(ii) C{3}H{6}O{3} + H·CO·NH{2} = CH{3}·CH(NH{2})·COOH + H·COOH. (iii) CH{3}·CH(NH{2})·COOH + H·COOH = CH{3}·CH{2}·OH + CO{2} + H·CO·NH{2}. According to this scheme all the sugar fermented passes through the form of alanine, and the formic acid acts along with the enzyme as catalyst, passing into formamide in reaction (iii) and being regenerated in (ii). The alanine is in the first place derived from the hydrolysis of proteins, or possibly by the reaction of the C{3}H{6}O{3} group with one of the higher amino-acids:-- C{3}H{6}O{3} + C{n}H{2n+1}·CH(NH{2})·COOH = C{n}H{2n+1}·CH{2}·OH + CO{2} + CH(NH{2})·COOH. There is as little positive evidence for this course of events as for that postulated by Schade, and the theory suffers from the additional disability that the chemical reactions involved have not been realised in the laboratory. Direct experiments with yeast-juice, moreover, show that a mixture of alanine with formic acid or a formate is not fermented, whilst neither the added mixture nor formamide seriously effects the action of the juice on glucose. OTHER THEORIES. Among other suggestions may be mentioned that of Kohl [1909] who asserts that sodium lactate is readily fermented, whilst Kusseroff [1910] holds the view that the glucose is first reduced to sorbitol and the latter fermented, in spite of the fact that sorbitol itself in the free state is not fermented by yeast. The rapid appearance and disappearance of glycogen in the yeast cell at various stages of fermentation [see Pavy and Bywaters, 1907; Wager and Peniston, 1910] has led to the suggestion [Grüss, 1904; Kohl, 1907] that this substance is of great importance in fermentation, and represents a stage through which all the sugar must pass before being fermented. The fact that the formation of glycogen has been observed in yeast-juice by Cremer [1899], and that complex carbohydrates are also undoubtedly formed (p. 31), are consistent with this theory. The low rate of autofermentation of living yeast, which is only a few per cent. of the rate of sugar fermentation, renders this supposition very improbable (Slator), as does the fact that the fermentation of glycogen by yeast-juice is usually slower than that of glucose [see also Euler, 1914]. An entirely different explanation of the chemical changes attendant on alcoholic fermentation has been suggested by [p117] Löb [1906; 1908, 1, 2; 1909, 1, 2, 3, 4; 1910; Löb and Pulvermacher, 1909], founded on the idea that the various decompositions of the sugar molecule both by chemical and biological agents are to be explained by a reversal of the synthesis of sugar from formaldehyde. As the sugar molecule can be built up by the condensation of formaldehyde, so it tends to break down again into this substance, and the products observed in any particular case are formed either by partial depolymerisation in this sense or by partial re-synthesis following on depolymerisation.