search this blog

Tuesday, January 19, 2016

Ancient genomes from Iron Age, Roman and Anglo-Saxon Britain

Open access at Nature Communications:

The purported migrations that have formed the peoples of Britain have been the focus of generations of scholarly controversy. However, this has not benefited from direct analyses of ancient genomes. Here we report nine ancient genomes (~1 ×) of individuals from northern Britain: seven from a Roman era York cemetery, bookended by earlier Iron-Age and later Anglo-Saxon burials. Six of the Roman genomes show affinity with modern British Celtic populations, particularly Welsh, but significantly diverge from populations from Yorkshire and other eastern English samples. They also show similarity with the earlier Iron-Age genome, suggesting population continuity, but differ from the later Anglo-Saxon genome. This pattern concords with profound impact of migrations in the Anglo-Saxon period. Strikingly, one Roman skeleton shows a clear signal of exogenous origin, with affinities pointing towards the Middle East, confirming the cosmopolitan character of the Empire, even at its northernmost fringes.

Martiniano, R. et al. Genomic signals of migration and continuity in Britain before the Anglo-Saxons. Nat. Commun. 7:10326 doi: 10.1038/ncomms10326 (2016).

And another one:

British population history has been shaped by a series of immigrations, including the early Anglo-Saxon migrations after 400 CE. It remains an open question how these events affected the genetic composition of the current British population. Here, we present whole-genome sequences from 10 individuals excavated close to Cambridge in the East of England, ranging from the late Iron Age to the middle Anglo-Saxon period. By analysing shared rare variants with hundreds of modern samples from Britain and Europe, we estimate that on average the contemporary East English population derives 38% of its ancestry from Anglo-Saxon migrations. We gain further insight with a new method, rarecoal, which infers population history and identifies fine-scale genetic ancestry from rare variants. Using rarecoal we find that the Anglo-Saxon samples are closely related to modern Dutch and Danish populations, while the Iron Age samples share ancestors with multiple Northern European populations including Britain.

Schiffels, S. et al. Iron Age and Anglo-Saxon genomes from East England reveal British migration history. Nat. Commun. 7:10408 doi: 10.1038/ncomms10408 (2016).

See also...

The enigmatic headless Romans from York

Hinxton ancient genomes roundup


Hyarotle said...

What's up with the J2 roman sample being closest to Arabia?

Krefter said...

>7/9 samples are Gladiators from York and date to 200-400 AD who died of battle-injuries. One is a female from Late Iron age and lived near what became York. The last is an Early Medieval(600-900 AD) Anglo Saxon Male from York.

>6/7 Gladiators were native Britons. Isotope analysis confirms 5/6 were born and raised in or near York. FinStructure and IBS show that the Gladiators and Iron age woman are closest to modern Welsh. One was born and raised in SouthWest Asia and genetically closest to modern Levant or Saudi Arabia(hard to tell which).

Romulus said...

2 examples of R1b U106(M405) from Roman Era Britain, so much for that being an Anglo-Saxon marker.

Davidski said...

They were probably Germanics who arrived in Britain as captives before the Germanic invasions, and may even have been Saxons.

Romulus said...

If they were Germanics then why don't they show any greater affinity to the Anglo Saxon sample than the others? Obviously they weren't.

Davidski said...

Well they definitely had ancestry from around the Baltic, because they show high affinity to Lithuanians and Poles. The others not so much.

So they were either of Germanic or part Germanic origin, with ancestry from somewhere near Poland...or Balto-Slavs with R1b-U106(M405+).

Take your pick.

Romulus said...

I wasn't looking at that median identity by descent map I was looking at the PCA. On the PCA they're really close to the others and not the Saxon. Whoever they were, they were there for a long time and not Anglo Saxons. Weird that they show Baltic ancestry, different from the Celts but not Scandinavian.

Davidski said...

Not sure which PCA you're looking at, but 2-3 of the Roman samples appear more eastern than the Anglo-Saxon.

They look like they have some continental European ancestry from well east of the North Sea. These are probably the U106 individuals. So it's difficult to argue for now that they represent a native British population rich in U106, rather than the descendants of recent migrants from somewhere near the Baltic.

Romulus said...

This one

I think the best explanation is that the first Beakers to arrive in the British isles made a stop over in Iberia, and the later ones (Rhine/Dutch on this map came directly from Central Europe.

Davidski said...

That PCA doesn't have any Northeast Europeans.

ryukendo kendow said...

@ FrankN @ Rob

Guys, surely the paradigm must be two-way. The results that we find from aDNA and genetic studies must be given a certain amount of epistemological weight, sufficient to overturn or to adjust prior archaeological interpretations, just as archaeological findings are needed to contextualise the results. I mean, now we actually know some things about the geographical origin, the matrilineage or the patrilineage, the ancestry proportions, of some of these people, which is extremely valuable evidence on par with some of the most incredible archaeological discoveries; surely this must cause some kind of adjustment of the models and assumptions? It will be weird if our interpretations and beliefs were exactly the same before and after this evidence has emerged.

I'm not well versed in archaeology, but my general impression from the economics, politics and anthropology people is that there is a considerable amount of 'democratisation and pacification' of the past going on in that field; social structures are often postulated that go against what we know about human nature and social relations in small-scale societies, because the assumptions are often derived from marxist theories of 'individual agency' or 'symbolic networks' instead of empirical anthropological or economic observations of similar societies we have now or in the recent past, resulting in the social interpretations sounding impossibly unrealistic to political economists or anthropologists. Well, this is just a stereotype, but still I'm just suggesting to you guys that, since you are in possesion of detailed evidence from pop gen, both modern and ancient, that those in academia do not know, you already have an extremely important social window into the past that they did not have: who reproduced and who didn't. And you can use that to draw conclusions that are dissimilar from them, no?

Rob, just raising some questions here, why would migrating farmers with nuclear family households contain closely related Y-DNA from a single lineage? The Germanic peoples who came later had likewise family households with a semi-mobile lifestyle, but they were nowhere near as homogeneous as the Y-DNA in the Corded Ware. If the Corded Ware phenomenon did involve a massive homogenization of Y-DNA, but without the levels of social complexity required for elite dominance, this surely demands an explanation of some kind?

ryukendo kendow said...


Let me clarify what I mean with elite dominance. The economic historian Gregory Clark, using surname and patrilineage records from a gigantic array of studies, has shown that the correlation between the socioeconomic status of a man and his patrilineal descendants across a wide variety of early medieval to early modern societies, incl. Japan, China, Iceland, England, Sweden, France, India etc. is at 0.75-0.9. At this rate, it takes at least seven generations, generally more, which is on the scale of centuries, for the descendants of a prestigious ancestor to become statistically indistinguishable from the general population. Putting this together with research from around the world that shows, e.g. wealthy males in early medieval portugal having twice as many descendants as poor males, and marginal males in low-status jobs like woodsman or shepherd having no descendants at all, he gives us quite a good explanation as to why, e.g. norman surnames are in fact quite over-represented in the English population in all social classes today, so many centuries later. A side effect of this is that we also get a very good sociological model to explain many pop gen results.

Archaeological estimates for the size of the anglo-saxon migrations are tightly bounded between 30,000 people to 5000 people; estimates for the population of Britain at the time range from 0.75 million to 4 million persons. I.e, the migrating population ranges from 4% to 0.125% of the total, for the A-S migrations. However, the ancestry proportion from the AS in the British today, as posted in the above paper, are at ~40% to ~20%, and from Y-DNA, we have some evidence that this gene flow is male-biased. If there was a consistent reproductive difference between the AS and the locals over the course of centuries in harmony with social status differences, then this suffices. In fact, we know that the early legal codes treated ASs and locals differently, the distinction between them not collapsing until centuries later (enforced exogamy again!). So some people put out a paper on this, utilizing research from pop gen, econ, hist, arch, to create a coherent picture of the past: 'Evidence for an apartheid-like social structure in Anglo Saxon England', IIRC.

This kind of reproductive differential can only be maintained peacefully in a complex society where resources are not distributed evenly, but sufficient concentrated violence capacity exists to permanently enforce resource extraction and/or inequality. I'm not sure how this can occur with the current understanding of Corded Ware, Bell Beaker, with any of the social models we have at this time. In a multicultural society without status differences, its not clear to me how a small migrating population can result in large scale male-biased gene flow. My conclusion would be that, either society was not complex so the 'disproportionate reproduction' was not peaceful, or that the migration must have involved a much larger number of individuals than is usually assumed, but I'm open to other scenarios. The point being that we still need to look for better models for how the Bell Beaker and Corded Ware societies worked, that integrate all the evidence we have at hand. In fact I wouldn't be surprised if you guys, putting all this new discoveries with what you know about the primary evidence, arrived at a picture that was much closer to the truth years before the rest of the academy swings round to something similar, but that requires us to be less attached to the interpretations that already exist.

Romulus said...

Despite being blonde haired and blue eyed the Anglo Saxon is only heterozygous for the SLC24A5! A real WHG.

2.10.1 Pigmentation
All samples are homozygous for the selected allele of SLC24A5 SNP (rs1426654), except for the
Anglo­Saxon, which is the only heterozygous. For SLC45A2, all samples are homozygous for the
selected allele, except for 6DRIF­21 and 6DRIF­22, which are heterozygous. Regarding TYRP1, the
following samples are homozygous selected allele: the Iron Age sample and 3DRIF­26. All other samples
are heterozygous, except for 6DRIF­18, which is homozygous for the reference allele (Supplementary
Table 16).

ryukendo kendow said...

@ Davidski

David, could you try using Rarecoal? The url is included in the paper. This, more than anything else, can tell us about whether or not some ancient sample directly contributed to a modern group, as opposed to looking alike because of similarity in overall autosomal profile only. And its resolution is just incredible. Really impressed with the algorithm.

Used here 'rare allele' is specified as any allele with up to count four, so I suppose samples with say more than 10 individuals could theoretically be used, but in this tree each population was represented with ~30 indviduals. So I do wonder if the dataset is large enough for us to discover all these rare alleles at the required level of resolution. Maybe we can set 'rare allele' to count 2 only, which might help.

Davidski said...

I've thought about it, and my view is that Rarecoal won't help with the stuff I'm doing unless it's modified in the future to include mixture edges.

ryukendo kendow said...

Davidski, for the Indo European expansion, yes. On the other hand, could you try it for the roman genomes? The question of whether or not eastern influence can be found in some individuals will be easily settled, using the same publicly available datasets as the authors.

Rob said...


Thanks for the comments , and I agree - Maybe it's good to be a "jack of all trades, master of none" ;)

But I'm not quite sure you understood what I wrote ? So ill re-post - let's stay with CWC for now

2) "A question that naturally follows is, why would a culture of small household units have homogeneous Y-DNA patterns, indicating recent common patrilineage?"

Isn't that consistent with small family units led by patriarchs, expanding broadly but thinly from an original core locale - say for arguements sake, CWC expanded from NW Pontic region ?

"However, how can this be reconciled with elite dominance in a society, if the society at the time is comprised of mobile nuclear households with no social complexity?"

Of course there was some social complexity. It just operated differently. But the crux of CWC society is: migratory, family bands, cattle herding, lowland ecological niche. Is there something in this which might be inconsistent with what you've read ?


