1. It was hypothesized that passive movement of either the whole leg or its separate segments, in a manner mimicking human gait, leads to attenuation of the soleus H reflex. It was further hypothesized that this attenuation arises from presynaptic effects. Reflex amplitudes were observed in humans during natural bipedal and unipedal stepping on the spot, during passive stepping, during passive movement of the lower limb segments about the hip, knee, and ankle individually in a stepping fashion, and during passive movement with tonic contraction of the soleus muscle. 2. In natural stepping at a cadence of 54 steps/min, the reflex means were substantially depressed in the swing phase (P < 0.01). (Means, standing control 90.1%, unipedal 8.3%, bipedal 6.9%, of maximum M wave.) During the stance phase, reflex magnitudes were mildly and significantly elevated in four of six subjects, compared with standing controls (P < 0.05). 3. For passive stepping, subjects were dorsally tilted 20 and 90 degrees (lying supine) from the vertical position, to obtain quiet electromyograms (EMGs) in the postural muscles. Recorded during natural stepping, the right leg was manipulated to match the electrogoniometer traces of the three major joints. 4. At 20 degrees of tilt of the body, mean H reflexes were significantly lower, by 26.4%, compared with the supine position (P < 0.05). During passive stepping movement of the leg at 54 steps/min, the reflex was profoundly attenuated over the entire cycle (P < 0.01). The significantly attenuated reflexes during active stepping and during passive stepping movement of the whole leg were not significantly different at the point where the limb approached full flexion in the swing phase (P > 0.48). This was the case for measurements made at either body position, 20 degrees dorsal tilt or supine. 5. Passive flexion-extension, around either the hip or the knee, significantly inhibited the mean reflex magnitude close to full flexion, at either body position (P < 0.01). Such movement around the ankle resulted in significant inhibition of the reflex in two of the four subjects (P < 0.05). The numeric sum of the reflex depression arising from the flexion-extension of the individual joints was greater than that arising from movement of the whole limb. 6. With the ankle braced, the significant reflex attenuation remained when a tonic isometric contraction of the soleus muscle was introduced. This suggests premotoneuronal mechanisms for the inhibition.(ABSTRACT TRUNCATED AT 400 WORDS)