A group of 408 catheterised patients who had mild angina or myocardial infarction without angina was selected in conformity with the criteria for entry into a previously reported randomised trial. Medical treatment had been chosen initially by the cardiologist, referring physician, or the patient, although 27% had late operation. Five year survival rates were 91% and 72% for mild angina with high or low ejection fractions and 85% for those who had myocardial infarction without subsequent angina. Survival rates were 95%, 88%, and 80% for one, two, and three artery disease respectively. For patients who had ejection fractions of at least 0.50, five year survivals were 95%, 89%, and 83% for one, two, and three artery involvement respectively. Good left ventricular function, single artery disease, and a short history were favourable prognostic variables in multivariate analysis of patients who had angina pectoris. Statistical methods of dealing with patients who had late operation influenced calculated survival, especially for patients at relatively high risk. The lower survival rates for the whole group and most subsets compared with survival rates in the randomised trial may be of clinical importance.