EMG responses to sudden displacement of the forelimb were studied in Cebus monkeys tranquilized with Atravet, a phenothiazine tranquilizer. The monkey's forearm was strapped firmly to a manipulandum handle. A torque motor attached at the pivot point of the handle, under servo control, provided reproducible limb displacements. In response to a sudden maintained displacement three periods of EMG activation in biceps muscle occurred with peak latencies of approximately 25, 45 and 85 msec. These correspond to the latencies of the M1, M2 and M3 responses in the alert animal. Similar responses were observed in 'naive' animals which had not previously been used in experimentation. All three responses increased in magnitude with increasing background activity and all appeared to be associated with suppression of EMG activity in the antagonist muscle. M1 and M2 responses were position dependent, M1 being greater in extension than in flexion and M2 the opposite. The position-dependence of the M2 response was produced by a depression of activity following the M1. This depression of activity lasted up to 30 msec following M1 and was directly dependent on the M1 magnitude.