Human locomotory performance is dependent upon the ability of skeletal muscle to generate mechanical power, and sustain that power -that is, resist fatigue. Not surprisingly the factors influencing this capability have attracted the attention of many investigators dating back to and beyond the beginning of this century (see e.g. Benedict & Cathcart, 1913;Krogh & Lindhard, 1920;Hill, 1922;Lupton, 1923;Dickenson, 1928;Wilkie, 1960Wilkie, , 1981Carnevale & Gaesser, 1991;McNaughton & Thomas, 1996). However, although there are a number of studies which have examined the maximum power of human locomotory muscles, and many more which have examined the constraints and limitations to sustained exercise -especially with respect to aerobic and anaerobic energy supply -rather few data are available from studies which have examined, in the same subjects, the relationship between maximum power and the power delivered in sustained exercise, and there are almost no data on the effect that movement frequency has on that relationship. In part, this paucity of data is due to the technical difficulty of measuring maximal power output at a constant known movement frequency in human locomotion. In seeking to address this difficulty, one of us developed an isokinetic cycle ergometer which enabled the maximum power generated by the main locomotory muscles to be measured over a range of movement frequencies (Sargeant et al. 1981). Subsequently, the ergometer system was modified so that the power could be measured continuously at the foot-pedal interface either during submaximal exercise, or during a maximum effort with the system switched to its isokinetic The effect of different pedalling rates (40, 60, 80, 100 and 120 rev min¢) on power generating capability, oxygen uptake (ýOµ) and blood lactate concentration [La]b during incremental tests was studied in seven subjects. No significant differences in ýOµ,max were found (mean ± s.d., 5.31 ± 0.13 l min¢). The final external power output delivered to the ergometer during incremental tests (PI,max) was not significantly different when cycling at 60, 80 or 100 rev min¢ (366 ± 5 W). A significant decrease in PI,max of •60 W was observed at 40 and 120 rev min¢ compared with 60 and 100 rev min¢, respectively (P < 0.01). At 120 rev min¢ there was also a pronounced upward shift of the ýOµ-power output (ýOµ-P) relationship. At 50 W ÄýOµ between 80 and 100 rev min¢ amounted to +0.43 l min¢ but to +0.87 l min¢ between 100 and 120 rev min¢. The power output corresponding to 2 and 4 mmol l¢ blood lactate concentration (P[La]2 and P[La]4 ) was also significantly lower (> 50 W) at 120 rev min¢ (P < 0.01) while pedalling at 40, 60, 80 and 100 rev min¢ showed no significant difference. The maximal peak power output (PM,max) during 10 s sprints increased with pedalling rate up to 100 rev min¢. Our study indicates that with increasing pedalling rate the reserves in power generating capability increase, as illustrated by the PI,maxÏPM,max ratio (54.8, 44.8, 38.1, 34.6, 29.2%), the P[La]4ÏPM,max ratio (50.4...