ABSTRACT. Objective: The aims of this study were to examine the time-varying developmental associations between conduct problems and early alcohol use in girls between ages 11 and 15 and to test the moderating role of race. Method: The study is based on annual, longitudinal data from oldest cohort in the Pittsburgh Girls Study (n = 566; 56% African American, 44% White). Two models of the association between conduct problems and alcohol use were tested using latent growth curve analyses: conduct-problem-effect (conduct problems predict time-specifi c variation in alcohol use trajectory) and alcohol-effect (alcohol use predicts time-specifi c variation in conduct problem trajectory) models. Results: Girls' conduct problems and alcohol use increased over ages 11-15.Results provided support for a conduct-problem-effect model, although the timing of the associations between conduct problems and alcohol use differed by ethnicity. Among White girls, conduct problems prospectively predicted alcohol use at ages 11-13 but not later, whereas among African American girls, prospective prediction was observed at ages 13-14 but not earlier. Conclusions: Study fi ndings indicate developmental differences in the time-varying association of conduct problems and alcohol use during early adolescence for African American and White girls. Ethnic differences in the development of alcohol use warrant further study, and have potential implications for culture-specifi c early screening and preventive interventions. (J. Stud. Alcohol Drugs, 71, 544-553, 2010)