The last-mentioned experiments prove[89] that fermentation can take place _in vacuo_, when the conditions are more favourable than those which present themselves within the almost full tins containing provisions. The change in these latter cases cannot (in the great majority of instances) proceed far[90], because there is no adequate space into which residual gases may be emitted. When this emission (which is almost always one of the accompaniments of a fermentative change) has taken place to a slight extent, the meats are in the very best condition for preservation. There is an utter absence of light, an absence of free oxygen, and also an absence of that diminished pressure which my experiments seem to show is favourable to the promotion of many kinds of fermentative change. So that if fermentation does not take place in a closed flask which is _full_ of a boiled infusion of hay,[91] it may be owing to the fact that there is no space for the residual gases, and that undue pressure retards many fermentative changes. This is also perfectly compatible with the other fact that the same kind of fluid will undergo change when a small quantity of it is contained in a comparatively large flask--owing to there being, in such a case, plenty of room for residual gases to be effused, before that undue amount of pressure is brought about, in the presence of which such a fluid will no longer ferment or putrefy. Fluids, therefore, whose putrefaction is hindered by increased pressure and favoured by diminution of pressure, may be placed under conditions which are successively more favourable than the last, by putting a gradually smaller and smaller quantity of fluid into a flask, to which calcined air is admitted, and, better still--if the stimulus of oxygen is not absolutely needed in order to incite fermentation in the fluid employed--by only half filling the flask, and procuring a more and more perfect vacuum. [89] This proof is more severe in certain other experiments (not yet published) in which I had the benefit of Professor Frankland’s assistance. The vacuum in these cases was perhaps more perfect, having been procured by means of a Sprengel’s pump and a simultaneous ebullition of the fluid, during which the flasks were hermetically sealed. The closed flasks were subsequently exposed to a temperature of 293° F. for a short time. [90] An examination of tins of “perfectly good” meats has convinced me that a very small number of _Bacteria_ and _Leptothrix_ filaments are occasionally to be met with.