THE NCS IS THE LARGEST VICTIMIZATION SURVEY CONDUCTED IN THE U.S. FIVE PERCENT OF ALL INCIDENTS REPORTED ARE SERIES INCIDENTS, CONSISTING OF THREE OR MORE SIMILAR INCIDENTS FOR WHICH THE VICTIM IS UNABLE TO PROVIDE SEPARATE DETAILS. A LONGITUDINAL FILE WAS CONSTRUCTED TO STUDY THE PATTERN OF CRIME INCIDENTS AT SPECIFIC ADDRESSES (NOT SPECIFIC HOUSEHOLDS) OVER A 3-YEAR PERIOD. THE SAMPLE WAS SELECTED FROM THOSE HOUSEHOLDS ENTERING THE NCS SURVEY IN THE MONTHS OF FEBRUARY, MARCH, JULY, AUGUST, AND SEPTEMBER, 1974. A TOTAL OF 664 ADDRESSES REPORTED 812 SERIES INCIDENTS OF ALL TYPES. OF THESE, 205 INVOLVED THE VIOLENT PERSONAL CRIMES OF RAPE, ROBBERY, AND ASSAULT. THESE CRIMES WERE CATEGORIZED BY THE JOB OF THE VICTIM (POLICE, BUS DRIVER, ETC.), DOMESTIC VIOLENCE, VIOLENCE BETWEEN CHILDREN, AND MISCELLANEOUS INVOLVING ADULTS WHO WERE NOT WELL ACQUAINTED. THE MAJOR TYPE OF CRIME IN ALL CATEGORIES WAS ASSAULT, WHICH COMPRISED 100 PERCENT OF THE JOB-RELATED CRIMES. SERIES CRIMES WERE CONSISTENTLY FOUND TO BE NONREPETITIVE (76 PERCENT REPORTED ONLY ONE SERIES OF INCIDENTS), NO MATTER WHAT THE CRIME. SERIES CRIMES TEND TO BE REPORTED NEAR THE BEGINNING OF THE HOUSEHOLD'S STAY IN THE NCS SAMPLE (9 PERCENT IN THE FIRST OR SECOND REPORTING PERIODS). IT WAS CONCLUDED THAT THE USE OF A SPECIAL QUESTIONNAIRE DEALING WITH SERIES CRIMES WOULD FACILITATE THE INCORPORATION OF SERIES DATA INTO THE REGULAR NCS CRIME STATISTICS. NOTES ON SAMPLING PROCEDURES, SEVEN TABLES COMPARING THE OVERALL RELATIONSHIPS AMONG THE FOUR GROUPS OF PERSONAL SERIES CRIMES, AND ONE REFERENCE ARE INCLUDED.