The distinction between radiation and tumor brachial plexopathy may be difficult. The electrophysiological recording of myokymic discharges, frequently present in the former but rare in the latter type of plexopathy, can be helpful for the diagnosis. However, the pathophysiology and the site of origin of these discharges remain unclear. We describe a patient presenting with radiation brachial plexopathy, clinical myokymia, cramps and pain. In this patient, the myokymia--due to abundant myokymic discharges--and the cramps, were related to the existence of persistent conduction block of several years duration. Several findings suggest that the myokymic discharges were generated on blocked axons: voluntary activity did not influence their occurrence nor modify their course; the motor unit potentials involved in the discharges were not evoked by stimulation proximal to the site of the conduction block, whereas the stimulation distal to this site could evoke, modify the rhythm, or interrupt the course of the discharges; the latency of these evoked responses indicated that the site of reflection was proximal on the axon, and likely coincided with that of the conduction block. Recent observations (Roth and Magistris, 1987b) indicated that myokymia, produced by numerous single or grouped fasciculations generated on axon terminals, may be related to persistent conduction blocks of various etiologies. The present case demonstrates that myokymia provoked by myokymic discharges may as well be related to persistent conduction block. The reason why these blocks are accompanied by fasciculations in some situations and by myokymic discharges in others remains an unsolved question. The cramps observed in this patient were also of interest as they occurred in the muscle territory of blocked axons and were provoked by passive muscle shortening. Their origin, distal to the conduction block, is unknown. Finally, a neurolysis did not prevent the progressive transformation of conduction block into axonotmesis.