The incremental torque resisting rotation of the foot about the ankle joint was studied in normal seated subjects. Prior to each rotation, subjects were required to activate triceps surae (TS) muscles and maintain a constant plantar flexion torque (range 6-14 N X m) on a platform whose position was controlled by a torque motor. Subjects were instructed to increase torque as rapidly as possible once rotation commenced. Rotations ranged from 0.5 to 14 degrees amplitude and from 20 to 300 degrees/s maximum velocity. The torque in response to rotations stretching TS muscles and releasing tibialis anterior (TA) muscles increased steeply and then rapidly decreased with stretch velocity. At approximately 60 ms from stretch onset, the torque reduction terminated, torque then increased again until it began to level off at approximately 120 ms. A further large increase in torque occurred at 180 ms. A burst of short-latency (SL) electromyographic (EMG) activity in soleus (SOL) commenced at 40 ms, and was followed by a second burst at approximately 68 ms, provided that stretch deceleration started later than 20 ms after stretch onset. A period of sustained EMG activity in SOL commenced at approximately 130 ms (long-latency (LL) activity). Incremental torque in response to stretch of TA and release of TS muscles initially showed a step decrease followed by a reversal of the torque trajectory back toward base line. This change was arrested at 60 ms and torque then remained approximately constant until a large increase in torque at 180 ms. Ischemia was used to reduce SL EMG reflexes without significantly modifying the background EMG activity. A comparison between torque curves under control and ischemic conditions indicated that SL EMG activity in TS muscles recruited the force responsible for terminating the torque reduction coincident with decreasing stretch velocity. The torque response prior to the onset of force recruited by SL activity was attributed to the intrinsic properties of active muscle fibers. Thereafter, until the onset of LL activity, the torque response was attributed to intrinsic and reflex-recruited force. Torque in these two time periods was compared under a variety of stretch conditions in order to test the hypothesis that force recruited by segmental reflexes compensates for the non-linear stretch properties of active TS muscles. The relationships of SL EMG amplitudes and areas to stretch velocity and acceleration were also examined.(ABSTRACT TRUNCATED AT 400 WORDS)