The pH, oxygen tension, and carbon dioxide tension of canine brain tissue were experimentally examined during profoundly hypothermic cardiopulmonary bypass with and without pulsatile assistance. After core cooling, a 60-minute of circulatory arrest was performed in group 1 (n = 16), a 120-minute of low-flow perfusion (25 ml/kg/min) in group 2 (n = 16), and 120 minute of moderate-flow perfusion (50 ml/kg/min) in group 3 (n = 16). The core rewarming was done to the temperature above 32 degrees C. Each group was divided into two subgroups with and without pulsatile assistance (subgroup-p; n = 8, subgroup-c; n = 8). In group 1, progressive brain tissue acidosis and hypercapnea were recovered by use of pulsatile assistance. In group 2, brain tissue acidosis and hypercapnea were recovered completely with pulsatile assistance, but incompletely without it. In group 3, mild acidosis and hypercapnea were eliminated with pulsatile assistance. Brain tissue hypoxia was severe in group 1, slight in group 2, but not found in group 3. We conclude that a pulsatile assistance provides brain protection at any flow-ratio, and that the less flow-ratio and the longer perfusion period will make the pulsatile assistance the more necessary.