The participants were selected through a purposive, snowballing technique and were 13 to 20 years old. They completed survey forms while a research staff member read each item aloud. The analysis identified four types of gangs operating in each city: social gangs, party gangs, serious delinquents, and incipient formal criminal organizations. The gangs all had high involvement in drug use, but they varied in their drug dealing. The severity of collective gang crime was associated with the prevalence of drug use in a gang. Drug dealing occurred among gangs with both high and low involvement in violence and other crimes. Involvement in cocaine, opiates, and phencyclidine occurred among both violent and nonviolent gangs, as well as among gangs with different involvement in drug dealing. Findings indicate that the drug-crime relationship among gang members is skewed and spurious. The positive association between drug involvement and serious gang crime is similar to the drug-crime relationship among urban youths who do not belong to gangs. In addition, among both gang members and nongang members, violence is not an inevitable consequence of involvement in drug use or dealing. Tables, footnotes, appended offense list, and 76 references. (Author abstract modified)