We studied myoglobin excretion in 33 patients admitted to the coronary care unit with a provisional diagnosis of acute myocardial infarction. Sixteen proved to have definite and uncomplicated acute myocardial infarction and 17 possible infarction, using WHO criteria. For 5 days after admission, aliquots of every urine specimen voided by each patient were analysed for myoglobin using an immunochemical method able to detect a minimum urinary myoglobin concentration of 0-02 mg/ml. Myoglobinuria was detected in 14 of the 16 patients with definite infarction but was not found in any of the 17 patients with possible infarction. There were 3 patterns of myoglobin excretion. In 8 of the 14 patients it was excreted in one episode starting 10 to 40 hours after the onset of chest pain and lasting for 5 to 45 hours. In 3 of the remaining patients it was excreted over a much longer period (mean 83 hours) and in the final 3 patients myoglobinuria occurred in 2 or 3 intermittent episodes with periods of between 10 and 20 hours during which it was not detected. Total myoglobin excretion, which varied between 2 and 100 mg (mean 51 mg), did not correlate with peak serum enzyme levels. We concluded that in the appropriate clinical setting, the finding of myoglobinuria provides additional evidence for a diagnosis of acute myocardial infarction. The variable myoglobin excretion pattern suggests that in seemingly uncomplicated myocardial infarction there is considerable variation between patients in the pattern of evolution of the infarction process. This may be relevant to the assessment of measures directed towards limiting infarct size.