Our experiments have confirmed the fact that the so called bacterial allergies are dependent upon a mechanism which differs materially from that determining true protein anaphylaxis. Anaphylaxis to protein substances of the bacteria probably occurs but plays a relatively unimportant rôle in the phenomena of infection. The bacterial allergies, however, are of great importance since they develop rapidly and render the infected animal highly vulnerable to products of the bacterial growth which are relatively innocuous for the normal animal. Neither the type-specific carbohydrate "residue antigens" (the "soluble specific substances" of Avery and Heidelberger) nor the antibodies reacting with them play any part whatever in bacterial allergy, and since these type-specific substances represent the haptophore groups of the whole bacteria by which they react with the agglutinins, precipitins, sensitizers, etc., of immune serum, allergy, as previously determined by Mackenzie and Woo, is in no way related to that phase of resistance which is determined by these antibodies. This does not, however, preclude the possibility that allergic hypersusceptibility may not in some way be related to other factors of resistance more definitely associated with cellular rather than with intravascular reactions. Our previous studies with Jennings and Ward in tuberculosis point in this direction (20). Guinea pigs can be actively sensitized with all the bacteria with which we have worked when repeated injections of whole bacteria or of the protein (nucleoprotein) fraction are administered. Large amounts of the latter are necessary since these materials are indifferent antigens, possibly because of the severe manipulations necessary in their production. Sensitiveness develops usually within 10 days after the first dose and increases with continued treatment for 3 or 4 weeks. Sensitiveness is relatively specific, by which we mean that there is a definite specificity which, however, in highly sensitive animals is not absolute and shows considerable overlapping. Continued treatment with considerable quantities of the above substances leads to gradual desensitization in animals in which there are no chronic foci present, which, as in tuberculosis, tends to continue the sensitization. Attempts at passive sensitization have been irregular and inconclusive. When any degree of sensitiveness has developed after the injection of immune sera, it has appeared late and has been of doubtful specificity. Conversely we have failed in any case to neutralize the activity of the active allergic constituents of bacterial extracts by incubation with any type of immune serum. We have failed so far to show any increased fixation of tuberculin material on the part of tuberculous tissues or on that of living tuberculous animals. These failures, however, seem to us of relatively slight importance since quantitative experiments of this nature are extremely difficult in the case of a substance as delicately potent for the tuberculous animal. On the other hand we have obtained definite, though irregular evidence that the incubation of O.T. with fragments of tuberculous lung tissue (less clearly with other tissues) leads to the formation of a substance that produces allergy-like lesions in the skin of normal guinea pigs. With somewhat greater regularity, similar treatment of O.T. has enhanced the potency of the tuberculin for tuberculous animals. And, in these experiments there was evidence that the factor responsible for this action was not easily separable from the cells themselves. When these experimental data are analytically considered they appear in many respects confusing and contradictory. There has been so much work done on the tuberculin reaction, moreover, that, in the face of experimental inconsistencies it would seem foolhardy to formulate more than tentative suggestions to explain the mechanism of these reactions. Nevertheless there are a few outstanding and sufficiently reliable facts which compel a limited number of definite deductions. In the first place there is no question of the complete independence of the true allergic phenomena from the ordinary bacterial antigen-antibody reactions. We know, moreover, that the allergic substance is chemically separable from the carbohydrate "residue" or haptophore group of the bacteria (Mueller, Laidlaw and Dudley). Indeed it has been shown by Long and Seibert (21) that the active allergic substance is either a protein in itself, or at any rate closely associated with the bacterial protein. Furthermore, the distinct, though limited, specificity of the allergic sensitiveness compels the conclusion that we are dealing with an immunological process in which the tissue cells acquire an increased specific capacity to react with this nitrogenous material, a capacity which, in principle, is not far removed from the supposed "sessile receptor" apparatus which is conventionally held responsible for protein anaphylaxis; and this analogy is further amplified by the apparent desensitization which continued treatment produced in many of our own experiments as well as in those of Mackenzie and Woo. Here, however, the analogy with protein anaphylaxis ends. Passive sensitization with any form of immune serum or with the sera of highly sensitized animals is either feeble or entirely unsuccessful and indicates quite convincingly that, whatever the receptor apparatus of the cells may be, it is not easily given up to the blood stream as are ordinary antibodies. Further than this, our tissue-tuberculin experiments, irregular and occasional as they were, nevertheless convinced us that: 1. The contact with the tissues of tuberculous animals results in the production of a toxic factor, not unlike the autolytic toxic materials of some bacteria. 2. The active cell constituent by which this action is wrought, is not easily separated from the cells, even by energetic methods of extraction. This close association of the entire process with the cells themselves is particularly significant in view of the obvious cell injury in which these delayed allergic effects differ from the ordinary urticarial, evanescent reactions associated with protein anaphylaxis. The process of allergy, as far as we can approach it then, may be conceived as follows: A nitrogenous, probably protein, constituent of the bacterial growth or of its body substance stimulates a specific reaction in the tissue cell by which its specific capacity to establish contact with this constituent is enhanced. The cell is thereby enabled to exert a, probably, enzyme-like effect upon this material in consequence of which a toxic substance is liberated, largely upon or possibly within the cell itself. Both processes may be dependent upon one and the same reaction body. But it seems more likely that increased contact and the increased cell activity are separately developed, an assumption which is rendered probable by the association of the highest degrees of allergy with inflammatory cell reactions, and by the fact that moderate and less specific allergic sensitiveness follows 10 or more days after the administration of considerable amounts of indifferent protein substances to guinea pigs. We interpret this as signifying that such injections may non-speciffcally increase cellular activity, a change which many earlier workers have spoken of as "cell irritability." Both processes are closely associated with the altered cell itself and the factors by which the reaction is brought about are not easily given up to the blood stream as are the antibodies formed in response to injections of proteins or whole bacteria. We are confronted, therefore, with an immunological mechanism which has some close analogies to those others in which circulating antibodies are formed, but which differs from these mainly in the intimacy with which the entire reacting system is associated with the cells themselves. It is difficult to conceive that a functional cell alteration, as profound as this, should be entirely unrelated to the phenomena of susceptibility or resistance.