Data were obtained on all homicide victims who were drug users in Philadelphia in the years 1969-1972. The sample also included a random population of homicide victims who were not drug users. The findings showed that, while death by homicide is similar in some respects for drug users and nonusers, there are significant differences in the circumstances of their deaths. Drug users were more likely to be murdered because of their involvement with drugs, attempts to secure drugs, or attempts to secure money to buy drugs. As drug use increases, along with antidrug efforts, it is likely that the rates of both robbery and homicide among the drug-using population will increase. 5 tables and 27 references