THE RA TIO OF THE SPECIFIC HE A TS OF GASES. 275 



There are two principal methods used in the experi- 

 mental determination of y, Clement and Desormes' method, 

 and Kundt's method, but as they are fully described in all 

 the text-books it is unnecessary to go into much detail con- 

 cerning them here. 



In the former method the gas is slightly compressed in 

 a large glass globe, and the pressure read. A stopcock is 

 then opened for a moment to allow some of the gas to 

 escape. The pressure falls rapidly, partly in consequence 

 of the expansion alone, and partly because of the cooling 

 caused by this expansion, and the lowest pressure reached 

 is read quickly before the gas has had time to be warmed 

 again by contact with the sides of the glass globe. The 

 temperature then slowly rises, till it is the same as it was 

 at the beginning of the experiment, when the pressure 

 is read a third time. If it is assumed that the gas is 

 perfect the ratio of the specific heats is given by the 

 equation — 



= log A - log A 

 log/, - log A' 



All who have used the method hitherto have made this 

 assumption, and for some of the more permanent gases it 

 leads to no appreciable error, but in most cases the equation 

 is not accurate enough. Rontgen's value for the y of 

 carbon dioxide, for instance, would be raised about \ per 

 cent, if a more accurate formula were used, and in the case 

 of the more easily condensed gases and vapours the correc- 

 tion would be much greater. 



It will be seen from the formula that the density of the 

 gas need not be known. This is a great advantage of the 

 method, for a small quantity of impurity only affects the 

 result in so far as the impurity has not the same y as the 

 gas under investigation, and such an effect is very slight. 

 Unfortunately the experimental difficulties are great. The 

 vessel that holds the gas must have a capacity of 50 to 100 

 litres to avoid errors from conduction of heat through the 

 walls, and gases are not easy to prepare pure in such large 

 quantities. Further, no one has hitherto succeeded in 

 devising a pressure gauge that is sensitive, dead-beat, and 



