THE RATIO OF THE SPECIFIC HEATS OF GASES. 277 



tions can be made, which is a great disadvantage, for the 

 shape of the isothermals is known for only a few gases, and 

 the investigator is generally obliged to undertake a series 

 of troublesome subsidiary determinations, if he wishes to 

 secure accuracy. Jamin and Richard (C. R., 71, 336) have 

 devised a form of experiment that might be made inde- 

 pendent of the density of the gas, or of its characteristic 

 equation. A constant amount of gas is enclosed in a 

 vessel, and has a definite quantity of heat given to it by 

 means of a wire heated by an electric current, first at 

 constant volume, and then at constant pressure. If, then, 

 the rise of temperature is read in each case, the ratio of the 

 two rises gives at once the ratio of the specific heats. 

 Jamin and Richard in their experiments did not read the 

 temperature directly, but when the volume was kept con- 

 stant they read the increase of pressure, and when the 

 pressure was constant they read the increase of volume, 

 thus bringing in the old difficulty of knowing exactly how 

 the pressure and volume depend on the temperature. 



In practice the method is not likely to be accurate, for 

 time must be given to allow the gas to come to a uniform 

 temperature, and then heat will have been lost by com- 

 munication to the walls of the vessel, and uncertain cor- 

 rections will have to be introduced. 



Considering the theoretical value of the constant y, it is 

 surprising that so few determinations have been made, and 

 still more surprising is the extraordinary want of agreement 

 among the results obtained by different observers for the 

 same gas. The values found by any one observer are 

 generally fairly consistent with each other, so that the 

 experimental error appears to be small, but each form of 

 apparatus seems to have a constant source of error of its 

 own. This was especially marked in the case of experi- 

 ments made with Clement and Desormes' apparatus before 

 Rontgen showed that the size of the gas holder has a 

 serious effect, if it is below a certain minimum. Kundt's 

 apparatus has been largely used, and seems to be less 

 subject to such one-sided errors, but when the subsidiary 

 determinations of vapour density and isothermal curves are 



