Agnew's strain theory has life-course features in viewing anger resulting in delinquency as an early phase of a stress process and delinquency itself as an additional source of distress as a result of the responses it provokes from others. The subsequent stages of this stress process may take gender-specific forms that involve internalized feelings of depression among females and externalized drinking problems among males. In order to test this theory, the data had to be longitudinal; include male and female adolescents; and incorporate meaningful measures of anger, juvenile delinquency, and later adolescent drinking problems and depression. Ideally, such data would be nationally representative of adolescents and track these youth to emerging adulthood. Data that meet the teenage phase of these requirements are available in the first two waves of the National Longitudinal Study of Adolescent Health (Add Health), which tracks youth to the emergence of adulthood. The data analysis was ultimately focused on determining the role of delinquency as an additive and mediating factor in models of the anger, depressive symptoms, and drinking problems of male and female adolescents. The study considered the antecedents of anger and links between anger and other factors, with delinquency as the subsequent mediating variable. The findings are consistent with the integrated theoretical account proposed and parallel previous studies in indicating that anger follows from disadvantaged and stressful family circumstances and is a source of delinquent behavior. The findings further indicate that anger is a source of depression among both adolescent boys and adolescent girls. Anger leads to drinking problems among males, but not among females; for males, involvement in delinquency is a mediating link between anger and drinking. The connecting link delinquency has in the transition from anger to depression and drinking may be especially important for understanding the cumulative and gendered nature of downward trajectories in life, leading, for example, to variation in the kinds of marital and employment outcomes that Sampson and Laub (1993, 1997) have studied. Suggestions are offered for future research. 5 tables, 4 notes, 73 references, and appended items used in full, violent, and nonviolent delinquency scales, as well as descriptive statistics and variables description