In our Memoir of 1862, which we have quoted several times, we have explained that it is almost impossible to draw any serious conclusions from experiments made in a basin of mercury, because of the organic particles of dust which always exist in that metal, and which, without the knowledge of the operator, pass into the interior of the vessel, where they produce certain changes which one is tempted to impute to _heterogenesis_. In all classical works an experiment of Appert’s, reproduced by Gay-Lussac, is given. This, through a faulty interpretation, led to the hypothesis of the continuity of the causes of fermentation, if we may use such a term, in the atmospheric air.[39] When we decant bottles of must, which has been preserved by Appert’s method, into other bottles, all the latter soon set up a fermentation: this constitutes the experiment. If it were proved that the must, whilst being decanted, came in contact with atmospheric air alone, as Gay-Lussac believed, we should be compelled to admit, according to the theory of germs, that the must had come in contact with some particles of ferment in the air during decanting. And again, if it were shown that the experiment could succeed in any place whatsoever, we must come to the conclusion that germs of ferment exist everywhere in a state of suspension in the air. “I have taken,” writes Gay-Lussac, “a bottle of must that had been preserved for a year and was perfectly transparent, and have decanted the must into another bottle, which I then carefully corked and exposed to a temperature of 15° C. to 30° C. (59° F. to 86° F.). Eight days afterwards the must has lost its transparency; fermentation has taken place in it, and soon my must has become transformed into a vinous liquor, sparkling like the best champagne. A second bottle of must that had been preserved for a year, like the preceding one, but which had not been brought into contact with the air, has presented no signs of fermentation, although placed under conditions most favourable to its development.”