Knowledge of the Specific Volumes of Steam. 27 



For comparison of the experiments the following principles 



pv 

 have been used : — should never exceed its limiting value for 



indefinite superheating; for two experiments in which the 



pv 

 pressures and temperatures do not differ much, — should be 



pv 

 almost the same (it would be constant for a perfect gas) ; — 



at constant pressure should be greater at greater volumes or 



temperatures, at constant volume it should be greater at greater 



temperatures or pressures, and at constant temperature it 



should be greater at greater volumes or lower pressures. The 



pv 

 best determination of the limiting value of — as calculated 



from the chemical composition of water is that given by Yan 

 Laar,* but since he places the absolute zero at —273° -2 C. 

 instead of — 273 0, 7, the writer has modified his value to corre- 

 spond, the number obtained being 34673, under the assump- 

 tion that j? is in millimeters of mercury and v in cubic meters 

 per kilogram. 



In this manner the writer has made careful comparisons of 

 the various experiments. Those of Regnault were all at such 

 low pressures that the substance should behave as a perfect gas, 

 which probability is well corroborated by them. The single 

 determination of Meyer is evidently incorrect, as it gives a 



pv 

 value of — of 3*472 in a region where the deviation from a 



perfect gas should be fully two per cent. 



The experiments of Fairbairn and Tate have in general 



pressures far lower than those of the others, but temperatures 



pv 

 nearly as high ; accordingly — should be much greater, whereas 



it runs far less. Of their twenty-three experiments six are 



comparable by the preceding principles with those of Hirn, 



Horstmann, Herwig and Battelli, four more with those of 



Hirn, Horstmann and Herwig, four with Herwig and Battelli, 



and four either with Herwig or Battelli. In almost every case 



pv 

 the values of — run from two to five per cent less than they 



should for any agreement whatsoever, and accordingly there 

 must have been considerable condensation. 



Hirn's experiments yield great discrepancies among them- 



selves. In seven of his twenty-two experiments, in which — 



should steadily increase, the values run 



3-376 3-390 3-366 8-36.7 3337 3'345 3'398 

 *Zeitschr. fur Phys. Chem., xi, p. 433, 1893. 



