TWO MODELS THAT HAVE APPEARED IN RESEARCH LITERATURE ARE REPLICATED, USING THIS CLASSIFICATORY SCHEME. DATA ANALYZED WERE FOUND IN U.S. GOVERNMENT REPORTS, THE EFFECTS OF PARTITIONING HOMICIDE BY TYPE ON THE ESTIMATION OF THE DETERRENCE AND ETIOLOGICAL MODELS WERE EXPLORED. PRIMARY HOMICIDE INVOLVES FAMILY OR ACQUAINTANCES AND IS USUALLY AN ACT OF PASSION, WHILE NONPRIMARY HOMICIDES ARE OFEN INSTRUMENTAL IN NATURE SINCE THEY GENERALLY OCCUR IN THE COMMISSION OF ANOTHER CRIME (GANGLAND SLAYINGS, FELONY HOMICIDE, INSTITUTIONAL HOMICIDE). THE TWO TYPES OF HOMICIDE CAN BE VIEWED AS HAVING DIFFERENT ORIGINS: PRIMARY HOMICIDES BEING ESSENTIALLY UNREASONED ACTS OF PASSION AND NONPRIMARY HOMICIDE INVOLVING SOME DEGREE OF DECISIONMAKING OR MEDITATION. STUDY FINDINGS SHOWED LITTLE SUPPORT FOR THE DETERENCE MODEL AND CONSIDERABLE SUPPORT FOR THE ETIOLOGICAL MODEL. STATE PRIMARY-HOMICIDE RATES WERE FOUND TO BE RELATED TO POVERTY AND TO THE PERCENTAGE OF THE POPULATION AGED 20-34, WHILE NONPRIMARY HOMICIDE RATES WERE SIGNIFICANTLY RELATED ONLY TO THE PERCENTAGE OF THE STATE POPULATION LIVING IN URBAN AREAS. THE ANALYSIS SUGGESTS THAT PUNISHMENT APPEARS TO HAVE LITTEL EFFECT ON THE MAJORITY OF HOMICIDES SINCE THEY TEND TO OCCUR UNDER CIRCUMSTANCES IN WHICH LAWS, PRACTICES, AND PERCEPTIONS HAVE LITTLE IMPACT. RESULTS SHOWED THAT WHILE THE VARIABLES COMMONLY USED TO PREDICT HOMICIDE RATES WERE SUCCESSFUL IN PREDICTING PRIMARY HOMICIDE RATES, THEY WERE LESS SUCCESSFUL IN ACCOUNTING FOR NONPRIMARY HOMICIDE RATES. FURTHER STUDY IS NECESSARY TO DEVELOP A GREATER UNDERSTANDING OF THE FACTORS CONTRIBUTING TO NONPRIMARY HOMICIDE. TABULAR DATA, NOTES, AND REFERENCES ARE INCLUDED. (MJW)