The construction of a balance-sheet between the sugar fermented and the products formed is of special interest in the case of alcoholic fermentation by yeast-juice, because, there being no cell growth as in the case of living yeast, an opportunity appears to be afforded of ascertaining whether the whole of the sugar is converted into alcohol and carbon dioxide, or whether some fraction of the sugar passes into any of the well-known subsidiary products of alcoholic fermentation by yeast, such as glycerol, fusel oil, or succinic acid. Unfortunately the question cannot be settled in this way. When the loss of sugar during the fermentation is estimated directly, it is usually found to be considerably greater than the sum of the alcohol and carbon dioxide produced from it. This fact was first observed by Macfadyen, Morris and Rowland [1900], and was then confirmed by Buchner [Buchner, E. and H., and Hahn, 1903, p. 212], in one instance, the excess of sugar lost over products being in this case about 15 per cent. of the total sugar which had disappeared. The matter was then more thoroughly investigated by Harden and Young [1904]. [p031] The conditions under which the experiment must be carried out are not very favourable to the attainment of extreme accuracy. Yeast-juice contains glycogen and a diastatic enzyme which converts this into dextrins and finally into sugar. This process goes on throughout fermentation, tending to increase the sugar present and to make the apparent loss of sugar less than the sum of the products. In spite of this it was found that a certain amount of sugar invariably disappeared without being accounted for as alcohol or carbon dioxide, and this whether the fermentation lasted sixty or a hundred and eight hours, and independently of the dilution of the juice. This disappearing sugar amounted in some cases to 44 per cent. of the total loss of sugar, and on the average of twenty-five experiments was 38 per cent. Further information was sought by converting all the sugar-yielding constituents of the juice into sugar by hydrolysis before and after the fermentation. This process revealed the fact that when the glucose equivalent of the juice before and after fermentation was determined after hydrolysis with three times normal acid for three hours (and a correction made for the loss of reducing power experienced by glucose itself when submitted to this treatment), the difference was almost exactly equal to the alcohol and carbon dioxide produced. In other words, accompanying fermentation, a change proceeds by which sugar is converted into a less reducing substance, reconvertible into sugar by hydrolysis with acids. Similar results were subsequently obtained by Buchner and Meisenheimer [1906], who employed 1·5 normal acid and observed a small nett loss of sugar. Still more recently Lebedeff [1909, 1910, see also 1913, 2] has carried out similar estimations with the same result. It is doubtful