SUMMARY1. A number of metabolites have been proposed to control the vascular tone of skeletal muscle during exercise. The present study was designed to investigate the role of adenosine in this response by determining the effect of the adenosine receptor antagonist 8-phenyltheophylline.2. The gracilis muscle of anaesthetized cats was exposed and made to contract by stimulating the obturator nerve (at 1 Hz, 5 V, 041 ms) for 20 min. Gracilis muscle blood flow and tension were measured during exercise and for 20 min following exercise. Initially this was performed in each animal during the infusion of a vehicle solution (50% polyethylene glycol 400, 50% 0'1 M-NaOH, 01 ml min' i.v.). Exercise was then repeated during infusion of either further vehicle (group I), 8-phenyltheophylline (group II) or 3-propylxanthine (group III), both at 2-7 x 107 mol min' kg-.3. In group 1 (n = 4) gracilis muscle blood flow during the first exercise period increased by 47-5 + 11-3 ml min-' (110 g)-' and gracilis muscle tension by 8-6 + 1X3 kg (100 g muscle mass)-1 at 20 min of exercise. These responses were not significantly different when repeated. 4. In group II (n = 5), blood flow increased by 46-9 + 9.9 ml min' (100 g)-1 and tension by 6-5 + 0-7 kg (100 g muscle mass)-1 during vehicle infusion. Infusion of 8-phenyltheophylline at a rate which abolished the vasodilatation response to 2-chloroadenosine, significantly reduced the muscle blood flow increase to 19-8 + 2-7 ml min' (100 g muscle mass)-1 (P < 0-05) but the tension response was unaffected (increased by 7-0+0-8 kg (100 g muscle mass)-'). 8-Phenyltheophylline did not affect gracilis muscle blood flow or tension at rest.5. Administration of 3-propylxanthine, which did not modify the vasodilatation response to 2-chloroadenosine, failed to alter the vascular responses to muscle contraction.6. These results suggest that activation of adenosine receptors can contribute to up to 40 % of the vasodilatation observed during isometric twitch contraction of the gracilis muscle of cats.