The relation between receptor occlusion by a competitive neuromuscular blocking agent and fade of an indirectly elicited tetanus at frequencies of 30, 100 and 200/sec has been determined. Tetanic fade at 30/sec returned to normal when one-fifth to one-quarter of the receptors became available, and thus recovery approximated closely that of the single twitch response. The tetanic fade at 100/sec was not normal unless half the receptor pool was available, and a normal response at 200/sec required two-thirds of the pool. It is suggested that recovery of the response to tetanic stimulation at 100/sec would provide a practical index of adequate recovery from curarization