Death in Closed Vessels 509 



bell-jar, while maintaining a current of pure air. This precaution, 

 as will be seen later, permits animals to become accustomed to a 

 certain degree to rather low pressures, which they seem unable to 

 endure at first. When I needed diminutions of pressure which 

 the pump could not give me, as soon as it stopped working, I 

 closed cock D (cock M having been closed some time before) and 

 put cock M in communication with an ordinary air pump by a 

 thick rubber tube, being able in this way to secure a vacuum 

 of about 1 cm. However, I rarely needed to resort to this procedure. 



Now I must explain how I procured for analysis at a given 

 moment, especially after the death of the animal, a certain quantity 

 of the air contained in the bell-jar. 



For this purpose I used the little model of mercury pump con- 

 structed by MM. Alvergniat. I think I should give here a descrip- 

 tion of this instrument which will be mentioned often in this work. 



The mercury pump (Fig. 16) consists of a barometric tube 

 whose chamber A forms a large bulb and has on top a cock R, 

 which I shall discuss shortly, the cock being surmounted by a 

 little mercury bowl C. The barometric tube is connected below 

 by a very thick rubber tube with a reservoir B, whose capacity 

 exceeds a little that of chamber A. This reservoir is fixed on a 

 piece of wood which can slide up or down in a double groove by 

 means of a system of gears whose arrangement the figure shows. 



The whole operation of the apparatus really depends on the 

 different positions of cock R. This is a three-way cock; the glass 

 ring in which it turns communicates by three orifices with the 

 barometric chamber, the mercury bowl, and the lateral tube lead- 

 ing to the exterior. 



The cock itself is pierced by two channels meeting at right 

 angles. It is easy to understand the significance of its different 

 positions, which are represented above at the left in Figure 16. 

 In 1, all communication is cut off, and the barometric chamber is 

 hermetically closed; in 2, communication is made between the 

 chamber and the mercury bowl; in 3, communication between the 

 chamber and the lateral tube. 



This glass cock, when properly greased, keeps the vacuum per- 

 fectly. However, for fear that air bubbles, penetrating between the 

 cock and its ring, should vitiate results, I had the whole enclosed 

 in a jacket of iron and rubber, which is kept full of water. 



The last detail of construction is the graduated rule, which 

 allows the height of the mercury column to be measured, which is 

 often useful; the whole apparatus is mounted on a wooden case 



