1 members found this post helpful.
I have revised my principal hypothesis regarding the propagation of haplogroup G2a3. Its presence in mountainous areas of Greece, Italy and southern France originally led me to think that it was linked to the diffusion of goat and sheep herding in the Neolithic via the Cardium Pottery culture.
However, three novel elements made me change my mind (or rather think of an additional migration). Firstly, the most common subclade of G2a in Europe is G2a3b1, and this clade is estimated to be only approx. 4500 years old. It is too young for a Neolithic dispersal across Europe.
Secondly, this G2a3b1 has been found in India alongside R1b1b2. If my theory that the Proto-Indo-European speakers originated near the Caucasus is correct, then we are almost bound to find some G2a in places settled by R1b1b. The modern Ossetians and Georgians have very high levels of both haplogroups. I think that the two originally represented different ethnic and linguistic groups (Indo-European vs Caucasian family), but their proximity would have led to some blending of population in the Caucasus region over time.
Thirdly, I realised that G2a3 was also high in northern Portugal, Galicia, Cantabria, Wales, the Alps and Bohemia, and it occurred to me that it was in the same copper- and tin-rich regions that the Indo-European R1b1b2 would have favoured. Brittany, Cornwall and Ireland do not have much G2a3 though, but extremely high levels of R1b1b2 to make up for it.
G2a3 would therefore represent Indo-Europeanised Caucasian people who migrated with R1b1b2 during the Bronze Age. It is possible that G2a3 percentage in western and central Europe remained fairly stable over time, while an originally small ruling elite of R1b1b2 grew exponentially due to their higher birth rate and cultural Indo-European predisposition of favouring of sons.