Plots of the pressure di!erence ( P) applied to plant roots vs. the resulting volume #ow rate (Q T ) often exhibit an anomalous o!set that has been di$cult to explain. The present analysis suggests that solute build-up in two-and three-compartment models of the root cannot account for this o!set. The Ginsburg}Newman three-compartment model explains the o!set in terms of di!ering re#ection coe$cients for the membranes bounding the intermediate compartment. This model appears more promising, but it predicts a minimum in the plot of xylem-sap osmotic pressure vs. Q T which is not observed in practice. Fiscus hypothesized that an internal asymmetric distribution of non-mobile solutes is responsible for the o!set. In the present study, this hypothesis is incorporated into a four-compartment model of the root that is conceptually related to the three-compartment model of Miller. But according to the four-compartment model, the asymmetric solute distribution does not arise because of solvent drag. Rather the anomalous o!set is associated with a concentration gradient of photoassimilates (the non-mobile solutes) that exists in the absence of volume #ow, and which drives the di!usive transport of these solutes from the stele to the cortex via endodermal plasmodesmata. This model is consistent with the existence of radial symplastic osmotic-pressure gradients, and it appears to have greater explanatory power than the Ginsburg}Newman model. In particular, it suggests explanations for diurnal variations in P}Q T curves, as well as the e!ects of changing external solute concentrations. It also shows how the overall root re#ection coe$cient can be less than unity, even when the cell membranes are e!ectively ideally semipermeable, and there is negligible extracellular transport of water and solutes. The model makes a number of experimentally testable predictions.