Some exact experiments conducted by M. Raulin in our laboratory have established the fact that saccharine worts, like water, soon become saturated when shaken briskly with an excess of air, and also that they always take into solution a little less air than saturated pure water contains under the same conditions of temperature and pressure. At a temperature of 25° C. (77° F.) therefore, if we adopt the coefficient of the solubility of oxygen in water given in Bunsen’s tables, we find that 1 litre (1-3/4 pints) of water saturated with air contains 5·5 c.c. (0·3 cubic inch) of oxygen. The three litres of yeast-water in the flask, supposing it to have been saturated, contained less than 16·5 c.c. (1 cubic inch) of oxygen, or, in weight, less than 23 milligrammes (0·35 grains). This was the maximum amount of oxygen, supposing the greatest possible quantity to have been absorbed, that was required by the yeast formed in the fermentation of 150 grammes (4·8 Troy ounces) of sugar. We shall better understand the significance of this result later on. Let us repeat the foregoing experiment, but under altered conditions. Let us fill, as before, our flask with sweetened yeast-water, but let this be first boiled, so as to expel all the air it contains. To effect this we arrange our apparatus as represented in the accompanying sketch (Fig. 60). We place our flask, A, on a tripod above a gas flame, and in place of the vessel of mercury substitute a porcelain dish, under which we can put a gas flame, and which contains some fermentable, saccharine liquid, similar to that with which the flask is filled. We boil the liquid in the flask and that in the basin simultaneously, and then let them cool down together, so that as the liquid in the flask cools some of the liquid is sucked from the basin into the flask. From a trial experiment which we conducted, determining the quantity of oxygen that remained in solution in the liquid after cooling, according to M. Schützenberger’s valuable method, by means of hydrosulphite of soda,[112] we found that the three litres in the flask, treated as we have described, contained less than one milligramme (0·015 grain) of oxygen. At the same time we conducted another experiment, by way of comparison (Fig. 61). We took a flask, B, of larger capacity than the former one, which we filled about half with the same volume as before of a saccharine liquid of identically the same composition. This liquid had been previously freed from alterative germs by boiling. In the funnel surmounting A, we put a few cubic centimetres of saccharine liquid in a state of fermentation, and when this small quantity of liquid was in full fermentation, and the yeast in it was young and vigorous, we opened the tap, closing it again immediately, so that a little of the liquid and yeast still remained in the funnel. By this means we caused the liquid in A to ferment. We also impregnated the liquid in B with some yeast taken from the funnel of A.