The kinetic properties of wild-type rat brain IIa sodium channels in excised macropatches were studied using step depolarizations and ramp depolarizations to imitate the slow settling-time of voltage in two-electrode voltage clamp. Ramp depolarizations longer than 1 ms produce an increasing suppression of peak sodium current (I_Na). Two rates of inactivation can be seen in macroscopic sodium current records from excised patches following both step and ramp depolarizations. During slow ramp depolarizations, reduction in peak I_Na is associated with selective loss of the fastest rate of test-pulse inactivation. This change can be interpreted as resulting from inactivation of a separate sub-population of `fast mode' channels. The slow rate of test-pulse inactivation is relatively unaffected by changing ramp durations. These results are sufficient to explain the typically slow inactivation kinetics seen in two-electrode voltage clamp recordings of sodium channels in Xenopus oocytes. Thus, the kinetics of sodium channels expressed in Xenopus oocytes are not readily characterizable by two-electrode clamp because of the large membrane capacitance and resulting slow clamp settling time which artifactually selects for slow mode channels.