Three large U.S. epidemiological surveys researching the prevalence of comorbid psychiatric illness and "substance abuse disorders" are reviewed: the Epidemiological Catchment Area Study, (N=20,219); the National Comorbidity Survey (N=8098), and the National Longitudinal Alcohol Epidemiological Survey (N=42,862). Findings suggest that comorbidity is highly prevalent but that longitudinal information remains limited limited--which restricts understanding of its "natural history," stability, common risk factors, and causal relationships.