The 'elite' in this society were men, codified as the 'male hero' with a "CWC funerary set' in the only markers we have for CWC - the funerary mound (occasional fleeting dwelling structures). Not all men - but specially chosen men. *Amongst* this group - they were equal; i.e. no one CWC patriarch was more important than another. Given that we now know that these men were virtually all R1a-M417, it must have had a biological basis - patrilines, filial heirs, etc.

Contemporary to CWC in the Polish lowlands were the GAC communities, as well as TRB cultures - both of which lingered until 2000 and 2500 BC, resp. We have no real aDNA from them, so are in the dark for, say ~ 66% of the population in our case study area. In the above quoted paper, the GAC group was also studied. The enduring collective identity of internal coherence and opposition to CWC cultural templates speaks of different identities, even ethnicities. Their economy was different - primarily agriculture (vs mobile pastoralism), community was composed of several families, rather than single households, more permanent (but not 'fixed') settlement patterns, etc, etc

It is tempting to link GAC more to descendants of MN groups, given they appear 'native' to Poland, arising earlier (3600 vs 2900 BC). So will the upcoming aDNA study reveal them to be mostly haplogroup G2 ?

Are they the losers of a war vs CWC ? Maybe. But possibly not. Were they even necessarily in perpetual conflict ? - or did they simply coexist in different niches? The GAC lasted as long as CWC, and both groups became extinct, or 'transformed' c. 24/2200 BC, due to other, new, external impulses- BB and groups from the Carpathian basin (Unetice). Given that M417* doesn't exist in Poland today, CWC and GAC were likely both 'losers', that is to say, were yet another statistic in the series of population turnovers seen in northern Europe. "

I would be interested in specific critiques on this. It's just my thinking. I do not come from a particular "school of thought" and i have no career attachments to any theory. So, I cannot speak for Frank, but yes I do see clear signs of migration for CWC - maintaining close social links amongst a dispersed landscape, and bein strongly patriarchical, it would seem.

Also, I'm not sure I'd take those figures for AS England at face value . I find it difficult to believe that 1 million Britons simply turned over like sheep for 5000 Germani. Something is amiss in the figures

ryukendo kendow said...

@ Rob

Stop being snide. God, I'm inviting more opinions from you here. Genetics has overturned the diffusionist interpretation for the archaeological neolithic already, when nothing physically new has come out of the ground.

There are many other examples of such extreme disparities. The figures for Brazil are 600, 000 Europeans and 4 million Africans till 1875, after which indigenous increase accounts for most changes; today White brazilians are almost completely European and black Brazilians are ~72% European. Madagascar populations are half Asian despite the founding Asian population having ~30 females only.

Of course I understood what you wrote, it just doesn't answer the question. why should nuclear family households have all the same Y-DNA lineage, suggesting recent shared paternal ancestry? Neither the neolithic before nor the Germanics after had such homogeneous lines, despite the Germanics having a similar organisation. Unless of course you suggest that the corded ware once used to be a single kinship group at its genesis? Are there any signs of clan based patriarchal organisation, which may help to propagate male homogeneity? I'm asking you, if you go back to the primary evidence, what does the new genetics do when it is taken into account with the rest? Since the academy has not taken the new evidence up into their interpretative framework, surely some others can? Think, for goodness sakes.

Rob said...


I'm not being snide- I'm not sure where you're getting that from.
And, again, we need to be careful with analogies. One can't make a case from modern colonialist experiences, where 16th century Europeans advanced with near- modern technology upon natives from a wholly different ‘civilizational outlook’.
The case for LN Europe is very different. Corded Ware people did not have some massive advantage over, say, the Baden culture – which was equally or more so militant, and economic strategy differences existed along a continuum.

You appear to be asking general questions which I cannot answer – as to why archaeologists still stick to diffusionist interpretations. For a start, not all do. But some archaeologists simply do not know about genetics - it requires a lot of extra research. Some are wary of it. But I already indicated- some are up to play with it.

You say I'm not answering the question to "why should nuclear family households have all the same Y-DNA lineage, suggesting recent shared paternal ancestry”.
I wrote: “The 'elite' in this society were men, codified as the 'male hero' with a "CWC funerary set' in the only markers we have for CWC - the funerary mound (occasional fleeting dwelling structures). Not all men - but specially chosen men. *Amongst* this group - they were equal; i.e. no one CWC patriarch was more important than another. Given that we now know that these men were virtually all R1a-M417, it must have had a biological basis - patrilines, filial heirs, etc.” and “I do see clear signs of migration for CWC - maintaining close social links amongst a dispersed landscape, and being strongly patriarchical, it would seem”

But, Corded Ware is not *all LB/EBA Europe*. It is one network, amongst a few others. So CWC seems to have been exclusivist from a male perspective. Sure, this might have involved physical expungement, but that depends on whether other LB/EBA men actually *wanted to be in this network* ?

What is it in that which appears incorrect, or coy, and moreso please provide your answer.

Colin Welling said...


You are misreading a lack of diversity in the y chromosome to mean social dominance expressed by mating exclusivity.

The problem in your reasoning is that the ychomosome is not a neutral marker and so it doesn't reflect inherited social inequalities (or social inequalities at all). The ychromosome is itself fast evolving and accumulates many deleterious mutations that get weeded out by natural selection. These mutations may have impacts on sperm count or the tendency to produce males, along with other factors that have nothing to do with social dominance. Most importantly, the natural selection on the y chromosome has not resulted in a relatively smaller mating pool of men relative to women, which is what you are postulating.

The only thing that adna, autosomal dna in particular, has told us is that MEN AND WOMEN tended to stick to their own kind. Elite men were not taking wives from the 'other' side. When migration did take place, and changed the local genetics, there was still barely any mixing that occurred. Thats why neolithic and LNBA types have a consistent meaning across many miles and many generations.You should appreciate the social relevance in todays multicultural societies. So while the marxist academics against migration/replacement are wrong, so are the advocates of elite men going around and taking local wives.

What you say about the rich folk having more kids is interesting but that doesn't imply the men had many more wives. It is certainly the case that the men and women just had more kids and those kids were more likely to survive. This explanation makes sense given the apparent near equal mating pool of men and women.

I hope people come to realize that ydna lineages that dominate don't need to have anything to do with male breeding rights.

[natural selection on y chromosome]

[female to male breeding ratio]

Colin Welling said...

btw, good observation david. Its amazing that the ydna and autosomal dna have corresponded so well in these samples.

I agree with your first suggestion that the two eastern romans where part germanic rather than any non germanic group from the baltic. Didn't the study not use danes and swedes as a reference group?

ryukendo kendow said...

@ Collin Welling,cs_srgb,w_620/MTI4ODMwNDQ5ODU3MzY5MzYy.png

Collin, where did I talk about male breeding rights? I am not referring to the reduced Neff for the Y chromosome across its entire history, or the breeding ratio for the autosome vs X chromosome. Approximately 8000-5000 ybp, an extremely powerful bottleneck in the diversity of Y chromosomes occured, which did not occur for the autosome and mtDNA. This bottleneck only manifested itself in populations in the Middle East, Europe, East and SE Asia, and S Asia. Neither selection, nor any of the other suggestions raised by the papers you posted, such as longer generation length for males, can explain this phenomenon.

Neither is it as simple as males taking multiple wives--it never is. No polygamous society actually had a 1:17 reproductive differential, and I'm not saying that. I suggest you read the paper on the models of Anglo-Saxon social structure. The selection coefficient needed to get the transformation in the Y-DNA landscape is relatively small, but it must be consistent over multiple generations. The argument for the early medieval societies relied on differential survival of elite and non-elite children, and remarriage after the spouse dies in childbirth, not polygamy. However, this requires social complexity, like I said.

@ Rob
Why are these people 'masters of none'?

I am not referring to colonialist experiences. The thing is, we have very good accounts of the kinship networks, power structures, economic exchanges, ideas of property and inheritance, that take place within small-scale societies within the historical period. We see that many small scale societies are very, if not extremely, conservative, with all kinds of idiosyncratic prejudices and taboos, and highly specific and tenacious ritual systems and methods of assigning social status. They do not just change because of the presence of an outsider. We see that, without mass anonymity, without third-party enforcement, only the simplest economic transactions can occur; no service transactions or economic guilds. We understand the role that violence and violence capacity plays in redistribution and in promoting cooperation in complex chiefdoms, vs. simple kin-based societies, for which just kinship norms are sufficient. We observe the generalized lack of trust in persons from outgroups, and the universal presence and expectation of arms and self-defense in males. It just seems that many of these ideas go out of the window the moment we talk about ancients.

For example, do you really think that migrating craftsmen can explain the transformation in this period? That a group of persons can just travel from group to group and violate ingroup-outgroup boundaries just like that? Or, in the Bell Beakers, that a transformation of the ritual system can occur through diffusion, or even more implausibly, by imitation of people from outside your community, without social upheaval that interrupted established traditions?

Either way, I suspect that there is a great space for the understanding to catch up, and that you and FrankN are well placed to bring up some new directions, if you would shed some of the old conventions and look at the primary evidence and interpret these yourselves. I'm just trying to get you guys to see new possibilities, though there are methodological issues since our backgrounds are so different, and say this with full respect for your knowledge and attention to detail, though this might not have been communicated properly.

Alberto said...

@Rob, RK

I also doubt the figure of 5,000-30,000 Germanic people moving to Great Britain. The figures we have for Iberia, at least by the old books (not sure if more recent studies have changed this, maybe Maju would know), were some 200-250 thousand Visigoths entering Iberia, that had a native population of around 9 million people. And that's only Visigoths (arguably the biggest and more long lasting migration), but before them Iberia was invaded by Suebi, Alans and Vandals too. That would make at least 300.000 people (many of them carrying Y-DNA lineages not previously present in Iberia, like I1 and R1a).

These Germanic tribes ruled Iberia for some 3 centuries before the Muslim conquest. Yet today their genetic impact, be it autosomally or by Y-DNA, is very low (~3%? Quite accordingly to their ratio of population).

Rob said...


I meant I am a master of none. I.e it was a joke with some truth

As for the rest, I'll leave it till tomorrow

ryukendo kendow said...

@ Alberto

Alberto, Harke is the one who holds a maximalist interpretation of the migration, which he estimated at 100, 000. His is not the most recent estimate. However, this is still 11% to 2.5% of the native population, a far cry from ~40% - ~20% that we get now.

ryukendo kendow said...

@ Rob

I tend react very quickly to what I read. My sincere apologies.

Rob said...

It's ok, RK

"Jack of all trades , master of none"
Has advantages as is not fixed to ones own niches and specialties, but has broad understandings of others etc etc

André de Vasconcelos said...

@ Alberto

I've been told similar figures to yours, with some slight differences such as a population of 6M in Hispania by the V century, but I never actually saw a source for them. Do you happen to have any? I'd really like to check them myself.

Still, the migration wasn't homogenous throughout Iberia, some areas were apparently heavily settled (NW Portugal, W Galicia) especially by the Suebi - and later "Visigoths" during the rule of Chindasvinto 642-655 - other areas not so much (S Portugal). For instance there is a very clear excess of Germanic toponyms in these areas, although most seem to be Gothic rather than Suebic.

