suMMARY Nine hundred and sixty-nine coronary care patients with acute myocardial infarction were followed for one year. Atrial fibrillation was documented in 107 patients. Compared with patients without atrial fibrillation, those with this arrhythmia were older, had clinically more severe infarction, and had a higher frequency of ventricular fibrillation or tachycardia, and right bundle-branch block. They had similar past histories of ischaemic heart disease and coronary risk factors. Patients with atrial fibrillation had a higher total mortality at 3 months and 12 months. The presence of atrial fibrillation was not associated with any significant increase in mortality within any decade of age or within any subgroup of clinical severity of infarction. The frequency of atrial fibrillation was similar in anterior and inferior infarction. Multiple episodes of atrial fibrillation occurred in 52 patients and episodes usually lasted for over 1 hour. In 50 per cent of patients with single episodes of atrial fibrillation the initial ventricular rate was greater than 120 beats per minute.The association of atrial fibrillation with acute myocardial infarction is known to carry a high mortality (Stannard and Sloman, 1967;Klass and Haywood, 1970;Helmers et al., 1973;Cristal et al., 1976;Liberthson et al., 1976). It is unlikely that this high mortality is the result of the arrhythmia itself but is more likely the result of the clinical situation in which atrial fibrillation develops. Thus atrial fibrillation is usually associated with older patients and with clinically more severe infarction (Stannard and Sloman, 1967;Klass and Haywood, 1970;Helmers et al., 1973;Cristal et al., 1976;Liberthson et al., 1976 as necessary and intranasal oxygen given at a rate of 2 1/min for the first 12 hours. The patients' rhythms were displayed on multichannel oscilloscopes in a central nurses' station. Ten-second electrocardiograph strips were recorded every hour for the first 2 days and more often if arrhythmias were detected. A 4-channel tape recorder was available for part of the period under study. Blood was taken and measurements made of creatine kinase and aspartate transaminase on admission and then daily for at least 2 days. Routine 12-lead electrocardiograms and 2 metre posteroanterior chest radiographs with the patient sitting on the side of the bed were made on admission, daily for the first 2 days, and then according to need.