The poisons of bacteria are, according to Sidney Martin, of a kind which cannot be fully expressed chemically, but only pathologically. They may be of a ferment nature in diphtheria and tetanus. The arguments in support of that view are--(1) that they act in infinitesimal doses, (2) that they may act slowly and produce death after many days by profoundly affecting the general nutrition, and (3) that they are sensitive to the action of heat in a way that no chemical poisons are known to be. If they are considered as ferments, they must be substances which have a peculiar affinity for certain tissues of the body on which they produce their special toxic effect. As for the products of digestion, they are formed either by the bacillus ingesting the proteid and discharging it as albumose, or the digestion occurs by means of a ferment secreted by the bacillus in the body of an individual or animal suffering from the disease. Sidney Martin suggests that _anthrax_ produces albumoses and an alkaloidal substance, the former producing fever, the latter stupor. In _tetanus_ the bacillus produces a secretion of the bacillus which causes the convulsions. The albumoses present in this disease are probably due to the secretory toxin. In _diphtheria_, too, we have a secretory poison in the membrane and in the tissues, and an albumose which is possibly the result of the secretion. It will be seen that these views differ in some particulars from those to which we have already referred. However the details of the _modus operandi_ of the formation of toxins are finally settled, we know that there comes a time when the disease symptoms vanish, the disease declines, and the patient recovers. Many of the older schools of medicine explained this satisfactory phenomenon by saying that this disease exhausted itself after having "gone through" the body. In a sense that idea is probably true; but recently a large number of investigators have applied themselves to this problem, and with some promising results.