Prolongation of clotting times produced by hematin was investigated both in vitro and in vivo. Hematin-derived anticoagulant (HDA) was found to be due to a degradative product or derivative of hematin, and was generated in vitro in standing (aging) aqueous solutions of the parent compound. Generation of HDA in vitro was inhibited by antioxidants. The anticoagulant effect of HDA was inhibited by freshly prepared hematin, fresh Sn-protoporphyrin, imidazole, or the iron chelator desferrioxamine. Ferrioxamine did not inhibit HDA, and inhibition by imidazole was reversed with ferric citrate, suggesting a role for iron in the mechanism of HDA activity. HDA activity was dissociated from hematin in plasma by clotting with thrombin. HDA segregated into the clot fraction, whereas hematin remained largely in the serum fraction, suggesting that HDA may preferentially bind to fibrinogen. TLC and HPLC showed a single peak of HDA activity that was not associated with the parent compound. Evidence for HDA generation in vivo was found when rats were injected with fresh (no HDA) hematin. Prolongation of clotting times appeared after hematin appeared in the plasma, and anticoagulant activity persisted after a fall in plasma hematin concentration. Thus, there was a temporal dissociation between hematin and HDA, suggesting that a modification of hematin must occur in vivo before an anticoagulant effect is produced. Generation of HDA in vitro has implications for hematin preparation and administration. Generation of HDA in vivo suggests that similar modifications of endogenous heme or other porphyrins may occur to produce HDA under physiologic or pathophysiologic conditions.