Regional cerebral blood flow was measured repeatedly in anesthetized and immobilized cats under various experimental conditions by recording the clearance of inhaled hydrogen with inserted platinum electrodes and by recording the distribution of 15 mu microspheres labeled with 3 different radioisotopes. The values for both methods in normal cortical tissue were comparable (75.7 +/- 23.5 ml/100 g min for H2-clearance; 67.6 +/- 26.2 ml/100 g min for microsphere technique), but were below those recorded in awake cats. With both methods the values could be reliably reproduced (correlation coefficient between measurements: 0.903 for H2-clearance, 0.754 for microsphere technique). During ischemia induced by temporary occlusion of the middle cerebral artery the microsphere technique usually yielded higher flow values than the H2-clearance, and did not indicate severe ischemia in 6 out of 20 instances. After restoration of flow, hyperperfusion was observed by the microsphere technique in 2 cases only while H2-clearance indicated hyperemia in 6 instances. This limited comparability between the 2 methods was also expressed in a low correlation coefficient (0.486) calculated from 139 flow values obtained simultaneously with both methods. The discrepancy between the methods under pathological conditions might be due mainly to the different recording volumes: while Pt-electrodes record H2-clearance from a few mm3 or less, tissue samples of 300-700 mg were necessary for the microsphere technique and inhomogeneities of flow may thereby escape detection. The technique for measuring cerebral blood flow in an experimental setup should be selected according to the requirements of the study and according to the limitations of the various methods.