Several anatomical and electrophysiological studies demonstrated the possibility of an ulnar-to-median anastomosis in the hand and the forearm--a "reversed Martin-Gruber anastomosis". Observations supporting this possibility were obtained in the course of routine clinical testing of nerve conduction velocity in upper extremities. The existence of an ulnar-to-median anastomosis was suggested when, in absence of clinical signs of the pronator syndrome or the anterior interosseus nerve syndrome, the amplitude of compound muscle action potential in the opponens pollicis was larger at stimulation of the median nerve at the wrist than at the elbow. When an ulnar-to-median anastomosis is suspected, special care should be exerted in evaluation of motor distal latency of the median nerve with a gradual and slow increase of the stimulus voltage. Use of high voltage from the beginning can simultaneously activate both the median nerve and a collateral branch of the ulnar nerve, the later bypassing the carpal tunnel and evoking a short latency response in the thenar muscles, with the simultaneous long latency response to the stimulation of the median nerve being masked by this fast response. Based on the same anatomical considerations, a decreased amplitude of the compound muscle action potential of the opponens pollicis at proximal stimulation, as compared with that at distal stimulation, should not be interpreted as an indication of a damage of the median nerve at the forearm.