Brain damage in profound hypothermia was studied by means of electroencephalogram, brain blood barrier and estimation of activity succinic dehydroganase in dogs. Hypothermia was induced by incorporating a heat-exchange mechanism in pump-oxygenator system and total circulatory arrest was maintained for from 3 to 50 minutes after esophageal temperaure had been reduced to the range of 5℃ to 17℃. Rewarming followed by circulatory arrest was discontinued at about 36℃ of esophageal temperature. Two types of pump-oxygenator, bubble and Kay-Anderson, were used and their advantages were compared. Poor recovery of electroencephalogram and brain damage were found even in many cases where perfusion time for cooling and rewarming is relatively short, ranging from 45 to 55 minutes, in bubble type group. These changes were found more frequently and much more exaggerated in proportion to increase of total perfusion time. Brain damage considered referable to perfusion per se was found only where duration of perfusion extended over 89 minutes in Kay-Anderson group. Perfusion less than 80 minutes is cnosidered safe in this type of oxygenator. Changes of temperature in the jugular vein and brain surface were very closed and the difference between those and esophageal temperature was around 8℃ at the maximum when the latter was aronnd 10℃ and it decreased slightly at the end of cooling.The three closely approximated at around 20℃ on rewarming and the gradient between the former two and the latter became maximum at around 30℃ (esophageal) and it decreased at the end of rewarming. There was found gradient in temperature between brain core and brain surface and the former was closed to esphageal temperature and the latter colosed to temperature in the jugular vein. No marked difference in electric brain activity was noted between the cases with and without additional perfusion during circulatory arrest. It is anticipated that circulatory arrest of 30 minutes duration cause no harmful effect on brain when brain temperature was below 20℃. Esophageal temperature rose regularly compared to temperatue in the jugular vein or brain surface during circulatory arrest. The rapidity of temperature drift was in order of esophagus, brain and jugular vein. At first, decrease in amplitude was noted on cooling, though decrease in frequency was minumum, in electroephalographic changes. Occasionally, reactivation of electroencephalogram was found in spite of general tendency toward flattening, in the coures of cooling. The same phenomenon was observed occasionally after circulatory arrest, though electroencephalogram disappeared compleltely ten minutes after arrest. On rewarming, electroencephalogram reappared at 20℃, which is about 3℃ higher in average temperature when it disappeared on cooling. Changes in electroencephalogram on rewarming was not consistent. Burst at first and then increase in frequency were noted in some cases. and wave of low voltage at first, secondly rapid increase of frequency and then gradual increase of amplitude were noted in another cases. Occasionally, suppression of electroencephalogram recurred after esophageal temperature reached to over 30℃, Decrease in both frequency and amplitude of brain waves were generally noted in most caces at the end of rewarming and return of amplitude to the initial level usually lagged behind that of frequency. Brain damage was found in cases where recovery of brain was poor when examination was performed 20 minutes or more later after completion of the experiment, though it was sometimes unsuccessful to detect brain damage where examination was done earlier than 20 minutes. Brain damage was found mostly in the cortical area of the brain and occasionally in nucleus and brain stem. There were few cases where brain damage was found despite of good recovery in brain acivity. No changes were noted usually in cortical area in those cases and minimum when existed, too. It is remarkable that good recovery of brain activity in electroencephalogram does not always guarantee absense of brain damage. Estimation of the activity of succinic dehydrogenase is considered an excellent method for detection of brain damage, because it makes passible to disclose changes in brain tissue in early period after their occurrence. Bubble type oxygenator is considered disadvantageous to profound hypothermia with rapid blood cooling.