Changes in cerebral blood flow (CBF) in response to changes in Pa were measured by intraaortic injection of133Xe in 12 patients during hypothermic (23–30°C) cardiopulmonary bypass. In each patient, CBF was determined at two randomly ordered levels of Paco2obtained by varying the rate of gas inflow into the pump oxygenator (Group I, n = 6) or by varying the percentage of CO2added to the gas inflow (Group II, n = 6). Nasopharyngeal temperature, mean arterial pressure, pump-oxygenator flow, and hematocrit were maintained within a narrow range. In group I, a Paco2(uncorrected for body temperature) of 36± 4 mmHg (mean ± SD) was associated with a CBF of 13 ± 5 ml.100 g−1·min−1, while a Paco2of 42 ± 4 mmHg was associated with a CBF of 19± 10 ml · 100 g−1·min−1. In group II, a Paco2of 47 ± 3 mmHg was associated with a CBF of 20± 8 ml. 100 g−1·min−1, and a Paco2of 53± 3 mmHg was associated with a CBF of 26 ± 9 ml. 100 g−1·min−1. Within group I, the difference in CBF was significant (p < 0.05); within group II, the difference in CBF was significant at the P < 0.002 level. All CBF measurements were lower than those reported for normothermic, unanesthetized subjects of similar age. The response of the cerebral circulation to changes in CO2tension was well-maintained during hypothermic cardiopulmonary bypass. CBF increased by an average of 1.07 ± 1.19 (SD) ml. 100 g−1·min−1·mmHg−1increase in temperature-uncorrected Paco2in Group I, and by 1.05 ± 0.54 ml · 100 g−1·min−1· mmHg−1increase in group II.