The effect of physical exercise on internal carotid artery (ICA) blood flow in conscious man was studied with the aid of electromagnetic flowmetry. A flow probe was implanted on the ICA in 25 patients after reconstruction of the artery. ICA mean blood flow and brachial artery mean blood pressure were continuously monitored in supine (25 patients) and sitting (24 patients) position at rest, during 5-6 minutes exercise on a bicycle ergometer and at rest after exercise. Arterial carbon dioxide tension (PaCO2) was studied in 6/25 work tests in supine and 7/24 in sitting position. Cardiac output was measured at rest and during exercise in 10/25 patients in supine and 8/24 patients in sitting position. In the supine group, ICA flow increased significantly within 1 minute and reached a maximal flow 15% above control flow within 2 minutes after the onset of exercise. The ICA flow then gradually declined, but remained almost significantly elevated, 7.5% above control, on termination of exercise. At rest, after exercise, the ICA flow decreased almost significantly to a level of 5% below the control flow within 5 minutes. There was a significant PaCO2 increase of 2.6 mmHg during exercise and a highly significant increase (72%) in cardiac output during exercise. The ICA flow at rest, before exercise, was about 15% lower in the sitting group than in the supine group. It increased in average 11.5% with 2 minutes of exercise and then gradually diminished. At rest, after exercise, ICA flow decreased further to a level of 8% below control flow within 5 minutes. PaCO2 increased significantly in average 1.6 mmHg during exercise. Cardiac output increased highly significantly (85%) during exercise. The ICA flow changes obtained during exercise in the present study indicate the presence of a regulatory mechanism counteracting the increasing perfusion pressure, but it is unable to compensate the decreased perfusion pressure when the body position was altered from supine to sitting. The cerebral vascular bed in the present patient material seems to operate above and below the lower limit of its pressure range for an adequate autoregulation.