This would theoretically imply some very clear regional differences, but I'm not sure/aware they exist. Unless, of course, later population movements (northwards during the 700s, southwards during the Reconquista) made this more homogenous on a North-South axis.


Alberto said...


Unfortunately I've never seen anything very updated about this matter. I checked the Spanish version of Wikipedia in case it had any reference to modern material, but it doesn't. It still goes by the traditional figures (my translation):

"In any case, the Goths must have been a minority that is supposed to have started to integrate in the Hispanic-Roman society. Their number has not been precisely established by any historian, but the most reliable estimates mention between 150,000 and 200,000 Visigoths settled in the peninsula, over a population that didn't reach 9 million, according to Saint Isidore of Seville. Other sources mention 80,000-100,000 Visigoths over a population of 6 million of Hispanic-Romans.

Recently [note: 1989!] an archaeological study about the Visigoth settlements estimated a figure for the Visigoth population between 130,000 and 150,000, what would represent between 3% and 4% of the total population of Hispania"

So we're basically stuck with the old figures, but I think they should be in the ballpark.

Cossue said...

From Javier Arce "Bárbaros y Romanos en Hispania" (2005) My translation:

"How many [Visigoths] came? It's impossible to know. We have debated the number of the Goths in Barcino in the times of Walia and Athaulf. There we spoke of some 12,000 or 15,000 considering the quantities of wheat assigned to them by the Romans in the agreement of 418. Those were the ones who settled in Aquitania. Maybe they number grew until the time they occupied Hispania [c. 500] But we can not stretch the number too much. 20,000, 40,000? Let's not forget that the Suebi were 25,000, the Vandals 80,000, the Alans maybe some 30,000".

For the population of Hispania, the only numbers I know are those given by Pliny to Gallaecia and Asturia at the beginning of out era, which where some 700,000 free people (in less than 70,000 sqr. km). Probably by 400 AD these numbers have dropped a lot, and I doubt that they can be extrapolated to the whole peninsula: Galicia and northern Portugal have been traditionally well populated areas (Galicia had 2M inhabitants by 1,900 in 30,000 sqr km with an almost exclusive farming economy).

In Galicia and northern Portugal is where we found more toponyms referring to Germanic settlements, as Sá/Saa < Sala, Lobio/Lóbio < laubj-, Boiro/Bouro < bûrj-... Also more toponyms derived of the genitive of Germanic names: Guimarães < Wimaranis, Gondomar < Gundemarii, Mondariz < Munderici, Forcarei < Fulcaredi, Sandiás < Sindilanis, Arulfe < Arulfi... the late Jospeh Piel identified these names as Gothic names; the problem is that the names used by the Suebi were also Gothic names: Ermeric, Reckila, Reckiar, Frumar, Andeca, Ilderic, Maldra, Frantan, Wittemir... Any way, by 900 AD most people in Galicia and N Portugal used a Germanic name, so these last toponyms, which were formed in between 500 and 900 AD are not really useful for knowing where the Germanic peoples settled, though they are a hint telling that the NW of Iberia was a little more Germanized than other areas.

FrankN said...

@Dave, Romulus: In relation to the Lithuanian-like ancestry showing up in some of the Roman samples, let's recall Tacitus' description of the Aestiones speaking a language similar to the Britains. Tacitus was of Gaulish ancestry, and has most likely travelled to Britain (his father-in-law was Roman governor of Britannia), so the fact that he uses the term "British" instead of "Gaulish" should indicate quite a difference between both languages.
Tacitus places the Aestiones at the central Baltic coast, in an amber-rich region with a sea almost void of agitation. This is commonly equated with the Curonian Lagoon. Alfred the Great mentions the Aesti dwelling near Truso, the predecessor of today's Elblag (60 km SE Gdansk). Hence, the Aestiones shouldn't be mistaken for today's Estonians, but rather for (Proto-)Prussians.

Note that Roman authors (e.g. Strabo) report another group also named Aestiones, living in the Kempten area as part of the Vindelici.

FrankN said...

CORRECTION: The Baltic Aestiones shouldn't be associated with the Curonian Lagoon, but with the Vistula lagoon (Aīstinmari in Lithuanian).

André de Vasconcelos said...

@ Cossue, Alberto

I've seen a website that has data of family names in Spain, shown to me by a bloke from Valladolid.

We noticed that most of the Germanic names (Rodríguez, González, Fernandez, Gomez, etc..) seemed to be more common in what used to be the old kingdom of León, thus in NW Spain, than in the rest of the country. Here in Portugal all of those names are also very common, particularly Rodrigues. Isn't Rodríguez the most common family name in Galicia?

Cossue said...

Yep. Rodríguez, then Fernández, then probably González and Álvarez. They are the "castillianized" version of the medieval form which were mostly identical to the Portuguese ones (Rodrigues, Gonçalves, Alvares: But, anyway, these are the most common family names because they were the most common personal names in the 13th, 14th, and 15th centuries, when patronymics became fixed as family names. But in the 9th centuries practically each person in a given document have an exclusive name (example: a Galician original charter from 883 AD, with some 100 personal names, +80% Germanic):

"(Christus) In Dei nomine. Colmellus diuisionis qui factus est inter filios Placenti et Romarici, licet inter pacificas mentes definito, sola constat uerborum tamen pro memoria temporum testimonium adiuandum est litterarum. Igitur, dum inter nos intemptio uertitur ad diuidendum mancipia uel parentum nostrorum Guntine et Rosule, de neptos senatoris Siserici et Esmorice et de suos iermanos. Ideoque euenit in portione de filios Gunterodis, id est: Argiuitus, Gentius, Trastiuegia, Recedruda, Gaudiosus, Tequilo, Iulia, […], Stoildi, Sitiuidis, Gluscudilum, Eramildi, Ruderigus, Sonobrida, Sabaricus, Argeleuba, Ostosia, Guntedrudia, Uitiza, Eyleuba, Guntildi, Iulia, Ragesindus, Ranildi cum sua filia Ermegundia, Seniorina, Uisteuerga, Sisulfus, Branderigus, Astruildi, Br[o]lis, Ranemirus, Goda, duas filias Gaudiosi: Maria et Felicia, Trudina, Quindiuerga, filio Frumildi, Astrulfo, Anulfo, Ranulfo, Teodeuerga, Sindileuba, Ermedrudia, filia Sonifrede maiore, alia filia Sonifrede minore, Aciulfus, Recedrudia, Uandi, Trudildi, Ausendus, filie Guduigie. fili Ostosie, Sabeg[oto...] inde duos filios Tequiloni, Uimaredus, filia Argeleube mino[re, filios] de Adaf qui est post Genitibo, Tedildi, Leodo cum duos filios, id est: Gundilo, Fargalosus, Sisuita cum suo filio.
Euentum in portione de filios Rosule, id est: Elias, Spintilo, Placida, Gundinum, Ebrildi, Guntuigia, Metoi, Tita, Bonosa, Aroildi, Fagildus, Sabaredus, Domnelius, Astrogoto, Leobegoto, Uanagildi, Seb[astianus], Trudilo, Genlo, Sisiuertus, Nunilo, Ansuildi, Dauid, Sparuildi, [...] cum suo filio Ranemiro, Teodilo, Guisenda, Ariulfus, Leouegildus. [...] Sabegoto, Stefanus, Saroi, filios de Spintilo duo: Guntemirus et Osoredus, Leodefredus, Gudigeba, filio Sonifrede minore, Ausendus. Gudileuba, Brunildi, Ebragundia, Nunnina, Astrulfus, Eldoara, Adosinda, filio La[...]que, Odorio, Leoueredus, Sisuertus, Manosinda, filia Guduigie, [...] Edonia et Elias, duos filios Tequiloni, Sisuigia, filia Leouegoto, [...] filio Adaf, Papinum, Genetibus, Uitiza, [...s]tromirus, [...,] Guendulfus.
[...] quod accepit firmiter obtineat.
Factus colmellus VIIII kalendas octobris era DCCCCXXIª.
Silocri colmellum diuisionis a me factum.
Munius anc colmellus divisionis a me factum.
Uimara in anc colmellus diuisionis quo ac persona de meos germanos uel de meo sobrino Spasando diuisi manu mea.
Adefonsus anc colmellum diuisionis a me factum [signo]
Odoarius in anc colmellum diuisionis a me factum.
Gundulfus anc colmellum diuisionis a me factum [signo]
Munnia anc colmellum diuisionis a me factum [signo]
Gunterode anc colmellum diuisionis a me factum [signo]
Ermegildus anc colmellum diuisionis a me factum [signo]
[Chistus marginal] Pelagius ad persona de meos neptos, filios Petri, in anc colmellum diuisionis manu mea [signo]"

FrankN said...

@rk and others: Numbers of migrating Anglo-Saxons

We have a pretty good idea of the Anglo-Saxons that emigrated from the Elbe-Weser region towards England - Feddersen-Wierde is among the archeologically best documented sites along the whole North Sea coast. Being densely populated, and a major iron-making centre during the Roman IA, the settlement was abandoned by the 5th cent., most likely because of sealevel rise. The archeological record points at 8 sub-settlements, each with some 45 houses, corresponding to in total some 2,000 inhabitants on an area of 23,5 km².

Excavation on the Schleswig-Holstein coast has so far yielded seven more settlement areas of the Feddersen-Wierde type (see linked map). This is most likely incomplete, as a substantial part of that coast, the area between Amrum and Eiderstedt, is today submerged and poorly archeologically documented. Here, settlement of coastal marshes was also given up in the 5th century AD and only recommenced with the 8/9th century Frisian expansion. Hence, from the German coast north of the Weser mouth alone, some 20,000 or more people should have migrated to England.


FrankN said...

.. con't

Recent excavations in Eastern Holstein linked to the A20 motorway construction have uncovered a large urnfield (1cBC - 4cAD) at the outskirts of Bad Segeberg, which indicates a town population around 1,000 by that time. A smaller urnfield has been identified at Schwissel, 7km to the south, while at Wittenborn, 6 km to the west, a hamlet of four large farmhouses was recovered. These finds demonstrate a late IA population of at least 1,200 on a 10x10 km area, i.e. a population density of 120 p/km² or more.
We are clearly talking a peri-urban centre here, so average densities should have been lower. Nevertheless, a similar-sized urnfield is known from Groß-Timmendorf, some 30 km east, close to the Baltic Sea shore, and the Neumünster area, 30km NW, has been demonstrated as a major center of bog iron smelting and processing in quasi-industrial extent. Pollen diagrams show a similar degree of landscape opening as today. Moreover, Eastern Holstein has excellent soils - Schleswig Holstein today has the highest per ha cereal yield of all German states, 20% above EU and 10% above French average. Note in this context also Roman IA evidence for plough (not ard) usage, mineralisation with crushed shell, rye planting, and use of the three-field system, which all point to Germanic agriculture in the Central European plain being more advanced, and better climate-adapted than contemporary Roman wheat monoculture north of the Alps.
Hence, it is fair to assume a late IA population density of 60 p/km², which is in line with worldwide figures established for pre-industrial agrarian societies. Eastern Holstein plus Anglia (the peninsula between the Schlei and the Flensburg Firth) comprise a bit over 4,000 km² and should thus have housed some 250,000 people.

