Coronary flow regulation in man is incompletely understood. We addressed the hypothesis that endogenous adenosine and prostacyclin (PGI2) affect basal coronary tone and/or the coronary flow response to increased myocardial work. In healthy volunteers coronary sinus flow and cardiac oxygen extraction were measured at rest and during leg exercise (40–50 W), before and after i. v. administration of theophylline (to block adenosine receptors) and oral administration of ibuprofen (to inhibit prostaglandin synthesis). Before drug, the basal coronary flow was 118 ± 23 ml min‐1and the cardiac extraction of oxygen 122 ± 3 ml 1‐1. Leg exercise elevated coronary flow by 89 ± 16 ml min′. Theophylline, at a dose blocking the coronary flow response to dipyridamole (an adenosine‐dependent mechanism), induced a moderate increase in myocardial oxygen extraction (by 11%,P<0.05), but failed to affect either the basal coronary flow (105 ± 16 ml min‐1) or the increase during exercise (88 ± 25 ml min‐1). Ibuprofen inhibited the urinary excretion of PGI2metabolite by about 90%, but did not affect basal coronary tone or the coronary response to exercise. These data indicate that endogenous adenosine is of limited importance for normal coronary regulation in man, and that