Five subjects exercised with the knee extensor of one limb at work loads ranging from 10 to 60 W. Measurements of pulmonary oxygen uptake, heart rate, leg blood flow, blood pressure and femoral arterial-venous differences for oxygen and lactate were made between 5 and 10 min of the exercise. Flow in the femoral vein was measured using constant infusion of saline near 0 degrees C. Since a cuff was inflated just below the knee during the measurements and because the hamstrings were inactive, the measured flow represented primarily the perfusion of the knee extensors. Blood flow increased linearly with work load right up to an average value of 5.7 l min-1. Mean arterial pressure was unchanged up to a work load of 30 W, but increased thereafter from 100 to 130 mmHg. The femoral arterial-venous oxygen difference at maximum work averaged 14.6% (v/v), resulting in an oxygen uptake of 0.80 l min-1. With a mean estimated weight of the knee extensors of 2.30 kg the perfusion of maximally exercising skeletal muscle of man is thus in the order of 2.5 l kg-1 min-1, and the oxygen uptake 0.35 l kg-1 min-1. Limitations in the methods used previously to determine flow and/or the characteristics of the exercise model used may explain why earlier studies in man have failed to demonstrate the high perfusion of muscle reported here. It is concluded that muscle blood flow is closely related to the oxygen demand of the exercising muscles. The hyperaemia at low work intensities is due to vasodilatation, and an elevated mean arterial blood pressure only contributes to the linear increase in flow at high work rates. The magnitude of perfusion observed during intense exercise indicates that the vascular bed of skeletal muscle is not a limiting factor for oxygen transport.