It is to be remarked that in this case, as in the case recorded § IV. p. 70, in order to detect with certainty any alcoholic ferment, the contents of the basins were transferred to a long-necked flask; since where, as in the basins, a liquid has a large surface exposed to spontaneous impregnation, the strictly alcoholic fermentation may escape observation. The reason of this is that, when a liquid of large surface but small depth is exposed to the air it affords a suitable medium for the active development of moulds, which, by absorbing the oxygen which would dissolve in the liquid, checks the growth of the ferment, or even prevents its germination altogether. As a matter of fact we shall see that for their growth and multiplication ferment _cells_, and still more ferment _germs_ (the difference between the two will appear in Chap. V.), require a larger supply of oxygen just in proportion to their age, state of desiccation, and distance from the budding condition. Now, if the spores of moulds be present and effect a settlement in the liquid, the increase of the ferment, or even its actual germination, is prevented. But by collecting the liquid in a deep and narrow vessel, such as a long-necked flask, after it has been exposed to spontaneous impregnation, we deprive the moulds almost completely of oxygen, and so allow the ferment to exert its peculiar energies. The mere act of transferring enables the liquid to take up a sufficiency of oxygen, and a liberation of gas very speedily shows that fermentative action is going on. We must add further that sometimes in a liquid of large surface and shallow depth, in which but little ferment is formed, the evolution of carbonic acid gas may fail to be detected, by reason of its diffusing itself into the air slowly as it is formed. Footnote 46: Here we had to seek for a most minute quantity of alcohol, that no alcoholometer could have indicated. A certain sign of the presence of alcohol is contained in the first few drops distilled; these always assume the form of little drops or _striæ_ or, better still, oily tears, when alcohol is present in the distillate. The distillation should be effected with a small long-necked retort and a Liebig’s condenser. We must carefully watch the neck of the retort at the moment of boiling; should the liquid contain 1/1000 part of its volume of alcohol, we shall observe the indications given above for a short, but appreciable time. 1/10,000 of alcohol is difficult to judge, but with care and practice we may do it without failing. Collecting a third of each distillate, and supposing the limit of appreciation to stop at thousandths, in three distillations we may easily detect the presence of 1/10,000 of 1 c.c. of alcohol in a total volume of 100 c.c. Footnote 47: