Wednesday, March 18, 2015

Peterborough Ware, DNA

The apparent resurgence of WHG profiles by the end of the Middle Neolithic has been variously commented on and I think putting one corner of Europe in an archaeological context might give a little clarity for why this happened.

I had meant to post this paper on Peterborough Ware as it relates to the food vessels of British Beakers, but got sidetracked with other news.

Fengate Ware, Berkshire, 1990,1009.31, British Museum (order FI-000715026)

The British Middle Neolithic Peterborough Ware may have its roots in Southern Sweden or generally the Baltic coast, having plausibly spread via the similarly named Peterborough Ware of the Dutch lowlands.   History repeating itself.

If this was a population movement by stone-axe wielding boatsmen, it may have been part of a flood of similar Middle Neolithic 'pre-Vikings' that were floating down the rivers throwing axes in the faces of Farmers.  It's possible that the rise of certain genetic profiles we see in the Middle Neolithic from Hungry and Sardinia to the Languedoc and Northern France, were subject to similar throngs of northern barbarians coming over the walls of the palisades.  Environmental conditions again changing the board of history.

The Peterborough People seem to have arisen about 3,330 B.C. and probably overlap the introduction of Bell Beakers (more on that).  But before the Beakers, the Peterborough tradition appears to have lived in a system of apartheid or avoidance from the Grooved Ware peoples, reasonably assuming the two wares weren't used differently by the same people.  These two peoples seem to have been total opposites in every conceivable way, preferring different living arrangements and beliefs.

The fluffy version of Grooved Ware origins is that it is an independent development in Britain with people being attracted to a cool culture based on communal feasting and alcohol, except for those that avoid it.  (Whenever you hear the words "independent development" come from the mouth of an archaeologists, a good rule of thumb is to immediately disregard and begin looking for population movements).  Regardless, Grooved Ware culture is intrusive to most or all of the Isles, having come several hundred years after Peterborough Ware.  When Bell Beaker Culture arrives, it begins consolidating and/or replacing various separate traditions into a more uniform Early Bronze Age.

How exactly food vessels and urns emerge is another question mark.  Clearly food vessels are heavily influenced by islander impressed potteries, like Peterborough, but overlaid with Beaker motifs.  According to historical dating schemes, this shouldn't have been possible given the end of one and the beginning of the other.  But it may be that Beaker pottery is both earlier than presumed and Peterborough Ware continues strongly with hillbillies.

To add to the weirdness, is it possible that some Peterborough pottery, like the Fengate bowl above, are already beginning to copy certain Beaker stylistic motifs?  Perhaps the relationship between Beaker immigrants and hillbillies was agreeable given a common enemy, or Beaker cattle drivers were more frequently in contact with the marginal areas.

The question of Scandinavian influence from the Middle Neolithic is a good one.  Given its strong influence on the Isles prior to the Bronze Age, it would certainly seem that British profiles may be shifted towards Fennoscandia and the Baltic Coast.  As such, we may expect some British haplogroups to descend from these early migrants, but more interesting is the abundance of certain maternal haplogroup H and V in Britain.

It would truly be the most bizarre situation ever if the maternal lineages in Britian were the recombination of Fennoscandian and Ibero-Saharan sister lineages, both making their way ultimately from the Northern Middle East and both generally pared with members of the R1 paragroup in opposite directions?

Total PeterboroughWare



Summary. Peterborough Ware is now recognized as the dominant ceramic tradition of the middle Neolithic in southern Britain during the period 3400–2800 BC, part of a wider north European family of Impressed Wares. Drawing on an extensive inventory of 600 recorded assemblages constructed by enriching previous lists with the results of development-driven research carried out over the last 20 years or so, this paper reviews the production, distribution and use of Peterborough Ware. Support is found for the traditional sub-division of the Peterborough Ware series into three sub-styles: Ebbsfleet, Mortlake and Fengate Wares on the basis of the materials used, forms, and the decorative
schemes preferred in each. The overall distribution of Peterborough Ware focuses on south-eastern Britain although there are important assemblages from areas to the west and north, especially those composed of Mortlake Ware. The range of contexts in which Peterborough Ware was deposited is wide, but suggests a backward-looking attitude in which the users of this style of pottery were trying to connect with their past.
Alex Gibson [Link] 

