Data were obtained from the ADAM programs of Maricopa and Pima Counties (Arizona) for 2000-2003. Only male arrestees who provided a urine sample and had a valid residential zip code within either of the counties were included in the analyses. Individual-level data pertained to race/ethnicity, education, and employment status, as well as the most serious offense at arrest and prior arrest history. Community-level measures were unemployment, educational level, and racial/ethnic heterogeneity. In order to capture how drug markets influenced methamphetamine use, data were obtained on acquisition methods, purchase sources, location of purchases, and dollar amounts of purchases, units of methamphetamine purchased, and frequency of purchases. Data were obtained on 6,042 arrestees. Findings showed that both individual-level and community-level data significantly influenced methamphetamine use; and the predictors of methamphetamine use differed significantly from those related to marijuana, cocaine, and opiate use. At the individual level, age had a negative effect on methamphetamine use, and Whites were more likely than Hispanic, Black, and Native-American arrestees to test positive for methamphetamine use. At the community level, racial/ethnic heterogeneity had an insignificant role in methamphetamine use, and unemployment, college education, and Spanish-speaking households had negative effects on methamphetamine use, suggesting that communities with higher methamphetamine use tend to have relatively high levels of social organization. Findings on methamphetamine markets suggest that the users purchase the drug indoors from a regular dealer; and all movements of the drug, including the final transfer to the consumer, occur in closed networks. Policy implications are drawn. 6 tables and 44 references