1956 7th grade students of high schools and secondary modern schools in three cities with 110,000 to 180,000 inhabitants were asked about cigarette, alcohol and drug consumption using questionnaires. Subsequently 1080 of the students took part in a newly invented non-smoking curriculum which was held during 8 lessons within 4 weeks (intervention group (I)). The lessons were performed by non-smoking physicians who had passed a training programme before. 876 students served as control (C) group. 2 years later the questioning was repeated. The number of students who started smoking within the two years was significantly lower in the intervention group. In the high schools a significant effect was seen in boys (I: 13.0%, K: 22.4%; p < 0.01) and girls (I: 21.4%, K: 28.4%; p < 0.05) whereas in secondary modern schools a significant effect was seen only in boys (I: 17.4%, K: 25.2%; p < 0.05) not in girls (I: 18.0%, K: 22.0%; n.s.). First-time consumers of drugs like cannabis, marihuana and organic solvents were found less often in the intervention group among boys in high schools than in the control group (I: 14.9%; K: 23.6%; p < 0.05). No effects of the intervention were seen in girls of high schools and in students of secondary modern schools. No effects of the intervention were seen on alcohol consumption. Students who drank beer, wine or liquor several times a week were found at the same rate in the intervention and control groups. We conclude that a non-smoking intervention of 8 lessons in the 7th grade of high schools and secondary modern schools may lower the rate of first-time consumers of cigarettes but not of alcohol and drugs.