To test the effect of cooling on EMG-activity of muscles working as an agonist and antagonist in the lower leg, 12 men dressed in shorts and jogging shoes performed a drop-jump exercise after 60 min exposures to 27 degrees C and 10 degrees C. Cooling decreased mean skin temperature 5.6 +/- 0.4 degrees C (mean +/- SD, P < 0.001), whereas rectal temperature was unaffected. The muscle temperature measured from m. gastrocnemius medialis decreased 4.1 +/- 0.3 degrees C (P < 0.01) at the depth of 30 mm below skin surface. To find the optimal stretching velocity for potentiation of elastic energy, the drop-jump exercise was performed from six different bench heights (10, 20, 30, 40, 50, and 60 cm). The optimal velocity was not altered on account of cooling. In cooled subjects during the stretch phase of the drop jumps the EMG-activity of m. triceps surae complex (agonist) increased (P < 0.05-0.001) while the activity of m. tibialis anterior (antagonist) remained unchanged. After cooling during the shortening phase of the jumps the EMG-activity of m. triceps surae complex decreased (P < 0.05-0.001), whereas the activity of m. tibialis anterior increased (P < 0.05-0.001). In addition, after cooling the peak EMG-activity appeared on the average 28 ms earlier, which shifted the peak activity from the shortening phase (at 27 degrees C) to the stretch phase (at 10 degrees C). Cooling increased the mean duration of stretch and shortening phases by 28 +/- 3 ms (P < 0.001) and 23 +/- 2 ms (P < 0.001), respectively. The average force production during the shortening phase was 26% less (P < 0.05) after cooling, which resulted in a decreased rise of body centre of gravity (P < 0.05-0.01). It is concluded that during a stretch-shortening cycle cooling alters the EMG-activity of agonist and antagonist muscles on a contradictory manner and results in an earlier peak EMG-activity. Therefore, alterations in motor unit recruitment could be responsible for the prolonged muscle contraction and decreased force production on account of cooling.