STUDY OBJECTIVE: To define the contribution of domestic violence (DV) to homicides in women in New Mexico and to examine differences in ethnicity, mechanism, previous documented injuries, incidence of sexual assault, and use of alcohol or illicit drugs between DV- and non-DV-related homicides. METHODS: We carried out a retrospective analysis of reports of the state office of the medical investigator (OMI) reports from all female homicides from 1990 to 1993 in New Mexico. A homicide was defined as being related to DV if the perpetrator was a current or former male intimate partner. The chi-squared and Mann-Whitney tests were used to analyze data. RESULTS: The OMI investigated 134 homicides in women for an overall fatality rate of 4.3 per 100,000. A male intimate partner was the perpetrator in 62 cases (46%). The rate of DV homicide among American Indians (4.9 per 100,000) was significantly higher than that among Hispanics (1.7) and non-Hispanic whites (1.8)(RR=2.8; 95% confidence interval (CI), 1.5 to 5.1). Firearms were almost two times as likely to be used in DV homicides as in non-DV homicides (RR=1.8; 95% CI, 1.2 to 2.6). Evidence of old injuries was found more often in DV homicide cases (35.5%) than in non-DV cases (83%) (RR=4.3; 95% CI, 1.8 to 9.8). The presence of alcohol or other drugs was higher among non-DV homicide victims (69%) than DV homicide victims (54.3%) (P=.03). CONCLUSION: American Indian women are at particularly high risk of homicide, including DV homicide. Firearms were overrepresented in DV homicides, suggesting that removing firearms from the homes of previous DV perpetrators would be a useful public health strategy. Alcohol or illicit drugs were found in approximately two thirds of New Mexico women who were victims of homicide. The high prevalence of history of previous injuries among DV homicide victims indicates that early identification of DV victims in the emergency department and other health care settings is an important point of intervention.