In previous work with the method of multiple indicator dilution (MID), we have established that a spatially distributed model of transcapillary exchange proposed by Goresky, Ziegler, and Bach (GZB) accurately describes, at the in vivo whole-organ level, the handling of extracellular indicators in the canine renal cortex. To date, however, it has not been possible to assess the key hypothesis that GZB corresponds to the actual local mechanism of exchange in vivo and is not just a compact summary of the kidney's average whole-organ behavior. By adapting the MID method to high speed computed tomography (CT), we are now able to report that the GZB mechanism is an accurate description of renal cortical transcapillary exchange down to volumes of cortical tissue comprising no more than a few per cent of the total cortical mass, i.e., containing no more than a few thousand nephrons. A small bolus of iohexol (radiopaque extracellular indicator) or iodipamide ethyl ester microparticles (radiopaque plasma indicator) injected into the renal artery was followed by CT as it passed through the kidney and into the renal vein. Time-attenuation value curves of the two contrast media obtained from the renal vein and from regions of interest in the cortex were then modeled with the GZB mechanism and with a more complex formulation that includes GZB as a limiting case. When applied to the data, the models converged to GZB as the best fit for each region examined. The GZB mechanism is found to provide excellent agreement with the regional data.