BACKGROUND--In patients with chronic heart failure there is no relation between cardiac output and symptom limited exercise tolerance measured on a bicycle or treadmill. Furthermore, the increase in cardiac output in response to treatment may not be matched by a similar increase in exercise tolerance. More important in determining exercise capability is blood flow to skeletal muscle. This implies that the reduction in skeletal muscle blood flow is not directly proportional to the reduction in cardiac output and that there are regional differences in blood flow in patients with heart failure. METHODS--Cardiac output and regional blood flow measured in 30 patients with chronic heart failure were compared with values obtained from 10 healthy controls. Measurements were made at rest and in response to treadmill exercise and were all made non-invasively. RESULTS--Cardiac output was lower in the patients at rest and during exercise. Blood flow in the superior mesenteric and renal arteries was also lower in the patients and represented a different proportion of cardiac output than in the controls. In response to exercise the increase in blood flow to the calf and therefore to skeletal muscle, was reduced in the patients. In the patients there was no correlation between resting cardiac output and blood flow in the superior mesenteric artery, renal artery, or calf. CONCLUSIONS--Because blood flow to skeletal muscle and to the kidneys is likely to be important in determining patients' symptoms this factor may explain why central haemodynamic variables do not correlate with the exercise tolerance in patients with chronic heart failure.