1. The effects of infusion of adenosine and its analogues on arterial and venous resistance have been studied in the vascularly and sympathetically isolated hind limb of chloralose-anaesthetized dogs. Resistance changes have been assessed by monitoring changes in perfusion pressures at constant flow through the femoral artery and metatarsal vein.2. With sympathetic stimulation (10 V, 2 msec, 0.5-2 Hz applied to the cut peripheral end of the lumbar sympathetic trunk), continuous infusion of adenosine, to produce a concentration of approximately 2 x 10(-5)m, resulted in a near maximal sustained decrease in arterial perfusion pressure of 35.3+/-3.6%, and a decrease of about half this in venous perfusion pressure.3. Bolus injections of adenosine into the artery, to produce a concentration of about 7 x 10(-5)m, caused a transient decrease in resistance similar to that observed with continuous infusion. However, the venous response was smaller than that induced by continuous infusion.4. Withdrawal of sympathetic stimulation to the limb had little effect on adenosine-induced vasodilatation in the artery, but abolished the small response of the vein.5. Less than 5% vasodilatation was produced in the artery and vein by 2-deoxyadenosine, inosine, guanosine, xanthosine, cytidine or uridine when infused in amounts up to ten times, or by sodium phosphate (pH 7.4) in amounts one hundred times the maximal amounts of adenosine given.6. These results suggest that adenosine caused vasodilatation, at least in arterioles, largely by acting directly on vascular muscle rather than via presynaptic inhibition of noradrenaline release.