1. Potassium currents were characterized in turtle cochlear hair cells by whole-cell voltage clamp during superfusion with the potassium channel antagonists, tetraethylammonium (TEA) and 4-aminopyridine (4-AP). The estimated resonant frequency, f0, was inferred from tau, the time constant of deactivation of outward current upon repolarization to -50 mV, according to the empirical relation, f0 = k1 tau-1/2 + k2. 2. Dose-response relations for TEA and 4-AP were obtained by exposing single cells to ten concentrations exponentially distributed over four orders of magnitude. Potassium current in cells tuned to low frequencies was carried by a single class of channels with an apparent affinity constant, K1, for TEA of 35.9 mM. Half-blocking concentrations of 4-AP were correlated with the time constant of deactivation and varied between 26.2 and 102 microM. In cells tuned to higher frequencies, K+ current was carried by a single class of channels with high affinity for TEA (K1 = 0.215 mM) and low affinity for 4-AP (K1 = 12.3 mM). This pharmacological profile suggests that K+ current in low frequency cells is purely voltage gated and in high frequency cells, it is gated by both Ca2+ and voltage. 3. For each current type, the voltage dependence of activation was determined from tail current amplitude at -50 mV. The purely voltage-gated current, IK(V), was found to increase e-fold in 4.0 +/- 0.3 mV (n = 3) in low frequency cells exposed to TEA (25 mM). The Ca(2+)- and voltage-gated current, IK(Ca), was more steeply voltage dependent, increasing e-fold in 1.9 mV (n = 2) in high frequency cells exposed to 4-AP (0.8 mM). 4. IK(V) was found to inactivate slowly during prolonged voltage steps (approximately 10 s). Steady-state inactivation increased with depolarization from -70 mV and was incomplete such that on average IK(v) did not fall below approximately 0.39 of its maximum value. 5. Superfusion of 4-AP (0.8 mM) reversibly depolarized a low frequency cell and eliminated steady voltage oscillations, while TEA (6 mM) had no effect. In a high frequency cell, voltage oscillations were abolished by TEA, but not by 4-AP. 6. The differential pharmacology of IK(V) and IK(Ca) was used to measure their contribution to K+ current in cells tuned to different frequencies. Both currents exhibited a frequency-dependent increase in maximum conductance. IK(V) accounted for nearly all K+ current in cells tuned to less than 60 Hz, while IK(Ca) was the dominant current in higher frequency cells. 7. Mapping resonant frequency onto epithelial position suggests an exponential relation between K+ current size and position. IK(V) appeared to be limited to the apical or low frequency portion of the basilar papilla and coincided with maximal expression of a K(+)-selective inward rectifier, IK(IR). This finding is consistent with the notion that low frequency resonance is produced by interaction of IK(V) and IK(IR) with the voltage-gated Ca2+ current, ICa, and the cell's capacitance. The ionic events underlying higher frequency resonance are dominated by the action of IK(Ca) and ICa and include a contribution from IK(IR).