1 We have investigated the vasodilating eects of D-erythro-C2-ceramide (C2-ceramide) in methoxamine-contracted rat mesenteric microvessels. 2 C2-ceramide (10 ± 100 mM) caused a concentration-dependent, slowly developing relaxation which reached maximum values after &10 min and partially abated thereafter. 3 Endothelium removal or inhibitors of guanylyl cyclase (3 mM ODQ), protein kinase A (10 mM H7, 1 mM H89) and various types of K + channels (10 mM BaCl 2 , 3 mM tetraethylammonium, 30 nM charybdotoxin, 30 nM iberiotoxin, 300 nM apamine, 10 mM glibenclamide) had only small if any inhibitory eects against C2-ceramide-induced vasodilation, but some of them attenuated vasodilation by sodium nitroprusside or isoprenaline. A combination of ODQ and charybdotoxin almost completely abolished C2-ceramide-induced vasodilation. 4 A second administration of C2-ceramide caused a detectable but weaker relaxation. L-threo-C2-ceramide (100 mM), which should not be a substrate to ceramide metabolism, had no biphasic time course. The ceramidase inhibitor (1S,2R)-D-erythro-2-(N-myristoylamino)-1-phenyl-1-propanol (100 mM) alone caused some vasodilation, indicating vasodilation by endogenous ceramides, and also hastened relaxation by exogenous C2-ceramide. The late-developing reversal of C2-ceramideinduced vasodilation was absent when a-adrenergic tone was removed by addition of 10 mM phentolamine. 5 We conclude that C2-ceramide relaxes rat resistance vessels in an endothelium-independent manner which is prevented only by combined inhibition of guanylyl cyclase and charybdotoxinsensitive K + channels. The vasodilation abates with time partly due to desensitization of the ceramide response and partly due to metabolism of C2-ceramide to an inactive metabolite.