Reciprocal relations between self-reported substance use and delinquency were examined through the analysis of 4-wave panel data collected from 1,218 high school students. A longitudinal, latent-variable framework was used to investigate interrelationships between changing patterns of generalized involvement in these 2 problem behaviors, while simultaneously accounting for specificity in drug use (e.g., marijuana use) and delinquency (e.g., theft). Analyses revealed that a model of bidirectional effects between polysubstance use and general delinquency was plausible for boys but not for girls. For boys, the effect of delinquency on substance use was small but consistent over time, whereas the effect of substance use on delinquency was larger but restricted to the earlier waves of the study. Implications for the question of whether or not substance use and delinquency have causal connections are discussed.