Monday, November 10, 2014

"Apparent Cline" vs. "Real Cline"

To many, it is evident or probable that the paternal transformation of Western Europe is owed to the Bell Beakers.  There's not much room left for alternate candidates at this point given the growing stack of genetic studies from either side of the horizon.

It is comfortably demonstrable that the material culture of the Beakers emerged from Western Iberia.  It's not a global synthesis of a "bunch-o-cultures" and it's not a product of an archaeologically non-existent "unicorn culture".  Proper Bell Beaker culture congealed from a slurry of Western Iberian and North African influences.

However, when it comes to R1b, Beakers and Iberia, a lot of people have been getting mired into a seeming paradox (basically this map)>

Expansion of R1b (Michael Hammer)

Here's the fundamental problem with a Danubian R1b expansion (and this holds for the entire R1 family):  It is subtly predicated on the demic diffusion rate of a farmer's descendants as they sprawl at a rate of 1 km a year.  (Ammerman, Cavalli-Sforza, 1972, also 1984) or again (Pinhasi, Fort, Ammerman, 2005)

At this speed, sufficient time is allowed for branching in which founder twigs have time to move away from the base so that a cline is perceivable.  But we know R1b was absolutely not relevant in Europe during the Neolithic and we also know that the Bell Beaker and Corded Ware folks didn't expand at the Neolithic rate of 1 km a year.

Instead, their expansion rate was very rapid and absolutely involved folk movements (again, we are not yet swimming in genetic studies from pre-historic Europe, but at this point it is much more likely that they were exchanging spit rather than attractive pottery and new ideas.)

So what are we seeing in the map above?  Well, M269 is probably very old in the Middle East and its immediate periphery.  That's about it.

For argument purposes, let's say that the above picture accurately depicts a bona fide, demic cline of R1b movement from the Middle East through Europe.  What post-Neolithic phenomenon would explain the picture above?  Hold that thought...

Before you imagine metal-age, sword wielding barbarians riding chariots, carefully consider a diffusion rate that would give us a perceivable cline.  (in other words, takes-a-lot-of-time-for-mutations-to-occur-so-we-can-see-it-on-a-map)

The movements of R1a and R1b are going to be tough coconuts to crack.  Not only did these people move very rapidly, they are both characterized by regional super-lineages which are deceptive to their structure. 

It boggles my mind that there haven't been more testing at this point on Beaker cemeteries!


  1. There's something strange going on. Word is that the Bell Beakers from Brandt et al. have now had their genome-wide DNA tested and they do carry that eastern stuff related to Mal'ta boy (MA-1), although not as much as the Corded Ware samples.

    It's difficult to imagine that this influence radiated out of Portugal at any stage, so it was most likely picked up by the Bell Beakers in Central Europe. The question is, what else did they pick up in Central Europe?

    In fact, if we look at the mtDNA from the Kromsdorf site, it looks very Eastern European overall, and could easily come from a Corded Ware or Unetice site, rather than an early Bell Beaker one.

    So I don't know what to think anymore. But the new Lazaridis/Reich paper on the ancient genome-wide data from eastern Germany and the Samara Valley is likely to be an eye opener. I've been told that it might be posted as bioRxiv within a few weeks.

    1. But then the question is where did Samara culture originate and its relationship with cultures of the Northern Zargos/Southern Caucasus?

      I know the majority opinion is that R1a and R1b reflect discrete populations separated by time and space, but I have a hunch they were a single population in the Northern Middle East and 'star-bursted' around 5,000-4,500 B.C. to the various steppes. I think lithics may be one diagnostic indicator. Maybe ANE is another indicator.

      I think North Africa is completely in everyone's blind spot. Do you know what the ANE is for White Berbers or Chadic speaking populations by any chance?

    2. To clarify, I don't suggest a splitting of various clades around 5000 B.C., I understand the age of the R1 split is older. Just a split of people, a sort of founder effect.

    3. I'm pretty sure there's no ANE at all in Northwest Africa, and it's at a very low levels in Tunisia and Egypt.

      In the northern Near East it reaches around 13% among ethnic groups that show relatively a lot of R1, like Armenians. But it's generally at low levels even there, and then suddenly peaks in the Northeast Caucasus at almost 30%.

      I suspect part of the problem is that the ANE in the Caucasus and the Near East doesn't have much to do with the early Indo-Europeans, but is rather linked to the southward expansion of the Kura-Araxes culture. This is also perhaps where a lot of the Near Eastern R1b comes from.

      I suspect the proto-Yamnaya people carried a lot of the recently described EHG (Eastern Hunter-Gatherer) ancestry, which was in large part ANE, but also significantly WHG, and their descendants mostly spread into Europe, Central Asia and Siberia along the steppe. They probably mixed with the farmer groups, but perhaps mostly the farm girls, near the steppe quite a bit though, hence the high frequencies of the Neolithic-like mtDNA among the Yamnaya, Catacomb and Corded Ware samples tested to date.

      The Y-hg R has to come from up north, in any case. I feel that an origin in Siberia or on the steppe, then migration to the Near East, and then a back migration onto the steppe where EHG peaks, doesn't make for a very parsimonious hypothesis.

    4. I'd agree that R1 came from Siberia or Central Asia, though I'm somewhat skeptical that it was north of the Black Sea before 5,000 B.C. or before the Samara Culture. Anyhow, any DNA is a good day.

