Three-and-a-half-month-old infants already seem to exhibit the other-race effect. In a study at the University of Kentucky, white babies were very good at distinguishing faces with 100 percent Caucasian features from faces that had been graphically morphed to include features that were 70 percent white and 30 percent Asian. They couldn’t do the reverse: They could not tell 100 percent Asian faces from those that were morphed to include 30 percent white features. In other words, they could detect small differences between white and not-quite-white faces, but not the same kinds of differences between Asian and not-quite-Asian faces. Lawrence A. Hirschfeld of the University of Michigan did some of the pioneering work on how early in life children begin to understand race. He showed children of ages three, four, and seven, a picture of “Johnny:” a chubby black boy in a police uniform, complete with whistle and toy gun. He then showed them pictures of adults who shared two of Johnny’s three main traits of race, body build, and uniform. Prof. Hirschfeld prepared all combinations—policemen who were fat but were white, thin black policemen, etc.—and asked the children which was Johnny’s daddy or which was Johnny all grown up. Even the three-year-olds were significantly more likely to choose the black man rather than the fat man or the policeman. They knew that weight and occupation can change but race is permanent. In 1996, after 15 years of studying children and race, Prof. Hirschfeld concluded: “Our minds seem to be organized in a way that makes thinking racially—thinking that the human world can be segmented into discrete racial populations—an almost automatic part of our mental repertoire.” When white preschoolers are shown racially ambiguous faces that look angry, they tend to say they are faces of blacks, but categorize happy faces as white. “These filters through which people see the world are present very early,” explained Andrew Baron of Harvard. Phyllis Katz, then a professor at the University of Colorado, studied young children for their first six years. At age three, she showed them photographs of other children and asked them whom they would like to have as friends. Eighty-six percent of white children chose photographs of white children. At age five and six, she gave children pictures of people and told them to sort them into two piles by any criteria they liked. Sixty-eight percent sorted by race and only 16 by sex. Of her entire six-year study Prof. Katz said, “I think it is fair to say that at no point in the study did the children exhibit the Rousseau type of color-blindness that many adults expect.