Abstract. The aim of this study was to clarify the mechanism of isotonic fluid transport in frog skin glands. Stationary ion secretion by the glands was studied by measuring unidirectional fluxes of 24Na+, 42K+, and carrier-free 134Cs+ in paired frog skins bathed on both sides with Ringers solution, and with 105m noradrenaline on the inside and 104m amiloride on the outside. At transepithelial thermodynamic equilibrium conditions, the 134Cs+ flux ratio, JoutCs/JinCs, varied in seven pairs of preparations from 6 to 36. Since carrier-free 134Cs+ entering the cells is irreversibly trapped in the cellular compartment (Ussing Lind, 1996), the transepithelial net flux of 134Cs+ indicates that a paracellular flow of water is dragging 134Cs+ in the direction from the serosal- to outside solution. From the measured flux ratios it was calculated that the force driving the secretory flux of Cs+ varied from 30 to 61 mV among preparations. In the same experiments unidirectional Na+ fluxes were measured as well, and it was found that also Na+ was subjected to secretion. The ratio of unidirectional Na+ fluxes, however, was significantly smaller than would be predicted if the two ions were both flowing along the paracellular route dragged by the flow of water. This result indicates that Na+ and Cs+ do not take the same pathway through the glands. The flux ratio of unidirectional K+ fluxes indicated active secretion of K+. The time it takes for steady-state K+ fluxes to be established was significantly longer than that of the simultaneously measured Cs+ fluxes. These results allow the conclusion that in addition to being transported between cells K+ is submitted to active transport along a cellular pathway.Based on the recirculation theory, we propose a new model which accounts for stationary Na+, K+, Cl and water secretion under thermodynamic equilibrium conditions. The new features of the model, as compared to the classical Silva-model for the shark-rectal gland, are: (i) the sodium pumps in the activated gland transport Na+ into the lateral intercellular space only. (ii) A barrier at the level of the basement membrane prevents the major fraction of Na+ entering the lateral space from returning to the serosal bath. Thus, Na+ is secreted into the outside bath. It has to be assumed then that the Na+ permeability of the basement membrane barrier (PBMNa) is smaller than the Na+ permeability of the junctional membrane (PJMNa), i.e., PJMNa/PBMNa 1. The secretory paracellular flow of water further requires that the Na+ reflection coefficients (Na) of the two barriers are governed by the conditions, BMNa 0, and BMNa JMNa. (iii) Na+ channels are located in the apical membrane of the activated gland cells, so that a fraction of the Na+ outflux appearing downstream the lateral intercellular space is recirculated by the gland cells. Based on measured unidirectional fluxes, a set of equations is developed from which we estimate the ion fluxes flowing through major pathways during stationary secretion. It is shown that 80% of the sodium ions flowing downstream the lateral intercellular space is recycled by the gland cells. Our calculations also indicate that under the conditions prevailing in the present experiments 1.8 ATP molecule would be hydrolyzed for every Na+ secreted to the outside bath.