We have developed the integral-mass balance (IMB) method to measure the solvent drag reflection coefficient (sigma f) for transcapillary macromolecular transport in skeletal muscle and other organs. Of course, sigma f is calculated from the cumulative amounts of water and macromolecule that move convectively across the microvascular membrane as determined from changes in hematocrit and plasma macromolecule concentration over a period of fluid filtration. We have investigated the effects of both theoretical and experimental factors that affect the validity and accuracy of the method. The effect of the following factors on sigma f determination by the IMB method were explored: low Peclet number; random-measurement errors; and systematic errors due to vascular leakage, hemolysis of red blood cells, evaporation, and osmolality changes. We found that all of these factors produced overestimations of sigma f, but their effects could be corrected. Also, appropriate experimental design could minimize these effects. Experiments using the IMB method in the isolated, perfused cat hindlimb preparation to determine sigma f for albumin and plasma proteins resulted in mean values of 0.82 +/- 0.08 (SD) (n = 7) and 0.83 +/- 0.02 (n = 4), respectively.