The Science of Racism is an odd title for an article in which Henry Louis Gates identifies James Watson’s view as something he calls “racialist” and distinct from “racist”.
Gates is pessimistic about the consequences of possible discovery of racial variations in the distribution of some real inheritable quality of “intelligence” but I think the point Watson is trying to get at is that if and when such variations are discovered that discovery would entail the ability to determine an individual’s genetic endowment directly and such determination would override any tendency to judge people based on a statistically but not causally linked characteristic. (If it were true that blue eyed people were statistically more likely to be taller than average, that wouldn’t cause anyone meeting me to overestimate my height, and even if blonde jokes really do have a statistical basis in fact, if intelligence were measurable my own could easily be established as whatever it is and to select on the basis of that measurement would be more cost effective for an employer than to do so on the basis of hair colour)