The archeological record, pollen diagrams, and also contemporary reports by Anglo-saxon monks point at an almost complete exodus during the 5th century, with the possible exception of maybe a handful coastal trading places (e.g. Oldenburg/Holstein). The causes are mysterious, especially as west of the line Kiel- Bad Oldesloe - Lüneburg, no drop in settlement and cereal planting becomes evident; land opening in Dithmarschen, near the mouth of the Elbe, even increased, as if people had qeued up there waiting for the next boat to England.
A substantial number of Anglians and Warnians (Mecklenburg) migrated to the Elbe-Saale area, as betrayed by the 8th C "Law of the Anglian and Warnians", and frequent toponyms on "-leben" (Anglian for "property", Danish equivalent is "-lev", English possibly "-low"). Hence, out of the ~250,000 Anglians/ East Holsteiners, maybe only 80,000 (1/3) migrated to England. Still, combined with the migrants from the West Coast, we should be talking about some 100,000 people arriving from Schleswig-Holstein alone.

PiT said...

Archaeological estimates for the size of the anglo-saxon migrations are tightly bounded between 30,000 people to 5000 people

On what basis do you make that claim? Who were the 'anglo-saxons'? And over what period did the migrations occur?

Estimates of the indigenous population do vary between 0.75 and 4 million. Pick your agenda....

As a lay reader of this site, with an amateur but thorough interest in the transition from Roman Britain to Germanic England, then 7 of the 9 samples offer a fascinating inset into Roman York, but neither these nor the 10 from Cambridgeshire are going to answer the questions of population size for either group.

Romulus said...

Wasn't the actual location of the Visigoth settlements in France? Would expect to see more Visigoth Y DNA there than Iberia.

André de Vasconcelos said...

@ Romulus, only in the early years, before the big fights with the Suebi for supermacy in Iberia took place. At the time the Suebi had control of over half the peninsula, even if they were actually based in western Gallaecia and northern Lusitania for the most part.

"The reign of Alaric II, who succeeded Euric in 484, marks an important turning point in the history of the Visigoths, for this was the time in which these people established themselves en mass in Hispania, abandoning their most important positions in southern Gaul to the Franks, whose supermacy here grew without stop for the following 20 years, until the enemies were reduced to Septimania. Omitting details about these events, we should keep in mind, however, as facts of greater consequence to the peninsular west, that the greatest movement of Gothic hosts to this side of the Pyrenees took place after 494..."

In "História de Portugal" (1992) by prof Henrique Mattoso, loosly translated by me.

André de Vasconcelos said...

*José Mattoso, I can't believe I actually got the name wrong lol

epoch2013 said...


"Not all men - but specially chosen men. *Amongst* this group - they were equal; i.e. no one CWC patriarch was more important than another. Given that we now know that these men were virtually all R1a-M417, it must have had a biological basis - patrilines, filial heirs, etc. "

We have historic examples of that, which must have had its origin in this: Be is the king of France who once was chosen from his peers or the German king and Holy Roman Emperor who was chosen by the electorat. Even after the new king was only selected by heritage they had to maintain the ritual.

Do we know if European nobility has a common genetic marker? Doesn't need to be 100%, as there were people ascending to high nobility (Sforza comes to mind).

epoch2013 said...


That "Law of the Anglian and Warnians" is actually the law of the Thüringians. It literally states: Lex Angliorum et Werinorum hoc est Thuringorum. It seems rather odd to separate the Thüringians from the Hermandurii so it looks like the three became the Thüringians.

Inklings of what have happened may be found in the myths of the Anglo-Saxons. They consistently seem involved with the Danes: Beowulf, the Fight at Finns Burg song in Beowulf, Ingeld myths.

Simon_W said...

Regarding the allegedly Brythonic character of the language of the Aesti, Tacitus wrote:
"ergo iam dextro Suebici maris litore Aestiorum gentes adluuntur, quibus ritus habitusque Sueborum, lingua Britannicae propior. matrem deum venerantur"

My translation: "on the right hand coast of the Suebian (Baltic) sea, there are the tribes of the Aesti, who have the customs and dress of the Suebi, and whose language is closer to the Brythonic language; they venerate the mother of the gods."

I'm not sure if it makes sense to seriously consider this, because we know Insular Celtic language and we know Baltic language, both are alive and spoken today. We also know the Baltic character of the language of the Old Prussians.

Romulus said...


I would still expect to find most of them where they settled rather than where they ruled.

Simon_W said...

It may have been an uneducated case of perceived phonetic similarity, like Basque and Spanish sound similar to the amateurish ear.

Matt said...

Trying to think of some worthwhile things to say...

3DRIF-26 seems worth testing in a world context, when we can, to see how he behaves in terms of African admixture. It's be interesting to see if he is like a modern Palestinian minus African admixture. Or if he already shows signs of African admix, that's interesting (as it may mean African admix in Levant already). If not, could be a more "African free" Levantine than any of the present day people we have (including Druze etc) which might be useful for certain admixture comparisons? He seems nearly equally close by IBS to Druze, Sardinian, Cypriot and Saudi.

Seems like this marks the first old genome (even if it's not near as ancient as others) which is closest to the present day Levant.

I wonder if there's some small effect size here boosting IBS to Anglo-Saxon by Welsh / Irish, Baltic to some of the Roman Britain samples (Basque and Sardinian higher up the scales than expected etc.)?

Grey said...

"Archaeological estimates for the size of the anglo-saxon migrations are tightly bounded between 30,000 people to 5000 people"

I think this is very silly.

Bede says it started with Saxon pirates raiding the east coast causing it to become depopulated (like parts of Iberia and Italy from later Moorish raids) and a Romano-British leader invited some Saxon tribes to *settle* along the east coast to defend it from others in a similar way to how the Normans were given Normandy later and the main war of conquest started some time after this - I'd assume after the allocated land was filled up and there were still more people wanting to come.

If that was the case - and I don't see any reason not to at least consider it - then the A-S population would have been much larger and *already there* along the east coast plus some unknown number of reinforcements from across the north sea.



"The causes are mysterious, especially as west of the line Kiel- Bad Oldesloe - Lüneburg, no drop in settlement and cereal planting becomes evident; land opening in Dithmarschen, near the mouth of the Elbe, even increased, as if people had qeued up there waiting for the next boat to England."

If they were queuing over a few generations to settle depopulated land along the east coast of England/Scotland then there could have been both a mass exodus and no change i.e. 100,000 gradually left and replaced themselves in Anglia over the same time period.


(over simplified labels ahead)

I think there is psychological tic in this area due to history dividing tribes into celtic and germanic leading to people thinking the genetic split is celtic vs germanic.

Personally I think
- the cultural division is celtic vs germanic
- the genetic division is celtic vs nordic
where germanic culturally is genetically a combination of celtic + nordic.

So what happens is people try and isolate what they think is the germanic genetic component but it's really only the nordic half of germanic.

For example Holland and Denmark are c. 50% R1 and c. 25%/50% I. If you assume only I is the germanic part then you'd underestimate the germanic *cultural* input by 3/4 to 1/2.

I think this is what happens with the A-S and possibly also some older groups like Belgae.


Back to Bede*, his sequence was
- pirate raids (east coast)
- depopulation (east coast)
- A-S settlement (east coast)
- rapid successful war for eastern part of the island (population replacement)
- stalemate
- slow piecemeal conquest of the west (mainly elite replacement)

so far nothing seems to be contradicting that.

(*or Gildas I forget)

Rob said...

We shouldn't read too much into supposed Baltic like character of the Iron Age samples. Suffice it to say, they might have had some continental origins which are best preserved today in north-Eastern Europe

Also, the allegedly Brythonic character of Aesti simply shows that Tacitus didn't know too well what he was describing, nothing more

Grey said...

Back on the samples they seem to me like further evidence there were two streams into Britain an earlier more southern weighted stream and a later more northern weighted stream with the latter gradually pushing the earlier backwards to the west over time.

Were gladiators always slaves or did some people do it as a job? If so you'd imagine they either had low prospects or some prior training (ex-soldiers).

"2 examples of R1b U106(M405) from Roman Era Britain, so much for that being an Anglo-Saxon marker."

Ex-auxiliries or captives? (early pirates even?)

Grey said...

then again

"If they were Germanics then why don't they show any greater affinity to the Anglo Saxon sample than the others?"

so scrub the pirate idea - unless there's some welsh like population in the Baltic they'd need to have been in Britain long enough to pick up Romano-British ancestry.


Grey said...

Apols, thinking aloud rather than after reflection but wasn't there a big argument on here about how some of the Hinxton "Celts" seemed too northern and wouldn't fit modern population without a lot of Welsh-like mixture?

Maybe this gladiator is the end result of one of those northern celts mixing with the more southern celts who were there before.

Rob said...


Hi again
I appreciate and share the enthusiasm you have for developing deeper understanding through a multi disciplinary approach

But I think we're misunderstanding each other a little yesterday. Eg you stated "For example, do you really think that migrating craftsmen can explain the transformation in this period? That a group of persons can just travel from group to group and violate ingroup-outgroup boundaries just like that? Or, in the Bell Beakers, that a transformation of the ritual system can occur through diffusion, or even more implausibly, by imitation of people from outside your community, without social upheaval that interrupted established traditions?"

I never said anything to that effect.
And I think I've made my theory as to why CWC might have been all R1a. I'm afraid I've run out of intelligence . And as per Collin, we can't wholly exclude biological aspects

epoch2013 said...


Tacitus is speaking about far away people and may have been merely pointing out that they spoke a different language. Overall, I think the text of Germania manages to mention a lot of European people for the very first time, in pretty much the right context (Aesti being Non-Germanic speakers). I find that quite remarkable.

Grey said...


I think about it in terms of ecological niches and divide terrain into optimal, average and marginal for neolithic farming.


Stage 1) neolithic
LBK settle on all the optimal and average farming land leaving marginal land for HGs (this marginal land possibly including large swatches of the Atlantic west).

Stage 2) (maybe) atlantic coast chalcolithic
small numbers of artisan cattle herders spread along existing networks as a minority until they reach lots of unused marginal land in the west and expand into the gaps swallowing up remaining HGs. (Hence concentration in the west). (With any violence happening after the population expansion.)

Stage 3) CWC (regardless of stage 2)
Peak LBK settled all the optimal and average land (leaving marginal for HGs) but some climate related change led to the abandonment of the average spots and a retreat to the optimal.

R1a cattle herders spread into the gaps - the abandoned average terrain - in a kind of infiltration-migration (IIRC something like that is speculated for Akkad/Sumer) so you end up with a distribution by niche, farmers on lighter soils like hill sides and river loess and the herders on the heavier soils in the gaps (also swallowing up remaining HGs).

Question 1) If so then how come the herders are all the same ydna?

"A question that naturally follows is, why would a culture of small household units have homogeneous Y-DNA patterns, indicating recent common patrilineage?"

The critical juncture is where marriages are contracted - so are marriages contracted between those small household units or in some central tribal gathering place?

