Surgical denervation of rabbit ear blood vessel beds was combined with the isolated perfused rabbit ear technique to investigate the mechanism of atropine's vasodilator action. Intramuscular injection of atropine 0.2 mg/kg dilated the denervated blood vessels in the rabbit ear like innervated ones in vivo. Atropine at the maximal concentration (Cmax) of 3 x 10(-6) to 3 x 10(-4) M did not increase effluent flow of the isolated perfused denervated rabbit ear under constant perfusion pressure, but chlorpromazine at a Cmax of 10(-6) M and acetylcholine (ACh) at 2.5 x 10(-7) M significantly increased it and noradrenaline (NA) at 10(-7) M significantly decreased it. Atropine at Cmax of 3 x 10(-7) M did not affect, but at 3 x 10(-6) M it abolished the increase of the effluent flow induced by ACh 2.5 x 10(-7) M. Atropine at 3 x 10(-7) M did not affect it, but at 10(-6), 3 x 10(-6), and 10(-5) it significantly alleviated the decrease of effluent flow induced by NA 10(-7) M. Because the increase of effluent flow of rabbit ear under constant perfusion pressure reflects vasodilation of the ear to some extent, the study suggests that atropine has no direct vasodilator action; its vasodilator action is not attributed to blockade of M-cholinoreceptors located on the vascular wall; however, the alpha1-adrenoceptor might be a target site mediating atropine's vasodilator action in vivo.