May 4, 2012

Ancient DNA from Chalcolithic Thuringia: R1b and headaches

Note: partly edited hours after first publication - reason: see comments.

New archaeogenetic research for our delight or confusion, we'll see:

This new study (which I could read in full already) has analyzed the DNA of several individuals buried in Kromsdorf (near Weimar, Thuringia, Germany) c. 2600 BCE. The burials are said to belong to the Bell Beaker "culture" (better described as phenomenon in fact) and the results are as follow:

  1. Grave 1 - 18-21 y.o. man - no DNA reported - laying on his left side, w/ cup
  2. Grave 2 - 17-20 y.o. woman - no DNA reported - laying on her left side, w/ bowl and bone needle
  3. Grave 3 - c. 35 y.o. man - mtDNA reported as U2e - laying on his right side, no grave goods
  4. Grave 4 - a couple:
    1. 4a - 21-25 y.o. woman - mtDNA reported as W5a - w/ chert flake.
    2. 4b - a man - no DNA reported - no grave goods
  5. Grave 5 - 35-50 y.o. man - mtDNA reported as I1, Y-DNA reported as R1b1b2 - no grave goods
  6. Grave 6 - 6-12 y.o. child (gender unknown) - no DNA reported - no grave goods
  7. Grave 8 - 21-26 y.o. man - mtDNA reported as K1, Y-DNA reported as R1b - w/ cup and flake
  8. Grave 9 - a couple:
    1. 9a - 25-45 y.o. woman - mtDNA reported as U5a1 - w/ loom weights
    2. 9b - 45-55 y.o. man - mtDNA reported as T1a  - no grave goods reported

Edit: I came to accept the reported mtDNA haplogroups as surely correct

As G. Horvat noticed (see comments) the four dubious haplotypes have the 12705C marker, which indicates that they belong to R.

Actually it looks like someone made an error in the tabulation of the results in fig. S7 and both columns 12705 (defining R) and 10873 (defining N) are identical. Considering the reported results and that it seems that the HVS sequences are common today, the 10873 column seems to be in tabulation error, leading to my mistaken previous critical read.

But... coding region markers say it's not even N!

Hat tip to Gail for the heads up here.

In the supplemental material, table S7, the sequence of several coding region markers is reported for the above individuals who yielded mtDNA. The problem is that all sublineages of macro-haplogroup N (and all lineages reported above are that) carry the 10873 site as T. However the burials number 3, 8, 9a and 9b have it as C instead. 

This implies that they are L(xN), whatever it is. Sadly the typing of the M-defining site 10400 failed for all samples, so we are a bit amiss. Another anomalous site is 12372: while analysis for this locus failed in the case of nos. 3, 9a and 9b, it was typed as A for the resident of grave 8. I could only find this transition among modern samples (out of N) in M12 and M7c1d. However I tried to track the haplogroup using the HVS markers and I failed miserably (it seems not to be L0 however what is not much help).

So the case for the correct adscription of the mtDNA haplotypes, at least in the four aforementioned cases, is extremely confusing and may well be totally wrong. All I can say is that I doubt that they are U2e, K1, U5a1 or T1a as reported because all these are N subclades and the data says L(xN) quite clearly. 

Bell Beaker?

I have been searching for this Kromsdorf site oline and could find almost nothing, sadly enough. However if I know anything about the Bell Beaker phenomenon is that:
  • It began just slightly before these burials, c. 2700 BCE. However it is true that the mainstream theories suggest a very nearby origin in Bohemia. 
  • They used a very standardized set of grave goods which consisted in most cases of: bell beaker, bone ornaments, V-pierced buttons, moon-shaped collars, flint arrow points, archer's bracelet, triangular copper knife and gold spiral (money equivalent possibly).
  • They used to bury their men on their left side and their women on the right side (not 100% predictable but most common), while Corded Ware did the opposite. Burials in fetal position are typical of the region since Neolithic times (this does not change). 

Demographic chronologies of Central and Northern Europe
with Bell Beaker and Corded Ware time-frames annotated.
Lowest graph is Great Britain.
(From a previous entry at Leherensuge).

It's clear to me that not a single burial of those has the grave goods typical of the Bell Beaker phenomenon. Not even the cups are unmistakably described as bell-shaped beakers but then again no arrows, no gold, no knife, no archer's bracelet...

And the chronology overlaps with the Corded Ware period. However the burials of Corded Ware also have standard grave goods, even if less impressive ones: cup, amphora, flint or bone tools and the famous "combat ax" (actually a ritual one) that once described the culture (Corded Ware culture = Combat Ax people in the past, although this name has been dropped nowadays, mostly because it became obvious that the axes were ritual and essentially useless for combat).

So what are they? Can't say but neither one nor the other they seem to me: they are chronologically in the Corded Ware and Bell Beaker time frame but the adscription to either culture seems pretty much not demonstrated to me. Feel free to correct me but here there are some true Bell Beaker grave goods:

Bell Beaker grave goods from Valencian Country (source)
Bell Beaker grave goods from Zamora, Spain (source)

Four Bell Beaker burials from Bohemia (source): 1-fem, 2 male, 3-child, 4-male

The origin of R1b in Europe?

