The foregoing and earlier data taken together demonstrate that an active immunity lasting several years can be produced in guinea-pigs, by the injection of toxin-antitoxin mixtures which have no recognizable harmful effect either immediate or remote. They also show, what might have been anticipated, that under the same conditions mixtures which produce local lesions and which, therefore, contain an excess of toxin produce a much higher degree of immunity than the neutral mixtures, and that an excess of antitoxin reduces the possibility of producing an active immunity, and may extinguish it altogether. There is, therefore, a certain definite relation between the components of the mixture and the degree of immunity producible. Furthermore, toxin-antitoxin mixtures do not change materially within five days at room temperature. They are apparently more efficacious at the end of forty-eight hours than immediately after preparation. The experiments finally prove that a relatively high degree of active immunity can be induced by a harmless procedure, whereas the use of toxin alone leading to very severe local lesions is incapable of producing more than an insignificant protection.

The method, therefore, invites further tests in regard to its ultimate applicability to the human being. Unless the subcutis of the guinea-pig reacts to toxin-antitoxin mixtures in a manner peculiar to itself, a practical, easily controlled method for active immunization can be worked out which should afford a larger protection than the serum alone and avoid the complications associated with horse serum. That proportion of toxin and antitoxin which would produce the highest desirable immunity consistent with the least discomfort would have to be carefully worked out for the human subject. From the nature of the immunity induced it is obvious, however, that such a method of immunization cannot take the place of a large dose of antitoxin in exposed individuals who must be protected at once. It would be applicable only as a general protective measure without reference to any immediate danger, since it would take several weeks, perhaps longer, to perfect the attainable immunity.

Passing to the theoretical aspects of the facts observed, we find no publications bearing directly upon the subject before us. Madsen has, however, approached it very closely in his experiments on the immunization of animals with mixtures not fully balanced, or, in other words, in which the "toxones" were still free. He found that the injection of such mixtures in rabbits, goats and horses produces an active immunity. He makes the significant remark that perhaps in the immunizing capacity we may possess the keenest reagent for a poison which is not able to exert any toxic action in the body. This is fully borne out by the experiments described, for in these we pass beyond the visible spectrum, so to speak, of the toxin-antitoxin effects, and we are able to recognize toxic action only by the lasting immunizing effects.

Another publication which touches upon some phases of the same problem is that of Morgenroth on the union between toxin and antitoxin. Morgenroth brought out the fact that a given toxin-antitoxin mixture is more toxic when injected directly into the circulation than when injected under the skin. Thus, an L₊ dose of 0.78 c.c. toxin + one unit antitoxin applied subcutaneously was of the same toxicity as 0.68 c.c. toxin + one unit antitoxin injected into the circulation. When the mixture had stood twenty-four hours this (L₊) dose was still 0.78 c.c. subcutaneously, but it had risen to 0.74 c.c. when introduced by the intracardiac route. The author makes two deductions from these results. He assumes that the …