On May 29th, 1873, eighteen flasks, free from air and with necks drawn out to a fine point, containing must, were taken into one of our rooms at the _École Normale_. A jet of gas from an ordinary burner was passed over the surface of the glass down to the surface of the liquid, with the object of burning any particles of dust that might have been deposited from the atmosphere of the laboratory; the points of the flasks were then broken with a pair of ordinary scissors that had been passed through the flame of a spirit lamp; and, lastly, the tops of the flasks were taken off, just above the surface of the liquid, and the eighteen flasks thus became transformed into eighteen basins, each containing about 100 c.c. (about 3-½ fluid ounces) of grape must. These eighteen basins were placed on a table in the room, and left there for five days, precautions being taken to prevent any one from entering the room. The basins were examined on June 2nd: they all contained little flakes of floating _mycelium_, but none of them had any white streaks on their sides—a proof that they were destitute of _torulæ_—the liquid had remained very bright. With the contents of nine of the basins we filled two long-necked flasks that we had prepared for our purpose—that is to say, had heated to a certain point, just before using them, with the object of removing from their sides any foreign germs which they might have picked up in the laboratory. Up to July 10th, when we deemed it useless to carry our observations further, these flasks presented no signs of fermentation whatever. With the contents of the remaining nine basins, we filled two other long-necked flasks, but before doing so, we kept the basins for twenty-four hours (June 2nd to 3rd) in the basement of our laboratory. These two flasks soon set up an active fermentation, and deposited an abundance of yeast—an additional proof of the great difference in character of the germs floating about the room and those floating about the laboratory. On June 3rd we exposed, simultaneously, in the aforementioned room and also in our laboratory, seven basins prepared as just described. On June 8th all the basins in the room showed signs of fungoid growths, without any trace of _mycoderma vini_, or lines indicative of the presence of _torulæ_, whilst six of the seven in our laboratory had their sides covered with a white precipitate, and on the surface of the liquids a layer of isolated patches of fungoid growth. The liquid in these latter basins was poured into a long-necked flask, which it nearly filled, and in the course of forty-eight hours it began to show signs of alcoholic fermentation.[45] This is another striking proof of the difference between the number of germs of ferment and _torulaceæ_ in the air of our laboratory and that of an ordinary room. [Illustration: Fig. 12.]