Census information was gathered on 24 metropolitan areas having a population greater than 100,000. The dependent study variable was the total reported homicide rate per 100,000 population. Main indicators of social organization included migration, marital dissolution, urbanization, racial and ethnic heterogeneity, and age composition. Trends in average metropolitan area homicide rates revealed a substantial increase from 1967-1971 to 1972-1976 and a marginal increase from 1972-1976 to 1977-1981. Several social-structural variables changed markedly over time. There was a dramatic shift in family income dissimilarity in metropolitan areas as compared to average Canadian family incomes. Employment increased in variability, but the effects of inequality and social disorganization variables on homicide were inconsistent. Of all social disorganization variables, only "percent divorced" showed a significant relationship and only in 1976. The authors conclude there is a convergence in murder rates between eastern, central, and western Canada in areas where higher levels of inequality and social disorganization exist. This finding, however, is not consistent over time. An appendix tabulates homicide rates for the 24 metropolitan areas. 53 references and 2 tables (Author abstract modified)