Recent preliminary evidence suggests that anabolic-androgenic steroids, drugs widely used by athletes to gain muscle mass and strength, may cause serious psychiatric effects, including manic syndromes, major depression, and psychotic symptoms. In addition, some individuals may become uncharacteristically aggressive and even commit violent crimes while exposed to steroids. Given recent studies showing widespread steroid use among American men, these effects may represent a major public health problem. The study proposed in this application seeks to replicate the investigators' uncontrolled pilot study, which found high rates of psychiatric symptoms associated with steroid use in athletes, while minimizing the methodological limitations of that earlier investigation. In the proposed study, 80 steroid users and 80 non-users will be recruited from gymnasiums in three metropolitan areas, using a recruitment process designed to minimize selection bias. Subjects will be interviewed, using an instrument refined from the investigators' pilot study, and will submit urine to be tested for both steroids and other drugs. They will also receive a physical examination, including caliper determinations of body fat, and have blood drawn for standard chemistry and hematology measures. Users and non-users will be compared on basic demographic variables, medical history of possible steroid-related violence or criminal activity. Possible predictors of steroid=associated psychopathology, such as dose and type of drugs used, presence of other drug use, and family history of psychiatric disorder, will be assessed by comparisons within the user group. Although the proposed study is naturalistic, and relies on retrospective reports by subjects, this method represents the only design capable of assessing the psychiatric effects of the very large doses of steroids used by actual athletes in the field.