The question of the causal relationship between substance abuse and criminal offending has proved to be problematic for researchers. Common sense suggests that substance abuse might promote criminal behavior and vice versa. Although a positive correlation is one of the most reliable results obtainable in criminology, the causal relationship is difficult to demonstrate. This study represented an effort to examine the causal relationship between substance abuse by undertaking a longitudinal study of substance abuse and criminal behavior in young men. Subjects were 596 males aged 16 to 19 from Buffalo, New York. Two interviews were conducted 18 months apart. Cross-lagged and synchronous structural equation models were fit for both early and late onset of delinquency groups with five types of delinquency: minor, general, serious, property, and violent offenses. The results showed no causal relationship between substance abuse and delinquency. The late-onset model showed that minor offenses had significant lagged and synchronous positive effects on drug use; drug use exhibited significant lagged and synchronous positive effects on general offenses; and drinking had significant lagged and synchronous negative effects on property offenses. The findings of the study suggested that the relationship on substance use and delinquency depended on the type of offender and on the type of offense. The alcohol/drug/crime nexus was contingent on the characteristics of the group under study. Second, the relationship between substance abuse and delinquency was reciprocal rather than unidirectional. Substance use and delinquency may influence each other, depending on the nature of the substance and the delinquent acts. Finally, the dynamic between substance abuse and delinquency was less pronounced for more serious types of delinquency. Serious offenses may have other connections with drugs, such as drug trafficking rather than drug use. Tables, references