India, 1901. The greatest ever exercise in human nose-counting had just been undertaken. Her Majesty's Indian subjects were being counted, sifted and sorted. The Census of India was on. Census commissioner Sir Herbert Risley noticed that upper caste Hindus were fair and had sharp noses. Since these are among the distinguishing features of the "white man"-caucasoids-he figured there may be a relationship between the two. To make this scientific, he took a few measurements. The methodology: measure nose length, divide by nose breadth, and call the number arrived at the "nasal index". The conclusion, of course, was that upper caste Hindus are distant relatives of Englishmen who have been out in the sun a few centuries.

Much water has flowed down the Thames and Ganga since. Adolf Hitler has made infamous the theory of Aryan supremacy. Computers and genetic engineering have been invented. Accepted wisdom on the question of an Aryan invasion of India has veered from history book standard to disbelief after archaeologists found evidence that the Aryans did not ride into India, subjugate the native population and set themselves up at the top of the caste heirarchy.

And now, this happens.

Eighteen scientists from India and the US led by human geneticist Michael Bamshad of the University of Utah compared genetic signatures of modern-day Indians of various castes with those of today's Europeans and east Asians. Using genetic markers, they traced back the paternal lineage of the Indians through the Y-chromosome. Maternal lineage was traced through mitochondrial DNA (see box). The results of the study, published in an American journal called Genome Research, conclude that the upper castes are genetically closer to Europeans and the lower castes to Asians. Also, that we have a common maternal ancestry but different paternal stock.

"This paper is clearly a landmark," says Partha Pratim Majumdar, head of anthropology and human genetics at the Indian Statistical Institute in Kolkata. "It uses a large battery of genomic markers to show that the observed trend among different castes matches expectations (about caste differences)." Dr Peter Forster of The McDonald Institute for Archaeological Research, University of Cambridge, says that "the conclusion that higher castes have greater genetic relatedness to west Eurasians is well founded on the basis of their data".

So then: are Brahmins and Kshatriyas really Europeans inside? They may be, but the final verdict on that is far from out yet. The paper has unleashed the academic equivalent of fistfights among historians and anthropologists across the country. Controversies once buried have returned from the grave and theories are being tossed about on all sides. How the European genes got here at all is among the most interesting points of the debate. The paper infers that caste Hindus are descendants of Aryans to explain the genetic data. The counter is quick. "Dravidian and Aryan are linguistic, not racial terms. There is no specific Aryan race," says social anthropologist V.N. Shrivastava. But the Aryan invasion theory was based, among other things, on linguistic grounds: the similarity between Sanskrit and European languages was taken as evidence of people from Europe having migrated to India. The Rig Veda is the oldest known Sanskrit text. It was dated to around 2000 b.c. which is about when the Indus Valley Civilisation is supposed to have ended. "But it was an oral genre-how can it be dated?" asks Nayanjot Lahiri, reader in history at Delhi University. Moreover there was no mention in the Rig Veda of migration or any other homeland, which would have been natural if a great journey had been undertaken or a war won. The people of Israel still talk with familiarity of their journey to the "Promised Land" 2,500 years ago.

The Aryan invasion theory takes its hardest blow from the skeletons found at Indus sites. They show the same racial mix as any population of South Asia. There was no evidence of the carnage that would have accompanied an invasion. "Of the skeletons found, only three showed any signs of injury," says S.P. Gupte of the Archaeological Society of India. "And even those had wounds that had healed-the people did not die of those wounds."

Yet there is the evidence. "Their conclusions are based on large sample sizes and on three independently inherited types of genetic loci-paternally inherited Y-chromosomes, maternally inherited MTDNA, and biparentally inherited Alu insertions," says Forster. Translation: the research team did their homework. Which means they are probably right about the European genes being there. "Our results clearly show that there are differences between upper and lower castes, and the upper castes are closer to Europeans," says B. Bhaskara Rao, an Andhra University anthropologist and one of the authors of the paper. And the fact that Sanskrit is close to European languages holds.

So, the plot thickens. If there was no Aryan invasion like the archaeologists say, and our upper castes are genetically closer to Europeans like the anthropologists say, what could possibly have happened?

In their paper, Bamshad and colleagues conclude that it was a migration, mainly of males, that brought the west Eurasians here, not invasion. The white men didn't come in plundering and pillaging. They just drifted in and, it must follow, became sufficiently popular to begin cohabiting with the local women here. That there was migration to India is not disputed. "India was never an ethnic vacuum," says Shrivastava. "Palaeolithic hunters arrived in the Old Stone Age, there was a stream of migrants-builders of the Indus Valley-around 10,000 b.c., and the Indo-Aryans arrived around 3,000 b.c. ." And caste began? "There are more than 100 theories on how caste originated, but one can be sure it didn't emerge because Aryans subjugated the native population." So how did they then end up at the top of the caste hierarchy? And what, in any case, is caste?