Arrest data for individual cities were obtained from the FBI, and arrests for homicides were disaggregated according to the Uniform Crime Reports data structure. Demographic-specific homicide arrest rates were calculated for the years 1969 to 1971, and the arrest/offense ratio was introduced to control for possible criminal justice system effects on the arrest rate. Analysis indicates that the percentage of blacks in the population has strong effects on aggregate homicide and age- and sex-specific homicide rates, but has no effect on black and white homicide once other structural characteristics are considered. Both the percentage of blacks in the population and population size have strong effects on the number of aggregate murders, which in turn affects white homicides. Additionally, racial income inequity and unemployment generally have a negative effect on demographic-specific homicide; whereas, poverty has a positive effect. All else being equal, black homicide rates tend to increase as black and white incomes converge. In conclusion, the data suggest that blacks in cities with large black ghettos do not have higher violent offending rates than blacks in cities with a small black population. Two appendixes and 64 references are included.