( 679 ) 



Fis. 1. 



Fig. 2. 



This ridge — the liquid ridge — carries the liquid branch of the 

 coiinodal of the liquid vapour plait ; the isotherms run o^•er the 

 ridge from the vapour to the liquid side and from higher to lower 

 entropy (1. c. PL I, fig. 1.). In the vicinity of the critical temperature 

 the ridge becomes broader and smoothes down into the double 

 convex surface (1. c. PI. I, fig. 3) which further forward is everywhere 

 double convex. For lower reduced temperatures the ridge passes 

 nearly into an ij e plane. The projection of the ridge on this plane 

 is a curve along which the inclination {tang-'^ =: absolute temperature 

 T) decreases towards negative values of the entropy (see fig. 2). The 

 projection of a cross section of the ridge shows the rapid change 

 of the inclination in the z; f plane {tang-'^ = pressure ^J) for a small 

 change of volume. The correspondence of the properties expansion, 

 compressibility and specific heat, for liquid and solid shows imme- 

 diately that the representation in the i], e, v coordinatesystem of the 

 experimentally determined conditions, belonging to one of the solid 

 aggregations of a material, can be supposed to belong to a ridge 

 corresponding to the liquid ridge. Also that other solid varieties 

 require further ridges. So long as we exclusively keep to the 

 experimentally determined values only narrow strips of these sup- 

 posed ridges are given for a short distance to the side of the tops 

 and thus form themselves isolated parts, not connected with the 

 vapour and liquid regions, of the whole Gibbs' surface for the 

 special substance. 



The various ridges, if we for a moment admit their existence, 

 will be more or less shifted, according to the densitj', towards zero 

 volume {v) and according to the fusion and transformation heats 

 more or less to zero entropy iji). The difference in specific heat of 

 the modifications, will be given by a variation in curvature. Looking 