Doubt it. Many are claiming rather happily that these burials represent the origin of R1b in Europe or something like that. But, if anything, these findings are a terminus ante quem (or most recent possible date) and nothing else.

However we do know that some local precursors were of other patrilineages: F*, G2a3 (local Danubian) and R1a1 (from Eulau, right across the Elbe). Al this may tell us something about the Elbe basin but do we dare to extrapolate to all Europe? I wouldn't (but many do without any shame).

Update (May5): Where is the mtDNA H?

A big part of the mystery surrounding ancient DNA is about mtDNA haplogroup H and variants. H today makes up some 40% of Western and Northern European genetic pools, but, excepted Atlantic parts of Iberia (and some arguable cases here and there: Neolithic Alps, Gravettian Russia, Oranian Rif, etc.), it has been reluctant to show up in the research in sufficiently high numbers to account for modern genetic pools.

As I say, Western Iberian and Basque genetic pools appear normalized very early in the Neolithic (or even before), with levels of H above 40%, while in Central Europe getting even 20% is rare and elusive. In this case even such low figures are missing.

This unlike what happens just some journeys downstream the river some 1500 years later at Liechtenstein Cave (Dorste, Low Saxony), when we get a quite modern mtDNA pool with plenty of H and no ultra-rare clades like N1a or U2e. The Y-DNA pool in this case leans heavily towards the regionally important I2a1b but this can be justified on grounds of patrilocality, arguing that the deceased were all close relatives (the haplotypes are actually identical or closely related).

So where is all the H? While this site is more recent than Megalithism in the area, it is also from a non-Megalithic zone. It is possible (but would have to be demonstrated) that genetic pools were already close to normal in Low and High Germany (Megalithic areas) but not in Middle Germany (non-Megalthic), homogenizing only later with Urnfields or something like that. However a lot remains to be clarified and the answer may well be different or more complex.

Update (Nov 8): co-author Johannes Müller has made the paper now available at


  1. Actually, from what I have read, in Germany Bell Beaker is known to have few grave goods except in rare cases, and it has been argued from that that Bell Beaker was a much more stratified society than anything contemporary there, or even for another ~2,000 years to come (e.g., in much of Northern Germany). Also, east of the Rhine, many of the very first Bell Beaker graves had grave goods more similar to Corded ware. So it appears that at later times, they tried to adhere to the western rituals more strictly, or were only then able to do so (could make/ had access to the materials).

    At any rate, they must have offered something that was difficult to refuse by their neighbors and by people who had lived there long before. Perhaps they just had great taverns and made really good beer...

    1. Yes, I can't claim to be any expert on the matter but there seems to be a diffuse overlap between Corded Ware, Bell Beaker and ill-defined cultural types like these tombs (Danubian/TRBK successor commoners?) What I question is the "absolute" adscription of this necropolis to Bell Beaker and not any of the other categories in this cultural blurry zone: why not say just "Late Chalcolithic"? That would be the correct description.

      I'd question that there is Bell Beaker without Bell Beaker cultural standards, seriously. That's why Bell Beaker is a phenomenon or subculture and not a full fledged culture: it is almost nowhere and no-when mainstream but secondary and what is not Bell Beaker is locally rooted (and in this case too: the burials can well be compared with Danubian ones from previous times: only their chronology makes them "something else").

      "At any rate, they must have offered something that was difficult to refuse"...

      Sure: an offer they could not refuse. :D

      Seriously now: it'd be really interesting to know for sure what the BB network was about but I have always imagined them as some sort of traders' guild with a religious component. The BB people were obviously wealthy archers of Neo-Danubian roots (probably Indoeuropeans by language initially, I guess, but Central Europeans above all).

      But these people do not seem to be members of the privileged BB guild: they may be their clients, their slaves or have no relation at all.

      (Btw and for the record: I replied to your comment at Dienekes' but got "lost". Wasn't too important anyhow, I think).

  2. The classifications are correct. Only site no. 10873 gave an erroneous result. The 12705C is indicative of 'R'.

    These sequences are very common in Europe today.

    1. Hmm. You must be right.

      In fact I'm now realizing that the 10873 column is probably an edition error and not the real results: the results of that column are exactly the sames as for the 12075 column, what looks like someone made a mistake in fact.

      Thanks for the remark: it does help.

  3. I noticed that but think it may be due to some other reason. The same was reported for Paglicci-12:

    "Paglicci-12 shows the motifs 00073G, 10873C, 10238T, and AACC between nucleotide positions 10397 and 10400, which allows the classification of this sequence into the macrohaplogroup N , containing haplogroups W, X, I, N1a, N1b, N1c, and N*." (Caramelli, 2003)

    1. One possibility: There are C's on both sides of site no. 10873 in the rCRS:


    2. 10873C is L(xN), 10873T is N. That's it. The site does not appear again in all PhyloTree (and that's my "bible" on the matter unless clearly proven otherwise).

