Many sports events require participants to exert a maximal effort in the closing stages-that is, after prior fatiguing exercise. Peak and mean pedalling rate during 30 s of high-intensity cycle ergometer exercise was manipulated by altering the applied resistance or the initial exercise intensity so that the effect of three contrasting strategies on subsequent peak power output could be examined. Seven female students cycled for 30 s in one of three conditions: (1) all-out effort against an applied resistance of 7.5% of body weight (test 1); (2) at a constant pace of 55% of the peak pedal rate of test 1 against a resistance of 10.9 +/- 0.4% of body weight (test 2); (3) all-out effort against the greater resistance (test 3). A 6 s sprint against the lesser resistance was performed 3 s after each test. Total work was greater (P < 0.01) in test 3 than in test 1, while mean pedal rate was higher (P < 0.01) in test 1 (mean +/- S.E.: 10.0 +/- 0.4 rad s-1) than in tests 2 and 3 (7.2 +/- 0.4 and 7.8 +/- 0.3 rad s-1 respectively). The peak power output in the subsequent 6 s sprint was similar following tests 2 and 3 (516 +/- 37 and 534 +/- 41 W respectively), but was lower following test 1 (420 +/- 37 W) (P < 0.01, test 1 vs tests 2 and 3). These results indicate that the number of muscle actions during 30 s of fatiguing exercise may exert a considerable influence on one's ability to subsequently produce peak power output. In sports such as cycling where the same external velocity is attainable at different muscle action speeds, then appropriate gear selection during the race will impact on the rider's ability to sprint in the latter stages.