The effect of Ehrlich's work has been clearly to distinguish the chemical changes involved in the production of fusel oil and succinic acid from those concerned in the decomposition of sugar into alcohol and carbon dioxide, and to bring to light a most important series of reactions by means of which the yeast-cell is able to supply itself with nitrogen, one of the indispensable conditions of life. [p096] CHAPTER VIII. THE CHEMICAL CHANGES INVOLVED IN FERMENTATION. It has long been the opinion of chemists that the remarkable and almost quantitative conversion of sugar into alcohol and carbon dioxide during the process of fermentation is most probably the result of a series of reactions, during which various intermediate products are momentarily formed and then used up in the succeeding stage of the process. No very good ground can be adduced for this belief except the contrast between the chemical complexity of the sugar molecule and the comparative simplicity of the constitution of the products. Many attempts have, however, been made to obtain evidence of such a series of reactions, and numerous suggestions have been made of probable directions in which such changes might proceed. In making these suggestions, investigators have been guided mainly by the changes which are produced in the hexoses by reagents of known composition. The fermentable hexoses, glucose, fructose, mannose, and galactose, appear to be relatively stable in the presence of dilute acids at the ordinary temperature, and are only slowly decomposed at 100°, more rapidly by concentrated acids, with formation of ketonic acids, such as levulinic acid, and of coloured substances of complex and unknown constitution. In the presence of alkalis, on the other hand, the sugar molecule is extremely susceptible of change. In the first place, as was discovered by Lobry de Bruyn [1895; Bruyn and Ekenstein, 1895; 1896; 1897, 1, 2, 3, 4], each of the three hexoses, glucose, fructose, and mannose is converted by dilute alkalis into an optically almost inactive mixture containing all three, and probably ultimately of the same composition whichever hexose is employed as the starting-point. This interesting phenomenon is most simply explained on the assumption that in the aqueous solution of any one of these hexoses, along with the molecules of the hexose itself, there exists a small proportion of those of an enolic form which is common to all the three hexoses, as illustrated by the following formulæ, the aldehyde formulæ [p097] being employed instead of the γ-oxide formulæ for the sake of simplicity:--