Third, lytic agents in the blood serum may play the chief rôle in the liberation of the toxic agent from its non-toxic combination. An immune individual would then be one whose immunity was not the positive one of antibody formation, but the negative immunity of failure to metabolize. An immune lesion in such a case might be conceived as presenting no eosinophilia, since no toxin is liberated. If the liberation of the toxin is dependent upon lytic agents present in the serum rather than in any cellular elements, a rational explanation would be available for the apparent results (subject to confirmation) of the experiment with sensitive and immune sera. In this experiment it will be recalled that the sensitive serum seemed to bring out the toxicity of the ground flies, and the serum itself seemed even to contain some of the dissolved or liberated toxin. The slowness with which a lesion develops in the case of the black-fly bite supports the view of the initial lack of toxicity of the injected material. The entire absence of early subjective symptoms, such as pain, burning, etc., is further evidence for this view. It would appear as if no reaction occurred until lysis of an originally non-toxic substance had begun. Regarding the toxin itself as the chemotactic agent which attracts eosinophiles, its liberation in the lytic process and diffusion through the blood stream attracts the cells in question to the point at which it is being liberated. Arriving upon the scene, these cells assist in its neutralization. The last view presented is the one to which the author inclines as the one which most adequately explains the phenomena. A fourth view is that the initial injection of a foreign protein by the fly (i.e., with the first bite) sensitizes the body to that protein. Its subsequent injection at any point in the skin gives rise to a local expression of systematic sensitization. Such local sensitization reactions have been described by Arthus and Breton, by Hamburger and Pollack and by Cowie. The description of such a lesion given by the first named authors, in the rabbit, however, does not suggest, histopathologically at least, a strong resemblance to that of the black-fly. Such an explanation of many insect urticariæ deserves further investigation, however, and may align them under cutaneous expressions of anaphylaxis to a foreign protein injected by the insect. Depending on the chemical nature of the protein injected, a specific chemotactic reaction like eosinophilia may or may not occur. Viewed in this light the development of immunity to insect bites assumes a place in the larger problem of anaphylaxis. [Illustration: 174. Experimental lesion produced from alcohol-fixed flies, dried and ground into a paste with glycerin. After Dr. J. H. Stokes] SUMMARY In order to bring the results of the foregoing studies together, the author appends the following résumé of the clinical data presented in the first paper.