Net hepatic uptake and release of K+ were studied in the isolated-perfused rat liver subjected to a 10-min period of hyperosmotic stress by addition of 80 mM mannitol or sucrose to the perfusing solution. Bile flow and effluent Na+, K+, and Ca2+ activities were monitored throughout. Upon initiation of hypertonic stress, a sharp transient dilution of effluent ion activities indicated hepatic water losses that were larger and occurred more rapidly with sucrose than with mannitol. During continuous hyperosmotic perfusion, portocaval differences in K+ uncovered a steady net influx of the ion into the liver that reached a higher maximum and led to a greater accumulation in experiments with sucrose compared with mannitol. This hepatic K+ uptake was completely blocked by 1 mM ouabain. Upon return to isotonic conditions, a sharp transient concentration of effluent ion activities suggested hepatic water uptake that was again more rapid and pronounced in sucrose- than mannitol-treated livers. This was followed by a transient phase of net hepatic K+ release whose magnitude and duration were proportional to the water movements induced by the removal of each carbohydrate. Administration of 2 mM Ba2+ abolished this K+ efflux. These results indicate that mannitol equilibrates between extra- and intracellular compartments, whereas sucrose apparently does not. The data also suggest that net movements of K+ may be involved in the regulatory volume responses induced by hyperosmotic stress and return to normal tonicity, respectively.