The goal of this study was to evaluate the relationship between hair lead levels of children and their attention-deficit behaviors in the classroom. Scalp hair specimens were obtained from 277 first-grade pupils, teachers completed the abbreviated Boston Teacher's Rating Scale for rating classroom attention-deficit behavior, and parents completed a short questionnaire. The children's hair lead concentrations ranged from less than 1 to 11.3 ppm (microg/g). The striking dose-response relationship between levels of lead and negative teacher ratings remained significant after controlling for age, ethnicity, gender, and socioeconomic status. An even stronger relationship existed between physician-diagnosed attention-deficit hyperactivity disorder and hair lead in the same children. There was no apparent 'safe' threshold for lead. Scalp hair should be considered a useful clinical and epidemiologic approach for the measurement of chronic low-level lead exposure in children.