The English language literature on the epidemiology of dual diagnosis is reviewed. The literature shows mental disorders to be significantly related to alcohol and drug use disorders. The strongest associations involve externalizing mental disorders and alcohol-drug dependence. Mental disorders are associated with alcohol-drug use, problems among users, dependence among problem users, and persistence among people with lifetime dependence. These dual diagnoses are associated with severity and persistence of both mental and alcohol-drug disorders. A wider range of mental disorders is associated with nicotine dependence. Most people with dual diagnosis report their first mental disorder occurred at an earlier age than their first substance disorder. Prospective studies confirm this temporal order, although significant predictive associations are reciprocal. Analyses comparing active and remitted mental disorders suggest that some primary mental disorders are markers and others are causal risk factors for secondary substance disorders. The article closes with a discussion of ways epidemiologic research can be used to help target and evaluate interventions aimed at preventing secondary substance use disorders by treating early-onset primary mental disorders.