Julian Thomas [Link]

*Update 1*  This big study from Nature genetics hit the wire literally moments after my post - via Dienekes.  I consider my above comments as "pre-criticism".  Most notable is red square country of England clustering with Scandinavia which overlays Peterborough world.  Only large-scale dna from ancient remains can properly resolve the human mosaic over time.


  1. I do find very interesting your take on pots first time I heard of Petersborough and I have many questions about Grooved Ware too, never mind the suggestive situation of ethno-cultural segregation you mention. However I do not really find your genetic categories "Fennoscandian" or "Ibero-Saharan" meaningful nor really relevant here:

    1. Fennoscandians seem to fall in three groups: Finns (high Y-DNA N1 and R1a), Swedes (high in I1 particularly) and Danes-Scanians (more normal in relation with Western Europe, with mild amounts of R1b and other lineages). Norwegians are probably mostly in the line of Danes but can also be differentiated. I'm emphasizing Y-DNA because it's a very obvious signature of different origins in this case.

    2. There's nothing in genetics that can be classed as "Ibero-Saharan", in general terms the Strait of Gibraltar has been more a barrier than a transited passage. The main exception here is mtDNA H, which is so strongly shared with all the rest of Europe that it's extremely forced to make an "Ibero-Saharan" connection as such, although obviously the NW African H (also present in some "Saharans" like Tuareg) arrived there from Iberia/France per all accounts. Other less remarkable exceptions are mtDNA U6 and Y-DNA E1b-M81 which surely originated in Morocco affecting only parts of Iberia (Western Third) and have very limited presence elsewhere in Europe (so can't relate to Britain unless you're talking of very exceptional cases).

    Instead I'd talk of NW and SW continental influences, being the former from the area between North France and Netherlands, maybe up to Saxony, and the latter from South France and less likely Iberia. This is very apparent especially in Y-DNA R1b, where we see the two major subclades S116 (from South France surely, otherwise Iberia) and U106 (from somewhere in NW Europe, most common in the Netherlands) overlapping in Britain. Nothing Swedish, Finnish or NW African is to be found, at least not at significant levels.

    1. These are kind of made up names for something difficult to frame. I'd separate Scandinavia as a Western section and the Eastern Baltic Sea up to the White Sea. I think those are the same categories you have as well. The first group you mention being Swedes, Karelia and the Baltic coast is, as you've said, probably more R1a which hovers at a mild frequency over the East of Britain. However there have been so many immigrations from Dutch Beakers to Vikings, its probably impossible to know where the modern frequencies come from.

      The maternal side I kind of guess is a cluster around Finland that appears more like Western Iberia than Central or Southern Europe. I'm not totally sure how to package it other than the outside closing in around the center of the continent.

    2. To clarify, I meant in Scandinavia (I) to separate from the Eastern Baltic, Karelia, White Sea (R1a, N)
      Also that I have no idea whether R1a in Britain came from Vikings, Saxons, Dutch Beakers or Peterborough, other than the situation may be a bit more complex

    3. Truly it's a bit labyrinthine but seems, according to the latest studies, that, in Germany at least, the population was roughly like modern at the end of the Chalcolithic (Bell Beaker sites are around the first ones to fit this description). It's hard to know how it works elsewhere in Europe (I have my ideas but better wait for further aDNA data).

