A nonelectrolyte model of proximal tubule epithelium has been extended by the inclusion of a compliant tight junction. Here "compliance" signifies that both the junctional salt and water permeability increase and the salt reflection coefficient decreases in response to small pressure differences from lateral interspace to tubule lumen. In previous models of rat proximal tubule, there has been virtually no sensitivity of isotonic salt transport to changes in peritubular oncotic force. With the inclusion of junctional compliance, decreases in peritubular protein can open the junction and produce a secretory salt flux. Thus the model can represent the "backflux hypothesis," as it was originally put forth (J. E. Lewy and E. E. Windhager, Am. J. Physiol. 214: 943-954, 1968). Additional calculations, simulating a tight junction with negligible water permeability, reveal that the quantitative impact of peritubular protein can be realized whether or not there is substantial junctional water flux. The epithelial model of proximal tubule has also been incorporated into a model of the proximal nephron, complete with glomerulus, peritubular capillary, and interstitium. The interstitial compartment is well mixed and interstitial pressure and osmolality are determined iteratively to achieve balance between tubule reabsorption and capillary uptake. For this model, two domains of operation are identified. When interstitial pressures are low, junctions are closed, and filtration fraction has no effect on proximal reabsorption. When interstitial pressures are relatively elevated, epithelial junctions are open, and proximal salt reabsorption changes in proportion to changes in filtration fraction. In neither domain, however, does the model tubule augment salt flux with isolated increases in luminal flow rate (at constant filtration fraction). The absence of a separate effect of tubule fluid flow on salt transport precludes perfect glomerulotubular balance.