Compression of Air and other Gases. §29 



tions, it was supposed that the dilatations which it experien- 

 ced were proportional to the quantities of air introduced into 

 it. 



If the reservoir had not been dilated, when 101 .2 grammes 

 of air were forced into it, the density of the air would have 

 been 113.5 times that of the atmosphere; but, when calculat- 

 ed, by applying the dilatation of the reservoir, it amounted 

 only to 110.5. In Fig. 17 we have shown the method in 

 which we made the experiments on the expansive force of air 

 compressed in such a reservoir. AB represents the reser- 

 voir, that is the culasse of an air-gun, CD is a plate with 

 a pillar CE, FH is a piece of iron, whose upper part receives 

 the axis round which the lever FG turns. This lever is ba- 

 lanced by the mass F. At I the lever carries a tooth or pin 

 with which it presses upon the valve !M. A slide N with 

 the scale L of a balance serves to determine the force ne- 

 cessary to open the valve. As this is kept in its place by a 

 spring, we at first examined what was the force required to 

 open it when the air inclosed had the same density as the at- 

 mospheric air. When this was done, we charged the reser- 

 voir as much as possible, and after having determined the re- 

 sistance which the introduced air gave to the valve, we dis- 

 charged the reservoir by degrees, always determining by the 

 balance the quantity of air which remained, and by the ap- 

 paratus in Fig. 17, its expansive power. 



This class of experiments, however, is not susceptible of any 

 rigour, because the valve does not always apply itself in the 

 same manner. When the valve is covered with leather, in 

 order to make it shut better, the inequality is very great, 

 and it is on that account that we have made a series of ex- 

 periments with a valve of steel, which was well ground into 

 the aperture, but we have not been able by this means to ob- 

 tain sufficiently great pressures. 



The first table shows the results obtained by a valve cover- 

 ed with leather, and the other the results obtained by a ground 

 steel valve. 