      Anyhow R1a is not too important in Britain (~10% compared with ~25% in Germany/Denmark and ~50% in Poland), so I tend to de-emphasize it, not just in Britain but also in SW Europe (basically the North Sea-Rhine-Alps line seems to be the main ancient Western border of this haplogroup, which can be attributed to IE migrations). The most important British patrilineage is R1b, which is split in the two subclades I mentioned before. Incidentally Ireland, which experience major demic growth in Bell Beaker times, has almost only the "southern haplogroup" S116, almost certainly arrived from France or Brittany (via SE Britain I guess, where it's also present in significant amounts).

      For European R1b details, see HERE, particularly this map (note: M529/L21 is wrongly written as M259; also circles represent only R1b subclades within the overall R1b-M269 fraction not their overall frequency in the various populations, which varies from near 100% in the Far West to near 0% in the East).

    4. I hope to see some ancient Atlantic DNA this year. There's so many questions

    5. Me too. Absolutely! It's of key importance and not everyone seems to understand this.

  2. I don't mean to sound like a broken record.

    But I remain far from convinced that all of the R-M269 found across Africa arrived in its entirety from Iberia/France.

    And yes, that is R-M269 - as opposed to R-V88. Many people seem to assume that R-V88 is the only form of R1b found in Africa.

    1. Probably correct. It seems to increase further down the Nile and into the SW. At this point I'd be happy with modern dna studies.

    2. M269 is definitely not original from West Europe unless it belongs to some specific subclades. Instead it must have originated either in West Asia or the Balcans for all I know.

      However I'd like to know on what do you base your claim that M269 is found "across Africa". My memory of studies of R1b in Africa, detected a peak in Sudan but the exact classification of the subclades was not clarified. Is there anything "new" on this matter that I have missed?

      Central African R1b seems to be all V88, so R1b(xV88) in Africa seems (if it can be confirmed as such) restricted to the Nile (Sudan, Egypt) and a few other areas (NW Africa mostly but rather low frequencies).

    3. "However I'd like to know on what do you base your claim that M269 is found "across Africa"."

      I am at work right now. I'll dig out the references over the weekend. This was from my own research, reading several papers and skimming through extensive Family Tree DNA data. R-M269, completely separate and distinct from R-V88, is found in the following African regions/populations (ones that I know of SO FAR):- Algeria; Libya; Egypt; Sudan; Chad; Ethiopia; Nigeria; Equatorial Guinea; Khoisan; Biaka pygmies (C.A.R.). Someone really needs to do a meta-analysis of all Y DNA and mtDNA data from studies in Africa over the years.

    4. Chris: the very rare "colonial" R1b (and sometimes also R1a, I) is just that: colonial, product of 500 years of systematic exploitation of that part of Africa by Europeans. You wouldn't think that such a dramatic socio-economic and political period of history would leave zero genetic legacy, would you? But in frequency terms it is anecdotal, mere "erratics" and I'm sure you track those lines to Europe.

      Finding R1b in Bioko (ex-Fernando Poo) is just normal, as the island was a European colony for some five centuries. Even Khoisan seem affected, surely by mediation of Coloureds, who are partly Khoisan themselves.

      Another thing is North Africa, Chad and the Horn, where European colonial legacy is more negligible and West Asian or pre-colonial European impacts are more reasonable to expect. I'd appreciate any detail on frequencies and haplogroups you could provide for those areas, really, because they could indeed shed light on the spread of R1b in Africa. Anyhow my main work hypothesis is to consider Sudan (and to lesser extent Egypt) as a hub of West Asian expansion, what rings true for J1 and R1b-V88 and may also include other lineages. NW Africa can harbor a more varied origin of R1b though (Europe here is a possible partial source), but should be analyzed before jumping to conclusions.

    5. @ Maju -

      Here are some references related to R-M269 in Africa:-

      "Human Y chromosome haplogroup R-V88: a paternal genetic record of early mid Holocene trans-Saharan connections and the spread of Chadic languages"

      Table 1 gives frequencies for various subhaplogroups of R1b [R-V88; R-V88*; R-M18; R-V8; R-V35*; R-V7; R-V69; R-M73; R-M269].

