In the course of some preliminary experiments (commenced by the late Allan Macfadyen, but subsequently abandoned) on the production of anti-ferments by the injection of yeast-juice into animals, the serum of the treated animals was tested for the presence of such antibodies both for the alcoholic and proteoclastic enzymes of yeast-juice, and it was then observed that the serum of normal and of treated animals alike greatly diminished the autolysis of yeast-juice. As the explanation of the comparatively rapid disappearance of the fermenting power from yeast-juice had been sought, as already mentioned (p. 20), in the hydrolytic action of the tryptic enzyme which always accompanies zymase, the experiment was made of carrying out the fermentation in the presence of serum, with the result that about 60 to 80 per cent. more sugar was fermented than in the absence of the serum [Harden, 1903]. This fact was the starting-point of a series of attempts to obtain a similar effect by different means, in the course of which a boiled and filtered solution of autolysed yeast-juice was used, in the hope that the products formed by the action of the tryptic enzyme on the proteins of the juice would, in accordance with the general rule, prove to be an effective inhibitant of that enzyme. This solution was, in fact, found to produce a very marked increase in the total fermentation effected by yeast-juice, the addition of a volume of boiled juice equal to that of the yeast-juice doubling the amount of carbon dioxide evolved [Harden and Young, 1905, 1]. This effect was found to be common to the filtrates from boiled fresh yeast-juice and from boiled autolysed yeast-juice, and was ultimately traced in the main, not to the antitryptic effect which had been surmised, but to two independent factors, either of which was capable in some degree of bringing about the observed result. Boiled yeast-juice was indeed found to possess a decided anti-autolytic effect, as determined by a comparison of the amounts of nitrogen rendered non-precipitable by tannic acid in yeast-juice alone [p042] and in a mixture of yeast-juice and boiled juice on preservation [Harden, 1905]. The anti-autolytic effect, however, appeared to vary independently of the effect on the fermentation, and the conclusion was drawn, as stated above, that the increase in the alcoholic fermentation was not directly dependent on the decrease in the action of the proteoclastic enzyme but was due to some independent cause. The property possessed by boiled yeast-juice of diminishing the autolysis of yeast-juice has now been carefully examined by Buchner and Haehn [1910, 2] and ascribed by them to a soluble antiprotease (p. 65).