      The Paglicci-12 sequence hence looks like L(xM,N) because the defining marker for M 10400T is also in the ancestral state "C".

      Most of the other mentioned markers say little:
      73G, 10238T, 10397A and 10399C are all RSRS.

      However 10398A is not RSRS (would be G). The sequences showing this transition are:
      · L1c1a1
      · L3e1a3
      · N (but N must be 10873T).

      So most probably L3(xM,N) in fact.

  4. Can the individual mtdna types be associated with localities in any way that might suggest a migration path (albeit without a known direction)?

    1. I don't know. I don't think I'd be able to make such analysis but the kind of FTDNA and the such do manage the kind of data (lots of individual haplotypes with, often, known location) that would allow them to at least reply to your answer.

      With my limited knowledge I'd say that only up to a point, that the same haplotype is often found across half Europe but that, in case of rare haplotypes, it may be possible - I guess.

      MtDNA is generally much less precise than Y-DNA because the DNA sequence is much much shorter and therefore mutations only happen every many generations. In Y-DNA you have several mutations each generation but the problem is the inverse: sequencing the whole chromosome is extremely costly and impractical.

      Anyhow, something I am pondering right now is that the gene pool as a whole does not look very modern: H nowadays a most common lineage (c. 40% in all West and Northern Europe) and not a single individual had it in this necropolis. So overall the myDNA pool looks strange.

      This is much unlikely the Urnfields era site of Liechtenstein Cave, near Dorste, just some 100Km or so downstream the Elbe but 1600 years later, which are almost perfectly modern in all aspects of their genetic pool.

      But it may be a fluke or, again, a very local issue. It's not like the Elbe basin has been a quiet place through history, right? Neither through prehistory, it seems. But for some reason lots of aDNA data is coming from there.

    2. I, pretty much, agree with what Maju wrote. Direction of movement is often quite difficult to determine in mtDNA studies because we are working with identical or virtually identical sequences. That did not stop me from trying, however. :-) As I was determining the distribution, I noticed that half of the sequences were, coincidentally?, also reported in an ancient Kurgan sample (url below). I doubt that any of this helps but since I've gathered the information...

      U2e - 16051G 16092T 16129C? 16182C 16183C 16189C 16362C 73G 152C 217C 263G (ignore 16182 & 16183) Bell Beaker, Germany Grave 3 Lee 2012
      Distribution: Ireland, Finland, Germany, France, Hungary, Estonia
      Outside Europe: Tibet (Chamdo), Xinjiang (Uygur), Kurgan 1800-1400 BC (S10), India

      W5a -16223T 16292T 16362C 73G 189G 194T 195C 204C 207A 263G Bell Beaker, Germany Grave 4a Lee 2012
      Distribution: Portugal, Catalonia, Germany, Poland

      I1 - 16129A 16172C 16223T 16311C 16391A 73G 199C 203A 204C 250C 263G Bell Beaker, Germany Grave 5 Lee 2012
      Distribution: Shetland Isles, Scotland W. I., Ireland, Iceland, Britian, France, Germany, Poland, Bosnia, N. Greece, Cress Islander, Estonia,
      Vlax Roma, Hungarian Roma, Transylvania
      Outside Europe: Mari, Mordvinian, Telenghit

      K1 16093C 16224C 16311C 16319A 73G 152C 195C 263G Bell Beaker, Germany Grave 8 Lee 2012
      Distribution: Scotland W. I., Spain (Roma), Andalusia, Transylvania
      Outside Europe: Kurgan 1800-1400 BC (S15), Kuala Lumper

      U5a1 16256T 16270T 16399G 73G 263G Bell Beaker, Germany Grave 9a Lee 2012
      Distribution: Scotland W. I., Karelia, Viking AD 700-1100, France, Poland, Hungary (& Transylvania), Estonia, N. Greece, Russia,
      Outside Europe: Anatolia, Xinjiang (Kazak), India (Kanet)

      T1a -16126C 16163G 16186T 16189C 16294T 73G 152C 195C 263G Bell Beaker, Germany Grave 9b Lee 2012
      Distribution: Shetland Isles, Iceland, Orkney, Ireland, British, Wales, Cornwall, Finland, Karelian, E Norway, Portugal Andalusia, Spain, France, Germany, Bavaria, Austria, Poland, N. Greece, Estonia, Cress Islander, Ingrians, Seto, Hungarian Roma, Georgia, Belarus, Russia, Ashkenazi Jew, Vlax Roma, Transylvania,

      Outside Europe: Mansi, Komi, Khanty, Altai, Telenghit, Mongolian, Tibet, China (Sichuan), Iraq, Armenia, Ossetes, Syria, Algeria, Berber, Iran, Adygeis, Adygeis, Tarim Basin 2500-2000 BP, Kurgan (S09), Kazakhstan 14th-11th c. BC, India

      Ancient DNA provides new insights into the history of south Siberian Kurgan people
      Keyser et al. (2009)
      (See Table 5)

