Knowledge of the Specific Volumes of Steam. 33 



condensation on Battelli's part, but the remainder of course 

 cannot. 



So far we have assumed Regnault's p and — to be correct, 



at 



and we see it is impossible to get any agreement with Battelli. 



clt) 

 If we allow an error in Regnault's p and — , we must take them 



dt 



according to the formula Battelli gives to fit his own experi- 

 ments on corresponding saturation pressures and temperatures, 

 for we are going on the supposition that his work is correct. 



cl'D 10V 



The change in -j- changes the saturation volumes, so — is 



doubly changed. The deviations then become 



+ •023 +-006 +-018 + '007 4- '013 +'019 



and things are made worse rather than better. 



It is thus a question of whether to discard Regnault's data 

 entirely, or Battelli's. Regnault's results, with the exception 

 of total heats at low temperatures, have received essential 

 corroboration from almost every one who has gone over the 

 same ground, and this is especially so concerning saturation 

 pressures and temperatures. On the other hand, Battelli's 

 saturation pressures deviate considerably from those of all 

 other observers, we have seen that in some of his volumetric 

 experiments at low pressures errors exist, and that finally there 

 is practically no corroboration between his volumetric experi- 

 ments and those of other investigators. 



The writer therefore concludes that all our knowledge con- 

 cerning the density of steam is limited to the saturation line, 

 the experiments on superheated steam presenting discrepancies 

 which cannot be reconciled. 



Am. Jour. Sci.— Fourth Series, Yol. VII, No. 37.— January, 1899. 

 3 



