Reduction of pump flow during cardiopulmonary bypass (CPB) reduces the formation of microemboli and trauma to the blood components, reduces both rewarming of the heart and the noncoronary collateral flow, and improves surgical exposure. Recent studies indicate that a reduction in pump flow, even at normothermia, does not increase the incidence of postoperative cerebral dysfunction. We examined the cerebral consequences of 2 h of normothermic CPB in pigs carried out at pump flows of either 70 ml/kg per min or 50 ml/kg per min, and compared the results with those of a nonperfused control group. We measured the regional cerebral glucose metabolism and the regional capillary diffusion capacity simultaneously in ten different brain regions. Brain morphology, the blood-brain barrier permeability to serum proteins and the regional cerebral water content were also determined in the same animals. Glucose metabolism decreased significantly in both CPB groups (P < 0.001), and significant differences were found between the capillary diffusion capacities of the three groups (P < 0.05), with decreases in eight out of ten brain regions examined in the 50 ml/kg per min group. The results indicate that a reduction of pump flows from 70 ml/kg per min to 50 ml/kg per min is deleterious to the brain, and that a pump flow of 70 ml/kg per min itself has an injurious effect, when normothermic CPB is carried out for 2 h without the use of vasoactive drugs to maintain the blood pressure. Mean arterial blood pressure (MAP) rather than pump flow seemed to determine the adequacy of the cerebral perfusion.