Forty-four studies of regional cerebral blood flow (rCBF), fractional oxygen extraction (rOER) and oxygen consumption (rCMRO2) were made on twenty-five patients with recent internal carotid artery territory infarcts. The purpose was to study flow-metabolism relationships in the contralateral hemispheres, and to investigate whether contralateral rCMRO2 was depressed as a result of the recent infarcts. Two groups of controls were included for comparison--seventeen normal volunteers, and ten patients with proven extracranial cerebrovascular disease but without evidence of cerebral infarction. The results demonstrated that: contralateral hemispheric rCMRO2 was less variable than regional oxygen availability (the product of rCBF and arterial oxygen content). This was due, in part, to the effect of individual variations in PaCO2 on rCBF, but other uncontrolled factors, such as intracranial pressure, may have had influences. As a result, rCMRO2 did not correlate with rCBF; mean rCMRO2 in the contralateral hemispheres was 12% lower than normal (a significant difference), but was not different from the value found in patients with extracranial vascular disease in whom there was no evidence of infarction or ischemia; contralateral rCMRO2 did not correlate with the size of the infarct in the opposite hemisphere. It is concluded that rCMRO2 cannot be inferred from rCBF measurements in uncontrolled human studies (as frequently done in the past), and that depression of contralateral rCMRO2 may have preceded infarction in the opposite hemisphere, a consequence of the previous influences of diseases that predispose to stroke.