302 STTiniES ON' FERMEXTATIOX. 



A, (Fig. 73) in which it had taken place had never been with- 

 drawn from under the mercury. We kept the liquid under 

 observation daily, and saw it gradually become brighter ; this 

 went on for fifteen days. We then filled a similar flask, B, 

 with the solution of lactate, which we boiled, not only to kill 

 the germs of vibrios which the liquid might contain, but also 

 to expel the air that it held in solution. When the flask, B, 

 had cooled, we connected the two flasks, avoiding the intro- 

 duction of air *, after having slightly shaken the flask, A, to 

 stir up the deposit at the bottom. There was then a pressure, 

 due to carbonic acid at the end of the delivery tube of this latter 

 flask, at the point a, so that on opening the taps r and s, the 

 deposit at the bottom of flask A was driven over into flask B, 

 which in consequence was impregnated with the deposit of a 

 fermentation that had been completed fifteen days before. Two 

 days after impregnation, the flask B began to show signs of 

 fermentation. It follows, that the deposit of vibrios of a com- 

 pleted butyric fermentation may be kept, at least for a certain 

 time, without losing the power of causing fermentation. It 

 furnishes a butyric ferment, capable of revival and action in a 

 suitable, fresh, fermentable medium. 



The reader who has attentively studied the facts which we 

 have placed before him cannot, in our opinion, entertain the 

 least doubt on the subject of the possible multiplication of the 

 vibrios of a fermentation of lactate of lime out of contact with 

 atmospheric oxygen. If fresh proofs of this important proposi- 

 tion were necessary, they might be found in t*he following 

 observations, from which it may be inferred that atmospheric 

 oxygen is capable of suddenly checking a fermentation pro- 

 duced by butyric vibrios, and rendering them absolutely 

 motionless, so that it cannot be necessary to enable them to 

 live. On May 7th, 1862, we placed in the oven a flask hold- 

 ing 2*580 litres (4| pints), and filled with the solution of 



* To do this, it is sufficient first to fill the curved euds of the stop- 

 cocked tubes of the flaslvs, as well as the india-rubber tube c c, which 

 connects them, with boiling water that contains no air. 



