I 131 -labeled purified human fibrin clots were introduced into the circulation of dogs at various time intervals after an infusion of human plasmin. One hour later, the clots were removed and incubated in saline at 37 C. In most cases, at the time when these clots were in the circulation, there was no significant "free" plasmin activity demonstrable in the blood. Nevertheless, the clots removed from these animals exhibited in vitro lysis upon incubation. Plasmin-antiplasmin complexes were formed in vitro with no demonstrable caseinolytic activity. These complexes were as effective in dissolving fibrin clots in vivo as the same amounts of plasmin were alone. The activator (streptokinase) alone used to activate plasminogen to plasmin was significantly less effective than streptokinase-activated-plasmin, alone or in complex form with antiplasmin. These experiments were interpreted as pointing toward the ability of fibrin to compete effectively with antiplasmin for plasmin. Thus, antiplasmin may serve as a transport form or reservoir of plasmin, releasing it when fibrin clots are available, but protecting other plasma proteins from its proteolytic effect. Other phenomena apparently confirming this hypothesis have been mentioned.