The study collected 4 years of longitudinal data from the Pittsburgh Youth Study for 506 local male adolescents. The sample was approximately 16.5 years old in the first of the four assessments and 19.5 years old at the end. The analyses included only adolescents who had used alcohol (beer, wine, or hard liquor), marijuana, and/or other drugs at least once during the 4-year period. This limited the analysis to 454 of the 506 subjects. The interviews solicited self-reports on the frequency of the commission of 19 types of illegal acts within the past year. Participants reported committing offenses against persons more often than general theft under the influence of alcohol or drugs. Aggressive acts were more often related to self-reported acute alcohol use than to marijuana use. Those who reported committing illegal acts under the influence of alcohol or drugs indicated they were committed with other people and frequently resulted in arrests. Offenses under the influence of alcohol or drugs were more prevalent among the heavier users, more serious offenders, more impulsive youth, and youth with more deviant peers. There were no significant interaction effects of alcohol and drug use with impulsivity or deviant peers in predicting whether illegal acts were committed under the influence. The association between drug use and illegal activity during adolescence is concluded to be a complex link, since no one single model can explain this relationship for all adolescents. 6 tables, 3 notes, and 49 references