Footnote 143: In what way are we to account for so great a difference between the two fermentations that we have just described? Probably, it was owing to some modification effected in the medium by the previous life of the bacteria, or to the special character of the vibrios used in impregnation. Or, again, it might have been due to the action of the air, which, under the conditions of our second experiment, was not absolutely eliminated, since we took no precaution against its introduction at the moment of filling our flask, and this would tend to facilitate the multiplication of anaërobian vibrios, just as, under similar conditions, would have been the case if we had been dealing with a fermentation by ordinary yeast. Footnote 144: In this case the liquid was composed as follows:—a saturated solution of lactate of lime, at a temperature of 25° C. (77° F.) was prepared, containing for every 100 c.c. (3-½ fl. oz.) 25·65 grammes (394 grains) of the lactate, C_{6}H_{5}O_{5}_Ca_O [_new notation_, C_{6}H_{10}_Ca_O_{6}]. This solution was rendered very clear by the addition of one gramme of phosphate of ammonia and subsequent filtration. For a volume of 8 litres (14 pints) of this clear, saturated solution, we used [1 gramme = 15·43 grains]:— Phosphate of ammonia 2 grammes. Phosphate of 1 grammes. potassium Phosphate of 1 grammes. magnesium Sulphate of ammonia 0·5 gramme. Footnote 145: [1 millimetre = 0·039 inch: hence the dimensions indicated will be—length, from 0·00039 to 0·00117, or even 0·00176 in.; diameter, from 0·000058 to 0·000078, rarely 0.000117 in.] Footnote 146: The carbonaceous supply, as we remarked, had failed them, and to this failure the absence of vital action, nutrition, and multiplication was attributable. The liquid, however, contained butyrate of lime, a salt possessing properties similar to those of the lactate. Why could not this salt equally well support the life of the vibrios? The explanation of the difficulty seems to us to lie simply in the fact that lactic acid produces heat by its decomposition, whilst butyric acid does not, and the vibrios seem to require heat daring the chemical process of their nutrition. Footnote 147: To do this, it is sufficient first to fill the curved ends of the stop-cocked tubes of the flasks, as well as the india-rubber tube _c c_, which connects them, with boiling water that contains no air. Footnote 148: