Facing race.

Race: The Reality of Human Differences, by Vincent Sarich and Frank Miele (Westview, 304 pp., $27.50)

THIS is a strange book about a strange debate. On one side of the debate are legions of left-leaning academics telling us that race is meaningless--a "social construct," divorced from physical reality and poisonous in its implications. They argue that racial categorization leads to judgments about which groups are inferior and which superior, and ends up justifying colonialism, slavery, and discrimination. Now a standard litmus test of political correctness, this anti-race perspective has been promoted by the American Anthropological Association (whose website tells us that "race is not a scientifically valid biological category"), by the American Sociological Association, by such Marxist Harvard luminaries as Richard Lewontin and the late Stephen Jay Gould, and by public television (whose three-part documentary last year, sponsored by the Ford Foundation, was titled Race: The Power of an Illusion). It is not absolutely necessary to be an intellectual in order to believe that race is meaningless, but--as with those long-ago pronouncements about the Soviet Union's being a democracy--it plainly helps. In an era when racial differences in DNA are increasingly identifiable, it takes ever more ingenuity to argue that the differences are phantasmal, especially when the arguers are simultaneously carrying on about the wonderfulness of race-based "diversity."

On the other side of the debate are several billion human beings who assume that race is an obvious and significant reality, and tend not to realize that this idea is controversial. Children at age three, without coaching, readily distinguish people by race. The U.S. legal system endlessly takes race for granted, does not bother to define it, and proceeds on the plausible assumption that we all know what we're talking about. Those assuming that race is real and consequential tend not to do much actual debating, so this new book is a breakthrough of sorts. If only it were a better book.

Vincent Sarich, professor emeritus of anthropology at the University of California, Berkeley, and Frank Miele, a senior editor of Skeptic magazine, are here undertaking to put down the correctniks and establish that racial differences are real and worth studying. Their definition of race--strangely, not delivered until page 207--refers to "populations ... within a species that are separated geographically from other such populations ... and distinguishable from them on the basis of heritable features." The reference to heritability tells us that the traits in question--both physical and behavioral--have some genetic basis. It was long held by race-deniers that genetic data cannot be used to identify racial differences. In the PBS documentary, some biology students were shown registering surprise when they were unable to identify the races of other students based on six DNA markers they had been given. But the students had been shortchanged. They did not have enough markers, and those they had reflected only maternal heritage. With 73 appropriate DNA markers, one learns from this book, it is possible to state with close to 100 percent accuracy whether the ancestors of the individual who supplied them came from Europe, Africa, Asia, or the Americas.

The objections to race come from all directions, and the Sarich-Miele team generally does an effective job of shooting them down. In one familiar line of argument, racial categories are dismissed on the ground that they do not account for racial admixtures. It appears that, biologically, American blacks are on average about 25 percent white, and many other racial boundaries are undeniably blurry. But categories can be real without being discrete. There is no discrete boundary between red and yellow--both blend into orange--yet we have no reason to regard those colors as in any sense unreal. Races, say Sarich and Miele, can be thought of as what mathematicians call "fuzzy sets."

Stephen Jay Gould persistently invoked evolutionary logic to argue that race differences could not be terribly meaningful. He observed that the human species was relatively young--it may have begun migrating out of Africa as recently as 100,000 years ago, in evolutionary time a mere eyelid flicker, not nearly long enough to generate major differences among the races. In other words, there are no major differences because there can't be major differences --a curious position for a man who prided himself on being an empiricist. Sarich and Miele turn the evolutionary logic against Gould. They note, for openers, that the latest data suggest the migrations from Africa began only around 50,000 years ago. Then, pointing to the observed heritable differences among the resulting populations--differences not only in skin color but also in body size, cranial capacity and brain size, intelligence, physical ability, and personality--they argue that so much adaptation in so little time means that racial differences had to be enormously important for survival. The differences were not trivial, and could not have been driven by chance.

A curious fact about Race is that much of its text--and, on my scorecard, its most interesting material--is not directly concerned with racial differences. It deals instead with "pre-racial" issues. One question has to do with the date at which the earliest hominids separated from the apes. Sarich himself played a major role in answering this question, in the process establishing that it was far later than previously assumed. The standard answer used to be 20 million years ago. The answer today is about 5 million years ago. That was the point at which our ancestors branched off from the chimpanzee/gorilla line.

Another question on which Race is interesting centers on the riddle of the Neanderthals, who clearly coexisted for a time with Homo sapiens, yet remain totally unrepresented in our own DNA. How this came about remains a riddle, although Sarich leaves you thinking that matings between the two groups were for some reason infertile--which would itself be a riddle.

The book is at once fascinating and irritating. It exposes the race-doesn't-exist arguments as involved nonsense. It is a treasure trove of memorable evolutionary insights, as in its explanation of Polynesians' over-representation in the National Football League: Given the vast Pacific distances their ancestors had to travel in primitive outriggers, they were "selected," in the Darwinian sense, for size and upper-body strength.

And yet the book is a bumpy ride. I cannot recall reading a work that had so many major merits and yet was so sloppily written. Sentences do not parse, explanations fail to explain, weirdness abounds. Cost-benefit analysis is defined as "opportunity costs." Race is "loaded with the onus of past events." Among the sentences triggering "huh?" in the margin: "Those sympathetic to Coon believed his personal dislike of Montagu was because he thought everyone else should dislike him as well." A special embarrassment is the ninth and final chapter, titled "Learning to Live with Race," in which the authors advocate meritocratic social policies but then, by way of exemplifying their approach, add that race can and should be a factor in college admissions because "a diverse student body ... is part of the educational experience." Coming as it does toward the end of a book that has mercilessly pilloried political correctness, this formulation got a "How sad" review in the margin.

Mr. Seligman, a contributing editor of Forbes, is the author of A Question of Intelligence: The IQ Debate in America.