Neuroendocrine activity was studied in 60 consecutive untreated patients with dyspnoea and a clinical suspicion of heart failure. On the basis of the so-called Boston clinical criteria the diagnosis of heart failure was regarded as unlikely in 26 patients, possible in 15 patients, and definite in 19 patients. These groups were studied before any drug treatment was started and were compared with a control group of 69 healthy individuals. Plasma atrial natriuretic peptide concentration was clearly raised in patients with definite heart failure and slightly raised in patients with possible heart failure. Plasma adrenaline concentration was somewhat raised in patients with definite or possible heart failure, whereas plasma noradrenaline concentration was raised only in patients with definite heart failure. Plasma renin activity was not increased in any of the patient groups and plasma aldosterone concentration was slightly increased only in patients with definite heart failure. In the total patient series there were significant correlations between plasma atrial natriuretic peptide concentration and markers of the severity of left ventricular dysfunction. There was some evidence of neuroendocrine activation in untreated heart failure: plasma concentrations of atrial natriuretic peptide and catecholamines were increased but the renin-angiotensin-aldosterone system showed little or no activation.