1. The contractile responses of spiral strips of human digital arteries to samples of a suspension of human platelets aggregated by thrombin have been studied at different time intervals after aggregation. 2. Platelets added to the arterial strips 1 min after aggregation of the platelets produced contractile responses which were significantly greater than those produced by corresponding platelets added 20 min after aggregation. 3. When platelets were aggregated in the presence of indomethacin or the thromboxane synthetase antagonist 1-benzylimidazole, contractile responses produced by the platelets 1 min after aggregation were significantly reduced. They were then not significantly different from those produced by addition of the aliquots 20 min after aggregation, which were unaffected. 4. Pretreatment of the arteries with the serotonin antagonist ketanserin nearly abolished the contractile responses produced by addition of the platelets 20 min after aggregation, and significantly reduced those produced by addition of the platelets 1 min after aggregation. 5. Ketanserin did not affect the contractile responses of the arteries to potassium chloride, prostaglandin F(2α), or the endoperoxide analogue U-46619, but antagonized the contractile effects of exogenous serotonin. 6. Combination of pretreatment of the arteries with ketanserin and aggregation of the platelets in indomethacin or 1-benzylimidazole virtually abolished contractile responses to platlets added both 1 min and 20 min after aggregation. 7. Tensions developed to different dilutions of platelets added 1 min after aggregation to arteries pretreated with ketanserin were not significantly different from those obtained to the same dilutions added 20 min after aggregation to arteries not pretreated with ketanserin. 8. It is concluded that there are two components to the contractile responses of human blood vessels produced by the aggregation of human platelets. One is labile, not produced when aggregation occurs in indomethacin or 1-benzylimidazole and is probably produced by thromboxane A2. The other is stable, markedly reduced by ketanserin, and probably produced by serotonin. Each of the components makes an approximately equal contribution to the contractures of human blood vessels produced by platelet aggregation. 9. It is also concluded that serotonin antagonists, in combination with a thromboxane synthetase inhibitor, may be of therapeutic benefit in clinical syndromes possibly mediated by platelet aggregation, such as cerebral transient ischaemic attacks and coronary spasm.