Using a step-wise, reductionist approach we characterized the time course and degree to which mechanical, vasodilatory and cardiac mechanisms contribute to the increase in leg blood flow (LBF) at the onset of dynamic knee-extensor exercise. Heart rate (HR) and LBF (ultrasound Doppler) were evaluated during (1) voluntary and (2) passive exercise in the seated position, (3) passive exercise in the supine position with the leg above the heart, and (4) passive exercise with measurements made in the non-moving leg. In trials 2 and 3, the degree of change and time course of peak DeltaHR (8.7 +/- 2 bpm, seated; 10 +/- 1 bpm, supine) and peak DeltaLBF (518 +/- 135 ml min(-1), seated; 448 +/- 179 ml min(-1), supine) were similar, supporting the concept that the skeletal muscle pump was minimized. Even with the reduction of skeletal muscle pump and metabolic influences (trials 2, 3 and 4) a significant cardio-acceleration and hyperaemia was seen. In the first 5 s of seated passive exercise, the retrograde component of the blood velocity profile was significantly greater than rest or the 5-20 s interval, which may suggest an arterial inflow that initially exceeded leg vasodilatation. Steady-state LBF (minutes 2 and 3) remained elevated during voluntary exercise, but returned to near baseline during passive movement. Taken together, these data suggest that cardio-acceleration (i.e. tachycardia) and mechanical forces other than the skeletal muscle pump play a role in reducing vascular resistance and ultimately increasing LBF at the onset of exercise, followed by steady-state LBF which matches muscle metabolic demand.