Thermodynamic Properties of Air. 5 



the more desirable, as the thermometrie ?cale is now better 

 defined, thanks to the investigations of the International 

 Bureau of Weights and Measures — and some light has also 

 been thrown on the variations of the specific heat of water. 

 I may mention now that my result, as regards the specific 

 heat of air, is nearly identical with that of Regnault. 



To detect any variations in the specific heat of air at atmo- 

 spheric pressure it was desirable to enlarge the limits of the 

 experiments on the side of low temperatures, because on the 

 ground of RegnauhVs work we may be sure that at high 

 temperatures such variations do not manifest themselves; in 

 this respect I have been able to reach the temperature of 

 -170°. 



§ 5. In order to adapt the well-known apparatus employed 

 usually to determine the specific heat of gases to the particular 

 problem of low-temperature calorimetry it was necessary to 

 to modify it in several respects. I wished also to make it 

 more reliable, so as to lessen as far as possible the limits of 

 experimental errors, because of the smallness of the expected 

 variations in the specific heat. 



In studying Regnault's memoir on the specific heat of 

 gases, one comes to the conclusion that the chief cause of 

 error in determinations of this kind is to be sought in the 

 uncertainty as regards the true temperature of the gas at the 

 moment when it enters the calorimeter, i. e. the cooler im- 

 mersed in it. The influence of this cause of error manifests 

 itself clearly by the fact that the calculated value of the 

 specific heat depends in a marked degree on the velocity with 

 which the gas-stream is made to pass through the apparatus : 

 on increasing this velocity the result tends towards a maximum 

 value. It is known that Regnault considered only those de- 

 terminations as valid which corresponded to this maximum. 



In my own experiments I used at first an apparatus similar 

 to that of E. Wiedemann (/. c.) . I soon became convinced 

 that in order to do away with the uncertainty as to the initial 

 temperature of the gas, and at the same time to diminish as 

 far as possible the direct influx of heat (or cold) from the 

 heating apparatus into the calorimeter, it was indispensable 

 to improve the mode of connexion of these two parts. The 

 following conditions were to be kept in view in devising this 

 most important part of the apparatus : — (1) The connecting 

 piece ought to be perfectly air-tight, and at the same time a 

 bad conductor of heat ; (2) it ought to transfer the gas from 

 the heater into the calorimeter with the least possible loss of 

 heat; (3) some thermometrie arrangement ought to be intro- 

 duced in it in order to measure directly the temperature of 

 the gas-stream where it enters the calorimeter. 