So possible CWC sequence - herder tribe with same ydna move into a gap between LBK farmers, set up their central meeting place where the chief lives and then everyone else splits up into households around that place. However marriages are all contracted at the tribal meeting place hence remain endogamous. Population expands and a younger son takes a part of the tribe and moves to a new spot and repeats the process. That sequence might explain how they could have kept the patrilineage.

Question 2) Why aren't the farmers still there - where did all the G go and when did the steppe people mix with the farmers?

My guess would be violence but instead of a sudden big invasion and conquest more of a gradual infiltration of IE into gaps in LBK lands with the violence coming at the *end* not the beginning that is after the IE population had maxed out in those lands.

FrankN said...

To continue with emigration estimates to England: I had hoped for some concise statistical analysis of English toponymy, but couldn't find anything on the internet. However, for quite a number of typical place endings, information & mapping is available here:

Many of these names are undiagnostic, i.e. shared by Franks, Saxons, Anglians, Frisians, Jutes, and Norse, respectively. This includes, e.g. -borough/-burg/-bury, -beck (creek), -bourne (well), -thorpe, and -ing/-ings (posessive marker, but note that it has been mixed with -t(h)ing, which has a completely different etymology, namely "assembly place"). However, several of them can be indicative of the source of the settlers (no. of British place names in brackets):

-wich (44) [bay, landing place (also on rivers)]
-wick (613) [should in many cases be of Norse Viking origin, but may include Anglian, c.f. Schleswig, Brunswick, Narvik, Vikings]
-heath (740) [infertile land, c.f. Heide/Holstein, Haithabu]

Coastal Saxon (&Frisian?):
-hoe, -hoo (76) [hook-shaped elevation, c.f. Itzehoe]
-worth(y) (483) [elevation surrounded by water, c.f. Ilienworth n. Cuxhaven, Werder (Bremen)]
-ham (1288) [enclosure, landing place, c.f. Hamburg, Nordenham]

Inland Saxon, Frankish:
-ste(a)d (203) [typical for 5.-10. cent. AD establishment]
-field (1960) [c.f. Bielefeld]
-ford (1960) [c.f. Bremervörde, Herford/Westf.]
-ley, -leigh (2120) [fire-cleared area, c.f. Hengelo, Gütersloh]
-ton (7047) [elevation, enclosure]
-ingham, -ington (1070)
-don (699) [often from Celtic dunum "elevation, enclosure", but some constructions may reflect its Saxon cognate "ton", e.g. Grindon "green hill (dune)"].

Though far from being comprehensive, the above shows Saxon/ Frankish placenames being much more widespread than Anglian ones. This points at a substantial higher Saxon/ Frankish than Anglian immigration. If my estimate of 80,000 Anglian setllers is correct, Saxon and Frankish settlers together might easily have accounted for half a million or more, which would take us close to the 40% share estimated in the second paper. Let's furthermore not forget Danish 9th century immigration, signified a.o. by -by (village, 712) and -ness (promontory, 634) toponyms.

Interestingly, typical Frisian toponyms on -um and -ens seem to be rare in England.
Even more intriguing is the fact that the most frequent English-Saxon toponym, namely -ton, is seldomly found on the continental coast - Emden and Norden being the exceptions that confirm the rule. Instead, it clusters in SE Lower Saxony (c.f. the Hannover suburbs of Gehrden, Anderten, Ahlten, Rethen, Lehrte, Sehnde); another cluster is in/around Twente.
The second paper linked by Dave points at a possible second migration wave during the 8th century. The 772-804 Frankish-Saxon war should have provided sufficient reason for such emigration, which might explain the prevalence of inland Saxon toponyms such as -ton, -field, -ford and -ley in England.

ryukendo kendow said...

@ FrankN
Frank, nobody who I've read considers anything near half a million to be even remotely plausible...

At the end of the 600s, two centuries later after the migration began, village density was 1 village every 2-5 km in areas where the people were of A-S background, which judging from the estimated pop size for each village, approx 50 people, translates to a total population of immigrant background of only 250 000.... The reason why Harke made this work with a migrant population of 100 000 was because he assumed population growth of less than 1% per year, close to zero, for the populations at the time...

ryukendo kendow said...

@ PiT

The middling figures for migration size, 20,000 - 30, 000, come from Robert Hedges. The smallest figures come from Catharine Hills.

The largest figure in the English language literature, 100, 000, comes from Heinrich Harke. In fact it was him who proposed the 'apartheid model', because it was very apparent to him that the results that geneticists kept producing, comparing English Y-DNA with Frisian and Welsh Y-DNA, were far disproportionate to even the largest plausible figures that even a pro-migrationist like him could come up with. He teamed up with Stumpf and Thomas, two geneticists who helped him model the social dynamics required to produce the disparity, to come up with the 'apartheid' paper. FrankN, he's based in Tubingen, you might have read him before.

We're talking about orders of magnitude here, as the estimates will always be off for one reason or another; but all point to a similar conclusion, whichever ones we pick.

@ Rob
Rob, not you, but many sources I've read so far basically reduce it to that, or even to a trade goods package, which I find very implausible. Many cultural innovations were supposed to have diffused. Take individualist vs collectivist mindset, as inferred from burials; Scandinavian-Americans are still measurably more collectivist than Anglo or German-Americans in social behaviour, centuries after their settlement in the United States. Peoples living side by side in India, or other places in the world where ingroups are quite small, such as Papua New Guinea, often find the slightly differing burial practices in other groups to be abhorrent. Cultural outlook is not something that spreads so easily, much less entire ritual systems.

I think such a transformation of the ritual system could only have happened if the local social or political structures were 'decapitated', and pre-existing traditions interrupted, which requires some kind of disruption of the social fabric.

The only person I've found so far to have a realistic apprehension of the inertia that needs to be overcome for the changes that we see to occur is BellBeakerBlogger. I find that he appreciates how cultural transmission across the boundaries of 'ethnicity' (at this stage, really just ingroup-outgroup boundaries) is extremely difficult, that the appearance of finished goods like beakers across a large area, and the changes in burial norms, have more to do more with vertical rather than horizontal transmission, occurring through the physical movement of a people who did have some sense of self-identity, in line with the genetic evidence we see.

@ Grey
Is there any evidence of such central meeting places?

@ Matt
Matt, what evidence do we have that he comes from the Levant? Looking at the paper, he clusters among Middle Easterners today, but since he is removed from his geographical locality, I'm thinking its quite difficult to say which historical snapshot he provides where.

Rob said...

@ Ryu

"I think such a transformation of the ritual system could only have happened if the local social or political structures were 'decapitated', and pre-existing traditions interrupted, which requires some kind of disruption of the social fabric. "

Ok, so this is the crux of it. I though I clarified already, but lets go into it again. I believe what you're seeing is correct, but the explanation is not.

Yes, CWC is intrusive to central Europe. It speaks of a very different ideological language c.f. preceding middle Neolithic groups. As I stated several times, I do see it as intrusive 'from the East'. All's we're debating now is where exactly did it come from - a Repin offshoot ? Some EHG-heavy late CT group ? These are more or less minutii which don't change the fact that, IMHO, CWC was intrusive to Netherlands, Germany and Scandinavia from eastern Europe, and that this intrusion was more than just a 'cultural one', but also clearly demic/ biological, because few people would argue R1a-M417 emerged in western Europe.

But did it wholly replace pre-existing communities ? I don't think so, because - as I've also mentioned - earlier 'cultural models' continued. This is the key. I highlighted to you that CWC co-existed in time and place with GAC and TRB groups (although not necessarily in perpetual harmony :) ), whose origins can be argued to have been 'native' to central Europe. So it's not so much a matter of 'decapitation' as 'transplatation'. A whole new cultural model ('ethnic group") arrived, and did not replace, but co-existed. Whilst TRB does indeed demise a couple hundred years after CWC arrived, GAC continued till the end, when c. 2300 BC, both end. Perhaps the changes in subsequent, bell Beaker period in the eastern aspect of the European plain brought together communities which had remained distinct, more or less, for hundreds of years. I think this is what the evidence is pointing to.

Does that make sense to you?

Rob said...

Of course, that's still a rather simplistic picture

Grey said...


"Is there any evidence of such central meeting places?"

Well historically there's a lot hence suggesting the idea as a model of how a group who lived scattered *might* remain endogamous without trying that hard.


"For example, do you really think that migrating craftsmen can explain the transformation in this period? That a group of persons can just travel from group to group and violate ingroup-outgroup boundaries just like that?"

"They are feared in some societies for their skill in metalworking, which is considered a form of magic, but universally revered by for their technological pioneering. While common people fear the power of the blacksmith, they are highly admired and hold high social status."

"The Mande blacksmiths hold important positions in society. Blacksmiths are often called upon by the chief for guidance in major decisions regarding the village. The power of the blacksmith is thought to be so great that they are also feared."

"The Bamana society is very similar to the Mande. Bamana society is also endogamous, so blacksmith families are the only Blacksmiths in the village and they hold a very high status, due to the extreme power and responsibility that they possess."

Magic breaks the usual rules.



The overwhelming place name evidence is why I've never believed the low estimates.



Maybe relevant to something

Three Rastlin BB samples - S21 and LP (2000-1500 BC)

- two roman era R1b-S21 (plus earlier iron age woman) possibly representing BB ancestry
- three positive LP
- two negative
- four blank

Two of the three positive LP results are the iron age woman and one of the roman era S21 with the other S21 no result.

So maybe early BB version of "Celtic" S21 and LP pushed back by later varieties of less LP R1b "Celtic"?

Rob said...


For CWC, the 'central place' was the tumulus / burial mound itself. A permanent & prominent mark in the landscape to venerate one's ancestors

Grey , can you tally the various R1b subgroups seen in British and European aDNA samples thus far ?

ryukendo kendow said...

@ Rob

Thanks, this makes more sense. So the CW simply outbred the rest?

@ Grey

Grey, this is a within-ingroup phenomenon, which is very unlikely to spread. We, or other people who are not Mande, with very different attitudes to metallurgy, will not accept the high status of these foreigners. And metallurgy is present before BB already. The question is why such cultural and ideological packages can spread across ingroup-outgroup boundaries; they generally can't.

I raised the issue of village density two centuries later, which gives us an extremely hard boundary.

Rob said...

@ Ryu

"So the CW simply outbred the rest?"

Are you basically asking why haplogroup G 'died out' ?

ryukendo kendow said...

@ Rob

Yeah, the point of my questions boils down to this: how do we reconcile the genetic findings and their implications about differential reproductive rates (on both the uniparental and autosomal sides), and our understanding of the texture of the societies at the time?

Grey said...


"Grey , can you tally the various R1b subgroups seen in British and European aDNA samples thus far ?"

I think a much more efficient person than me already did it.


also, interesting to me if no-one else, 3/9 carry the red hair gene, 3/7 if you exclude the east med guy and the A-S

Creative said...

Makes me think of the Roman fortress in South Shields called Arbeia, and Barates a merchant from Palmyra.

Fort of the Arab(ian) troops.

Tombstone of a Roman woman

It was set up outside the Roman fort at South Shields in north-east England and records a British woman called Regina, who originally came from south-east England, and a man called Barates, who came from Palmyra in Syria. Regina was a slave, but Barates freed her and married her, and when she died aged 30, had this expensive tombstone made for her. It is Roman in style and has a Latin inscription, but also, uniquely in Britain, a second inscription in his own language, Aramaic, reading 'Regina, freedwoman of Barates, alas'.

