Experiments were carried out in the isolated short-circuited skin of the toad Bufo marinus ictericus. 42K influx and efflux experiments were carried out with skins bathed on both sides by NaCl-Ringer's solution. Those fluxes showed very similar kinetics of equilibration with time and the results could be fitted by equations of a model of two intraepithelial compartments and the bathing solutions. In the steady state K influx is 3.99 +/- 0.36 nmol cm-2 hr-1 (n = 7) and efflux 3.62 +/- 0.38 nmol cm-2 hr-1 (n = 7) and are not statistically different, indicating that no net K flux is present across the epithelium. Different kinds of perturbations affecting the rates of 42K discharge into the bathing solutions were studied. Immediately after addition of amiloride (10(-4) M) to the outer solution, a sharp decline is observed in the rate of 42K discharge into the bathing solution, JK21, which falls from 3.62 +/- 0.38 nmol cm-2 hr-1 to 2.02 +/- 0.04 nmol cm-2 hr-1 (n = 7) 2 min after addition of the drug, followed by a partial recuperation with time. A complete Na by K substitution in the outer bathing solution induces a prompt and marked decline in JK21 which is similar to that induced by amiloride. Increase in the outer bathing solution Na concentration from zero Na concentration induces a nonlinear increase in JK21 and a linear relationship was observed between JK21 and short-circuit current in the range of 0 to 115 mM external Na concentration. The decline in JK21 induced by amiloride or by lowering external Na concentration was interpreted as being caused by electrical hyperpolarization of the external barrier of the epithelium induced by these procedures. Depolarization of the epithelial barriers by inner Na by K substitution in the short-circuited state (when the potential barriers are equal) drastically interfere with the rate of 42K discharge from the epithelium into the bathing solutions. Thus, transient increases are observed both in the rate of 42K discharge to the outer and to the inner bathing solutions upon depolarization of the barriers. These results indicate that at least the most important component of transepithelial K unidirectional fluxes goes through a transcellular route with a negligible paracellular component. Addition of ouabain (10(-3) M) to the inner bathing solution induces a transient rise in the rate of 42K discharge to the outer bathing solution with a peak on the order of 200% of the stationary value previous to the action of the inhibitor, followed by a return to new stationary values not statistically different from those observed previously to the effect of ouabain. The behavior of JK21 upon the effect of ouabain, as suggested by comparison with predictions from computer simulation, strongly supports the notion of a rheogenic Na pump in the inner barrier of the epithelium against the notion of a nonrheogenic 1:1 Na--K pump.