( 693 ) 

 model with the position of the ridges chosen, wliile it is at the same 



dp 



time easy to see that by another arrangement the — for the three 



dt 



lines will not agree so well either with Tammann's values or witli 



our desire to make the course of the temperature line the simplest 



possible. We hence consider that the arrangement of the ridges which 



we have chosen agrees with the experiments and tliat thus the specific 



volume of the modification A is larger than that of B. 



On this model the binodal for the liquid and gaseous condition 

 (line GL) is shown and also the gas branch of the binodal for the 

 gaseous state and the modification B (line GLb). The three points 

 belonging to the triple-point gas-liquid-solid B are joined two and 

 two by steel wires. The dashed line which passes through these 

 points is the isotherm of the triple-point, and close to this runs the 

 critical isotherm. The dotted line is the pressure line of the triple- 

 point. According to Tammann's values the solid phase of this triple- 

 point belongs to the modification B, while the ridge of the JL modi- 

 fication (see PI. Ill fig. 5) lies below the fundamental plane of the 

 triple-point determined by Kuenen and Robson. A tangent plane that 

 touches at more than three points, cannot be placed on our model 

 which is in agreement with the phase rule. In addition to the just 

 mentioned triple-point the existence of two others on our model, the 

 triple-point gas-liquid-solid A and the triple-point liquid with the two 

 solid modifications A and B is rendered probable by Tammann's 

 experiments. 



On comparing our model witli one constructed after the equation 

 of VAN DER Waals, one sees that on ours the liquid ridge rises 

 more steeply from the critical temperature to the lower temperatures. 

 It hence follows that the specific heat of the liquid is too small on 

 the model after van der Waals. The slow rise of the liquid ridge 

 in the latter has also the result that the heat of vajiorisation is 

 too small. 



III. The Gibbs' surface for CO^ at greater densities. 

 {Detail model of the liquid and solid states). 



For the general model the specific heat at constant volume was 

 assumed to be constant in the ideal gas state. But for the construction 

 of the detail model, to be now described, we have been obliged to 

 consider the variability of Cc. Regnault and E. Wiedemann have 

 measured Cp for CO^ at one atmosphere pressure and various tempe- 

 ratures, and have expressed the change with 7' by empirical formulae. 



46* 



