THE SUBJECTS WERE REFERRED FROM SCHOOLS IN THE BOSTON, MASS., AREA, FROM PRIVATE PHYSICIANS, AND FROM JUVENILE COURT AUTHORITIES. RANGING IN AGE FROM 5 1/2 TO 14 1/2 YEARS, THE GROUP INCLUDED 3 GIRLS AND 14 BOYS, ALL OF WHOM SUFFERED PERIODIC EPISODES OF MISBEHAVIOR CHARACTERIZED BY VIOLENT OUTBURSTS. IN ADDITION, SEVERAL HAD HISTORIES OF FIRE-SETTING, CRUELTY TO ANIMALS, HOMICIDAL ASSAULTIVENESS, STEALING, AND BIZZARRE FUGUE-LIKE STATES. THE SUBJECTS WERE ADMINISTERED PSYCHOMETRIC AND NEUROLOGICAL TESTS, ALONG WITH ROUTINE ELECTROENCEPHALOGRAMS, AND WERE THEN PLACED ON A DOUBLE-BLIND PLACEBO-CONTROLLED TRIAL OF DIPHENYLHYDANTOIN (DPH). BLOOD LEVELS OF DPH WERE OBTAINED, AS WERE SELECTED POSTTRIAL MEASURES, INCLUDING THE CONTINUOUS PERFORMANCE TEST, THE PORTEUS MAZE TEST, AND PARENT QUESTIONNAIRES. ELEVEN SUBJECTS WERE FOLLOWED WITH SCHOOL QUESTIONNAIRES. NO STATISTICALLY SIGNIFICANT GROUP CHANGES WERE ATTRIBUTABLE TO THE DRUG EFFECT. TABULAR DATA AND REFERENCES ARE INCLUDED. (KBL)