In previous studies the tissue injection, residue detection method failed to provide results of diffusional capillary permeability comparable to those of other methods. For this reason we reconsidered the kinetic theory and found that it is necessary to take into account the apparent (restricted) diffusion coefficient of the indicator within the tissue, D', in order to determine the permeability-surface area product, PdS = Vev.D.D'-1.tev-1 = Vev'.klo where D is the diffusion coefficient in free aqueous solution, Vev is the physical interstitial water volume of distribution, Vev' is the virtual (apparent) interstitial volume of distribution, tev is the mean transit time of the indicator, and klo is the recorded fractional initial washout rate constant. In experiments on open chest dog hearts we examined capillary permeability for 51Cr-EDTA and 99mTc-DTPA with the tissue injection, residue detection method and the single injection, residue detection method. Blood flow was measured independently with local 133Xenon washout. D and D' were measured by a true transient diffusion method. We found that the tissue injection, residue detection method gave results for capillary extraction and PdS-product similar to those obtained with a number of other methods based on indicator diffusion, so the tissue injection, residue detection method in its new kinetically correct formulation could prove useful in clinical studies of capillary permeability since it is applicable to determination of relative changes of permeability during interventional procedures.