The survival of 61,769 patients diagnosed as having colorectal cancer in the total Swedish population in 1960-1981 was analyzed. The 5-year relative survival rates during the total period were approximately 40%; patients with colonic carcinoma had a somewhat better prognosis than patients with rectal carcinoma, and men had a generally less favorable prognosis than women. Comparison of patients diagnosed in 1960-1964 and in 1975-1979 showed a slight improvement in survival during the more recent study period, i.e., an improvement rate of 7.0% for both cancer of the colon and cancer of the rectum. The improvement occurred mainly during the first months of follow-up. However, the excess mortality due to colorectal cancer 1-5 years after diagnosis as compared to the expected mortality in the general population remained virtually unaltered. A multivariate analysis in which adjustment was made for the possible confounding variables of sex, age, and site (of colonic carcinoma) gave consistent results. Reduced postoperative mortality was considered to be the most likely major explanation for the temporal trend toward a more favorable prognosis. Thus it seemed that the treatment of colorectal cancer in Sweden during the last two decades has not improved significantly except in terms of reduced postoperative mortality. Thus the challenge of improving survival in patients with colorectal cancer must be given high priority.