      For R-M269:- Moroccan Arabs [N=55] 1.8%; Morocco Asni Berbers [N=54] 1.9%; Northern Egyptians [N=49] 2.0%; Egyptian Siwa Berbers [N=93] 1.1%; Egyptians from Baharia [N=41] 2.4%; Ethiopia Amhara [N=83] 1.2%.

      "Inferring population structure and demographic history using Y-STR data from worldwide populations."

      I have the data from this paper and am emailing it over to you in a minute in an excel file.

      Hausa [Zaria, North Central Nigeria] R-M269 1/16 = 6.25%.

      "Chad Tribes - Y-DNA Colorized Chart"

      R-M269 in Chad = 1/38 = 2.63%

      "R1b and Subclades Gateway Project for R1b (M343+) Y DNA Haplogroup (includes M269) - Y-DNA Colorized Chart"

      Several pages of R1b data if you can be bothered to trawl through. I found R-M269 from people in Algeria, Libya, and one from the mountains of Cameroon.

      "R-Arabia Y-DNA Project --- في العالم العربي R مشروع السلالة - Y-DNA Classic Chart"

      R-M269 in Egypt, Libya, Tunisia shown.

      "البحث الجيني لبلاد المغرب الكبير Maghreb DNA Project - Y-DNA Classic Chart"

      R-M269 in Morocco, Algeria, Tunisia, Libya.

    6. Thanks for providing your sources.

      First of all, it's not "all across Africa" but we see R1b1a2(M269) limited to North Africa, Chad and the Amhara and with no characterization downstream of this generic marker, all regions where West Asian genetic influence is apparent (lots of J1, surely mediated by Egypt/Sudan). It's possible that it's also found in Sudan, where R1b1-P25 is important but has been never researched for downstream markers AFAIK.

      In general I'd say that the main origin (as with R-V88 and J1) is West Asia via the Nile. It's possible that in NW Africa there's also some West European R1b but this needs to be confirmed.

    7. Thanks Maju. On an unrelated matter, what are your thoughts on some of the data on the spreadsheet which I emailed over to you? Ones which caught my attention include C-M217 in Mbuti pygmies and C-M130 in Biaka pygmies [which also have R-M124 (R2) and H*-M69xM82]. I had already previously mentioned Q-MEH2 in Tanzania Sandawe; now that I can see that this is also present in Druze it probably implies migrations from West Asia into both Eastern Africa and Central Asia, as also seen with mtDNA N1a etc.
      Q might also be present in Fulani as in another study P*(xR) was found at 4% in a sample of Fulani in Bongor.
      A while ago I pointed out the presence of N1c-tat in Equatorial Guinea, however you were adamant that it must be of recent origin, eg Scandinavian sailor mating with local woman or some such.
      But if other Y DNA haplogroups typical of Central Asia, North Asia etc. have actually entered Africa from West Asia in antiquity [eg haplogroups C and Q, maybe others] then can we assume that N1c-tat didn't also arrive in this manner?

    8. Q is relatively common in West Asia, where it originated (NE Asian & Native American Q1 is derived from that origin, just as mtDNA X2 is as well).

      I didn't notice those C individuals in cursory look but I did notice H* and IJ*, which do suggest very old back-migrant erratics from the Western parts of Asia (surely in the LSA genesis, which is somehow linked to the West Eurasian Upper Paleolithic). In brief we are probably talking of minor flows some 50 Ka ago. Quite more important among those Asiatic flows (arguably from the same era) are J1 and R1b, which in Africa seem to have spread first to the Nile.

      From memory P* also seems important in Cape Verde, which is a mixed colonial population but it was not R1b, so it may correlate with the P* among Fulani. Another thing is whether that is of Paleo- or Neolithic arrival, as Fulani seem to have some non-negligible West Eurasian genetic element, which has been argued to be from Neolithic herders of the "Green Sahara".

      As for N1c it could be Dutch, German, Danish or Norman, all European populations involved in the exploration or exploitation of West Africa in the last 500 years or so.

