To investigate the impact of low-level lead exposure on children's behavior within a natural setting, we assessed the school performance of a sample of 141 elementary school children classified according to dentine lead level. Twenty-two children had "elevated" levels (≥20 parts per million), 71 had "midrange" levels (10.0-19.9 parts per million), and 48 had "low" levels (<10 parts per million). Four types of outcome data were collected: (1) scores on a standardized group intelligence test, (2) teachers' ratings, (3) incidence of academic failure (assignment to remedial aid, grade retention), and (4) observations of classroom behavior. In general, higher dentine lead levels were associated with less favorable school performance, with most covariance-adjustedp-values in the range of 0.05-0.15. The incidence of grade retention was the outcome most strongly related to lead level. The pattern of results suggests a consistent, though weak relationship between children's dentine lead levels and elementary school performance.