      The following website appears to contain all of the ancient western Eurasian ancient mtDNAs retrieved thus far (including the Bell Beaker):

  5. I left this comment over at the Harappa Project, on the new DIYHarappaWorld:

    I just ran my data - I'm 100% Irish from Ireland, and here's what I got:
    0.00% S-Indian
    11.10% Baloch
    5.81% Caucasian
    50.58% NE-Euro
    0.08% SE-Asian
    0.03% Siberian
    0.00% NE-Asian
    0.00% Papuan
    0.61% American
    0.00% Beringian
    31.64% Mediterranean
    0.02% SW-Asian
    0.11% San
    0.02% E-African
    0.00% Pygmy
    0.00% W-African

    I'm mtDNA T1a1 and I've been looking into this haplogroup for a while, and note that:
    Top 10 locations for T1 by highest frequency:

    1. Romania - 8.51%
    2. Parsis (Pakistan) - 6.80%
    3. Bulgaria - 6.38%
    4. Portugal (Northern) - 6.38%
    5. Azerbaijan - 6.25%
    6. Armenia - 5.76%
    7. Brahui (SW Pakistan) - 5.30%
    8. Mazandarian (N Iran) - 4.8%
    9. Macedonia - 4.5%
    10. Pathan (NW Pakistan) - 4.5%

    Here's what Doug McDonald predicted:
    Most likely fit is 97.5% (+- 0.5%) Europe (all Western Europe)
    and 2.5% (+- 0.5%) S. Asia (various subcontinents)
    The following are possible population sets and their fractions,
    most likely at the top
    Basque= 0.205 Irish= 0.765 Sindhi= 0.031
    Basque= 0.207 Irish= 0.758 Pathan= 0.035
    Basque= 0.201 Irish= 0.775 N_India= 0.025
    Basque= 0.198 Irish= 0.781 S_India= 0.021
    Basque= 0.080 English= 0.901 S_India= 0.019
    Basque= 0.084 English= 0.895 N_India= 0.021
    Basque= 0.088 English= 0.891 Sindhi= 0.021

    So here's my prediction, that mtDNA T1 is somehow involved in the spread of the Baloch component, as it is found in a path starting in Mehrgarh - - and going into Southern Afghanistan, across Northern Iran, and into the Caucasus, and from there to Europe. IMO it spread both into South East Europe by land mostly, but also by boat to South West Iberia, and then kicked off the Bell Beaker revolution. I now wonder if R1b also followed this path?!

    1. This is a bit labyrinthic for me, Conroy. I have low trust in comparison with zombies, much less when these have been designed to measure rather remote regions like South Asia (Harappa) or the Aegean (Dodecad). It might be interesting to you to develop your own set of zombies for Western Europe: that would be informative, I guess.

      I do not understand well how McDonalds arrived to the figures below but the first ones make total sense considering what I know of your ancestry (Irish with some Gascon, right?)

      In any case there is no objective "Baloch" population that actually spread: it is just an approximation developed by someone (Zach) for some reason (analyzing South Asians). However it may reflect some Neolithic flows - unsure.

    2. Maju,

      My ancestry on my father's side is 100% Native Irish AFAIK, on my mother's side it's mostly Native Irish, with is a small amount of Huguenot, Northern English and Norman ancestry. My last non-Irish ancestor came to Ireland in the early 1700's.

  6. Maju and pconroy,

    I guess you've seen the Maria Pala paper?

    This has mtDNA J and T expanding into Europe from 19,000 to 12,000 ya.

    Have a look at the countries of association for mtDNA T that you mention, above. Most are dominated by pastoral economies, at least up to the recent past.

    I think it is a little premature to associate this pastoralism with y chromosome R, but that's my bet.

    However, I think that more good papers like this, with rigorous DNA analysis combined with rigorous archaeological work, are needed to make these inferences.

    All the early y chromosome R DNA in Europe is probably up in the hills with the sheep. (Buried with their bagpipes, of course. (Just joking.))

    1. I've only read the abstract (it's PPV for six months). Anyhow it looks like is just another paper about molecular clock estimates, what for me mean little to nothing.

      AFAIK no mtDNA T, much less T1, has ever been reported in a pre-Neolithic European. It's just lack of evidence but at the very least suggestive. Same for J. However I've seen a HVS-I sequence that looks totally JT* in Nerja and the same rare paragroup (modernly found only in Italy and North Africa, AFAIK) was reported by Kéfi in Morocco, although I could not confirm it based on his sequences.

      But no T nor J in any case: they still look Neolithic arrivals to me - but I have to read the paper, of course.

    2. Marnie,
      Oh I totally agree that R1b seems to track pastoralism.

      Here is my comment from Nov 2011 on the issue:

      Hi I'd like to post a comment on - see full text below:

      In term of cattle in Western Europe, the Kerry cattle breed is a fairly
      unique breed. Kerry is in the extreme South West of the country.

