The result of this experiment, when roughly made, is correctly described by Gay-Lussac; in other words, it may be proved that if, at the time of vintage, we prepare some bottles of must, after Appert’s process, and, in the course of time, open them and decant their contents into other bottles, we shall soon see the must ferment and deposit yeast. It is, nevertheless, equally certain that the inferences which have been drawn from this celebrated experiment have been founded on error, and that the germs of yeast are very rarely derived from the particles of dust floating in the air with which the must comes in contact. The germs in question are, in our opinion, generally derived, not from the air, but from the sides of the bottles, from corks, from the string employed in corking, from corkscrews, and from a variety of other things. The reason for this is that any room, vault, cellar, or laboratory where the grapes, or must, or vintage, are handled—unless special precautions, of which Appert and Gay-Lussac certainly never thought, are taken—all the utensils, as well as all articles of clothing, and all the sides of the bottles which the hands have touched, are contaminated by cells of ferment derived from must that has fermented, or by germs from the surface of grapes and clusters. Thus, at the moment of decanting the must, a thousand accidental circumstances may lead to the introduction of those germs, the origin of which, as we have seen, may be actually traced to the very grapes which served for the manufacture of the must. In other words, we believe the inference that the germs of yeast which cause the experiment to succeed, are derived from particles of dust floating in the air of the place where we decant the contents of our bottles, to be altogether an erroneous one. Since the preceding remarks were written, we have endeavoured to repeat this experiment of Gay-Lussac’s in such a manner as we believed would confirm our views, by varying the conditions in such a way as would cause it to succeed or fail, according to the circumstances of the manipulation employed.