A multicenter randomized single-blind study was performed to evaluate the effects of propranolol administered during the evolution of myocardial infarction. Five centers enrolled a total of 269 patients, with 134 receiving propranolol and 135 placebo. Propranolol or placebo was given intravenously upon randomization (0.1 mg per kilogram of body weight) and then orally for nine days to keep the heart rate between 45 and 60 beats per minute. Less than 2 per cent of patients were treated within 4 hours after the onset of symptoms, but 50 per cent received therapy within 8 hours of onset of chest pain, and the remainder between 8 and 18 hours. The heart rates in the propranolol-treated group were significantly lower than those in the placebo group (P less than 0.001). Base-line characteristics, including the mean heart rate (79.6 vs. 81.3) and the left ventricular ejection fraction (49.0 vs. 49.5), were similar in the two groups. The primary end point evaluated--infarct size as estimated from plasma MB creatine kinase activity--was virtually identical in the two groups, averaging 13.3 and 13.6 gram-equivalents of MB creatine kinase per square meter of body-surface area. Peak plasma levels of the enzyme were also similar in the two groups. No significant difference was observed between the propranolol and placebo groups in the change in left ventricular ejection fraction, extent of area involved in pyrophosphate uptake, R-wave loss on electrocardiograms, or mortality (after three years). These results do not support the use of propranolol administered four or more hours after the onset of symptoms to limit infarct size.