      I read decades ago that the Kerry breed has some similarities with cattle
      in Nigeria. Reading Sam's comments prompted me to look into this more, and
      here's what I found:

      Title: "Shorthorn cattle of West and Central Africa I. Origin,
      distribution, classification and population statistics"


      The African cattle population is a result of three major introductions from
      centres of domestication in Asia (Epstein, 1957; Faulkner and Epstein,
      1957; Payne, 1970; Williamson and Payne, 1977; Oliver, 1983), which mostly
      followed the Nile Valley through Egypt or came through the Horn of Africa
      (Figure 1). Further migrations resulted in heavy concentrations of cattle
      in the East African Highlands, present-day Ethiopia and Kenya (Payne,
      1970). Humpless Hamitic Longhorns arrived about 5000 BC. They were followed
      by the humpless Shorthorns about 2500 BC and the humped zebu, first around
      1500 BC, then in large numbers around 670 AD.

      Only a few publications exist on the Shorthorns of West Africa, for
      example, Jeffreys (1953) and Ferguson (1967). Although they are currently
      termed trypanotolerant, humpless Shorthorn cattle were historically
      distributed in almost all ecological zones. Not much is known about the
      actual movements of Shorthorn-type cattle, however. Despite their wide
      distribution, humpless Shorthorns are not represented in rock paintings to
      the same degree as are Longhorns.

      It is speculated that increasing aridity at the time of their introduction
      may have made it impossible for the Shorthorns to follow the southwestward
      migration route across the Sahara. This may explain the route taken
      westward along the North African Mediterranean coast. Modern cattle in
      Egypt exhibit characteristics that resemble those of Shorthorn-type cattle,
      while existing breeds in the Libyan Arab Jamahiriya, Tunisia, Algeria and
      Morocco are almost entirely of the Shorthorn type (Payne, 1970; Epstein,
      1971). It is highly likely that, like the Longhorns before them, the
      Shorthorns split into two routes around Morocco. One stream moved north
      into present-day France and the British Isles - the Jersey, Guernsey and
      Kerry breeds are partly derived-from this stock - and the other moved south
      and westward, eventually coming into the rain-forest zone where, through
      exposure to trypanosomiasis, they became tolerant.
      So the suggestion seems to be that pastoralists with Bos Taurus type cattle
      spread out from the Middle East/West Asia - in Irish Gaelic Bull=Tarbh
      (pronounced TOR-OBH) - and maybe from just West of the Taurus mountain
      region, spreading R1b haplotypes with them.

      Here's a map of possible migration routes:

      Paul Conroy

    3. Thanks, Paul! I didn't know that the Kerry breed were thought to be related to African short horn cattle. They are thought to have been domesticated in the Ethiopian highlands, as you mention. And yes, they are thought to have been domesticated after longhorns, late enough that they missed the Wet Phase and ended up being more concentrated in North Africa. (Roger Blench)

      I'll make a point of checking and posting your material, above. (Thanks!)

    4. Maju,

      I had a hard look at the supplemental material of the Maria Pala paper, which is available. I have to say, they really went all out on their maps and trees. Really beautiful.

      From this material, after studying it for a while, there seem to be several trends:

      In Europe, J and T seem to be concentrated in the Alps, the Balkans and the UK. (I'm not sure why there is no ancient DNA for J and T, but as you have previously pointed out, it might be due to different burial customs.)

      Overall, the widely distributed geographical concentrations of the branches of J and T speak for a pre Neolithic dispersal into Europem, North Africa and points East.

      From the maps, you can see the dispersal into Egypt and Chad, as you would expect if branches of J and T followed R-V88 into Chad. (R-V88 coalescence is estimated to have occured in the early mid Holocene.)

      The concentration of J and T in the Alps is very interesting, as the Alps seem to be at the nexus of a dispersal for R1b also. (see Myers).

      So, no I don't think it is just another paper on molecular clock estimates.

      PS. The appearance of J and T in Northern Spain could still be due to a Neolithic migration. The maps hint at several candidates for the origin of this Neolithic migration (if it was a Neolithic migration, rather than one in the pre-Holocene.)

    5. In truth the supplementary material of Pala is very interesting, regardless of age estimates. I certainly fail to see T1 as "European-specific" at any level (all subclades are full of West Asian branches) and the age estimates may well be again a mirage.

      T2 on the other hand has branches that appear as essentially or even totally European (T2b, T2f1 and several other sublineages). The same can be said of subclades of J1 and J2 but they might well have expanded in the Neolithic for what I know.

      What you say about the Alps may well be a mirage, reflecting the broader Central European region, specially the Upper and Mid Danube basin but also the Rhine, etc. If so, there are a number of unrelated prehistoric processes that follow that pattern: (1) Aurignacian, (2) Gravettian, (3) Danubian Neolithic and (4) Western Indoeuropean expansion (specially Celto-Italics). But there's no obvious Late UP nor Epipaleolithic process which may serve as explanation.

      What I see in the maps and networks/trees is:

      J1: West Asian originated ·> J1c mostly European star-like clade of Neolithic spread. Focus in Balkan-Danubian Neolithic with spillover to Po basin, Ukraine and Vasconia.

