THE ANALYSIS ENCOMPASSES TWO LEVELS: A MACROLEVEL (CENSUS TRACT) EXAMINATION OF HOMICIDE RATES; AND A MICROLEVEL EXAMINATION OF INDIVIDUAL CASES OF HOMICIDE. THE STUDY SAMPLE INCLUDES ALL 1, 871 HOMICIDES (EXCEPT EXCUSABLE AND JUSTIFIABLE HOMICIDE) RECORDED BY THE DETROIT POLICE DEPARTMENT FROM JANUARY 1, 1970, TO JANUARY 1, 1973. MAJOR FINDINGS ARE AS FOLLOWS: BLACK HOMICIDE OFFENDERS WERE INVOLVED IN HOMICIDES BETWEEN STRANGERS MORE FREQUENTLY THAN WHITE HOMICIDE OFFENDERS; THE FREQUENCY OF HOMICIDES COMMITTED BY FEMALES WAS GREATER FOR BLACKS THAN WHITES; RACE WAS RELATED TO USE OF WEAPONS OTHER THAN HANDGUNS IN HOMICIDE; JUVENILES WERE MORE LIKELY THAN ADULTS TO USE WEAPONS OTHER THAN HANDGUNS IN HOMICIDES; BLACKS WERE MORE LIKELY THAN WHITES TO BE INVOLVED IN HOMICIDES WITH MORE THAN ONE OFFENDER. THE MACROANALYIS SHOWS A MODERATE RELATIONSHIP BETWEEN THE PERCENTAGE OF BLACKS IN A CENSUS TRACT AND THE TRACT'S HOMICIDE RATE, AND A STRONGER CORRELATION BETWEEN HOMICIDE RATES AND PERCENTAGE OF CENSUS TRACT RESIDENTS WITH INCOMES BELOW THE POVERTY LEVEL. THE MODERATE RELATIONSHIPS BETWEEN HOMICIDE AND THE VARIABLES EXAMINED ARE SAID TO INDICATE THAT SUBCULTURES REFER NOT SO MUCH TO REGIONAL STRUCTURES OR SUBSOCIETIES AS TO THE TENDENCY OF PEOPLE WHO LIVE TOGETHER TO SHARE A SET OF VALUES TOWARD LIFE AND TO TRANSMIT THOSE VALUES TO SUCCEEDING GENERATIONS. THE DATA ALSO SUGGEST THAT THERE IS NOT A SINGLE, GENERALIZED TYPE OF SUBCULTURE CHARACTERIZED BY HIGHER RATES OF VIOLENCE, BUT RATHER A VARIETY OF SUBCULTURES, SOME MORE VIOLENT THAN OTHERS. A THIRD IMPLICATION IS THAT VIOLENCE IN MANY SITUATIONS IS FUNCTIONAL AND SERVES TO PROVIDE SUPPORT FOR EXISTING NORMS. FURTHER IMPLICATIONS ARE DISCUSSED. SUPPORTING DATA AND A BIBLIOGRAPHY ARE PROVIDED. (LKM)