The New York Times has an odd, but fascinating story about DNA services that purport to determine whether one's ancestors were of such heritages as Native American, North African, British Isles, or (apparently) European Jewish:

Alan Moldawer's adopted twins, Matt and Andrew, had always thought of themselves as white. But when it came time for them to apply to college last year, Mr. Moldawer thought it might be worth investigating the origins of their slightly tan-tinted skin, with a new DNA kit that he had heard could determine an individual's genetic ancestry.

The results, designating the boys 9 percent Native American and 11 percent northern African, arrived too late for the admissions process. But Mr. Moldawer, a business executive in Silver Spring, Md., says they could be useful in obtaining financial aid. . . .

Genetic tests, once obscure tools for scientists, have begun to influence everyday lives in many ways. The tests are reshaping people's sense of themselves — where they came from, why they behave as they do, what disease might be coming their way.

It may be only natural then that ethnic ancestry tests, one of the first commercial products to emerge from the genetic revolution, are spurring a thorough exploration of the question, What is in it for me?

Many scientists criticize the ethnic ancestry tests as promising more than they can deliver. The legacy of an ancestor several generations back may be too diluted to show up. And the tests have a margin of error, so results showing a small amount of ancestry from one continent may not actually mean someone has any. . . .

One Christian is using the test to claim Jewish genetic ancestry and to demand Israeli citizenship, and Americans of every shade are staking a DNA claim to Indian scholarships, health services and casino money.

"This is not just somebody's desire to go find out whether their grandfather is Polish," said Troy Duster, a sociologist at New York University who has studied the social impact of the tests. "It's about access to money and power."

Driving the pursuit of genetic bounty are start-up testing companies with names like DNA Tribes and Ethnoancestry. For $99 to $250, they promise to satisfy the human hunger to learn about one's origins — and sometimes much more. On its Web site, a leader in this cottage industry, DNA Print Genomics, once urged people to use it "whether your goal is to validate your eligibility for race-based college admissions or government entitlements." . . .

DNAPrint calls the ethnic ancestry tests "recreational genomics" to distinguish them from the more serious medical and forensic applications of genetics. But as they ignite a debate over a variety of genetic birthrights, their impact may be further-reaching than anyone anticipated. . . .

Ashley Klett's younger sister marked the "Asian" box on her college applications this year, after the elder Ms. Klett, 20, took a DNA test that said she was 2 percent East Asian and 98 percent European.

Whether it mattered they do not know, but she did get into the college of her choice.

"And they gave her a scholarship," Ashley said.

Pearl Duncan has grander ambitions: she wants a castle.

The article has many interesting stories of people using these new tests to determine ancestry.

But the New York Times story is very weak on the science behind the tests. The last time I reviewed the genetics literature behind racial classifications was about a decade ago, so I am not up to speed on recent developments, though I recall discussing with a genetics scholar the possibility of such testing for the general public. Here (in my lay opinion) is where things stood a decade ago.

First, the consensus is that race is a social construct that maps very crudely onto real genetic (and thus physical) differences, of which skin color is only one example.

Second, whether there are genetically three races or five races or twenty races depends more on whether the racial classifier is a "lumper" or a "splitter." It's a bit like determining how many major fields there are in law. Are there just two: private law and public law? Or are there twenty fields, or fifty fields, or more? (Though for law, while there are real differences between fields, there is little physical or genetic reality lying beneath the classifications.) At the margin, any classifications are going to be (or appear to be) arbitrary.

Third, there are few (if any) genetic markers that everyone in one traditional racial category has and everyone in other traditional racial categories lack. (I take it that there may be a few markers that are found only in some racial or ethnic groups, but not all people in those groups have the markers.)

If one looks for genetic markers, one could draw something like the isobars one sees in weather maps. In one area (even without recent migration), 90% of those living there have a particular marker, in another area 80% have the characteristic, and so on. This is obviously true for physical characteristics, such as hair color, eye color, and skin color. For example, there is a real difference in skin color between those whose families lived for the last 600 years in southern Sweden and those whose families lived over the same period in sub-Saharan Africa. Yet there is no one line that could be drawn between light skin and dark skin. If one drew a map of mean skin color between southern Sweden and central Africa, there would be the equivalent of isobars marking the gradual differences, with the isobars closer together in some regions and farther apart in others. Wikipedia has a fairly crude version of such a skin color map and another from the dust jacket of Cavalli-Sforza's The History and Geography of Human Genes (1994), a population genetics text that I consulted in my research a decade ago. Further, the maps would be somewhat different for skin color, hair color, or one genetic marker or another.

Genetic Variation (Cavalli-Sforza)

At the genographic project, they trace genetic markers that identify groups that migrated around the world tens of thousands of years ago, thus creating genetic diversity in regions long before the age of exploration.

Coming back to these new genetic services described in the New York Times, I would have liked more of an explanation in the article about what markers allow them to determine that someone is "10% British Isles" or "3% Native American" or "2% East Asian." Because I wasn't aware that the genetic signature of British Isles ancestry was so specific that it permitted such a conclusion, I would have appreciated an identification of the scientific basis for such a claim. As for 3% Native American or 2% East Asian, I would have liked a confirmation that they found a marker present only in those groups, rather than one simply unlikely to be found in other ethnic groups.

Despite the frustrating absence of information about the scientific bases for the determinations, I recommend reading this Times account of the operation of what Justice Brennan called "benign racial sorting."