It must be remembered that proteins do not normally get into the circulation except by way of the alimentary tract. Here all proteins that are absorbed are first broken down to their constituent amino-acids, absorbed as such and these are built up into the proteins characteristic of the animal’s blood. Hence when protein as such gets into the blood it is a foreign substance to be disposed of. The blood contains proteolytic enzymes for certain proteins normally. It is also true that the body cells possess the property of digesting the proteins of the blood and building them up again into those which are characteristic of the cell. Hence it appears rational to assume that the foreign proteins act as stimuli to certain cells to produce more of the enzymes necessary to decompose them, so that they may be either built up into cell structure or eliminated as waste. If in this process of splitting up of protein a poison were produced, then the phenomena of “anaphylaxis” could be better understood. As a matter of fact Vaughan and his co-workers have shown that by artificially splitting up proteins from many different sources--animal, vegetable, pathogenic and saprophytic bacteria--a poison _is produced_ which appears to be the same in all cases and which causes the symptoms characteristic of anaphylaxis. On the basis of these facts it is seen that anaphylaxis is simply another variety of immunity. The _specific antibody_ in this case is an _enzyme_ which decomposes the protein instead of precipitating it. The enzyme must be specific for the protein since these differ in constitution. Vaughan even goes so far as to say that the poison is really the central ring common to all proteins and that they differ only in the lateral groups or side chains attached to this central nucleus. The action of the enzyme in this connection would be to split off the side chains, and since these are the specific parts of the protein, the enzyme must be specific for each protein. The pepsin of the gastric juice and the trypsin of the pancreas split the native proteins only to peptones. As is well known, these when injected in sufficient quantity give rise to poisonous symptoms, and will also give rise to anaphylaxis under properly spaced injections. They do not poison normally because they are split by the intestinal erepsin to amino-acids and absorbed as such. Whether Vaughan’s theory of protein structure is the true one or not remains to be demonstrated. It is not essential to the theory of anaphylaxis above outlined, _i.e._, a phenomenon due to the action of specific _antibodies_ which are enzymes. On physiological grounds this appears the most rational of the few explanations of anaphylaxis that have been offered and was taught by the author before he had read Vaughan’s theory along the same lines.