The homicide rates for various immigrant groups in Australia were calculated, and the influence of the rate in countries of origin on immigrant rates and the relationship between the birthplace of the accused and victim were examined. Age-sex standardised mortality ratios (SMRs) for homicide were calculated for the immigrant groups, based on 1984-1988 mortality data and 1986 census data. The Australian-born homicide rates were used as the standard. Standardised mortality ratios for countries of origin were derived from WHO data. A cross-tabulation of the birthplaces of the accused and the victim was compiled from 1989-1992 police records. Male SMRs ranged from 0.13 (P < 0.01) for immigrants from Africa and the Americas to 5.83 (P < 0.05) for Koreans. Several female groups had lower SMRs than the Australian-born, although none of these differences were significant. Indonesian females had the highest SMR (5.32, P < 0.01). There was a positive Spearman correlation between the ranking of homicide rates for the origin populations and the immigrants (males 0.64, P < 0.05; females 0.62, P < 0.05). Overall, 51.3 per cent of immigrants were killed by their compatriots. This ranged from zero for New Zealanders to 100 per cent for immigrants from the Middle East. In order to further identify factors contributing to the large differences in rates it is imperative to have information on the victim, the perpetrator and the circumstances surrounding the murder.