In isolated perfused organs, vascular protein reflection coefficients (sigma) can be calculated from the changes in hematocrit and perfusate protein concentration (CP) that occur during edema formation. This technique requires the assumption that transvascular protein flux by diffusion is negligible. To assess diffusion-induced errors in calculations of sigma, we derived an expression for CP that includes determinants of diffusive protein flux: protein permeability-surface area product (PS), transvascular fluid flux (J), true sigma, and transvascular protein concentration. We used this expression to obtain values of CP under various experimental conditions and then calculated values of sigma (measured sigma) for those conditions. Diffusion causes measured sigma to be lower than true sigma. The diffusion-induced error is larger and potentially substantial when J/PS is low and when true sigma is high. Diffusion-induced error is also larger when the amount of edema formation is greater. In recent isolated canine lung experiments where J/PS was approximately 2.7, diffusion-induced errors in measured sigma for albumin would have been approximately 0.06 (at true sigma = 0.5) and approximately 0.18 (at true sigma = 0.9). When J/PS was higher, the potential for diffusion-induced errors was much smaller. We conclude that diffusion causes underestimation of true sigma and that the error in measured sigma may be substantial when J/PS is < 5 and when true sigma is > 0.5.