A reduction in compound muscle action potential (CMAP) amplitude and area following proximal versus distal stimulation is the accepted clinical hallmark of conduction block; however, quantitative criteria for determining conduction block remain ambiguous. In this study, digitized records of individual motor unit action potentials (MUAPs) elicited by incremental stimulation in vivo were arithmetically combined in a computer simulation of CMAP generation. Through simulation of possible phase interaction patterns of individual MUAPs, we have shown that abnormal temporal dispersion alone can produce reductions in CMAP area of up to 50%, values that are commonly thought to represent conduction block. Furthermore, by simulating conduction block without excessive temporal dispersion in defined subpopulations of axons, we have demonstrated the importance of the fastest conducting (largest MUAP) axons in determining CMAP amplitude and area. In conclusion, measurements of CMAP amplitude and area in determining conduction block may be misleading if there is significant abnormal temporal dispersion, and quantitation of the degree of conduction block is difficult without knowledge of which subpopulations of axons are affected.