A study of the effectiveness of an intervention program designed to favorably modify behaviors hypothesized to be related to the future development of cancer was initiated among 1,105 eligible children in 15 schools in the vicinity of New York City. Schools were assigned to either an intervention or a nonintervention group. Subjects in schools in the intervention group received each year, from fourth through ninth grade, a teacher-delivered curriculum focusing on diet and prevention of cigarette smoking. After 6 years of intervention, the rate of initiation of cigarette smoking was significantly lower among subjects in intervention schools than among those in nonintervention schools. There was a significant net decrease in reported intake of saturated fat and a significant net increase in reported intake of total carbohydrate among subjects in intervention schools compared to those in nonintervention schools. These findings, if replicated, suggest that such programs are feasible and acceptable and may have a favorable effect on diet and prevention of cigarette smoking in children.