The control of pulmonary oxygen uptake (VO2) kinetics above the lactate threshold (LT) is complex and controversial. Above LT, VO2 for square-wave exercise is greater than predicted from the sub-LT VO2-WR relationship, reflecting the contribution of an additional "slow" component (VO2(sc)). Investigators have argued for a contribution to this slow component from the recruitment of fast-twitch muscle fibres, which are less aerobically efficient than slow-twitch fibres. Six healthy subjects performed a rapid-incremental bilateral knee-extension exercise test to the limit of tolerance for the estimation of VO2(peak), ventilatory threshold (VT), and the difference between VO2(peak) and VO2 at VT (Delta). Subjects then completed three repetitions of square-wave exercise at 30% of VT for 10 min (moderate intensity), and at VT + 25%Delta (heavy intensity) for 20 min. Pulmonary gas exchange was measured breath-by-breath. Surface EMG was recorded from m. rectus femoris; integrated EMG (IEMG) and mean power frequency (MPF) were derived for successive contractions. In comparison to moderate-intensity exercise, the phase 2 VO2 kinetics in heavy exercise were marginally slower than for moderate-intensity exercise (time constant (+/- SD) 25 +/- 9 and 22 +/- 10 s, respectively; NS), with a discernible VO2(sc) (VO2 difference between minutes 6 and 3 of exercise: 74 +/- 21 and 0 +/- 20 ml min(-1), respectively). However, there was no significant change in IEMG or MPF, either in the moderate domain or in the heavy domain over the period when the slow component was manifest. These observations argue against an appreciable preferential recruitment of fast-twitch units with high force-generating characteristics and fast sarcolemmal conduction velocities in concert with the development of the VO2 slow component during heavy-intensity knee-extensor exercise. The underlying mechanism(s) remains to be resolved.