In mice breathing 5% oxygen, pretreatment with the optimal dose of 200 mg/kg of phenytoin increased survival time 123%. This increase was somewhat less than that observed with certain barbiturates using the same model but significantly greater than that observed with diazepam which is more effective than phenytoin in suppressing hypoxemic convulsions in this model. In dogs maintained at an expired halothane concentration of either 0.87% or < 0.1%, phenytoin tended to decrease cerebral blood flow and had no effect on the cerebral metabolic rate for oxygen at 3 different doses. Assuming a similar effect in mice, there cerebral protection during hypoxemia observed with phenytoin cannot be explained by a reduction in metabolic rate, an increase in oxygen delivery, or by an anticonvulsant effect per se. In additional dog studies, pretreatment with phenytoin decreased the rate of potassium accumulation in cisternal cerebrospinal fluid following 20 minutes of anoxia. We speculate that phenytoin protection may be linked to this effect.