      J2: Most diverse and common in Egypt (good to know) ·> scattered founder effects in Europe, NW Africa, West and South Asia. There is a particularly notable founder effect of J2b1a in Galicia.

      The chronology they propose is Magdalenian or later (after 16 Ka) but that does not fit well in principle with archaeology, so again a Neolithic time frame seems more likely. However considering that the African and the East European number of basal branches are exactly the same, I would not totally exclude an initial Solutrean-Oranian founder effect.

      T1: Clearly West Asian with scatter in Europe that looks totally Neolithic or even post-Neolithic (and mostly T1a). The main center in Europe seems to be in Romania.

      T2: A star-like structure with smaller star-like structures in the branches, specially T2b. It looks totally Neolithic. Most relevant subclades:

      ·> T2b: Central European Danubian surely, with spillover to some other regions as mentioned above (Padania, Vasconia and Ukraine specially).
      ·> T2f: Looks like scattered founder effects in Germany/France and Ukraine maybe in the same process as its sister. Or may also be of Indoeuropean scatter, if we look at the lone South Asian individual located at the central node - hard to judge.
      ·> T2e: a naive sloppy observer could argue that it is an Icelander clade. It is but, of course, only as end-of-route founder effect. Otherwise it could be related to Cardium Pottery (in Europe) if we have to judge from its concentration in Albania, Padania and Iberia.

    6. Maju,

      Look a little closer at J.

      Take J1c, for instance:

      J1c1 is almost entirely European (coalescence 11.1 plus or minus 1.9 kya), while

      J1c2 is distributed in Europe, the Southern Caucasus and the Near East (9.5 plus or minus 1.4),

      J1c3 is almost all European (11.1 plus or minus 2.5)

      J1c7 is again almost all European (13.6 plus or minus 3.7)

      Interestingly, J1c12 is mostly Near East and Caucasus, with a slightly earlier coalescence date (15.3 plus or minus 2.1)

      At least for J1c, the data speaks to an expansion into Europe before 8.5kya, even when estimating the error with the latest arrival date.

      Clearly, the data speak to the need to consider subclades and sub-subclades.

  7. Yes, J1c is clearly European and it might even be pre-Neolithic. Cardoso 2011 argued J1c to be of Franco-Cantabrian origin... BUT in this paper the West European diversity of J1c is low compared with that of the Near East and even Eastern Europe (v. tale S5) so not really sure right now.

    This is the basis of my analysis: age estimates should be just discarded for the analysis. They may serve as interesting exploration if you understand the how of the estimates (and therefore the pros and cons of a system that is bound to be a very rough approximation at the best) but otherwise they are just confusing noise.

    Also, for whatever is worth, something I find quite suspicious in this chronology is that the JT node (est. 59 Ka ago) is older than the known colonization of West Eurasia (55 Ka ago at the earliest), regardless that the lineage is more distant from the ancestral R node than others like R0: beginning at the N node (coding region transitions only):

    ·T12705C·A11719G·> R0 (incl. HV)
    ·G1719A·T10238C·G12501A·> N1 (incl. I)
    ·G709A·G5046A·C11674T·T12414C·> N2 (incl. W)
    ·T6221C·C6371T·A13966G ·T14470C·> X
    ·T12705C·A11467G·A12308G·G12372A·> U
    ·T12705C··T4216C·A11251G·C15452a·> JT

    From this I get the strong impression that the first colonist lineage was R0, then N1 and that the other four (N2, X, U and JT) only expanded (first bifurcation: haplogroup defining node) at a second moment. I can't know with any certitude when was that but IMO some 5-10 Ka after R0, certainly not in the very first moment, much less before arrival to West Asia as Pala et al. defend.

    Then you have chronological leaps in their model that I make no sense of, for example between JT and J there are two coding region mutations but they make the time lapse to be more than 15,000 years (IMO 5-10 Ka for that tic-tac, not more) but later they make mutations accumulate at faster rhythm. I do not understand (nor in principle can share) that without due careful explanation (maybe is my fault for being ignorant of something - you tell me).

    And I fail to see it explicit in the paper (I was sent a copy today), although they may use this "calculator" designed by Soares:

    So IMO assuming 2.5 Ka per coding region mutation and a calibration of JT c. 40 Ka. And then:
    · J c. 32.5 Ka
    · J1 c. 30 Ka
    · J1b c. 27.5 Ka
    · J1c idem
    · J1d c. 25 Ka
    · J2 c. 27.5 Ka
    · J2a c. 23.5 Ka
    · J2b idem
    · T c. 17.5 Ka
    · T1 c. 15 Ka
    · T2 c. 12.5 Ka

    But of course these figures are just approximations because we can't control nor accurately evaluate the randomness behind these mutations, who are not enough in number to effectively nullify the randomness.

    To put an example, I just posted at my other blog a video of an artistic performance in which two guitars are played by the disintegration of atomic nuclei, which have a known half-life (statistic value) but is in fact random, just like our mtDNA mutations. If you listen (which is curious at the very least) you'll notice that there are bouts of hyperactivity and lapses of silence and that a disintegration rate can only be understood as an average, not a regular tic-toc.

