fermentable liquid of air; we simply have completely to fill a flask with the liquid, and place it in the oven, merely avoiding any addition of butyric vibrios, before the lapse of two or three days.  We may wait even longer; and then, if the liquid does become impregnated spontaneously with vibrio germs, the liquid, which at first was turbid from the presence of bacteria, will become bright again, since the bacteria, when deprived of life, or, at least, of the power of moving, after they have exhausted all the oxygen in solution, will fall inert to the bottom of the vessel.  On several occasions we have determined this interesting fact, which tends to prove that the butyric vibrios cannot be regarded as another form of bacteria, inasmuch as, on the hypothesis of an original relation between the two productions, butyric fermentation ought in every case to follow the growth of bacteria.

We may also call attention to another striking experiment, well suited to show the effect of differences in the composition of the medium upon the propagation of microscopic beings.  The fermentation which we last described commenced on March 27th and continued until May 10th; that to which we are now to refer, however, was completed in four days, the liquid employed being similar in composition and quantity to that employed in the former experiment.  On April 23, 1875, we filled a flask of the same shape as that represented in Fig. 11, and of similar capacity, viz., 6 litres, with a liquid composed as described at page 69.  This liquid had been previously left to itself for five days in large open flasks, in consequence of which it had developed an abundant growth of bacteria.  On the fifth day a few bubbles, rising from the bottom of the vessels, at long intervals, betokened the commencement of butyric fermentation, a fact, moreover, confirmed by the microscope, in the appearance of the vibrios of this fermentation in specimens of the liquid taken from the bottom of the vessels, the middle of its mass, and even in the layer on the surface that was swarming with bacteria.  We transferred the liquid so prepared to the 6 litre flask arranged over the mercury.  By evening a tolerably active fermentation had begun to manifest itself.  On the 24th this fermentation was proceeding with astonishing rapidity, which continued during the 25th and 26th.  During the evening of the 26th it slackened, and on the 27th all signs of fermentation had ceased.  This was not, as might be supposed, a sudden stoppage due to some unknown cause; the fermentation was actually completed, for when we examined the fermented liquid on the 28th we could not find the smallest quantity of lactate of lime.  If the needs of industry should ever require the production of large quantities of butyric acid, there would, beyond doubt, be found in the preceding fact valuable information in devising an easy method of preparing that product in abundance. [Footnote:  In what way are we to account for so great a difference