In a prospective cohort study of 249 children from birth to two years of age, we assessed the relation between prenatal and postnatal lead exposure and early cognitive development. On the basis of lead levels in umbilical-cord blood, children were assigned to one of three prenatal-exposure groups: low (less than 3 micrograms per deciliter), medium (6 to 7 micrograms per deciliter), or high (greater than or equal to 10 micrograms per deciliter). Development was assessed semiannually, beginning at the age of six months, with use of the Mental Development Index of the Bayley Scales of Infant Development (mean +/- SD, 100 +/- 16). Capillary-blood samples obtained at the same times provided measures of postnatal lead exposure. Regression methods for longitudinal data were used to evaluate the association between infants' lead levels and their development scores after adjustment for potential confounders. At all ages, infants in the high-prenatal-exposure group scored lower than infants in the other two groups. The estimated difference between the overall performance of the low-exposure and high-exposure groups was 4.8 points (95 percent confidence interval, 2.3 to 7.3). Between the medium- and high-exposure groups, the estimated difference was 3.8 points (95 percent confidence interval, 1.3 to 6.3). Scores were not related to infants' postnatal blood lead levels. It appears that the fetus may be adversely affected at blood lead concentrations well below 25 micrograms per deciliter, the level currently defined by the Centers for Disease Control as the highest acceptable level for young children.