Mortality data for the years 1981-1984 have been analyzed by marital status, sex and residence area. Results show high levels of mortality in widowers and widows, especially in areas with high mortality risks. In stomach cancer the widowers/widows mortality excess, compared to married and unmarried, is constant all over the country and does not change in the areas with high levels of mortality. Also for liver cancer mortality the widows show, everywhere, a constant excess compared to unmarried women. Widowers and unmarried men, if resident in north-west, show similar rates; while mortality levels for unmarried men approach those of married ones as the latitude of residence areas decreases and in the south they are equal. Mortality by ages confirms differences by marital status and shows, for many sites, larger variation in the age classes younger than 65-69 years.