The size of the long-latency component of the stretch reflex has been examined in the time interval between a signal to move and the required rapid voluntary contraction of triceps and flexor pollicis longus in 8 normal subjects. Bilateral movements of the elbow and thumb were made following an auditory signal. In 50% of the trials a torque pulse was applied unilaterally in order to elicit a stretch reflex response in one arm. The voluntary response in the contralateral arm was uncorrupted by a stretch reflex response, so was used as an indicator of voluntary reaction time. Control experiments, using an electrical stimulus to the fingers rather than muscle stretch, verified that both arms reacted almost simultaneously to the auditory cue, even when the reaction time was shortened by the presence of a unilateral electrical stimulus. Similarly, an interposed muscle stretch stimulus considerably reduced the reaction time to the audio signal. Because of this, the start of the voluntary EMG response frequently 'over-lapped' the end of the long-latency stretch reflex. Failure to take this shortening of voluntary reaction time into consideration can lead to the erroneous conclusion that reflex gain is increased prior to a rapid movement. If the 'overlap' of EMG responses is accounted for, very little change in the size of the long-latency stretch reflex is evident prior to activation of the muscles responsible for the movement of either the elbow or the thumb.