Findings show that the pattern of substance-use onset was similar for both girls and boys; adolescents first tried either alcohol or cigarettes, followed by the use of both, then dagga (cannabis), and then inhalants. The prevalence of lifetime cigarette use was slightly greater for girls; dagga (cannabis) and inhalant use were greater for boys. The similarity of factors related to the onset of substance use in the current sample with those related to previous international samples suggests that prevention programs across contexts can be effective. Study participants lived in a township established during the Apartheid era, approximately 15 miles outside of Cape Town. The area had high unemployment, and students lived in crowded neighborhoods of largely government-supplied housing. The longitudinal trial began in 2004. Four schools were randomly selected to participate in the HealthWise South African program (n=913), and 5 schools served as matched controls (n=1,291). This study focused on the members of the randomized no-treatment control group. Participants were followed longitudinally from the beginning of eighth grade until the beginning of the ninth grade. Dynamic variables of interest tracked lifetime use of alcohol, cigarettes, dagga (cannabis), and inhalants. Latent transition analysis (LTA) was used to model stages of substance use as latent variables based on measurement of the dichotomous manifest indicators of lifetime substance use. LTA is useful for testing stage-sequential developmental models. LTA estimates three sets of parameters: the proportion of participants in each latent stage, the conditional probability of transitions between stages, and the measurement precision of the model. 5 tables and 40 references