Brief comments on the "The Genetic Heritage of the Earliest
Settlers Persists in Indian Tribal and Caste Populations" by
Kivisild,Rootsi et al.(Am.J.Hum.Genet. 72:000-000,2003):
The article is important and many of its conclusions are surely
correct. However, some caution is appropriate both in presenting the
genetic data and in interpreting it. I shall confine my remarks on
the Y-chromosome data, in particularly on the haplogroups of the R1
clade, namely R1a and R1b. (as defined by the new Y Chromosome
Consortium Nomenclature, see details at
http://hpgl.stanford.edu/publications/GR_2002_v12_p339-348.pdf
First, the suggestion that is most likely to create stir and
discussion is the proposal by Kivisild, Rootsi et al. that
haplogroup R1a (often known also via the mutation M17) may have
originated in Southern Asia, perhaps even in the Indian
subcontinent. This is a dramatic reversal of the previous mainstream
opinion according to which R1a coalesced either in Central Asia or
in the Ukrainian refugium during the Last Glacial Epoch. As copies
of this mutation are widespread among Indo-European speaking people,
it has often been called an "Indo-European marker". To avert
unnecessary misunderstandings, it is worth reminding here that most
mutations studied by population geneticists in this context do not
even code proteins; hence all considerations abot "race" are
here singularly inappropriate (and it would be nonsensical to
connect particular genes to particular languages anyway). So, all
this talk about "Indo-European markers" is just shorthand for
"genes whose copies are widespread among people who now speak Indo-
European languages". Given the inertia of historical change, the
genetic data gives us, all the same, some very useful information on
possible population movements, both very ancient and some more
recent, too. And we do know that R1a is very prominent in Eastern
Europe, in many areas of Central Asia and in northern areas of the
Indian subcontinent.
Previous studies have found R1a1 in India mainly among the Indo-
European caste populations. Combined with the idea that this
haplogroup may have started to spread from Ukraine, many geneticists
(and others have followed them) have speculated that its spread may
have been speeded by the Kurgan people of Ukraine (often associated
with the Proto-Indo-Europeans) and eventually reached India thanks
to the Vedic Aryans. Estimates about the arrival times have very
large margins, but fairly recent estimates suggested that the
haplogroup R1a has certainly existed in India before the Common Era
but not probably before the Holocene era. In fact, although the most
likely arrival time has been around 3000 BCE, the likely arrival
time of Vedic Aryans, i.e., around 1500 BCE, would still belong
within the margin permitted by the admittedly uncertain estimates.
Kivisild, Rootsi et al. however overturn all these considerations.
The fact that R1a is found in relatively high frequency among a
Dravidian speaking tribe, the Chenchus, is remarkable enough. But
the high South-Asian haplotype diversity associated with the R1a,
compared with the the haplotype diversity shown by R1a in Central
Asia and in samples from Czech and Estonian populations suggests
that the haplogroup might have emerged in South Asia. The conclusion
has immediately been taken as evidence for their views by some
indigenists who believe that Indo-European languages are indigenous
to the Indian subcontinent: surely, now even the Y chromosome
evidence supports our case!
Well, there is no doubt that the conclusions of Kivisild, Rootsi et
al. are to some extent welcome to indigenists. But for those who are
not in the business of discovering, ignoring and rediscovering
genetic evidence as the need may call, will notice that
(a) Disappointingly, Kivisild, Rootsi et al. did not try to estimate
the coalescence times for R1a and R1b. Even more surprising than the
prevalence of R1a among Chenchus and the high haplotype diversity in
South Asia is that R1b, often taken as the "Basque marker" is
very well represented among the "Indian gypsies", the Lambadis. In
fact, their Y-chromosomes are farther from other studied Indian
populations and closer to Western Europeans than those of Greeks!
All this is very intriguing but also somewhat baffling as R1a and
R1b are many steps away from M9: M9>M45>M207>M173>M17 (R1a) and M269
(R1b) how could they be as ancient as the new study seems to
suggest? They must be ancient if Kivisild, Rootsi et al. are right
because R1b surely existed in the Western refugium already during
the Last Glacial Epoch. How and when did it then arrive in Europe?
And why did all R1b go to Western Europe and all R1a to Eastern
Europe? Morover, Kivisild, Rootsi et al. do not provide us any other
haplotype data on Eastern Europe than the Czech and Estonian
samples. They may well turn out to be representative, but I
certainly would like to see at least the Ukrainian data (as well as
the Basque data) before I make any firm verdicts here. Clearly, we
have not heard the last word on these issues.
(b) The new results cannot demolish the "genetic basis" of
the mainstream version of the arrival of Indo-Aryans to India
because the mainstream version is based on the model of elite
dominance, i.e., on the idea of rather small groups of immigrants
arriving India. True, the Greater Indus Valley may have then been
partially severely depopulated, but India itself must have been
populous even then. There is no way a small number of people can
substantially change the genetic composition of the pool, unless we
assume an invasion similar to the invasion of Europeans to Pre-
Columbian Americas. (And that is the reason why I intended to write
a message for this board pointing out the necessity of an earlier
migration from the North to the Indian subcontinent, but Kivisild,
Rootsi et al. have now challenged a major premiss of my intended
argument.) Anyway, that is not the mainstream view, hence there is
no need to revise the mainstream view in the light of the present
genetic evidence. In fact, if and when Indo-Europeans arrived in
India around 1500 BCE, they probably still carried copies of M17, no
matter what the original "expansion home" of this mutation
was. It is "just" that we cannot then distinguish cases that
involve "indigenous" and "back migrating" genes ceteris paribus,
of course.
Summa summarum: The new and exciting genetic studies will eventually
provide us a lot of information we cannot reach in any other way.
But they are still taking their first and often halting steps, in
the same way radiocarbon studies took their first and halting steps
since 1947. And it will take some time before we can achieve the
sort of sophistication that dendrochronologically calibrated high-
precision C14 data provide. But the Kivisild, Rootsi et al. article
is important: many interesting and illuminating studies will surely
follow later.
Best regards, Juha Savolainen