The study's goals were to determine if alcohol produced a consistent change in testosterone levels, to learn if the extent or direction of change was associated with observable changes in aggressive behavior, and to determine the distribution of testosterone level abnormalities in alcohol addicts during a period of sobriety. Nine male alcohol addicts volunteered for a study designed to measure aggressive behavior and plasma testosterone levels before, during, and after an 11-12-day period of chronic intoxication. Eight sober alcohol addicts and 34 alcohol addicts who were in a correctional institution during a period of sobriety were also studied with respect to plasma testosterone levels only. Results showed that intoxication was associated with suppression of plasma testosterone in eight of the nine subjects of the first studies. The suppression seemed to be independent of baseline testosterone values. During alcohol withdrawal, plasma testosterone levels returned towards baseline levels. There was no systematic relationship between testosterone level and aggressive behavior during baseline, intoxication, or withdrawal. Analysis and categorization of histories of violence and examination of plasma testosterone values in the sober subjects showed that the distribution of plasma testosterone values was comparable in each of the three violence levels. Figures, a discussion by others, and a list of 49 references are provided. (Author abstract modified)