The effects of amiloride on the cochlear responses in the guinea pig were compared with those produced by tetraethylammonium (TEA). Amiloride has been reported to reduce membrane permeability to sodium in a wide variety of ion-transporting epithelia. TEA has been documented to suppress the active potassium permeability increase during the repolarization phase of the action potential in mammalian excitable cells, and to reduce the resting potassium conductance in mammalian smooth muscle cells. Perilymphatic perfusion of 10-8 m amiloride or intravenous injection at a dose of 20 mg/kg suppressed the whole nerve action potential (AP) of the cochlea but did not significantly affect the cochlear microphonics (CM) or endocochlear potential (EP). Application of amiloride to endolymph by iontophoretic or perfusion techniques also produced no significant changes of CM and EP when compared with appropriate control procedures. Perilymphatic perfusion of 10-2 m TEA did not suppress CM or EP but the AP was reduced. Iontophoretic application of TEA to the endolymph caused a marked suppression of CM while the EP was significantly increased. The effects of endolymphatic TEA application are consistent with the concept that the normal EP recorded from scala media is the algebraic sum of a positive electrogenic potential and a negative diffusion potential, the latter component being sensitive to potassium permeability changes of the endolymph-perilymph barrier. Maintenance of normal cochlear microphonics also appears dependent upon the maintenance of normal potassium permeability properties of the endolymph-perilymph barrier. The functional importance of normal sodium permeability properties appears less certain.