THE ANALYSIS DISTINGUISHED VICTIMS KILLED BY PERSONS THEY KNEW FROM THOSE KILLED BY STRANGERS, SEPARATED HOMICIDES THAT OCCURRED IN A HOME AND HOMICIDES THAT OCCURRED IN A PUBLIC PLACE, AND EXAMINED HOMICIDE RATES FOR BOTH BLACKS AND WHITES LIVING IN AREAS OF LOW AND HIGH SOCIOECONOMIC STATUS. STUDY FINDINGS REVEALED THREE IMPORTANT TRENDS: LARGE INCREASES IN HOMICIDE RATES FOR RESIDENTS OF BOTH RACES, A RELATIVE SHIFT IN THE PLACE OF OCCURRENCE OF HOMICIDES FOR WHITE ATLANTANS FROM THE HOME TO PUBLIC PLACES, AND AN INCREASE IN THE FREQUENCY WITH WHICH HOMICIDES WERE COMMITTED WITH FIREARMS: HOMICIDE RATES FOR BLACKS AND WHITES IN 1971-72 WERE HIGHEST IN CENSUS TRACTS WITH LOW INDICES OF SOCIOECONOMIC STATUS: THIS ASSOCIATION WAS FOUND ONLY FOR HOMICIDES IN WHICH THE VICTIM AND ASSAILANT WERE RELATIVES OR ACQUAINTANCES. THERE WAS LITTLE GEOGRAPHIC OVERLAP BETWEEN AREAS WITH HIGH RATES OF HOMICIDE IN THE HOME AND HIGH NUMBERS OF HOMICIDES IN PUBLIC. IT IS SUGGESTED THAT USING POPULATION-BASED RATES FACILITATES COMPARISON OF HOMICIDE OCCURRENCE BETWEEN TIME PERIODS AND/OR PLACES. REFERENCES ARE INCLUDED. (AUTHOR ABSTRACT)