INFERENCES MADE ON THE BASIS OF COMPARING THE UCR AND NCS ARE LEGITIMATE ONLY IF IT IS KNOWN TO WHAT EXTENT THEY ARE MEASURING THE SAME CRIMES. THE RELATION OF UCR AND NCS CRIME RATES WAS ESTIMATED BY CALCULATING THE NCS CRIME-SPECIFIC RATES IN 26 AMERICAN CITIES TO BE DEFINITIONALLY AS SIMILAR TO UCR RATES AS POSSIBLE. THE CORRELATION BETWEEN THE RATES FOR EACH TYPE OF CRIME WAS THEN MEASURED. WHILE NCS DATA WAS BASED ON RANDOM SAMPLES OF HOUSEHOLDS AND BUSINESSES WITHIN SELECTED CITIES, UCR DATA WAS BASED ON PERSONS WHO HAD REPORTED CRIMES TO THE POLICE WITHIN EACH CITY. MOST SIGNIFICANTLY, PERSONS LIVING OUTSIDE A CITY'S BOUNDARIES COULD AFFECT UCR RATES BY BEING VICTIMIZED WITHIN THE CITY, BUT COULD NOT AFFECT NCS RESIDENTIALLY BASED CRIME RATES. THE STUDY FOUND THAT NCS RATES WERE MODERATELY TO STRONGLY RELATED TO UCR RATES FOR VICTIMIZATIONS CHARACTERIZED BY THEFT BUT WERE UNRELATED FOR PERSONAL VICTIMIZATIONS NOT CHARACTERIZED BY THEFT. ECOLOGICAL CORRELATES SUGGESTED THAT THE INDIVIDUAL CORRELATES OF BEING VICTIMIZED FOR NCS AND UCR TYPES OF DATA WERE LIKELY TO BE MORE SIMILAR FOR THEFT THAN FOR PERSONAL CRIMES WITHOUT THEFT. THIS INFERENCE, HOWEVER, CANNOT BE EASILY TESTED UNTIL MORE INFORMATION ON THE INDIVIDUAL CHARACTERISTICS OF CRIME VICTIMS ARE PUBLISHED IN THE UCR. THIS PAPER IS ONE OF 30 PRESENTED ON VICTIMOLOGICAL ISSUES AT THE 1978 MEETINGS OF THE AMERICAN SOCIETY OF CRIMINOLOGY. REFERENCES ARE GIVEN. (MJW)