Approximately 1,000 cases referred to the London Family Court Clinic over a 6-year period were reviewed, and a sample of 99 juveniles who had committed dangerous offenses in which physical harm was inflicted on the victim was obtained. A comparison group of 91 juveniles not charged with person offenses was matched to the first group according to three criteria: sex, age at the time of the referral, and time period when referred. The 190 offenders were further divided into four groups based on two factors: (l) the nature of the offense committed (dangerous or not dangerous) and (2) the clinical rating of dangerousness (dangerous or not dangerous). The four groups were compared on the basis of variables that included family background, behavioral symptomatology, court systems' response, and school-related variables. Chi-square statistics were computed for all the variables, producing a substantial number of variables which significantly differentiated the four groups. Results suggest that in the area of family background variables, only two variables (the existence of parental abuse and history of numerous out-of-home placements) differed significantly. Findings indicate that these groups were not homogeneous with respect to their personal histories or their potential for dangerous behavior. A clinical rating of dangerousness was found to be associated with numerous school difficulties, specific behavioral symptomatology, and clinical personality symptoms. Finally, results suggest that intervention strategies for disturbed, dangerous young offenders will be frustrated by the very nature of their problems. Three tables and 31 references are included.