1. The effluxes of sodium and chloride from unanaesthetised fish in sea water have been determined; also the salinity dependencies of the sodium influx and efflux and the electrical potential between the extracellular fluids and the water have been measured. 2. The ingested sodium (1.8 meq/kg · h) and chloride (2.1 meq/kg · h) is largely absorbed and excreted extrarenally as part of the effluxes of 8.84±1.29 (S.D.) and 5.60±1.68 (S.D.) meq/kg · h respectively in sea water. 3. The branchial net flux of salt is deduced to be almost linearly dependent on the external salinity over the range 1/3 to 1 2/3 sea water justifying the application of linear non-equilibrium thermodynamics to this phenomenon. 4. The salinity dependence of the net branchial salt flux is in good agreement with predictions for an efficient system of active transport, supporting the proposition that most of the branchial ionic fluxes pass through such a mechanism. The parameters would be optimal for salt transport with an efficiency of 37%, and this could not exceed 69% with these values. 5. The small branchial potential observed in sea water (2.9±4.4 (S.D.) mV) is deduced to be the result of a modest electrogenic potential (5.8 mV) partially offset by a concentration-dependent component (−2.9 mV in sea water), attributable to the larger magnitude of the Nernst potential for chloride relative to sodium, the transport numbers for anions and cations being similar. 6. No evidence was obtained for the existence of an exchange diffusion mechanism for chloride, but if such a phenomenon exists it cannot account for more than about 20% of the chloride efflux in sea water. 7. Changes in the sodium efflux in low salinities are not well explained by branchial potential changes, and obligatorily electrically neutral exchanges are possible. However exchange diffusion of sodium cannot exceed about 1 meq/kg · h (12% of the sodium efflux in sea water). Na−K exchanges could account for to 1.5 meq/kg · h of the sodium efflux, but other effects could equally well account for the efflux reduction in low salinities and its stimulation by low concentrations of sodium and potassium chlorides, and one such is discussed briefly.