The boys, who were initially in the first, fourth, and seventh grades of public schools, were monitored at regular intervals with low attrition. The study design, methods, and measures are described. Findings address the prevalence, frequency, and onset of various severity levels of self-reported and adult-reported delinquency, as well as court records of offending. Documentation of a typology of offenders identified the following categories: "stable nondelinquents," "starters," "stable moderates," "escalators," "stable highs," "de- escalators," and "desisters." Stability figures for delinquency and aggression are presented. Longitudinal findings show various pathways to serious delinquency, starting from nondelinquent problem behaviors. Results are presented on child risk factors for delinquency, including impulsivity, intelligence, personality, attitudes, and drug dealing. Other risk factors reviewed are hyperactivity, academic factors, family factors, macro factors (socioeconomic factors, neighborhood), and peer factors. The best predictors of self-reported violence and court- referred violence are presented. Results are reviewed on the clustering of problem behaviors, the cumulative effect of risk factors, and differences in risk factors with age. Findings on protective factors are also reported. The paper summarizes findings on parents' help-seeking for their boy's problem behaviors and delinquency, including the timing of such help- seeking in relation to the boy's delinquency career and contact with the juvenile court. Finally, the paper reviews several methodological issues, including dichotomization of constructs, the validity of measures, the yield of information from multiple informants, and comparisons between informants. Conclusions and priorities for future work are highlighted in this ongoing study. 1 table, 4 figures, and 58 references