Paul Gross reviews Race: The Reality of Human Differences.

Update: Biometrician Mike Miller has an interesting post over at the Evolutionary Psychology list:

Neil Risch is a great statistical geneticist and he does a much better job...You'll be seeing some very interesting papers from him during the next year or so. In one, unless there is some error in it, he shows that using about 350 markers from a genome scan...he can classify 3500 unrelated black, white and asian subjects (their self-reported races) into groups by multidimensional scaling only misclassifying 5 people. So he can place people with others of their self-reported race, based only on some of their DNA, and be correct 99.8% of the time, yet there is no "biological basis" for the self-identified racial classification! He also could work with a sample of Japanese and Chinese, identify two groups based only on these markers, and classify individuals with their correct group 95% of the time.

The Japanese vs. Chinese result surprised me-perhaps Jomon/Ainu admixture?

Update II: A critic responds to Miller.