DATA WERE OBTAINED, COVERING THE PERIOD BETWEEN 1970 AND 1975, FOR 55 CITIES THROUGHOUT THE UNITED STATES WITH 1970 POPULATIONS ABOVE 250,000. VARIABLES WERE INTERCORRELATED USING PRODUCT-MOMENT CORRELATION. MURDER RATES OF POLICE OFFICERS WERE COMPARED FOR 1970 TO 1972 AND 1973 TO 1975 BUT FAILED TO SHOW SIGNIFICANCE AS MURDER OF POLICE IS RELATIVELY RARE AND FLUCTUATES WIDELY FROM YEAR TO YEAR. ANALYSIS USING THE MURDER RATE FOR THE ENTIRE 6-YEAR PERIOD INDICATED A SIGNIFICANT CORRELATION WITH 3 URBAN VARIABLES: POLICE MURDER RATES WERE HIGHER WHERE THE PROPORTION OF BLACKS IN THE POPULATION WAS HIGHER, WHERE THE PROPORTION OF POPULATION BELOW THE POVERTY LEVEL WAS HIGHER, AND WHERE THE CRIME INDEX WAS HIGHER. ALTHOUGH STATISTICALLY SIGNIFICANT, THESE RELATIONSHIPS WERE ONLY MODERATE IN STRENGTH. IT IS SUGGESTED THAT VARIABLES WITHIN POLICE FORCES SUCH AS AGE, TRAINING, RACIAL STRUCTURE, AND FORMAL EDUCATION MIGHT CORRELATE WITH THE RATES OF MURDER AND ASSAULTS. HOWEVER, SUCH DATA ARE NOT READILY AVAILABLE. COLLECTION OF SUCH DATA MAY PROVE USEFUL FOR FURTHER RESEARCH. POLICE MURDER RATES IN THE CITIES STUDIED AND CORRELATIONS BETWEEN CITY VARIABLES AND MURDER RATES ARE PRESENTED IN TWO TABLES. (JAP)