Fiberoptic oximeters measure oxyhemoglobin saturation from the optical reflectance of whole blood, but the calibration of such oximeters is hematocrit dependent. Therefore, using photon-diffusion theory and an empirical approach, we have developed a new reflectance method that determines hematocrit and correspondingly corrects the oxyhemoglobin-saturation measurement. Our method employs four fiber-optic light guides, a photodetector, and three inexpensive light-emitting diodes (one with emissions at 660 nm and two at 813 nm). Hematocrit is determined from the ratio of reflectances from the differently spaced emitting fibers at 813 nm and is used to correct the 813-660 nm measurement of oxyhemoglobin saturation. In red cell suspensions, the mean difference between reflectance measurements of hematocrit and conventional determinations was only 2.09% (r = 0.99), and when compared with conventional gasometric measurements of oxyhemoglobin saturation, the reflectance method yielded the same calibration curve for different hematocrits and gave a mean difference of only 2.67%. Although the technique is demonstrated with a cuvette appropriate for an extracorporeal circulation in animal experiments, it could possibly be further developed for fiber-optic catheter oximeters.