Commonly encountered yet not widely appreciated sources of error in the study of motor nerve conduction velocities of the median and ulnar nerves include (1) spread of stimulating current to the nerve not under study, (2) recording a volume conducted potential from distant muscles, and (3) presence of anomalous crossover in the forearm between the median and ulnar nerves. Under these circumstances, it is desirable to block impulses in one nerve without affecting those in the other. This can be achieved simply by collision if a second stimulus is delivered distally to the nerve not being tested.