Participants were interviewed in 1994, during a 6-month time period in Cleveland and a 9-month time period in Florida. The 185 participants in Cleveland were all males; the 241 participants in St. Petersburg, Fla., included both males and females. Results were consistent with earlier reports on prevalence estimates that used a bioassay component. The participants generally reported infrequently and tested positive by assays at rates greater than the self-reported use. Urinalysis indicated more drug prevalence than did interviews; hair assays, which have a greater retrospective time window, revealed an even greater prevalence than did urinalysis. Findings affirmed results reported in 1994 by Feucht, Stephens, and Walker. Tables, author biographies and photographs, and 14 references (Author abstract modified)