A loss of motor units is the primary defect in many neurogenic disorders. The number of motor units remaining in a muscle or group of muscles can be estimated by a number of methods on either needle electromyography or as part of nerve-conduction studies. Comparison of the methods of motor unit number estimates (MUNE) and their underlying assumptions shows that each method can provide reliable and useful clinical information. Each method relies on obtaining an estimate of the average size of single motor units, and compares that size with the size of a maximal compound muscle action potential in which all motor units are activated. The potential errors of all the methods are similar and relate primarily to adequacy of sampling of the size of single motor units. The method of statistical estimates of MUNE does not attempt to isolate and measure individual motor unit potential sizes. Rather it estimates the size of single motor units from the variation in size of a muscle action potential when individual motor units in a group of motor units fire randomly.