THE RESEARCH PROJECT, WHICH WAS STARTED IN 1972, CONSISTS OF THREE PARTS: (1) THE STATISTICAL ANALYSIS OF ALL REPORTED CHILD OFFENDERS IN FULDA BETWEEN 1962 AND 1971, (2) THE QUESTIONING OF 560 REPRESENTATIVE STUDENTS TO EXPLORE UNREPORTED OFFENSES, AND (3) A SEPARATE INVESTIGATION OF REPEAT OFFENDERS. THE ANALYSIS OF REPORTED JUVENILE DELINQUENCY SHOWED 1,775 OFFENDERS FROM THE AGES OF 3-13. OF THE OFFENDERS, 90 PERCENT (94 PERCENT FOR REPEAT OFFENDERS) WERE MALE. OF THE REPEAT OFFENDERS, 40 PERCENT (31 PERCENT FOR ONE-TIME OFFENDERS) CAME FROM A LOWER-CLASS BACKGROUND. THEFT, AGGRAVATED LARCENY, AND VANDALISM WERE FOUND TO BE THE PREDOMINANT OFFENSES. THE ANALYSIS OF THE DARK FIGURES QUESTIONNAIRE REVEALED THAT THE PERCENTAGE OF UNREPORTED JUVENILE CRIME IS EXTREMELY HIGH, BUT THAT GIRLS CONFESSED TO FEWER UNREPORTED OFFENSES THAN BOYS. LOWER-CLASS LOWER-CLASS CHILDREN HAD COMMITTED MORE UNREPORTED OFFENSES THAN MIDDLE-CLASS CHILDREN. SUPPLEMENTARY PSYCHOLOGICAL TESTING (THE FREIBURG PERSONALITY INVENTORY) SHOWED NO SIGNIFICANT CONNECTION BETWEEN PARTICULAR PERSONALITY TRAITS AND THE NUMBER OR TYPE OF SELF-REPORTED OFFENSES. HOWEVER, CHILDREN HAVING PSYCHOPATHOLOGICAL ABNORMALITIES RATED CONSIDERABLY HIGHER IN UNREPORTED CRIMES. THE STUDY OF REPEAT OFFENDERS, WHICH IS STILL IN PROGRESS, IS DESIGNED TO TEST WHETHER SIGNIFICANT CHARACTER DIFFERENCES BETWEEN NONOFFENDERS AND REPEAT OFFENDERS ALREADY EXIST BEFORE THE OFFENSE OR WHETHER THESE DIFFERENCES DEVELOP LATER AS A RESULT OF THE LABELING PROCESS. PRELIMINARY RESULTS OF THE REPEAT OFFENDER PROJECT SHOW THAT (1) EARLY DELINQUENCY IS NO DEFINITE INDICATOR FOR ADULT OFFENSES; (2) CHILD OFFENDERS HAVING COMMITTED A VARIETY OF OFFENSES ARE MORE PRONE TO BECOME ADULT OFFENDERS; AND (3) CHILDREN HAVING PSYCHOLOGICAL ABNORMALITIES AND COMING FROM A DISADVANTAGED BACKGROUND ARE A PARTICULARLY ENDANGERED GROUP. THE ARTICLE INCLUDES NUMEROUS CHARTS AND A BIBLIOGRAPHY. --IN GERMAN. (SAJ)