The research sought to test the hypothesis that adult criminality can be predicted from juvenile delinquency. The birth cohorts (1942, 1949, and 1955) comprised 6,127 persons, of whom 4,079 had essentially continuous residence in Racine, Wisconsin, from at least the age of 6 years until the cutoff date for data collection (1974 for the first two cohorts and 1976 for the third). Data were gathered from police records and court dispositions and from interviews with 889 persons from the 1942 and 1949 cohorts. The results showed that from 5 to 7 percent of the persons in each cohort were responsible for over half of the nontraffic police contacts. About 20 percent of each cohort's members were responsible for 80 percent of the nontraffic contacts generated. In addition, between 8 and 14 percent of the members of each cohort were responsible for all of the cohort's felonies, and about 5 percent committed 75 percent of the felonies. Although the probability of having an initial police contact was very high (more than 80 percent of the males in the 1942 and 1949 cohorts and over 70 percent in the 1955 group), over half of the males with a first contact discontinued having police contacts before having a fifth contact and half of the females before having a second contact. The existence of a hard core group of continuers suggests that the justice system should focus on a relatively small group of persons. The most prevalent pattern, however, was one of declining seriousness and of discontinuation of police contacts after the teenage period. Other findings relate to inner-city versus suburban continuity, the disposition of police contacts, the effectiveness of sanctions, and interview results concerning the normalcy of juvenile misbehavior and responses to community disapproval. In addition, the results of a multivariate analysis of the correlates of frequency and seriousness are provided. Neither a reference list nor data tables are included in the summary; these appear in the full final report, NCJ 77744.