In this investigation we examined the changes in the pattern of activity in the medial gastrocnemius (MG) muscle in walking cats following transection of the nerves innervating synergist muscles (lateral gastrocnemius, soleus, and plantaris). Immediately following the nerve transections, there was a large increase in ankle flexion during early stance (from approximately 10 to approximately 30 degrees ) and a marked increase in the magnitude of the MG bursts during stance. We attribute this increase in the magnitude of the MG bursts to an increase in afferent feedback from the abnormally stretched MG muscle. During the week after the nerve transections, there was a progressive decrease in ankle yield. This improvement in ankle function was correlated with an increase in magnitude of two components of the MG bursts; the initial component starting during late swing and ending approximately 40 ms after ground contact, and a late component associated with stance. The time courses of the increases in the initial and late components of the MG bursts were different. Large and significant increases in the late component occurred the day after the nerve transections, whereas increases in the initial component occurred more gradually. This difference in time course was reflected in the kinematics of ankle movement. Over the first few days after the nerve transections, improvement in ankle movement occurred primarily late in the stance phase, and there was little change in ankle yield during early stance. At 1 wk, however, there was a significant reduction in ankle yield during early stance. This decreased yield was most likely due to an increase in stiffness of the MG muscle at the time of ground contact resulting from the increase in magnitude of the initial component of the MG bursts. The increases in the magnitude of the initial and late components of the MG bursts, as well as the improvement in ankle function, depended on use of the leg. All these changes were delayed by immobilizing the leg for 6 days in an extended position. We discuss possible mechanisms underlying the increase in the magnitude of the MG bursts and propose that proprioceptive signals from the stretched MG muscles provide an error signal for rescaling the magnitude of the centrally generated initial component. Our data support the concept that proprioceptive feedback functions to scale the magnitude of feed-forward motor commands to ensure they are appropriate for the biomechanical properties of the musculoskeletal system.