Matt said...

RK: Matt, what evidence do we have that he comes from the Levant? Looking at the paper, he clusters among Middle Easterners today, but since he is removed from his geographical locality, I'm thinking its quite difficult to say which historical snapshot he provides where.

Other than the autosomal dna, isotopes only (which are apparently consistent with the climate and geology there). I don't know of many other parts of the Roman Empire which would be plausible (are Turkey, Cyprus, North Africa also viable?).

ryukendo kendow said...

@ Matt

The supplementary material associates him with Egypt, but the indicators are all moveable, i.e. it states that many of these ratios may not exactly be the same in the past, though the general area is more or less there. I think this will affect our inferences abt African ancestry in the ME quite a bit. Too bad we have no context that will place him more securely.

Onur said...


Three Rastlin BB samples - S21 and LP (2000-1500 BC)

- two roman era R1b-S21 (plus earlier iron age woman) possibly representing BB ancestry
- three positive LP
- two negative
- four blank

Two of the three positive LP results are the iron age woman and one of the roman era S21 with the other S21 no result.

So maybe early BB version of "Celtic" S21 and LP pushed back by later varieties of less LP R1b "Celtic"?

No, the Rathlin genomes are R1b-L21, not R1b-S21, thus unlike the two R1b-S21 Roman British samples you speak of. BTW, one of the Roman British genomes is R1b-L21.

Onur said...


(continuing a discussion from the previous thread)

I think the Aryan element in the formation of Proto-Turks were mainly carriers of R1a-Z2124, but I am still undecided what type of Aryans they were. It all depends on when Proto-Turks received their Aryan element (and the Caucasoid part of their ancestry along with them). If it happened during the Bronze Age, they may be undifferentiated Indo-Iranians. If it happened during the Iron Age, then we should expect them to be Iranics such as Scythians. The oldest Indo-Iranian layer in the Turkic languages might give us a clue. Do you know to which phase of the Indo-Iranian languages does the oldest Indo-Iranian layer in the Turkic languages belong?

Davidski said...

I've got the English genomes. The Middle Eastern Gladiator clusters very close to Yemenite Jews and some Saudis.

Not sure if and how much Sub-Saharan ancestry he has yet. I'll check that later today.

Maju said...

I'm with Romulus: he's making the correct and very objective observations. You don't need NE Europeans to identify Saxon blood when you have actual Saxons and Dutch. Reality check time!

Davidski said...

Much of Saxony is well east of Holland. In any case, they may have had paternal ancestors from near the Danube or even Black Sea. Some of the isotopes for the other remains from the same burial ground showed that they belonged to people who ate millet as children. Millet wasn't grown in England at the time, and is most typical of Eastern Europe.

I've got the genomes, so I'll check this out.

Maju said...

We know from other studies that Dutch, Frisian, Saxon, Danish... are pretty much the same thing. I would accept that you wanted a Danish, Frisian or Low Saxon sample to be 100% sure instead of just 99.9% but why Polish or NE European?!

The Roman Britons look all them pretty much like modern Welsh, what is what we should expect, especially if we think that there was a Saxon (and Viking!) immigration episode, as this study suggests.

Kurti said...

The Middle Eastern dude was possibly Ghassanid.

Davidski said...

Here's the PCA.

Red = England Iron Age
Orange = England Anglo-Saxon
Yellow = British Romans with R1b-U106
Black = British Romans
Black star = Middle Eastern Roman

And here are a few Eurogenes K15 results. I'll do a proper analysis of all these samples soon.

Pop England_AS
ID NO3423
North_Sea 46.53
Atlantic 23.61
Baltic 10.28
Eastern_Euro 12.46
West_Med 2.79
West_Asian 1.85
East_Med 0
Red_Sea 0.4
South_Asian 0.01
Southeast_Asian 0
Siberian 0
Amerindian 0
Oceanian 0.27
Northeast_African 0.84
Sub-Saharan 0.95

Pop England_IA
ID M1489
North_Sea 37.21
Atlantic 26.14
Baltic 9.85
Eastern_Euro 6.57
West_Med 9.84
West_Asian 3.24
East_Med 2.86
Red_Sea 0.02
South_Asian 1.62
Southeast_Asian 0.24
Siberian 0.86
Amerindian 0.27
Oceanian 0
Northeast_African 0
Sub-Saharan 1.27

Pop England_Roman
North_Sea 29.11
Atlantic 31.3
Baltic 13.4
Eastern_Euro 11.15
West_Med 9.47
West_Asian 3.11
East_Med 0
Red_Sea 0
South_Asian 0.02
Southeast_Asian 0
Siberian 0
Amerindian 0.58
Oceanian 0
Northeast_African 0
Sub-Saharan 1.87

Davidski said...

Oh yeah, also, here are the K15 results for the Middle Eastern Roman.

Pop England_Roman
North_Sea 0.02
Atlantic 4.06
Baltic 0
Eastern_Euro 0
West_Med 11.24
West_Asian 10.99
East_Med 46.16
Red_Sea 20.98
South_Asian 0
Southeast_Asian 0
Siberian 0
Amerindian 0
Oceanian 0
Northeast_African 6.54
Sub-Saharan 0.02

Ariele Iacopo Maggi said...

Ok, then the "slave trade" story is confirmed, or what?.
Red_Sea 20.98
Northeast_African 6.54
Sub-Saharan 0.02

Davidski said...

Put it this way, ancient DNA from just before and during the Roman Empire Italy will be interesting.

rozenfag said...

This middle Eastern guy has less SSA than roman Britons?! Am I reading this numbers correctly?

Helgenes50 said...

One of the previous England_AS, the most southern, was probably of British( Celtic) origin

Davidski said...

The Roman Britons don't have any Sub-Saharan admixture. That's just noise from post-mortem damage. I'm sure you've seen it before in other ancient samples.

The Middle Easterner does have real Sub-Saharan admixture, because he's got well over 5% of East African admix.

Ariele Iacopo Maggi said...

The middle eastern sample is better and has more calls, the ssa in those britons is just noise.

Ariele Iacopo Maggi said...

"The Middle Easterner does have real Sub-Saharan admixture, because he's got well over 5% of East African admix."
Yes, also Saudi have 4,14 EA and 0,31 SSA. I guess you can say that those populations did not change by much in that respect over the last 2k years.
Btw, a comparison of bedouin with this ancient ME guy (k15)
bedouin - 3DRIF-26
East med: 38% 46%
West Asian: 13% 10%
Red Sea: 21% 20%
West Med: 7.7% 11%
Atlantic: 2,9% 4%
NE African: 6,9% 6.5%
Sub-Saharan 2,8% 0.02

Times of Grace said...

The K15 oracle results for 3DRIF-26 using Admix4 [least-squares]

Using 1 population approximation:
1 Palestinian @ 7.892861
2 Bedouin @ 9.735329
3 Jordanian @ 10.699103
4 Samaritan @ 10.882444
5 Egyptian @ 12.261471
6 Yemenite_Jewish @ 13.881736
7 Lebanese_Christian @ 14.134447
8 Syrian @ 15.455195
9 Saudi @ 15.633462
10 Libyan_Jewish @ 16.488793
11 Lebanese_Druze @ 16.849952
12 Tunisian_Jewish @ 16.947318
13 Lebanese_Muslim @ 17.665131
14 Cyprian @ 18.849272
15 Sephardic_Jewish @ 21.690977
16 Kurdish_Jewish @ 21.923758
17 Algerian_Jewish @ 21.983456
18 Iranian_Jewish @ 22.708555
19 Italian_Jewish @ 22.898936
20 Assyrian @ 26.329748
206 iterations.

Using 2 populations approximation:
1 Tunisian_Jewish+Yemenite_Jewish @ 4.730545
2 Samaritan+Yemenite_Jewish @ 5.433585
3 Cyprian+Yemenite_Jewish @ 5.435184
4 Libyan_Jewish+Yemenite_Jewish @ 5.54292
5 Syrian+Yemenite_Jewish @ 5.863652
6 Samaritan+Saudi @ 6.055764
7 Jordanian+Yemenite_Jewish @ 6.058507
8 Lebanese_Muslim+Yemenite_Jewish @ 6.412908
9 Palestinian+Yemenite_Jewish @ 6.546099
10 Sephardic_Jewish+Yemenite_Jewish @ 6.605741
21321 iterations.

Using 3 populations approximation:
1 50% Yemenite_Jewish +25% Samaritan +25% Tunisian_Jewish @ 3.790951
2 50% Samaritan +25% Egyptian +25% Yemenite_Jewish @ 3.835664
3 50% Yemenite_Jewish +25% Cyprian +25% Egyptian @ 3.970079
4 50% Yemenite_Jewish +25% Syrian +25% Tunisian_Jewish @ 4.141381
5 50% Yemenite_Jewish +25% Samaritan +25% Sephardic_Jewish @ 4.156833
6 50% Yemenite_Jewish +25% Algerian_Jewish +25% Samaritan @ 4.168106
7 50% Yemenite_Jewish +25% Kurdish_Jewish +25% Moroccan @ 4.197818
8 50% Yemenite_Jewish +25% Jordanian +25% Tunisian_Jewish @ 4.22339
9 50% Yemenite_Jewish +25% Kurdish_Jewish +25% Tunisian @ 4.26225
10 50% Yemenite_Jewish +25% Lebanese_Christian +25% Tunisian @ 4.274044
1550395 iterations.

Using 4 populations approximation:
1 Samaritan+Saudi+Tunisian_Jewish+Yemenite_Jewish @ 3.616255
2 Samaritan+Tunisian_Jewish+Yemenite_Jewish+Yemenite_Jewish @ 3.790951
3 Egyptian+Samaritan+Samaritan+Yemenite_Jewish @ 3.835664
4 Cyprian+Egyptian+Yemenite_Jewish+Yemenite_Jewish @ 3.970079
5 Libyan_Jewish+Samaritan+Saudi+Yemenite_Jewish @ 4.032339
6 Lebanese_Christian+Saudi+Tunisian_Jewish+Yemenite_Jewish @ 4.032568
7 Egyptian+Palestinian+Samaritan+Yemenite_Jewish @ 4.113074
8 Syrian+Tunisian_Jewish+Yemenite_Jewish+Yemenite_Jewish @ 4.141381
9 Samaritan+Sephardic_Jewish+Yemenite_Jewish+Yemenite_Jewish @ 4.156833
10 Algerian_Jewish+Samaritan+Yemenite_Jewish+Yemenite_Jewish @ 4.168106

Kristiina said...

Onur, your question is very interesting but, unfortunately, I am not aware of any papers that would list words adopted at different times from Indo-Aryan languages to Turkic languages. I would presume that Turkic linguists have made such lists and they exist at least in Turkish.

However, for my personal interest, I checked a few words. I noticed that there are many words that have been taken from Persian and often they do not exist/ are not common in other Turkic branches: e.g. Persian džoft, Turkish çift, ‘pair’; Persian wa, Turkish ve, ‘and’; Persian ...-īke, Turkish iken, ‘while’; Persian (na-)xejr, Turkish hayır, ‘no’; Persian hame, kāmel, Turkish kamu, ‘all’, ‘whole’; Persian kohan Turkish köhne, ‘old’. On their basis of their distribution and a clear-cut meaning, they should be quite recent.

