Frequencies of homicide victimization of wives and husbands, while cohabiting and when separated, are reported for all spousal homicides known to the police in Canada (1974-1990), in New South Wales, Australia (1968-1986), and in Chicago (1965-1990). In all three data sets, the degree to which spousal homicide victimization was female-biased was significantly greater when the couple were estranged than when they were coresiding. Victim counts and population-at-large estimates of coresiding and separated now-married spouses were combined to estimate differential homicide rates incurred by coresiding and estranged married persons. Wives in all three countries incurred substantially elevated risk when separated as compared to when coresiding.