We are presenting a detailed 20-year survival analysis by stage, age, calendar period of diagnosis, and treatment for cases of breast cancer diagnosed in the State of Connecticut between 1935 and 1954, and are including a review of statistical methodology in survival analysis. Observed survival is measured relative to expected survival in the general population. The 5-year relative survival rate for cases diagnosed during the first 10-year calendar period, 1935–44, was 50 percent. Mortality among the 5-year survivors was far in excess of that expected, based on general population experience. The 5-year relative survival rate for cases diagnosed during the second 10-year calendar period, 1945–54, was 56 percent. There was, however, no evidence of improvement in the 10-year survival rate. In general, there appeared to be little variation in the agespecific survival rates, except that among localized cases, women 75 to 84 years of age had a somewhat more favorable experience than women under 75, and among cases with regional spread, women under 35 had a lower survival rate than older women. The survival results in treatment of breast cancer in Connecticut are similar to those obtained in several well-known treatment centers.