Haematocrit in the high normal, as well as the pathological, range may be a risk factor for the development of stroke. As stroke is one of the most important causes of physical handicap attempts to reduce its incidence further are justified. Repeated small volume venesection without volume replacement is a simple method of haematocrit reduction which could be used in a controlled trial if shown to be safe and practicable. Forty three male patients (33 evaluable) with haematocrit above 0.46 without primary proliferative or secondary polycythaemia, were entered into an open study to assess the clinical, haemorheological and haematological sequelae of serial venesection. Haematocrit was reduced from a median of 0.49 to 0.385 which necessitated removal of 2.75 1 of blood (range 1.5 - 4.25 1). This resulted in a marked reduction in blood viscosity which correlated significantly with the change in haematocrit. Although haematocrit has been postulated to affect blood pressure through an effect on blood viscosity and hence peripheral resistance, no clinically significant reduction in blood pressure was observed. Despite producing a slight reactive thrombocytosis, serial venesection was not shown to increase thrombogenic activity. Bleeding time was prolonged significantly and there was a trend for a reduction in platelet adhesion. The reduction in platelet adhesion correlated significantly with the reduction in haematocrit. Maintenance of haematocrit below 0.45 necessitated further venesection on average once every 2 months during follow-up for one year. Two patients with a previous history suffered further strokes which were considered unlikely to be related to venesection. One patient had worsening of angina and 2 patients experienced intermittent palpitations when haematocrit was reduced to 0.40; these events were considered possibly due to an increase in cardiac output as a consequence of lowered haematocrit. Except in patients with cardiac disease, repeated small volume venesection without volume replacement is a safe and practicable method of reducing haematocrit.