Recent findings in humans and primates suggest that the long latency electromyographic (EMG) responses, evoked by opposing sudden limb displacements, might be mediated via a transcortical pathway. However, it is not clear whether these "late" responses are servo controlled or preprogrammed. We have addressed these questions in two interrelated experimental series. Firstly, the late EMG responses (termed the FSR) evoked in the ankle muscles of 8-12 normal human subjects were mapped as a function of displacement amplitude and velocity. Secondly, the FSR evoked by sustained ramp displacements (lasting 500 msec) were compared with those elicited by transient pulse displacements (lasting 60 msec) having entirely different amplitude and velocity profiles. The findings demonstrated a general lack of a systematic relationship between the characteristics of the initial component of the FSR and displacement parameters, although exceptions were noted in individual subjects. Furthermore, no marked difference was found in either the latency or the rise time of the FSRs evoked by ramp and pulse displacements. The similarity of the FSRs in spite of widely different input patterns is not consistent with the behaviour of a servo response. However, such a finding is in agreement with the concept of a preprogrammed pattern of activity which, once triggered, has to run its full course without the possibility of modification. It is therefore concluded that, at least the initial part of the FSR is generated by the triggered release of a preformulated pattern of intended movement relatively independent of the specific pattern of limb displacement, rather than acting in the manner of a servo mechanism.