IX. Summary: 1. Experimental data: 2. Theoretical conclusions: The available evidence indicates that in dogs the blood contains no toxic substance during anaphylactic shock. All of the symptoms of the shock appear to be traceable to the direct effect of the antigen upon the sensitized liver. This has been demonstrated of the incoagulability of the blood, and of the congestion of the liver itself. The latter phenomenon appears adequate to produce the fall in the general blood pressure by removing a large percentage of the blood from circulation. The evidences of proteolytic change in the blood are exactly the same as those which accompany other forms of hepatic injury, such as phosphorus or chloroform poisoning. The peptone theory of hepatic shock maintains that peptones are produced during the anaphylactic reaction, and that they are directly responsible for the symptoms. There are no experimental observations which establish this theory. It is based chiefly on the resemblances between peptone intoxication and anaphylactic shock, and these are adequately explained by the fact that in both conditions the same hepatic mechanism is affected. The peptone theory cannot be harmonized with the experimental data recorded in this paper. Anaphylactic shock in dogs, therefore, appears to be a cellular reaction, the site of which is the hepatic parenchyma. Other tissues may participate, but at present there is nothing to indicate that this is the case. In the guinea-pig the chief site of the cellular reaction is the smooth muscle tissue. The fact that in both species the fundamental mechanism is a cellular reaction, even though different tissues are involved, furnishes the basis for a uniform theory of anaphylaxis. The data herewith advanced indicate that the humoral theory of anaphylaxis (which is intimately bound up with the conception of a circulating toxic product, such as anaphylatoxin), is inapplicable to dogs, as previous studies have shown it to be inapplicable to guinea-pigs. The fact that peptone affects the liver in very much the same manner as does a specific antigen to which the organ is sensitized may serve to explain the therapeutic effects observed after the injection of this substance in certain infections, such as typhoid fever. These results introduce us to a new function of the liver, namely its participation in the immune reaction. This participation does not appear to be indirect, as had previously been believed, namely by virtue of the toxic degradation of the antigen. On the contrary, it is a direct and immediate reaction profoundly involving the functions and structure of the organ. The accompanying general symptoms appear to be merely accidental byproducts of this reaction, just as asphyxia in the guinea-pig is an accidental result of the contraction of the smooth musculature. It is perhaps needless to add that the hepatic reaction does not occur during the course of the infectious diseases in the exaggerated form induced by the anaphylactic experiment; but that it plays a more subdued and continuous rôle can hardly be doubted, in view of the direct evidence afforded by the study of the blood in human serum sickness. Similar effects upon coagulability as determined in anaphylactic guinea-pigs lead to the belief that the liver is probably a constant and important factor in the immune reaction throughout the mammalia.