Two forms of the microcathode Po2 electrode exhibit a hysteresis effect which is probably a source of discrepancy in previously reported results. Blood Po2 was measured, avoiding hysteresis, by exposing the electrode to nitrogen both before the calibrating gas and the blood sample so that all readings were made from the same zero-output base line. Calibration was performed by equilibration of blood samples with gases of known Po2. Significant differences in Po2 readings due to membrane material were demonstrated. The indicated Po2 of blood is near the gas Po2 in the physiological range when polypropylene membrane is used on either electrode, but this is not true of Teflon. Indicated blood Po2 is 10-24% lower than gas Po2 in the higher ranges. Standard error of estimate of samples in the range 0-105 mm Hg was 2. 5 mm Hg using polypropylene membrane on one of the electrodes. With blood samples of identical Po2, polypropylene gives a smaller error of measurement than Teflon.