Objective: This article expands on the results from a 1997 report from the Collaborative Study on the Genetics of Alcoholism (COGA), using a new phase of the protocol to evaluate the prevalence and characteristics of substance-induced and independent major depressive episodes (MDEs) in a population of alcoholics and nonalcoholics. Method: Data were evaluated from Phase II of the six-center-wide COGA investigation using information gathered beginning in January 1997. Data were generated through face-to-face evaluations using the updated version of the Semi-Structured Assessment for the Genetics of Alcoholism (SSAGA-II) interview, with distinctions between substance-induced and independent MDEs based on the chronology of development of full depressive syndromes. The analyses focused on the 2,548 men and women who were divided into 351 individuals who had only an independent MDE (Group 1), 238 subjects who experienced only substance-induced MDEs, and 1,959 individuals with no MDE history. Results: The two MDE groups were similar in age, marital status, and religion; but those with substance-induced depressions (Group 2) were more likely to be original alcoholic probands, be males, be nonwhite, and have less education. They were also more likely to have alcohol, drug, or antisocial personality diagnoses and to report higher maximum drinks. In addition, only Group 2 subjects reported an elevated family history of alcohol diagnoses compared with the nondepressed Group 3. Subjects with independent MDEs were different from the comparison Group 3 regarding the family histories of independent MDEs. However, symptoms during the worst depressive episode were quite similar across Groups 1 and 2. Conclusions: This study corroborates a high rate of substance-induced MDEs among alcoholics, with these disorders explaining about half of the lifetime depressive episodes. The results also support the validity of the distinction between substance-induced and independent depressions regarding external validators of gender, substance-use patterns, and family histories of independent MDEs.