We have used a sensitive direct radioimmunoassay to study the effects of exercise on plasma atrial natriuretic peptide (ANP) concentrations in man. Plasma ANP concentration increased three-fold in sixteen patients undergoing bicycle ergometer electrocardiographic tests for the investigation of chest pain. Resting ANP concentrations were higher in those patients in whom there was more evidence of heart disease, such as a positive exercise test, treatment with a beta blocker or history of myocardial infarction, although exercise resulted in increased ANP in both groups. We also confirm the increased plasma ANP concentration observed in patients with congestive cardiac failure and renal failure. In nine patients with renal failure routine haemodialysis was accompanied by a 30 per cent reduction in plasma ANP concentration. Plasma ANP concentrations were similar in treated hypertensive patients, untreated borderline hypertensive patients and normotensive subjects.