( 686 ) 



different in principle, we confine ourselves to the simplest group 

 of substances. 



For these we have constructed the representation which we shall 

 further develop in the following. 



In agreement with Tammann, we also assume, although under- 

 cooling only occurs to a limited extent, that the liquid ridge 

 continues to very low temperatures (at first we may assume to the 

 absolute zero) states with increasing times of relaxation up to the 

 glass condition being encountered in passing over the ridge towards 

 decreasing temperature. Thus we do not come into collision with the 

 observations. Still we do not mean that it is quite impossible for 

 crystalline properties to be found on the glassy ridge e.g. at very low 

 temperatures. Further we do not suppose that the existence of one 

 such a ridge would exclude the possibility of other amorphous condi- 

 tions where other equilibrium relations between entropy and volume 

 could equally be found. Moreover it is highly doubtful if the term 

 amorphous does not include very various structures in the solid state, 

 so that it is certainly not necessary that an amorphous condition 

 should be present on the ridge where liquid would be found at a 

 higher temperature. As to the crystalline ridge, our whole represen- 

 tation makes it appear more probable to us, that the crystalline 

 ridge in the simplest case should run next to the liquid ridge down 

 to very low temperatures, than that it should follow Tammann's ring 

 shaped form (c. f. § 4). 



The process of transformation from the crystalline to the gaseous 

 (below the liquid-gas critical temperature, liquid) state does not at all 

 disagree with the usual assumptions concerning the molecular forces, 

 but is immediately to be deduced from them. A very satisfactory 

 agreement with the suppositions would be obtained when the charac- 

 teristic difference between two ridges appeared to result from the 

 differences of density and entropy (specific heat) of the modifications, 

 the crystalline or amorphous form taken by each of these modifications 

 being thermodynaraically of secondary importance, so that for a first 

 investigation they would not come into account in comparison with 

 the cliange of properties of the solid phase due to difïerences of 

 density and entropy. 



However it may be, we must certainly assume that the crystalline 

 form will result from the molecules being by choice oriented and 

 arranged in a giA'en manner owing to the forces from the not 

 corresponding points. The directing and arranging forces will then 

 be different for different densities and entropies, whence the most 

 probably occurring orientation and arrangement will be different 



