Following the start of low-intensity exercise in healthy humans, it has been established that the kinetics of skeletal muscle O(2) delivery is faster than, and does not limit, the kinetics of muscle O(2) uptake (V(O(2)(m))). Direct data are lacking, however, on the question of whether O(2) delivery might limit (V(O(2)(m))) kinetics during high-intensity exercise. Using multiple exercise transitions to enhance confidence in parameter estimation, we therefore investigated the kinetics of, and inter-relationships between, muscle blood flow (Q(m)), a-(V(O(2))) difference and (V(O(2)(m))) following the onset of low-intensity (LI) and high-intensity (HI) exercise. Seven healthy males completed four 6 min bouts of LI and four 6 min bouts of HI single-legged knee-extension exercise. Blood was frequently drawn from the femoral artery and vein during exercise and Q(m), a-(V(O(2))) difference and (V(O(2)(m))) were calculated and subsequently modelled using non-linear regression techniques. For LI, the fundamental component mean response time (MRT(p)) for Q(m) kinetics was significantly shorter than (V(O(2)(m))) kinetics (mean ± SEM, 18 ± 4 vs. 30 ± 4 s; P < 0.05), whereas for HI, the MRT(p) for Q(m) and (V(O(2)(m))) was not significantly different (27 ± 5 vs. 29 ± 4 s, respectively). There was no difference in the MRT(p) for either Q(m) or (V(O(2)(m))) between the two exercise intensities; however, the MRT(p)for a-(V(O(2)) difference was significantly shorter for HI compared with LI (17 ± 3 vs. 28 ± 4 s; P < 0.05). Excess O(2), i.e. oxygen not taken up (Q(m) x (V(O(2))), was significantly elevated within the first 5 s of exercise and remained unaltered thereafter, with no differences between LI and HI. These results indicate that bulk O(2) delivery does not limit (V(O(2)(m))) kinetics following the onset of LI or HI knee-extension exercise.