16 



PROPERTIES OF STEAM AND AMMONIA 

 LATENT HEAT AND HEAT OP LIQUID, 2ia-6ooP. 



S*> 



600 



971.74 



1162.06 



mm 



MS. 18 



an B 



am :: 



M.A * 

 863.60 



JJI .43 



na ^ 

 373.71 



416.78 



1200.56 

 739.78 



;'- H 



004, .i 

 688.02 

 506.07 



1182.68 

 629.51 



1165.94 

 562.96 

 fa** 



Referring to the first of the preceding tables, the close agreement 

 of the two sets of values of r may be noted. The greatest difference, 

 \\hirh o<( urs at 32-40 degrees, is about 6 in 10,000. This agreement is 

 a decisive test of the validity of the analysis. The two sets of numbers 

 are obtained independently, <>nr from the characteristic equation, the 

 other from the heat-content equation, and the agreement between the 

 two shows the satisfaction of the Clapeyron relation. Of the two sets 

 the one obtained from the heat-content equation should be chosen, rather 

 than the set derived by means of the Clapeyron relation. The reason 

 for this lies in the slight uncertainty in the exact value of the derivative 



* at low temperatures. It was shown in connection with Fig. i that 

 dt 



the course of the Scheel and Heuse points indicates that the true value 

 of this derivative at 32 degrees is probably slightly greater than the value 

 obtained from the formula. The slightly lower values of r calculated 

 from the Clapeyron relation in the range of 32-80 degrees may be 

 ascribed, therefore, to a small error in the derivative. 



For the range 2i2-6oo F. the important result is the set of values 

 of /", heat of the liquid. In Fig. 6 curve A represents the new set of values 

 for the range 2i2-6oo F. It lies between Dieterici's curve and Callen- 

 dar's extrapolated curve and represents very well the Regnault experiments 

 as interpreted by Callendar. Above 400 degrees the curve runs from 

 I to 3 B.t.u. lower than the Dieterici points, a deviation of 0.2 to 0.6 

 per cent. Dieterici admits a possible error of 0.3 to 0.5 per cent in the 

 experiments to determine the mean specific heat c m and a further error in 

 the reduction of c m to the actual specific heat. It is likely that a possible 

 error of at least I per cent may be attached to Dieterici's points; hence if 

 the points are too high, as is indicated by Regnault's experiments and 

 Calendar's extrapolated formula, the curve probably represents the true 

 values fairly well. 



The values of the latent heat r given in the preceding table may be 

 compared with direct experiments within the range 32-212 degrees. 

 For this purpose four sets of experiments are available, those of Dieterici,* 

 Griffiths,t Smith, t and Henning. The following table gives the results 

 of these experiments expressed in a common unit, the mean B.t.u. 



Annalen der Physik, Vol. 37, pp. 494-508 (1889). 



* Phil. Trans., Vol. I86-A, pp. 261-341 (1895). 

 t Phys. Review, Vol. 25, pp. 145-170 (1907). 



f Annalen der Physik (4). Vol. 21, pp. 849-878 (1906). 