    Similarly DNA mutations are random and, unless LOTS of them are evaluated in order to near-nullify the randomness, we can't draw conclusions from the "molecular clock", whose tic-tac is irregular, random, inherently chaotic.

    1. Maju! Don't tell me you are going to meander down the twisted path of over generalization!

      First, regarding the guitar video, which I did listen to, it can be explained entirely by the phenomena of resonances, beats, cross coupling and loss. (That's my day job, so watch out!)

      I'm not going to get into a long discussion about molecular dates. Whether J and T arrived during the late ice age or in the very early Neolithic seems immaterial to me. However, I am quite sure, based on the distribution, that J and T were pastoralists, not farmers.

      Also, I had another look at Haak:

      J and T are in the mix at Derenburg with H, HV, I and W even 6000 kya. It is telling that Derenburg (early compared to the other LBK sites Haak looked at) has a level of J and T that is higher (and more strongly associated with the Taurus-Zagros arc) than the later LBK sites, indicating that the pastoralists remained separated from the "farmers" for a significant period of time.

    2. Whatever the exact mechanisms used to transmit the atoms' activity to the guitars' strings (something I did not bother checking), it's clear that when the atoms decay, they create sounds and when they do not, then it's silence.

      Or if you prefer a more classical example: it's a case of "Schrodinger's cat" once and again. However in each case we know the result but we do not know when exactly it happened: neither when the rhythm was randomly accelerated nor when the rhythm was randomly stopped altogether.

      I'm not just talking of random novel mutations but of random drift out or random fixation or random minority survival of these. All that is unmeasurable. Believe me: I've tried and it's pure chaos and there are no valid formulas, just generalization and simplifications which are bound to fail unless you gather lines of thousands or at least hundreds of mutations to neutralize randomness by mutual cancellation of errors.

      These lines never have more than 15 mutations from the top to the bottom. In many cases they have five or so.

      Let's be serious about the limits of molecular-clock-o-logy: it's a curious exercise but its reliability is too close to zero.

      "... J and T were pastoralists, not farmers".

      In Europe there's no such difference: mostly not. Why? Because there are not many marginal steppe/semi-desert lands which can only be exploited by semi-nomadic herders as in parts of Asia and Africa.

      Even where you have semi-steppe like Ukraine or Castile, those lands happen to be excellent also for the cultivation of cereals. Only in mountain areas herding is more central but still every valley is and has always been farmed as much as possible.

      I do not think that there is archaeological support for the existence of "pastoralist" peoples in Europe, although pastoralism has always been hand-by-hand with agriculture and in some very specific moments and areas it has gained some greater role.

      Also the distribution of T and J does not correspond with such ideas: when there is a European branch it's always important in Central Europe and/or North Italy: mostly farming areas (although with some cattle) according to the archaeological record and key passages for the Neolithic dispersal.

      "J and T are in the mix at Derenburg".

      J and T show up in Europe with Neolithic, including Derenburg. Other sites in the early Neolithic are: Seehausen, Flomborn, Halberstadt, Schwetzingen, Vedrovice, Cova Avellaner, Los Cascajos, etc.

      In every other early Neolithic site we see one or the other or both clades. Derenburg is just oversampled (and hyped it seems) - plus they were not "pastoralists" but an ample spectrum West Danubian people. They grew two types of wheat, barley, green peas, lentils, flax and opium poppies. They had cattle but also some sheep, goats, pigs and the occasional dog.

      But, as I say, almost at the same time the T lineage shows up in Catalonia in a Cardium Pottery context and the J lineage in Navarre soon after. They did not come from Central Europe for sure but may have arrived from the West Balcans via Italy instead in the context of Cardium Pottery. This means even less importance of cattle, although the "Neolithic package" is very similar overall (just the emphasis changes according to local needs). The main exception could be that these people were also able seamen, something that logically was not part of the culture of inlander Central Europeans.

      "... that the pastoralists remained separated from the "farmers" for a significant period of time."

      Not at all: get your archaeology straight please.

  8. Marnie,

    Where do you see these pastoralists? I could understand high in the mountains in the Balkans, and in the Alps, with sheep and goats, and it seems that a portion of La Huguette was a hint more oriented such with cattle. But other than that, pastoralism is not documented in most of neolithic Europe (excluding the East European steppes, much later). Note that in much of Europe you need stables for the animals and their food in the winter, so such sites would have left traces if they existed.

    I can definitely see some JT subclades entering Europe just before LGM or just after, but I would suspect them to be mostly limited to Greece and the Balkans, and mostly due to diffusion rather than large-scale migrations. Now, as I have stated elsewhere, the tricky part is to distinguish JT subclades brought into Europe with the neolithic, and those already present in SE Europe and being picked up in the neolithic wave (and we know the agriculturalists picked up pretty much anything they found on their way). In the end, we really are at a point where we need very high resolution or full genome ancient mtDNA to make proper comparisons.

    1. "... pastoralism is not documented in most of neolithic Europe (excluding the East European steppes, much later)".

