(FORTUNE Magazine) – Not all people are equally intelligent, and not all groups of people have the same average intelligence. Those statements are at once truisms and a center of raging controversy. They strike sparks because differences in intelligence lead to differences in pay and status, and therefore work against the ideal of equality, all of which leaves many Americans uneasy. The recent convention of the American Psychological Association featured an acrimonious debate on the ethics of studying these matters. People who flinch from such studies are advised to steer clear of Why Humans Vary in Intelligence (Paideia, $18), by Seymour Itzkoff. The author, professor of education and child study at Smith College, is attempting something enormously ambitious. Although the book stands on its own legs, it is part of a multivolume project on the evolution of human intelligence over many millennia. Itzkoff is trying to make a coherent whole of evidence drawn not only from standard psychological sources but from paleontology, physical anthropology, neuroscience, and the emerging discipline of sociobiology. The effort is not entirely successful. Itzkoff is a terrible writer. His book is disorganized, repetitious, often maddeningly obscure. Yet it is also fascinating. In assembling data from those diverse disciplines, the author has somehow put together a remarkably original guide to the present state of knowledge about human brainpower. The book is strewn with arresting details. Among kids precocious in math, 20% are left-handed, which is twice the national average; the same group is five times as likely as a sample of average students to suffer from asthma and other allergies. As the grandfather of a 3-year-old who for some reason cannot read yet, I was also fascinated to learn that on tests of ''reading readiness,'' the single best predictor is not the child's mastery of letters but his ability to handle numbers. Apparent explanation: Numbers far more than letters make kids think, which is also required for reading. Among the book's most gripping passages are some suggesting that researchers are at last closing in on the link between IQ test results and underlying neurological reality. For many years now, most serious students of intelligence tests have agreed that they were measuring something that was mostly heritable and ultimately rooted in the physical brain. (Like many other scholars, Itzkoff suggests that hereditary factors probably account for 70% to 80% of variability in IQ.) Until recently, however, it was hard to point to data linking test scores to observable activity in the brain. Now such data are becoming available. Some studies, for example, have related IQ to electrical impulses given off by the brain--its ''evoked potential.'' One study in Britain found strong correlations, some as high as 0.95, between certain tracings on the electroencephalograph and scores on several subtests of the Wechsler Adult Intelligence Scale (a standard IQ test). Plain implication: Your reasoning power depends heavily on the electrical power in the brain you inherited. Comments on racial differences in intelligence are always guaranteed to provoke a row. In a chapter headed ''The Irrelevance of Race,'' Itzkoff labors to minimize the row by placing the data in a long-term evolutionary framework, but he has trouble stating plainly why his readers should think race irrelevant. He accepts data indicating that Japanese, Chinese, and Korean IQs are between four and nine points higher than those of white Americans--a significant difference. He also points to the average black-white gap of 15 points or so and suggests it may be widening. While professional psychologists overwhelmingly accept the view of a heritable IQ, many Americans flee from this view. They flee because they think the view condemns them to believe that certain minorities are innately inferior. In fact, the idea of heritable IQs has no such implication, as Itzkoff repeatedly seems on the verge of explaining (he never quite makes it). What the idea does imply is that races or other gene pools will get lower IQs if, within the pool, fertility is greatest among those with the least intelligence. Heritability does not signify any irreversible disadvantage for members of the group. The most controversial, and also most fascinating, material in Why Humans Vary in Intelligence is not about race but about sex. A main point: Women's IQ is much less variable than men's. Women and men average about the same, but the women's scores tend to cluster more clearly around the mean. This has the unfortunate implication--especially unfortunate for a professor at all-female Smith College, one would imagine--that at the highest intellectual levels, women will inevitably be underrepresented. (Minor consolation: They will also be underrepresented at the lowest levels.) Which presumably also means that, % even in the absence of invidious discrimination, they will not get a proportional share of high-level, high-status jobs. THE EXACT ratios are a matter of dis--pute. Itzkoff cites calculations suggesting that men may outnumber women by 2 to 1 at 130 IQ and over 5 to 1 at 145 IQ. Those ratios are higher than estimates elsewhere (Arthur Jensen guesses around 7 to 5 at 140 IQ and above), but the general point is well supported in the literature, as is Itzkoff's explanation of the lower female variability. He attributes it to the fact that women have one more X chromosome than men, and it serves as a source of stability, limiting extreme effects that are both desirable and undesirable. Right now New York State is establishing a committee to deal with the regrettable fact that women's scores have been lower than men's on scholastic aptitude tests (which are, of course, taken mainly by above-average students). With no debate at all, the legislature seems to be requiring equal representation for women and men at all points on the bell curve. Perhaps the issue will be debated at next year's convention of the American Psychological Association.