Race falls from grace.

Martin Bernal's characterization of race as a social entity, rather than a biological one, echoes a growing trend among anthropologists to reject the usefulness of the three traditional racial categories -- Caucasoid, Negroid and Mongoloid -- for the study of human populations.

"We should abandon race designations and use geographic designations only, such as western Africans and northeastern Europeans," argues anthropologist C. Loring Brace of the University of Michigan in Ann Arbor, who first promoted this position more than 20 years ago. Racial characteristics, such as skin pigmentation, change rapidly in response to environmental conditions and thus serve as poor clues to anatomical relationships between human populations, Brace maintains. Minor anatomical features that mainly evolved randomly, such as tiny bones on the skull formed by cranial sutures, provide a better comparison for groups living in different parts of the world, he asserts.

Anatomical variations among people of the same race render such categories meaningless, Brace adds. For instance, Mongolians display marked anatomical differences from the Chinece. "To classify the people of China as Mongoloid is on a part with classifying the people of Europe as Swedish," Brace contends.

Others see a place for racial categories. With the advent of agriculture about 10,000 years ago, three general racial groups evolved in Africa, Asia and Europe, replacing small groups of hunter-gatherers, holds Donald E. Tyler of the University of Idaho in Moscow. No labels accurately denote the anatomical differences of the three races, Tyler says. The terms "black," "white" and "Asian" severe social purposes, but each encompasses people with diverse anatomical features, he notes.