Adenosine is a widely used pharmacological agent to induce a "high-flow" control condition to study the mechanisms of exercise hyperemia, but it is not known how well an adenosine infusion depicts exercise-induced hyperemia, especially in terms of blood flow distribution at the capillary level in human muscle. Additionally, it remains to be determined what proportion of the adenosine-induced flow elevation is specifically directed to muscle only. In the present study, we measured thigh muscle capillary nutritive blood flow in nine healthy young men using PET at rest and during the femoral artery infusion of adenosine (1 mg min(-1) l thigh volume(-1)), which has previously been shown to induce a maximal whole thigh blood flow of approximately 8 l/min. This response was compared with the blood flow induced by moderate- to high-intensity one-leg dynamic knee extension exercise. Adenosine increased muscle blood flow on average to 40 +/- 7 ml x min(-1) x 100 g muscle(-1) with an aggregate value of 2.3 +/- 0.6 l/min for the whole thigh musculature. Adenosine also induced a substantial change in blood flow distribution within individuals. Muscle blood flow during the adenosine infusion was comparable with blood flow in moderate- to high-intensity exercise (36 +/- 9 ml x min(-1) x 100 g muscle(-1)), but flow heterogeneity was significantly higher during the adenosine infusion than during voluntary exercise. In conclusion, a substantial part of the flow increase in the whole limb blood flow induced by a high-dose adenosine infusion is conducted through the physiological non-nutritive shunt in muscle and/or also through tissues of the limb other than muscle. Additionally, an intra-arterial adenosine infusion does not mimic exercise hyperemia, especially in terms of muscle capillary flow heterogeneity, while the often-observed exercise-induced changes in capillary blood flow heterogeneity likely reflect true changes in nutritive flow linked to muscle fiber and vascular unit recruitment.