From the experiments above described it is evident that the viscous metamorphosis of the blood platelets in shed blood is intimately associated with the early stages of coagulation and that the presence of calcium is a very important element, though perhaps not absolutely necessary. No constant differences in this phenomenon could be detected with platelets or coagulating elements from the blood of normal or diseased individuals or from man or rabbits. All the substances, with the exception of egg white, causing this phenomenon were derived from the blood either during the alteration of fibrinogen or after it had been acted on by thrombin. It seems that the substances or mixtures capable of producing the metamorphosis are especially those associated with the early stages of coagulation or capable in the test-tube of forming or liberating that active coagulating element known as thrombin. The substance in serum that is capable of metamorphosing platelets seems to be attached to the globulin fraction rather than the albumin fraction and is destroyed by heat at temperatures which destroy prothrombin, thrombin, and serozyme, and precipitate fibrinogen. The reaction is not caused by pure thrombin or a mixture of pure thrombin and calcium, though substances causing the metamorphosis are intimately related to thrombin. The metamorphosis seems to be caused by serozyme-like substances as shown both by the fact that barium sulfate absorbs the power of serum to cause the reaction and that a serozyme-like substance is probably to be recognized in all the substances or mixtures, including egg white, causing this change in the platelets, except perhaps when thrombin and fibrinogen are reacting. It is to be noted that in serozyme are contained antithrombin, calcium, and potential thrombin, and that a combination of these isolated factors mixed together occasionally allowed the viscous metamorphosis to occur and not infrequently an abortive metamorphosis. The pseudo- or abortive metamorphosis caused by the mixture of pure thrombin, antithrombin, and calcium may be interpreted on the supposition of a close approximation but not a real reproduction of the colloidal state known as serozyme. That such a reaction is related to real viscous metamorphosis is suggested, because it sometimes occurred with the above mixture and when thrombin and heavily oxalated plasmas were reacting, rather than a real metamorphosis. Simple agglutination of the platelets may occur independently of a viscous metamorphosis, though an agglutination of platelets is to be considered an integral part of the viscous metamorphosis phenomenon. The inconstant results seen with the cephalin-treated sera are probably due to the fact that exactly the same mixtures were not obtained. The results with cephalin-soaked and platelet-soaked sera and with dilutions of sera suggest that the reaction of viscous metamorphosis of the platelets is quantitative.