The effects of hypotension, hemodilution, and their combination on the relationship between concurrent brain electrical activity and resulting brain injury were studied in anesthetized monkeys. The authors compared changes in the electroencephalogram and somatosensory and auditory evoked potentials with eventual neuropathologic outcome. Our goals were: 1) to define the margin of safety for the monkey brain during hemodilution and hypotension under several simulated clinical conditions; and 2) to determine whether noninvasive measurements of brain electrical activity can predict ischemic brain cell damage. Forty-one monkeys were anesthetized with halothane (0.8 vol % inspired) and ventilated mechanically. Arterial hypotension was induced with trimethaphan (25 ± 8 mmHg mean arterial blood pressure [MABP] for 30 min). Hemodilution was induced by replacing blood with lactated Ringer's solution (14 ± 2% hematocrit for 1 h). Combined hemodilution and hypotension consisted of 30 min of hemodilution alone followed by superimposing hypotension for 30 min (16 ± 3% hematocrit and 29 ± 5 mmHg MABP). Ten monkeys died following severe hypotension alone or combined hemodilution and hypertension as a consequence of cardiac arrest or undetermined (possibly neurologic) causes. No histologic evidence of ischemic brain cell injury was found in surviving monkeys subjected to hemodilution or hypotension alone. Neuropathologic alterations in the cerebral cortex, cerebellum, hippocampus and globus pallidus as well as neurologic and behavioral deficits were found in seven of 16 surviving monkeys subjected to both hemodilution and hypotension. These findings resulted from combinations of hematocrit less than 20% and MABP below 40 mmHg. Only the degree of amplitude reduction in cortical components of the somatosensory evoked potentials during the stress period indicated a high probability of neuropathologic outcome.