glyceraldehyde can readily be converted into it by oxidation to glyceric acid followed by abstraction of water (Erlenmeyer), [p110] CH{2}(OH)·CH(OH)·CHO → CH{2}(OH)·CH(OH)·COOH → CH{3}·CO·COOH, + O −H{2}O and finally methylglyoxal CH{3}·CO·CHO is its aldehyde. 2. The decomposition of pyruvic acid into acetaldehyde and carbon dioxide has already been fully discussed (Chapter VI). The universality of the enzyme carboxylase in yeasts and the rapidity of its action on pyruvic acid form the strongest evidence at present available in favour of the pyruvic acid theory. Given the pyruvic acid, there is no doubt that yeast is provided with a mechanism capable of decomposing it at the same rate as an equivalent amount of sugar. 3. The final step postulated by the pyruvic acid theory is the quantitative reduction to ethyl alcohol of the acetaldehyde formed from the pyruvic acid. The idea that acetaldehyde is an intermediate product in the various fermentations of sugar has frequently been entertained [Magnus Levy, 1902; Leathes, 1906; Buchner and Meisenheimer, 1908; Harden and Norris, D., 1912] although no very definite experimental foundation exists for the belief. It is, however, a well-known fact that traces of acetaldehyde are invariably formed during alcoholic fermentation [see Ashdown and Hewitt, 1910], and this is of course consistent with the occurrence of acetaldehyde as an intermediate product. Important evidence as to the specific capability of yeast to reduce acetaldehyde to alcohol has been obtained by several workers. Thus Kostytscheff [1912, 3; Kostytscheff and Hübbenet, 1913] found that pressed yeast, dried yeast and zymin all reduced acetaldehyde to alcohol, 50 grams of yeast in 10 hours producing from 660 mg. of aldehyde 265 mg. of alcohol in excess of the amount produced by autofermentation in absence of added aldehyde. Maceration extract was found to reduce both in absence and in presence of sugar, whereas Lebedeff and Griaznoff [1912] obtained no reduction in presence of sugar, and observed that the power of reduction was lost by the extract on digestion, a circumstance which suggests the co-operation of a co-enzyme in the process. Neuberg and Kerb [1912, 4; 1913, 1] have also been able to show by large scale experiments that alcohol is produced in considerable quantity by the fermentation of pyruvic acid by living yeast in absence of sugar and that the yield is increased by the presence of glycerol. When treated with 22 kilos, of yeast, 1 kilo, of pyruvic acid yielded 241 grams of alcohol in excess of that given by the yeast alone, whilst in presence of glycerol the amount was 360 grams, the amount theoretically obtainable being 523 grams. The function of the glycerol is not understood but is probably that of lessening the rate of destruction of the yeast enzymes. [p111]