We report iridium optical constants fitted from synchrotron reflectance data. Specifically, we have used the NKFIT algorithm of D. L. Windt to derive (delta) (E) and (beta) (E) from 2 - 12 keV reflectance calibrations of AXAF witness mirrors. The model is applied at each energy separately, to fit four to nine data points from reflectance-versus-energy scans at selected grazing angles. The stability of the model in the presence of Gaussian noise has been tested extensively. We report the results of several bias studies, involving the generation and analysis of artificial data. Bias studies have been used to determine the optimal grazing angles to be scanned in the various x-ray energy ranges to condition the optical constants. They have also been used to investigate the effects of individual errant data points on the resulting fits and derived optical constants. The results will aid in eliminating systematic errors in the derived optical constants. We also present results of our investigation of the Debye-Waller and Nevot-Croce roughness correction algorithms as applied to our measurements. The Nevot-Croce method gives a much better representation of the data, however its rigorous justification in this experiment is lacking, and the roughness parameter derived is not constant with energy. A more self- consistent model for roughness correction is sought.