Anthropology of Eurasia. 2021;49(4):80-90.

  6. Ventresca Miller, A.R., Makarewicz, C.A. Intensification in pastoralist cereal use coincides with the expansion of trans-regional networks in the Eurasian Steppe. Sci Rep 9, 8363 (2019).

  7. Stolarek, I., Zenczak, M., Handschuh, L. et al. Genetic history of East-Central Europe in the first millennium CE. Genome Biol 24, 173 (2023).


  1. Thank you for this! And in general your content. As someone with Indo-Iranian, potential Sarmatian, paternal ancestry that forms a apparent cluster from no later than Medieval Northern Britain, I am particularly interested in this topic. I'm not a fan of the so-called Sarmatian Theory of Arthurian origins for exactly some of the issues you outlined here - there tends to be a fair amount of unnecessary, unconvincing circularity to it. Hoping you'll continue to be so active here and perhaps at Genarchivist too in the new year!


  2. I know that certain surnames in GB are an indication of a Sarmatian origin and also carry Z93 - Z2123, Michał posted it on Anthrogenica a long time ago.

    but that's true Offord Cluny 203645 would be misinterpreted and thanks for the tip

  3. We also know that Alans interacted with Vandals and migrated through Gaul, Hispania and into North Africa, this sample of Offord Cluny may have been a Caucasian mercenary

  4. Maybe he came from the Bosporan Kingdom? It was a Roman client kingdom at the time.