Cumulative concentration response curves to 5-hydroxytryptamine (5-HT; 10(-10)-10(-4) mol/l) were constructed using isolated rings of equine digital, facial, tail and coronary arteries (endothelium intact). 5-HT was 17.7 and 41 times more potent as a vasoconstrictor of digital arteries than facial and tail arteries respectively. Removal of the endothelium increased the vasoconstrictor potency of 5-HT in the facial artery by 3.7-fold (P<0.05) but did not alter the sensitivity of digital arteries to 5-HT. Coronary arteries failed to contract to 5-HT. Coronary arteries pre-contracted with U44069 showed concentration dependent relaxation to 5-HT, a response which was partially dependent on the presence of the endothelium. No vasorelaxant effects were found in the digital or facial arteries. The concentration of 5-HT in platelet poor and platelet rich equine plasma was found to be 6.70+/-1.1 x 10(-8) mol/l and 1.77+/-0.36 x 10(-6) mol/l (mean +/-s.e.) respectively by high performance liquid chromatography (HPLC). Plasma which contained no detectable platelets had a 5-HT concentration of 1.12+/-0.48 x 10(-8) mol/l. Isolated digital arteries constricted when exposed to dilutions of platelet poor and platelet depleted equine plasma. These plasma induced contractions were almost completely inhibited by 5-HT receptor antagonists, ketanserin and methiothepin. The change in isometric tension in rings of equine digital artery in vitro was therefore used as a bioassay for plasma 5-HT and the results obtained by this method showed an excellent correlation (r2 = 97.2%, P<0.001) with the concentration estimated by HPLC. Circulating free concentrations of 5-HT in normal horses may be sufficient to constrict digital blood vessels partially in vivo but are well below the threshold for contraction of other peripheral blood vessels examined.