      On the contrary: it is well documented but always in association with agriculture, AFAIK. Bovine, ovi-caprine and porcine cattle are known to have existed since the early European Neolithic of Thessaly (and before it in West Asia). I could even agree that pastoralist practices could even gain some weight as the Neolithic economy moved northwards (Starcevo, still in the Balcans, already shows some such trend) but there's never a decoupling of herding and farming as Marnie claims: the economy was all the time broad spectrum: cereals, legumes, flax and diverse cattle types (an in the West Danubian case also opium, and in Cardium Pottery also fishing and olives and rabbits in Spain since very early - although it's unclear if they were "domestic").

      "Note that in much of Europe you need stables for the animals and their food in the winter, so such sites would have left traces if they existed".

      I do not think that claim stands. In many cultures people shared the homes with animals (Bretons, for example, lived with their pigs until recently, yeah!) or maybe the cattle types were sturdy enough to withstand the winter on their own (maybe the farmers kept some fodder and barley reserves but not proper stables) or maybe we miss part of the picture (wood and adobe structures are quite light and may be multi-purpose).

      In any case, there was indeed cattle of all four types - but always along with agriculture.

    2. Maju,

      Obviously, responding to Marnie, I meant pure pastoralism -- as in: animal herding without other aspects of the rather broad typical European neolithic package.

      As to the other point, even if people were able to live in tents (like in parts of the East European steppes), animals were not in most of Europe, and significant storage was required for winter fodder. Thus the requirement for structures - and the only thing we see in the initial neolithic in much of Europe is the typical long houses (that likely included quarters for animals), or stone houses in parts of the south. My main point is that all such structures are associated with grains. Pure pastoralism would have necessarily left similar structures and settlements not associated with grains - and these simply do not exist in the archaeological record.

    3. Alright. Understood: we are in agreement then.

    4. Without going into a lot of detail, I'll say that there is some evidence that animals such as sheep and goats hug the margins of glaciers. There's water there and also a higher concentration of minerals to make the grass they would feed on more norishing.

      Following the retreat of the glaciers in Europe, there would have been many of these ice margin eco zones for both large mammals and humans to move into.

      For these ice margin pastoralists, their economy probably looked different than pastoralists of Europe today: the animals they followed were likely in a state of quasi domestication.

      As more domesticated breeds became available, these probably replaced the wilder ones.

    5. Marnie: the Neolithic is much later than the retreat of glaciers in Europe. The Ice Age ended before 10 Ka ago, long before the Neolithic, and did so quite sharply, as you can observe in this graph.

      Sure that there was a less warm period (similar to recent temperatures, which were lower than the Neolithic "climate optimum") known as the Boreal Age but the glaciers had already retreated.

      Glacier margins would only be found on mountains like the Alps and would be quite marginal economies in any case.

      I really do not make sense of your hypothesis.

    6. Here's a good source for understanding how, where and when the glaciers in Europe retreated:

      In Western Europe, between 15,000 and 11,000 years ago, glacier margins would have been found in the British Isles, the Pyrenees, the Alps, Iceland and Germany.

      "would be quite marginal economies"



      Turner, M. D. Zeller, E. J., Dreschhoff, G.A., and Turner, J. C. (1999). Impact of ice-related plant nutrients on glacial margin environments. In Ice Age People of North America: Environments, Origins, and Adaptation, ed. R. Bonnischen, and K.L. Turnmire. Corvallis, OR: Oregon State University Press and the Center for the Study of the First Americans, pp. 42-77.

    7. But Marnie: 11,000 years ago there was NO Neolithic in Europe. For Western and Central Europe the date is not before 7500 BCE and at that time there were only glaciers on the highest parts of the Alps and, sure, much of Norway.

      British Neolithic did not begin until much later: c. 4000 BCE.

      However the animation is quite cool. Thanks for the find. :D

    8. Maju, I'm not sure what we're discussing here now. (That's the problem with email and blog discussions!) I didn't say there was a Neolithic in Europe 11,000 years ago!!

      I'm saying that mtDNA J and T appear to have some sort of pre-Neolithic pastoral/glacial margin/herd following affinity, based on the geographic positions and coalescence dates in the Pala paper.

      Anyway, I think we've beat this one to death!

      I came across a great paper on Basque rock paintings:

      Les Ihizi: et si un mythe basque remontait à la préhistoire?

      If you don't have it or need a translation, I can work on that.

    9. Paleolithic "pastoralism"?

      "If you don't have it or need a translation, I can work on that".

      I can generally read French fair enough but I do not know the paper (sure: feel free to send it to me, even if just out of curiosity). I do not know of any "ihizi" Basque myth, which sounds strange and suspicious.

  9. Can someone send me a copy of the Pala paper?


Please, be reasonably respectful when making comments. I do not tolerate in particular sexism, racism nor homophobia. Personal attacks, manipulation and trolling are also very much unwelcome here.The author reserves the right to delete any abusive comment.

Preliminary comment moderation is... ON (your comment may take some time, maybe days or weeks to appear).