Cumulative addition of atropine to the organ bath containing endothelium-intact (+E) rat aorta, which was precontracted with phenylephrine (PE, 1 μM) and subsequently relaxed with carbachol (1μM), caused biphasic changes in the vascular contractility of +E rat aortic rings. Low concentrations of atropine (10 nM–1.0 μM) caused progressive restoration of contraction to PE; whereas at higher concentrations (1–100 μM), atropine caused progressive relaxation. Atropine-induced aortic relaxation was significantly inhibited upon endothelium removal by either rubbing or saponin treatment, but considerable relaxation still persisted in the range of 30–100 μM atropine. Similar findings were also obtained when the nitric oxide (NO) generation was inhibited with 300 μM NO synthase inhibitor, L-NAME. Atropine-induced relaxation was also observed when 5-hydroxytryptamine (5-HT) was used as the agonist and the atropine-relaxation was more potent at lower concentrations of PE and 5-HT. However, atropine had no effect on the contraction elicited by KCl or prostaglandin F2α. Also, atropine-induced relaxation was not affected by indomethacin (1–10 μM), nicotine (10–100 μM) or hexamethonium (30 μM). Pretreatment of +E aorta with tetraethylammonia (TEA, 3–10 mM) or 4-aminopyridine (4-AP, 1–3 mM) showed prominent inhibitory effect on atropine-induced relaxation; on the other hand, preincubation with glibenclamide (1–10 μM), BaCl2 (1–30 μM) or 2 μM charybdotoxin and apamin, had little effect on the relaxation induced by atropine. When added to tissues after relaxation to atropine, TEA and 4-AP concentration-dependently reversed the relaxation in -E aorta, whereas in +E aorta, TEA up to 30 mM and 4-AP up to 10 mM only partially affected atropine-induced relaxation. Although TEA and 4-AP potentiated the PE-contraction, such potentiation is unlikely to contribute to the change in sensitivity to atropine-induced relaxation, since in the presence of 15 mM KCl, which also potentiated PE-contraction to a comparable extent, the atropine-relaxation remains unchanged. Scopolamine also acts like atropine, except that the effect of scopolamine was smaller than that of atropine and is primarily endothelium-dependent. Atropine-induced relaxation also occurs in medium artery (renal artery) and small muscular artery (mesenteric artery). In conclusion, atropine-relaxation is mediated in part via voltage-dependent K+ channels in both smooth muscle and endothelium and forms the mechanistic basis for the observed vasodilation, reduced blood pressure and facial flushing following atropine overdose.