It will be noticed that this theory resembles the pyruvic acid [p115] theory in postulating the immediate formation of acetaldehyde but differs from it by supposing that the reduction is effected at the expense of formic acid produced at the same time. The acetaldehyde question has already been discussed. In view of the fact that formic acid is a regular product of the action of many bacteria on glucose [see Harden, 1901], Schade's theory of alcoholic fermentation may be said to be a possible interpretation of the facts. Formic acid is known to be present in small amounts in fermented sugar solutions and the actual behaviour of yeast towards this substance has been investigated in some detail by Franzen and Steppuhn [1911; 1912, 1, 2], who have obtained results strongly reminiscent of those obtained with lactic acid by Buchner and Meisenheimer (p. 102). Many yeasts when grown in presence of sodium formate decompose a certain proportion of it, whereas in absence of formate they actually produce a small amount of formic acid--the absolute quantities being usually of the order of 0·0005 gram molecule (0·023 gram) per 100 c.c. of medium in 4 to 5 days. Only in the case of /S. validus/ did the consumption of formic acid in 5 days reach 0·0017 gram molecule (0·08 gram). Somewhat similar but rather smaller results were given by yeast-juice, a small consumption of formic acid being usually observed. The possibility thus exists that formic acid may be an intermediate product of alcoholic fermentation and Franzen argues strongly in favour of this view. Direct experiment, on the other hand, shows that yeast-juice cannot ferment a mixture of acetaldehyde and formic acid, even when these are gradually produced in molecular proportions in the liquid by the slow hydrolysis of a compound of the two, ethylideneoxyformate, OHC·O·CH(CH{3})·O·CH(CH{3})·O·CHO, this method being adopted to avoid the inhibiting effect of free acetaldehyde and formic acid [Buchner and Meisenheimer, 1910]. Nor is the reduction of acetaldehyde assisted by the presence of formate [Neuberg and Kerb, 1912, 4; Kostytscheff and Hübbenet, 1912]. A modified form of Schade's theory has been suggested by Ashdown and Hewitt [1910], who have found that when brewer's yeast is cultivated in presence of sodium formate the yield of aldehyde, as a rule, becomes less. They regard the aldehyde as derived from alanine, CH{3}·CH(NH{2})·COOH, one of the amino-acids formed from the proteins by hydrolysis, which is known to be attacked by yeast in the characteristic manner (p. 87), forming alcohol, carbon dioxide, and ammonia. Fermentation is supposed to proceed in such a way that the sugar is first decomposed into two smaller molecules, C{3}H{6}O{3} [p116] (equation i), and that these react with formamide to produce alanine and formic acid (ii). The alanine then enters into reaction with formic acid, producing alcohol, carbon dioxide, and formamide (iii):-- (i) C{6}H{12}O{6} = 2C{3}H{6}O{3}.