This study analyzed existing data sets: three longitudinal birth cohorts for 1942, 1949, and 1955 (6,127 persons, of whom 4,079 had continuous residence in Racine, Wis.). Data encompass official police contact records, referral records, and court dispositions for all cohorts. All police contacts were coded according to 26 categories of offenses. Age at first felony or major misdemeanor before or after age 16 could not predict who would have later police contacts or the number of contacts they will have for major misdemeanors, felonies, or Part I offenses. Juvenile offenders with the more extensive criminal records did, as a group, account for a disproportionate share of the delinquent and subsequent adult criminal behavior, but the serious juvenile offender subsequently accounts for only a small portion of adults who commit serious offenses. These findings argue against any attempt by the criminal justice system to selectively incapacitate serious juvenile offenders so as to preclude their committing serious crimes in the future. The future behavior of individual juvenile offenders cannot be accurately predicted regardless of their early criminal records, and should such prediction be possible, it would address only a small portion of juveniles who subsequently commit serious crimes as adults. Tabular study data, 6 references, and a 29-item bibliography (on career criminals, prediction, and problems in manadtory sentencing) are provided.