      There's a label for these random findings: "erratics". We can't issue judgements based on them because there can be almost any private history behind them, without any relevance for historical population genetics; another thing would be if frequencies would be significant (but they are not). It might also be significant if the lineage appeared to be very different, suggesting an old erratic flow rather than a recent one, but, honestly, I doubt that will be the case with N1c.

    9. So African populations, which many understand only harbour Y DNA haplogroups A, B, and E, actually have a considerable diversity of markers:-
      A, B, C, DE, E, F, G, H, IJ, I, J, K, L, P, Q, R, and T [leaving out the N!].

      mtDNA haplogroups in African populations are also proving to be more diverse than just 'L' plus U6 and M1 as was previously believed.

      You estimate some of this Eurasian back-flow to be in the order of 50kya, potentially also including R1b.
      Is this due to your doubts about the reliability of Y for molecular clock analysis?
      [K2b-P331 estimated age is 50ky; P-45 between 27-45ky; R itself between 27-34ky; R1 between 12.5-26ky; and R1b <18.5ky]

    10. That depends on what you consider "African" and what you consider relevant. Often "Africa" is implied to mean Ultrasaharan Africa where no significant Eurasian backflow is present (i.e. excluding also NE Africa: Sudan, the Horn). North Africa is for all genetic purposes more Eurasia than Africa, the Horn is like 50-50. Of course there's also some African stuff in Europe and West Asia, most notably E1b (but not only).

      "Is this due to your doubts about the reliability of Y for molecular clock analysis?"

      It is, yes. But also when I use (others') full chromosome trees and recalibrate for age(CF)=100 Ka BP, as archaeology demands, I get those dates. For example HERE (scroll down to update), I got age(R1)=48 Ka (from recalibrating a previous work by T.D. Robb, see HERE and again scroll down to updates) and age(R1b)=34 Ka, while K2 (ex-MNOPS) would be around 76 Ka BP, what fits well with the theory that its expansion (as that of mtDNA N/R, strongly associated) are related to the Toba catastrophe of 74 Ka BP. Using Underhill's tree already, I get that European R1b could be Magdalenian or older (but a bit uncertain because branches are not exactly equal, so still allowing for a Neolithic founder effect - to be demonstrated).

      My doubts, beyond the problem of methodology, relate strongly to the impossibility that CF (the key Eurasian macro-haplogroup, or M in the case of mtDNA) need to be of around the time when H. sapiens reached to India and East Asia, which is c. 100 Ka BP. A secondary expansion from SE Asia (the Bengal-Sundaland arch) happened later c. 74 Ka BP, exploiting the Toba generated chaos. This seems to fit for Y-DNA as for mtDNA and also when I look at African DNA, I get c. 125 Ka for age(L3), etc.. Yes, a couple of early diverged Y-DNA branches seem pre-sapiens but why not? I doubt they could really differentiate between H. sapiens and a close cousin in Africa back then. There's even an X-chromosome lineage that seems Neanderthal, and Australasians have even small admixture from an even more remote cousin, H. heidelbergensis, alias "Denisovan".

      See also:

    11. Thought that I would just give a quick update. According to the Y DNA data which I emailed to you, N1c is found in the Druze. So if Q1a, T, G, J, and R1b R-M269, which are all found in the Druze and in north African and tropical African populations, I see know reason why the same can't also be true of N1c with it being found in Druze and Equatorial Guinea. Maybe it will turn up in Egypt and Sudan with improved sampling. Of course it could still be due to recent European admixture as you say.

  3. "Fennoscandian and Ibero-Saharan"

    I kind of agree with these labels except I see them as directions of travel rather than specific source regions if you see what I mean.

    I think the physical geography creates three primary east-west paths through Europe:
    1) north of Carpathians and along the Baltic/North Sea coast
    2) Danube to south Germany/northern France
    3) Med. coast (north and south)

    and then there is the north-south aligned Atlantic region joining all three together and thus allowing the possibility of what you describe.

