Patients presenting for carotid thrombo-endarterectomy have precarious brain perfusion. The anaesthetic approach must not be detrimental to adequate cerebral blood flow. Systemic arterial blood pressure must be maintained. Agents or techniques are avoided which may increase cerebrovascular resistance by affecting blood viscosity, cerebrospinal fluid pressure, cerebrovascular diameter, or body temperature changes. In the preoperative period, a light premedication is advocated to promote early awakening and to maintain a stable respiratory and cardiovascular status. Induction is smooth, with no straining or coughing. An armoured oro-endotracheal tube is used to ensure a patent airway at all times. Maintenance anaesthesia is nitrous oxide, oxygen, and halothane. Thus, cautery may be used. Cerebral vasodilatation tends to occur with halothane. The respiratory tract is not irritated. Early awakening is possible. Gallamine is used to ensure a light plane of anaesthesia and to offset the bradycardia possibly produced by halothane. Also, the carotid sinus reflex, which may be stimulated during surgery, is blocked with gallamine. Use of a muscle relaxant prevents abdominal muscle contractions, which inevitably occur during spontaneous respiration and which may increase central venous pressure and, consequently, cerebral venous pressure. Gallamine facilitates the use of controlled ventilation. By monitoring systemic arterial and internal jugular venous Po2, Pco2, pH, bicarbonate and base excess (Astrup), precise regulation of tidal volume, minute volume (Wright respirometer) and Pco2 at 40-45 mm. Hg is achieved. Moderate hypercarbia tends to increase cerebral blood flow and hyp erventilation; cerebral vasoconstriction and cerebral hypoxia are avoided. A polyethylene surgical shunt bypass is used during the thrombo-endarterectomy. Additional special monitors used are the electrocardiogram, electroencephalogram, central venous pressure, and oesophageal temperature. Low molecular-weight dextran is advocated to decrease viscosity and thus increase cerebral blood flow. Results of carotid thrombo-endarterectomy in the past five years in 60 patients undergoing 78 procedures showed 47, or 78 per cent, with clinical improvement. There were three deaths, giving a mortality rate of 5 per cent. One death was due to a myocardial infarct one month postoperatively. Another patient died from rethrombosis of the internal carotid artery in the early postoperative period. The third death resulted from intracranial rethrombosis with infarction, superimposed on a setting of severe generalized cerebral atherosclerosis.