TWO CONTRASTING HYPOTHESES WERE EXAMINED: 1) THAT FEMALES WHO ARE DRUG USERS SHOULD HAVE GREATER RISKS OF BEING KILLED THAN NON-DRUG-USERS, AND THAT THE CIRCUMSTANCES OF THEIR DEATHS WOULD REFLECT THIS PARTICIPATION, AND 2) THAT THE FEMALE ROLE WILL BE THE DOMINANT FACTOR IN DETERMINING FEMALE PATTERNS OF HOMICIDE, AND THEREFORE NO SIGNIFICANT DIFFERENCES WOULD BE FOUND IN THE CIRCUMSTANCES OF DEATH AMONG DRUG-USING AND NON-DRUG-USING FEMALES. DATA WERE DRAWN FROM THE RECORDS OF THE MEDICAL EXAMINER IN PHILADELPHIA. RECORDS OF ALL ILLEGAL DRUG USING HOMICIDE VICTIMS WERE COMPARED WITH A RANDOM SAMPLE OF NON-DRUG-USING VICTIMS. DATA WAS COLLECTED ON THE PLACE OF DEATH, IMPLEMENT USED TO KILL, THE ASSAILANT, THE REASON FOR DEATH, AND THE INCIDENT PRECIPITATING THE DEATH. DEMOGRAPHIC CHARACTERISTICS OF VICTIMS WERE ALSO DETERMINED. THE AUTHOR STATES THAT BOTH HYPOTHESES RECEIVED SUPPORT AS A RESULT OF THE DATA. IT IS CONCLUDED, HOWEVER, THAT THE IMPORTANCE OF THESE FINDINGS CAN ONLY BE ASSESSED WHEN MORE IS KNOWN ABOUT HOW SEX ROLE CONSTELLATIONS WORK IN DRUG-USING GROUPS. (AUTHOR ABSTRACT MODIFIED)