While I found many words adopted from Persian, with my quick check I came across only a few words that were shared with Sanskrit. The pan-Turkic word for milk is süt and this word is shared with Pashto, šode/šide, and with Sanskrit, sudhā. However, Sanskrit has even 8 words meaning milk, i.e. dugdha, avadoha, ūdhasya, sudhā, rasa, sara, jīvana, gorasa. Another word that I came across is Turkish dana ‘calf’. It has many cognates in IE and other languages, cfr. Lithuanian dienì/ Finnish tiine ‘pregnant animal’, Old Indian dhénā ‘female’, Old Irish dinu ‘lamb’; Saami deävene ‘female elk’; Chechen däχni cattle. This root has a much wider distribution and meanings are not so compact, so it is probably older. However, it is not found in other Turkic languages, so it is probably not proto-Turkic.

Then I took a look at some Proto-Turkic roots at the Tower of Babel and found out that there are roots that could be Nostratic, e.g. av (OUygh.), Proto-IE: *ag'-, Uralic aja, ‘drive’; aɣɨz, aɣaz (OUygh.) ‘mouth’, Proto-IE *okʷ- eye, Finnic avata, ‘open’, auki ‘opened’; Proto-Turkic: *Ăŋ(k)ɨt ‘wild duck’, Proto-IE: *anǝt duck, Mansi ǟŋghā̊, Khanty eŋχ/ ɨŋˁǩ ‘willow ptarmigan’. I stopped at 'A', so I probably missed many relevant candidates.

FrankN said...

@Ryu, Rob: I am still reading the CWC/GAC Poland and the Harke papers, and will come back on them later in more detail. Here already a few, unsorted remarks:

Artisans: The magical blacksmith, feared for his strength, socially apart, but with specific high status roles (e.g. executioner, i.e. the King's "right hand") is well evidenced from Germanic mythology (Celtic as well, I think).

Polish CWC: What has struck me is the distinctive CWC grave good package described in the Szmit e.a. paper: Bow/arrow, knife, hammer. This doesn't appear to be a pastoralist's toolset, it rather points to hunting. Looking at the strong CWC presence in Switzerland (which deserves some deeper analysis), and, IIRC, substantial finds of beaver bones there, I could actually assume fur trapping here, which might also explain the hammer for killing trapped prey with minimal fur damage. Specialised beaver, also mink, fox etc. hunting is already documented for Ertebolle and other Mesolithic cultures. Furs should have been a valuable trade article, especially during colder climate periods.

English population estimates:
a.) I miss the urban population. We have toponymic, archeological and textual evidence of various Roman towns continuing to be inhabited, e.g. London, York, Chester, Manchester, Colchester, Winchester, Chichester, Canterbury. Allowing for 25 such towns with an average population of 1,000 (London and York most likely surpassed that figure even during their darkest days) takes us already from 250,000 to 500,000 Anglo-Saxons.
Moreover, coastal settlement density and population should have been higher than in the inland, for trade opportunities and access to maritime ressources, and in view of the coastal origin of many of the incoming Anglo-Saxons, Frisians and Jutes. Obviously, England has quite a lot of coast...
b.) Later migration: Harke apparently assumes a background migration rate of 0.1%, i.e. ~100 persons/year from the 6th cent. onwards. This seems far too low, considering intensive relations across the English Channel and the North Sea. The German average 2000-2014 has ranged around 0.6% (mostly East European females), with war-inflicted migration rates (1990 ex-Yugoslavia, Syria since 2014/15) easily reaching towards 1% and above. In particular, the 0.1% assumed by Harke is clearly insufficient to accomodate possible immigration during the Frankish-Saxon wars, and the 9th/10th century Danish immigration.
c.) The Plague of Justinian is commonly assumed to have reached England. It coincided with the 535/536 cold event, caused by either meteor impact or a volcanic eruption in the Eastern Indian Ocean.
"Maelgwn, king of Gwynedd, was said to have died of the "Yellow Plague of Rhos" around 547 and, from 548 to 549, plague devastated Ireland as well. Saxon sources from this period are silent, as there are no sixth-century English documents."
An assumed 40% mortality rate would bring Domesday-book derived population estimates of around 1.9 m by 1086 in line with 0.6 m after, 1m before the plague (0,3% pa pop. growth), and a 40%, i.e. 0.4m AS immigration (or, alternatively, a somewhat lower initial immigration, and higher 8th-10th c. Saxon-Danish immigration).

The Justinian Plague might also explain some of the depopulation east of the Trave/ Ilmenau/ Elbe/ Saale: Some Anglo-Saxons (Warnians etc.) could have remained in the original homeland, but the Plague would have lead to such low densities that the survivors finally also left (possibly not to England, just to Saxon territories further west). This would in particular explain the fact that no re-population of those "cursed lands" occurred from the West, but they instead remained empty until Slavs arrived from the East more than a century later.

Rob said...

I mailed you an updated map
Btw, one of the R1a, "east-shifted" samples from Haak was from a LN site- not clearly CWC but rather Baalberg. It's attribution seemed difficult, so maybe we should look at it specifically

Grey said...


"No, the Rathlin genomes are R1b-L21"

Yeah I get L21 and S21 mixed up a lot - I should just call it the Irish one.

The intended point was the higher correlation with LP.

Grey said...


"The Middle Eastern dude was possibly Ghassanid."

I was thinking Syrian auxilia archer but that would fit Ghassanid as well.

Unknown said...

It seems millet was eaten to some extent in Roman Britain.

Rob said...

@ Ryu

“Yeah, the point of my questions boils down to this: how do we reconcile the genetic findings and their implications about differential reproductive rates (on both the uniparental and autosomal sides), and our understanding of the texture of the societies at the time?”

I’ll stick with the north European plain as a case study because Im relatively familiar with it, and it shows particular dynamism.

I don’t think CWC ‘outbred’ Neolithic farmers, at least not initially. In fact, as described, it appears to have been based on smaller social units dispersed throughout the landscape.

Again has to highlighted the demography of the north European plain, which shows a relative and absolute decline from c. 3400 BC – thus 600 years before the ‘arrival’ of CWC. Rather, this period seems to coincide with the rise of GAC in northern Poland, and clearly expansionist movements seen by the Baden culture centred in Hungary (which Frank has mentioned). We have no aDNA data from GAC yet (actually we have one mtDNA sample – K), so we’re not sure of how it looked genomically. We know that Baden sample from Gamba 2014 looked like a European mid-Neolithic (with greater WHG – similar to Copper Age Iberia and MN Germany), with little EHG and CHG (correct me if wrong), but no Y DNA samples (was just one female). But maybe it had greater I2a frequency ?

So then perhaps, the decline of the classical, haplogroup G dominated EEFs began considerably before steppic arrival, sandwiched as they were between rival cultures, and also facing soil exhaustion from their extensive agriculture, plague, bad health (as some studies indicate), ‘internal unrest’, etc.
Nevertheless, TRB continues into M3 in parts of Poland, before being ‘replaced’ by GAC. By 2800 BC, CWC also appears. As mentioned, it is synchronous with GAC, but the trend is that it appears later, and somewhat ‘succeeds’ it. But both cultures end c. 2400 BC.

Rob said...


Western Poland at this time sees the arrival of BB groups. In eastern Poland, CWC –like features continue in modified forms as the Trziniec culture. Later, c. 1300 BC we see the Luzatian culture – arising due to clear impulses from central Europe. This then collapses c. 700 BC, to see the spread of Pomoranian culture form the Baltic coast. This in turn collapses until the Przeworsk culture emergences again from central European “la Tene” influences. It too collapses c. 400 AD, after which ‘early Slavic’ cultures arise from Ukraine and the Carpathian region.

The point of the above paragraph is to illustrate that the CWC was not the ‘final chapter’ in the oder-Vistula plain. Its history is characterised by constant turnover. This is evident by the fact that the original R1a groups of CWC no longer exist, and the oldest clades – L664 – is found in far NW Europe, whilst older clades under Z282 are found in the east Baltic and forests of Russia. This matches the IBD analyses someone else posted that the CWC and Yamnaya –like impacts are best appreciable in peripheral European groups – like Irish, Baltic and certain Scandinavian groups.

Today, the form of R1a which predominates in Europe (L1029) has a TMRCA of 2000 years, and it expanded probably a few hundred years later. To conclude – all groups experienced turn over and loss. However, R1a probably had sources of renewal – adjacent ares of eastern Europe from where it originated, on the one hand, and experienced less turn over (e.g. when R1a-M417 groups associated with CWC moved in to the Baltic region, it expanded into thinly populated hunter-gatherer territory, and appears to have remained there ever since). By contrast, the G-farmers were a long way from home. And, if you note the scarcity of haplogroup G in the modern Levant- it appears their bridges were burnt  But that’s another story. So it probably a numbers and geography game, but this is not to deny agency of CWC people – who must have obviously had advantages – better health, ? taller, a more flexible and mobile social system, more carnivorous, ha ha.

The dynamics of R1b-L51 and BB are probably different. Italy, Iberia, Balkans are different still.

Naturally, this is a sweeping narrative. To go into the social dynamics would require specific case studies of individual cultures and micro-regions - which deserve monographs on their own right.
What are your thoughts ?

Rob said...

Maybe Frank can shed more like from the western end of the north European plain

Grey said...

Just an example of an empty meeting place: horse fairs, which might suit the steppic nature of the intruders

"Ballinasloe historically served as a meeting point, or hosting area, for clansmen from local tribes. Indeed its name derives from Béal Átha na Sluaıghe, the Irish for "Ford-mouth of the Hostings"."

Minimally - just a big field with a lot of horses in it

nb the bare back riding and a bit of a brawl over a ban on chariot racing

Appleby horse fair

21st century chariot racing

Romulus said...


In the Bell Beaker period men outnumber women, I've read this on Bell Beaker Blogger's blog a couple times, he's the kind of guy to be right. In the Neolithic women greatly outnumber men, one study found something like 7 women for every 1 man in a Neolithic grave site and noted the strange ratio in the study. Other Neolithic studies have also found more women than men in this kind of ratio. There are definitely genetic predispositions for having sons vs daughters, I don't know if this is the reason G2a is a rarer occurrence in Europe today.

Rob said...

@ Rom

That's very interesting.
Have they catered for sex-biased burial traditions ?

ryukendo kendow said...

@ FrankN

Frank, Harke did of course take into account towns in his calculations; the figure is still ~250 000 of people with immigrant background, 1:4 migrant vs native background, 2 centuries after the initial migration. Those of migrant background are indeed concentrated along the coast.

Do you know what an immigration rate of 0.6% means? An immigration rate of 0.6%-1.0% per year translates to ~20%-40% of the population being foreign-born after two generations. There's an issue of basic math here... I highly doubt we have any support for dramatic figures over a fifty year period in 6th-10th century England...

@ Romulus
Romulus, I think biased burials is the issue here... genetic influence on sex ratio at birth is weak, and if it exists its distributed over many polymorphic sites.

ryukendo kendow said...