      I look forward to see what comes out of Samara and East Germany.

  2. I don't think you need a cline to create clades when you have channels caused by physical geography.

    If the source region was Ukraine imagine the flow as like water then where would it go?

    One channel would be north of the Carpathians and another would be Danubian (and possibly a third sailing over the Black Sea or down the east coast of the Black Sea).

    So move that P311 circle up above Crimea and P312 southwest of that with U106 taking the north of the Carpathians route to Germany and U152 taking the Danube route up to the source of the Danube (and the source of central European Celtic).

    From the source of the Danube the U152 circle (assuming LBK was initially a barrier) branched into northern Italy and southern France (Z195) and via a maritime route to Iberia (SRY2627) from southern France.

    Eventually the U152 source turns pushes west over the top of LBK into Britain (L21) and Ireland (M22)

    (Although personally I think Britain and Ireland had both a southern maritime BB flow from Iberia, a central European Celtic flow and a northern Germanic flow.)

    Just move the circles to fit the barriers and channels created physical geography and Bob's your uncle.

  3. @Davidski

    "Word is that the Bell Beakers from Brandt et al. have now had their genome-wide DNA tested and they do carry that eastern stuff related to Mal'ta boy (MA-1), although not as much as the Corded Ware samples.

    It's difficult to imagine that this influence radiated out of Portugal at any stage"

    If maritime BB and river BB were two branches of traders and copper workers spreading respectively along the maritime and riverine trade networks from the same source region then I think that is easy to imagine.

    The question then is why would a small group of traders and metal smiths have a sudden massive population expansion once they arrived at the relatively unpopulated Atlantic coast?


  4. After reviewing the background evidence I could find for modern R1b distributions, I actually don't think that the supposed clines support the Anatolia -> Balkans -> C. Europe -> W. Europe route after all. If people were significantly using maritime routes, and I think all of us here would probably agree that they were, then we can ask the question: Is there a place in this standard haplogroup origin map that one of the dots could be moved by a maritime jump without bending things implausibly out of shape? Although my surmise could be instantly invalidated by ancient DNA or much greater detail in modern mapping, it looks to me like there is evidence to contradict the map you cite above, and no real evidence to contradict a Western European (Atlantic?) origin of haplogroup L51/M412 (a node between L23 and P311 on your map). Here's a review of the distributions of W. European R1b clades and their immediate ancestors that I think shows why this is, at least, possible:

    M269-: basically not in Europe
    M269+, L23-: centered in Balkans; also C. Italy, stray Swiss, Polish; Turkey, the South

    Caspian, and far East Europe:

    Note that, according to ISOGG, R1b1a2-M269 is estimated at between 4k and 8kya, which puts the following clades in a reasonable range for Bell Beaker spread:

    L23+, L51/M412- (not on your map): Wide, incl. W. Asia; within Europe, mostly S. Europe but scattered all over:

    All of the above likely have an origin either in the Balkans or points further S. and east.

    The following, from P311 on down, however, suddenly are all rooted in W. Europe--yes, that means that the map placing P311, etc., in the Balkans is, at the least, not inferrable from modern distributions!

    L51/M412+ (not on your map): basal examples in W. Europe + Poland (note Tagus!):

    Including all descendants, the pattern covers W. Europe, with (not very visible) outliers to the East:

    Western Europe seems to be the home for all the deep clades of L51/M412:

    PF7589/Z2118 (not on your map): rare, Britain and Scandinavia and ??? (couldn't find a map)

    P310/P311/L11 (equivalent so far): overall in W. Europe with slight eastern extension:

    Its subclade U106/S21:

    Its other major subclade, P312/S116:

    This subclade has three major subclades, DF27 (ancestor of your map's Z195), U152/S28, and L21:

    These are almost completely confined to W. Europe, although S28 lightly covers a much larger area, quite possibly as a result of the Roman Empire spreading genes from Italy (as well as bringing other genes to Italy).

    In sum, I think that, if we are going to try to use modern distributions as evidence at all, we can throw out the slow trek across Europe version, and think about how L51 jumps to the West. I think the only reason why the model and map that you show are so popular is because people didn't look at the other clades I include here, which are largely newer discoveries. Perhaps it is also a result of wave-of-advance thinking?

    More's the pity for African Bell-Beaker hypotheses, I can't find any evidence for these strands of R1b in N. Africa. Given how crappy coverage is there, this is unsurprising, and proves nothing.

    1. Great comment. The North African issue is plagued by two things. The Central massifs and Western Desert have an effective modern population of zero. These were respectably populated in the Middle Pastoral (theoretically pre-proto-beaker in my view).
      I suspect in time a 'grass fire pattern' will be visible with R1b in North Africa. To a degree this is visible in modern circum-saharan populations, even with M269 although it is dwarfed by V88.
      Also there is extremely limited testing. With the exception of a few low resolution studies of specific population, dna is basically nonexistent.

  5. Correction: L51 (xL11/P310) really does seem to consist only of a clade with z2118+, with a very sparse but wide distribution described here:

    While the distribution is wide, again, everything seems to be WEuropean (the single Hungarian is from the "Saxon" ethnic group, traced to German settlers).

    1. There are two new papers on ydna. One is on Hungarians, the other Slovaks.
      Unfortunately they are pay-per-view