Aims: To compute the common and specific genetic and environmental contributions to nicotine dependence (ND) alcohol dependence (AD) and cannabis dependence (CD). Design: Twin model. Participants: Data from 1874 monozygotic and 1498 dizygotic twin pair members of the Vietnam Era Twin Registry were obtained via telephone administration of a structured psychiatric interview in 1992. Measurements: Data to derive life-time diagnoses of DSM-III-R ND, AD and CD were obtained via telephone administration of the Diagnostic Interview Schedule. Findings: The best-fitting model allowed for additive genetic contributions and unique environmental influences that were common to all three phenotypes. Risks for ND and AD were also due to genetic and unique environmental influences specific to each drug. A specific shared environmental factor contributed to CD. Conclusions: These results suggest that the life-time co-occurrence of ND, AD and CD is due to common and specific genetic factors as well as unique environmental influences, and vulnerability for CD is also due to shared environmental factors that do not contribute to ND and AD. The majority of genetic variance is shared across drugs and the majority of unique environmental influences are drug-specific in these middle-aged men. Because differences between models allowing for specific genetic versus shared environment were small, we are most confident in concluding that there are specific familial contributions - either additive genetic or shared environment - to CD.