The purpose of this study was to assess the feasibility of utilizing low-current stimulation for F-wave generation, thereby avoiding the discomfort of repetitive supramaximal stimulation. We employed the same technique as is used for generating F waves in the conventional way, except for using a stimulating current that was just strong enough to evoke a motor response on the oscilloscope. This usually required a stimulus of about 10-15 mA at 0.2-ms duration. Both median nerves of 30 subjects were evaluated with this technique and with F waves generated by the conventional technique in the same subjects. Amplitudes were larger when using supramaximal current stimulation. However, there was no statistical difference between F-wave latencies, chronodispersion, and persistence (penetrance) elicited with maximal and low-current stimulation. This procedure should represent a significant improvement for patient comfort during electrodiagnostic procedures involving F-wave studies.