In 39 mongrel dogs, regional cerebral blood flow was measured during pulsatile and nonpulsatile deep hypothermic cardiopulmonary bypass with total circulatory arrest. Total circulatory arrest was performed at 20 degrees C cerebral temperature for 40 minutes in 15 dogs, 60 minutes in 12 dogs, and 80 minutes in 12 dogs. Cerebral blood flow in both groups decreased as cerebral temperature fell and there was no significant difference in cerebral blood flow between the two groups during the cooling period. After circulatory arrest for 40 minutes, as cerebral temperature increased to 35 degrees C, cerebral blood flow in both groups recovered to values as high as the respective initial values, which were measured just after the beginning of cardiopulmonary bypass for cooling (102.5% +/- 10.2% in the pulsatile group and 97.2% +/- 12.6% in the nonpulsatile group). After circulatory arrest for 60 minutes, cerebral blood flow in the pulsatile group increased to 141.8% +/- 16.1% of its initial value when the cerebral temperature became 35 degrees C, but it remained significantly lower (64.5% +/- 9.2%) in the nonpulsatile group (p < 0.01). After circulatory arrest for 80 minutes, cerebral blood flow in both groups remained lower than the respective initial values. These results suggest that pulsatile perfusion maintains cerebral blood flow even during profound hypothermia and that it may protect the brain from ischemic and hypoxic damage caused by profound hypothermia and total circulatory arrest in cardiac operations.