138 PROCEEDINGS OP THE AMERICAN ACADEMY. 



constant pressure, or the latter would be 0.00366256 — 0.00000134 = 

 0.0036612, a value but little above the value 0.00366092 found in the 

 present series of experiments. An objection to this comparison lies in 

 the fact that Regnault made his measurements of the change of pv at 

 4°, and not at 32°; but probably this difference would not affect very 

 seriously the results. 



Lord Rayleigh, in his recent careful determination of the compressi- 

 bility of hydrogen, has found that if the value ot'pv at one atmosphere's 

 pressure is taken as 1.00000, at half an atmosphere's pressure it will be 

 0.99974.* The results of Regnault given above show that this change is 

 nearly linear with the pressure, hence the slight extrapolation giving the 



value 1.000062 for 1.12 atmospheres is permissible. Now 



is 0.052 per cent, or the percentage difference between the coefficients of 

 expansion in constant volume and that under constant pressure; therefore 

 on the basis of the results of T ravers and others just recorded, the latter 

 value a p should equal 0.00366065, while we find 0.0036609. The 

 agreement is as good as could reasonably be expected, especially consid- 

 ering the fact that Lord Rayleigh's experiments were made at 11°, while 

 ours demand results for 32°. 



Thus if the usual value for the tension increase in constant volume is 

 assumed as correct, the application of Regnault's value for the relation 

 of pressure and volume would make our results appear slightly too low, 

 while the application of Lord Rayleigh's value would make it appear 

 slightly too high. While neither application is strictly precise, in any 

 case it appears that the new value for the coefficient of expansion is not 

 without support in other work, and may be not far from the truth. 



Further work upon hydrogen seemed to be aimless unless much time 

 was spent in the redetermination of the essential dimensions involved, — a 

 colossal task, involving several separate investigations. Therefore another 

 gas was studied. 



The Coefficient of Expansion of Carbon Dioxide. 



Carbon dioxide, made by the action of hydrochloric acid on marble 

 and purified by passing through two wash bottles containing acid sodic 

 carbonate solution, an Emmerling tower containing the same solution, 

 two towers with sulphuric acid and over phosphoric oxide, gave the fol- 

 lowing results : 



* Phil. Trans., A 204, 351 (1905). 