    So to me
    - your "Fenno-Scandian" represents the direction of travel towards the North Sea (and beyond)
    - your "Ibero-Saharan" represents the direction of travel to southwest Europe (and beyond)
    - with the third one Maju mentions being the direction of travel to the central Danubian route (and beyond) via Brittany

    so I see it as three streams (east, south and north-east) rather than three sources hence why each stream can have components from much further away.

    Apart from that quibble on the mechanics - very interesting thought.

    1. The Fennoscandian/Ibero-Saharan question is more of an illustration as how the genetic mosaic becomes less simple through the lens of archaeology. Here's two populations that in some combination or no combination contributed to similar mtdna of the Isles. It just makes understanding the current situation more difficult based on limited information.

  4. This is going to interest you:

    An interesting detail is that most English form a single cluster, impossible to break apart, apparently. But more interesting maybe is that the main continental source of ancestry in general is from NW France, followed by Germany and at a distance Belgium and Denmark. Neither Scandinavia proper nor Iberia are significant sources of ancestry in the UK, excepted the Scottish islands, where Norway has a minor impact.

    The main English cluster is mostly contributed by one but not the other French cluster (darker blue), which is also dominant in Cornwall and Devon. Secondarily but significantly contributed to also by Germans and Danes.

    On the other hand Welsh and Scots feed mostly from the alternative French cluster (lighter blue), but also from one of the German clusters (yellow, rather Renish, the orange one looks rather Saxon instead but it's much less influential).

    In general it looks like mostly Neolithic to me, although it's difficult to discern this from possible from Celtic sources. A possible interpretation could be:

    1. Light blue (FRA14) source: Neolithic Megalithic of the kind that went from Brittany/High Normandy along the Irish Sea.
    2. Dark blue (FRA17) source: Neolithic Danubian of the kind that affected mostly England.
    3. Yellow (Ger6) source: Celtic (first IE)
    4. Orange (Ger3) source: Anglo-Saxon
    5. Danish source: Viking or Anglo-Jute

    But just an idea anyhow.

    1. I'm still trying to make sense of the clusters in light of genealogy. I've read previously that most people in the island of Britain share the same progeny at about 1400 A.D. ( to the tune of 90%). So it's no surprising to see a lack of diversity.
      On the other hand some of the clusters they show may be explained by more recent marriages and alegiences.
      I may not fully understand their angle, it seems a little paradoxical that they describe Brits as monolithic but distinguishable by prehistoric population movements.
      They need to release the data so the genealogical and population genetics communities have a chance to scrutinize

    2. I'm systematically skeptic about chronological estimates, as the science behind them is very frail and subject to a long chain of scholastic vices. I don't see any reason to imagine a 1400 CE date at all, rather a founder effect should be much much older, maybe from 4000 BCE.

      You are right no doubt about endogamous tendencies explaining the clusters. I'm not even considering them but the evidence about continental origins, which is clearly not Anglosaxon in essence but "French" and secondarily "Rhenish". With my previous interpretation on hand about the possible origin of the various clusters I made some rough estimates for the three main British populations:

      ENGLISH (main cluster): 50% Danubian + 25% Celtic + 15% Danish + 10% Saxon

      Notice that the "Danish" reference might be closer to Anglos and Jutes than the "Saxon" one, so not necessarily all Viking in that source. Other English clusters seem in general terms to follow this pattern except "West Yorkshire", which seems dominated by the Celtic (Rhenish) component.

      The Celto-Germanic fraction (modern German and Danish sources) fits rather well when you compare English R1a with (West) German one. Judging on this Y-DNA clue, the overall "German" influence in England is around 40% and judging on this new autosomal data is around 50%.

      WELSH: 60% Armorican Megalithic + 40 % Celtic.

      The three Welsh clusters follow approximately this pattern.

      SCOTS (most clusters): 40% Armorican Megalithic + 30% Danubian + 30% Celtic.