@ FrankN
Frank, the reproductive advantage needed to get the ratio of A-S ancestry up from 5% to 50% in the 300 years where legal differences exist between saxons and natives, is something like 1.2. I.e. persons of A.S ancestry have 1.2 children for 1 child from people with native ancestry. I think you can agree that this is indeed well within the range of possibility, in fact considering the differing social status of A-S and natives in the laws at the time, its likely to *underestimate* the differential reproductive rates, which we know from empirical evidence are often very large in premodern societies where status differences occur. Once we have this factor established, and we realise that the actual contribution is ~35%, we simply do not *need* a very large migration at the beginning at all; anything very large will peg the reproductive rates very precisely at 1:1, which is v difficult to reconcile with what we already know about A-S society and how it treated locals for those three centuries where legal differences were observed.

On a purely physical basis this is simply the most parsimonious explanation.

Times of Grace said...

More K15 results for 3DRIF-26, this time the nMonte oracle by Huijbregts

fitted 0.183205
[1] "distance% = 2.142 %"
[1] "percentage ancestors by pop:"

Yemenite_Jewish: 42.80
Samaritan: 32.00
Sardinian: 6.80
Lebanese_Christian: 6.40
Saudi: 6.00
Ethiopian_Tigray: 4.60
Abhkasian: 1.05
Ethiopian_Amhara: 0.35

ghostnorris said...

Davidski, could you do a global PCA including these samples? The global PCA in the paper seems to suffer from projection bias. The ancient Middle Eastern individual seems to be among the least SSA-shifted Middle Easterners in the paper's analysis, and in a PCA without projection bias may be more clearly *the* least SSA-shifted ME individual.

Onur said...


Thank your for your elaborate assessment. Let me clarify, for people who may have questions about what exactly I am talking about, that by "Aryan" I meant Indo-Iranian, and not just their Indo-Aryan branch, so I was asking for the oldest Indo-Iranian layer in the Turkic languages, whether it is from undifferentiated Indo-Iranian or its Iranic branch (the Indo-Aryan branch has far less chance to be the source).

BTW, some of the Persian loanwords you mentioned are ultimately from Arabic (loaned in the order Arabic>Persian>Turkic or Turkish), those ones (e.g., xejr) cannot be older than the New Persian era (beginning from the 8th or 9th century AD) of the Persian language, when Arabic loanwords make their first appearance in Persian.

Grey said...


"On a purely physical basis this is simply the most parsimonious explanation."

If there were no casualties in the fighting.

What the written records say is losses were made up by new contingents from the continent.

FrankN said...

@rk: I am wondering if we both have been reading the same paper. This is what I will be referring to, but you may actually have had other publications in mind:

First of all, the a/m paper is neither putting forward explicit demographic models (with or without considering urban settlement), nor setting an upper limit on the size of AS immigration. In fact, it considers up to 200,000 early (5th/6th c.) AS immigrants as possible based on the archeological record, to be amended by later, especially Danish immigration (possible Saxon immigration during the Frankish-Saxon wars is not being discussed).
Taking the figures brought forward by you, without knowing their ultimate source, of 1 village every 2-5 km with some 50 inhabitants, this translates to a rural population density of 5 p/km² (50 / 1.75 km² times Pi). For England (130k km²), this yields a rural population of 650k, plus another 250k in urban settlements (see above). Close to the million I assumed before the Plague of Justinian (and the difference may well be found along the coast), and well in line with 600k after the Plague.
One of the main problems Harke had was the AS genetic share of 54% or more proposed in previous studies that were based exclusively on modern DNA. The second study linked by Dave has, based on aDNA, corected the AS share downwards to 38%.
Now, it is perfectly possible to build demographical models arriving at 38% AS (plus Danish, which is probably difficult to distinguish genetically) share without having to resort to AS reproductive advantage. I have tested two scenarios, which would both yield some 1,9m population by 1086 (Domesday Book) and 38% AS share from a 0.6m population by 550 AD (Plague of Justinian):
The first assumes 32% AS share already by 550 (equalizing some 320,000 AS immigrants before the Plague), and little migration thereafter (0.05% background migration, 15k extra Danish migration after 860, 0.2% p.a. pop.growth for both AS and Britons). The second one assumes only some 225,000 initial migration, but instead 0.1% background migration (as with Harke), and extra 30k Saxons coming in 770-800, plus 60k Danes 860-1030; natural growth for all groups is reduced to 0.15% p.a.
There is a variety of other demographic scenarios that will ultimately add up, including some postulating a reproductive advantage for AS. Which scenario comes closest is difficult to say - we would need more aDNA, and also more information about the extent of later Saxon and Danish immigration. The point is that the AS reproductive advantage postulated by you is purely speculative, unproven (but also not disproven) by the data at hand.


Alexandros said...

Maybe most have figured it out by now, but just to clarify following some initial posts regarding 'mystery' surrounding the closest modern population to the outlier Roman 'Briton' (3DRIF-26).. Figures 1a+b from the original paper pretty much clarify the picture. In the PCA plot (fig. 1a) this person sits nicely in a cluster comprising of modern Jordanians and Palestinians. In the ADMIXTURE analysis (fig. 1b), again he shows a genetic profile very similar to these two populations and different from modern Levantine (Lebanese, Syrians, Druze) populations and Cypriots. In fact, modern Levantines and especially Cypriots, have higher European-like (WHG/EHG-related) admixture and lower North African-like admixture than 3DRIF-26.

FrankN said...

Now the "apartheid" stuff: Harke considers it as one of several possible scenarios, within the context of regionally different patterns of AS-ization. He explcitely distinguishes three scenarios, noting that additional ones may well have occured, namely:

- "Kin-group model": AS, having arrived with females, and Britons living together in the same unit, also sharing cemetaries. The model is based on a late 5th to early 7th cemetary from Oxfordshire showing 64% burials with weapons (AS), average proportion of typical AS graves in the Upper Thames area is 53%.
- "Warband model": Predominantly male AS burials, but presence of Brithonic females nearby concluded from textile techniques, also Roman-British cemetaries in the immediate vicinity, exemplified from Warwickshire (82% of burials male AS).
- "Elite transfer" especially considered for Bernicia (Northern England), where other evidence points to a strong survival of "native" elements.

A fourth case discussed, but not developed into a model, is the followng:
"In Sussex, the earliest Anglo-Saxon finds dating to the second half of the 5th century are concentrated in an area between the rivers Ouse and Cuckmere where Romano-British settlement had been sparse, and avoid the area around the large town of Noviomagus Reginorum (Chichester); Anglo-Saxon finds only spread to the rest of Sussex in the 6th century. (..) In neighbouring Wessex,(..) burials, artefacts and settlements in diagnostic Anglo-Saxon style suggest the presence of intrusive settlers by the second half of the 5th century around Winchester (..)
There is widespread agreement that this situation reflects an early, possibly peaceful settlement of Saxons in Wessex, followed by the conquest by an ethnically mixed warband whose leaders and their descendants had British-sounding names (Cerdic, Cynric, Cædwalla) even though they were later called‘West Saxons’."

Apartheid under British leadership? In fact, Harke provides several arguments against it, most importantly body height homogenisation: "The main skeletal criterion distinguishing the two groups in the 5th/6th centuries was a stature difference, with men in weapon burialsbeing, on average, 20 to 50 mm (1 to 2 in) taller.
In this case, the differential isunlikely to have been the result of social and environmental factors because stress indica-tors such as dental enamel hypoplasia have the same incidence in both groups. Second,the stature differential between the two groups ceased in the 7th century, ie exactly at the time when social status differences appear to increase."
I am not sure whether we already have enough aDNA for final judgement, but the above points at a rather quick process of ethnic homogenisation by intermarriage.

The apartheid model is presented as "controversial" and "unnecessary assumption", and in fact only brought up as possible explanation for a AS genetic share above 54% - a figure now revised downward. Its presumed "evidence", the Law of Ine, may in fact be read just the other way round:
- Welshman are adressed in a variety of roles, including large landholders (above 5 hides~250 ha), farmers and smallholders, tenants, King's protegées (geneats), King's horsemen, burough citizens, and slaves - not less socially differentiated than Anglo-Saxons outside the nobility
- The weregeld for Welsh tended to be lower - 120S for a landowner/ tenant of a full hide, compared to 200S for a ceorl. However, ceorls were arms-bearing farmers, and owed military service to the King, which Welsh farmers didn't. This might explain the weregeld difference, especially when considering that Welsh horsemen bore a 200S weregeld equal to the ceorls. The weregeld for geburen, i.e. non-arms-bearing AS peasant in servitude to their overlords is unknown, otherwise, they are treated equally to Welsh land tenants.

Rob said...

@ Frank

"There is widespread agreement that this situation reflects an early, possibly peaceful settlement of Saxons in Wessex, followed by the conquest by an ethnically mixed warband whose leaders and their descendants had British-sounding names (Cerdic, Cynric, Cædwalla) even though they were later called‘West Saxons’."

I think many 5th century "Saxons" were Britons. This is suggested by the nature of the archaeological evidence south of the Thames- which continued Late Roman styled inhumations with Quoit brooch style etc.

"Anglian" settlement is more obvious north of the Thames - with their typical continental urn cremations

Obviously it was more complex than that

ghostnorris said...

@ Alexandros

"In the PCA plot (fig. 1a) this person sits nicely in a cluster comprising of modern Jordanians and Palestinians"

All of the PCAs are affected by projection bias. It's not always obvious, but we can assume it's there. If you look at the global PCA, it's clearer, because all of the British samples (excluding the ME outlier) cluster together, but quite far outside of the modern British range. The same is probably the case in the West Eurasian and Middle Eastern-specific PCAs in the paper.

Rokus said...

'2-3 of the Roman samples appear more eastern than the Anglo-Saxon'
Why this wouldn't be a more general trend among the Roman period natives of this geographic region? There is no need to jump to conclusions without knowing the actual age of the suspected Baltic admixtures in Britain. Don't forget the samples turned out quite native in the study, the evidence so far is that U106 was pre-Anglo Saxon in Britain and hence any eastern IBD markers would simply fit the more general Northern European affinity already mentioned in the Martiniano study. From Schiffers we know that the rarecoals method doesn't have any trouble to distinguish the Migration Period origin between Dutch and Danish over a time gap of almost 2000 years, so even affinities twice that age should still be feasible and may be expected.
As already mentioned elsewhere, according to some prominent investigators the ancestors of the Celts indeed came from a more eastern location than those of eg. Germanic and Italic, since they share 2000 BC linguistic innovations with Iranians and Slavs. Baltic IBD admixture thus may as well be reminiscent of the ultimate origin of Celts, or at least of the northern branch. This initial 'Celtic' expansion towards Britain was probably congruent to the expansion of Bell Beaker, and to the British distribution of L21 and U106 that may be deduced to fit the shared Bell Beaker origin from Armorica (L21) and the Lower Rhine (U106) neatly.

Simon_W said...

I guess around 38% Germanic admixture, with a wide range 25 to 50% should also be in the ballpark for Southwestern Germans / northern Swiss. Hopefully we'll soon see some western Hallstatt or La Tène DNA.