The distance between the two electrode tips can greatly inlfuence the parameters used for record-ing compound nerve action potentials. To investigate the optimal parameters for these recordings in the rat median nerve, we dissociated the nerve using different methods and compound nerve action potentials were orthodromically or antidromically recorded with different electrode spac-ings. Compound nerve action potentials could be consistently recorded using a method in which the middle part of the median nerve was intact, with both ends dissociated from the surrounding fascia and a ground wire inserted into the muscle close to the intact part. When the distance be-tween two stimulating electrode tips was increased, the threshold and supramaximal stimulating intensity of compound nerve action potentials were gradually decreased, but the amplitude was not changed signiifcantly. When the distance between two recording electrode tips was increased, the amplitude was gradually increased, but the threshold and supramaximal stimulating intensity exhibited no signiifcant change. Different distances between recording and stimulating sites did not produce signiifcant effects on the aforementioned parameters. A distance of 5 mm between recording and stimulating electrodes and a distance of 10 mm between recording and stimulating sites were found to be optimal for compound nerve action potential recording in the rat median nerve. In addition, the orthodromic compound action potential, with a biphasic waveform that was more stable and displayed less interference (however also required a higher threshold and higher supramaximal stimulus), was found to be superior to the antidromic compound action potential.