Anti-substances may be arranged, as has been done by Ehrlich, into three main groups. In the first group, the anti-substance simply combines with the antigen, without, so far as we know, producing any change in it. The antitoxins are examples of this variety. In the second group, the anti-substance, in addition to combining with the antigen, produces some recognizable physical change in it; the precipitins and agglutinins may be mentioned as examples. In the third group, the anti-substance, after it has combined with the antigen, leads to the union of a third body called complement (alexine or cytase of French writers), which is present in normal serum. As a result of the union of the three substances, a dissolving or digestive action is often to be observed. This is the mode of action of the anti-substances in the case of a haemolytic or bacteriolytic serum. So far as bacterial immunity is concerned, the anti-serum exerts its action either on the toxin or on the bacterium itself; that is, its action is either antitoxic or anti-bacterial. The properties of these two kinds of serum may now be considered.

The term "antitoxic" signifies that serum has the power of Antitoxic serum. neutralizing the action of the toxin, as is shown by mixing them together outside the body and then injecting them into an animal. The antitoxic serum when injected previously to the toxin also confers immunity (passive) against it; when injected after the toxin it has within certain limits a curative action, though in this case its dose requires to be large. The antitoxic property is developed in a susceptible animal by successive and gradually increasing doses of the toxin. In the earlier experiments on smaller animals the potency of the toxin was modified for the first injections, but in preparing antitoxin for therapeutical purposes the toxin is used in its unaltered condition, the horse being the animal usually employed. The injections are made subcutaneously and afterwards intravenously; and, while the dose must be gradually increased, care must be taken that this is not done too quickly, otherwise the antitoxic power of the serum may fall and the health of the animal suffer. The serum of the animal is tested from time to time against a known amount of toxin, i.e. is standardized.

The unit of antitoxin in Ehrlich's new standard is the amount requisite to antagonize 100 times the minimum lethal dose of a particular toxin to a guinea-pig of 250 grm. weight, the indication that the toxin has been antagonized being that a fatal result does not follow within five days after the injection. In the case of diphtheria the antitoxic power of the serum may reach 800 units per cubic centimetre, or even more. The laws of antitoxin production and action are not confined to bacterial toxins, but apply also to other vegetable and animal toxins, resembling them in constitution, viz. the vegetable toxalbumoses and the snake-venom group referred to above.

The production of antitoxin is one of the most striking facts of Action of antitoxin. biological science, and two important questions with regard to it must next be considered, viz. how does the antitoxin act? and how is it formed within the body? Theoretically there are two possible modes of action: antitoxin may act by means of the cells of the body, i.e. indirectly or physiologically; or it may act directly on the toxin, i.e. chemically or physically. The second view may now be said to be established, and, though the question cannot be fully discussed here, the chief grounds in support of a direct action may be given. (a) The action of antitoxin on toxin, as tested by neutralization effects, takes place more quickly in concentrated than in weak solutions, and more quickly at a warm (within certain limits) than at a cold temperature. (b) Antitoxin acts more powerfully when injected along with the toxin than when injected at the same time in another part of the body; if its action were on the tissue-cells one would expect that the site of injection would be immaterial. For example, the amount necessary to neutralize five times the lethal dose being determined, twenty times that amount will neutralize a hundred times the lethal dose.

In the case of physiological antagonism of drugs this relationship does not hold. (c) It has been shown by C. J. Martin and Cherry, and by A. A. Kanthack and Cobbett, that in certain instances the toxin can be made to pass through a gelatine membrane, whereas the antitoxin cannot, its molecules being of larger size. If, however, toxin be mixed with antitoxin for some time, it can no longer be passed through, presumably because it has become combined with the antitoxin.

Lastly it may be mentioned that when a toxin has some action which can be demonstrated in a test-tube experiment, for example, a dissolving action on red corpuscles, this action may be annulled by previously adding the antitoxin to toxin; in such a case the intervention of the living tissues is excluded. In view of the fact that antitoxin has a direct action on toxin, we may say that theoretically this may take place in one of two ways. It may produce a disintegration of the toxin molecule, or it may combine with it to produce a body whose combining affinities are satisfied. The latter view, first advocated by Ehrlich, harmonizes with the facts established with regard to toxic action and the behaviour of antitoxins, and may now be regarded as established. His view as to the dual composition of the toxin molecule has already been mentioned, and it is evident that if the haptophorous or combining group has its affinity satisfied by union with antitoxin, the toxin will no longer combine with living cells, and will thus be rendered harmless. One other important fact in support of what has been stated is that a toxin may have its toxic action diminished, and may still require the same amount of antitoxin as previously for neutralization.

This is readily intelligible on the supposition that the toxophorous group is more labile than the haptophorous. There is, however, still dispute with regard to the exact nature of the union of toxin and antitoxin. Ehrlich's view is that the two substances form a firm combination like a strong acid and a base. He found, however, that if he took the largest amount of toxin which was just neutralized by a given amount of antitoxin, much more than a single dose of toxin had to be added before a single dose was left free. For example, if 100 doses of toxin were neutralized by a unit of antitoxin (v. supra) it might be that 125 doses would need to be added to the same amount of antitoxin before the mixture produced a fatal result when it was injected. This result, which is usually known now as the "Ehrlich phenomenon," was explained by him on the supposition that the "toxin" does not represent molecules which are all the same, but contains molecules of different degrees of combining affinity and of toxic action. Accordingly, the most actively toxic molecules will be neutralized first, and those which are left over, that is, uncombined with antitoxin, will have a weaker toxic action.