The study used data from five repeated cross-sectional probability surveys of Ontario students enrolled in grades 11 and 13 (aged 16-19 years). Data were based on interviews with 2,037 students in 1983, 2,010 in 1985, 2,008 in 1987, 1,742 in 1989, and 2,100 in 1991. Only students in grades 11 and 13 were asked to report on delinquent behaviors. Each survey used a stratified (grade by region) single-stage cluster sample of homeroom classes. For each of seven delinquent behaviors, students were asked to report the number of times the behavior occurred during the 12 months before the survey. The study analyzed the association between delinquency and the use of 17 drugs. The measures referred to drug use that occurred at least once during the 12 months before the survey. Findings show that during the period under study, drug use and drug-selling declined significantly; whereas, reported non-drug-related delinquency increased. In addition to differing secular trends in rates of drug use and delinquent behavior, the association between the two behaviors weakened. There was a steady decline in the correlations between drug use and delinquency across time, and the typological pattern of drug use and delinquency changed significantly. 4 tables and 20 references