We studied the effect of environmental exposure to lead on children's abilities at the age of four years in a cohort of 537 children born during 1979 to 1982 to women living in a community situated near a lead smelter. Samples for measuring blood lead levels were obtained from the mothers antenatally, at delivery from the mothers and umbilical cords, and at the ages of 6, 15, and 24 months and then annually from the children. Concurrently, the mothers were interviewed about personal, family, medical, and environmental factors. Maternal intelligence, the home environment, and the children's mental development (as evaluated with use of the McCarthy Scales of Children's Abilities) were formally assessed. The mean blood lead concentration varied from 0.44 mumol per liter in midpregnancy to a peak of 1.03 mumol per liter at the age of two years. The blood lead concentration at each age, particularly at two and three years, and the integrated postnatal average concentration were inversely related to development at the age of four. Multivariate analysis incorporating many factors in the children's lives indicated that the subjects with an average postnatal blood lead concentration of 1.50 mumol per liter had a general cognitive score 7.2 points lower (95 percent confidence interval, 0.3 to 13.2; mean score, 107.1) than those with an average concentration of 0.50 mumol per liter. Similar deficits occurred in the perceptual-performance and memory scores. Within the range of exposure studied, no threshold dose for an effect of lead was evident. We conclude that postnatal blood lead concentration is inversely related to cognitive development in children, although one must be circumspect in making causal inferences from studies of this relation, because of the difficulties in defining and controlling confounding effects.