824 



ill fig. 4, they must be situated between tiie metastai)le parts of A" 

 and F' . Tiie position of />" and JJ' with respect to one another is, 

 however, not jet defined \)\ liiis; we shall refer to this later. 



Further it follows from (5) and ((i) that C' is situated at the left 

 of F' and A" ; consequently C is situated within the angle which 

 is formed bv the stable part of E' and the metastable part, of F' . 



For the position of A' it follows from (5) and (6) that A' must 

 be situated at the right of F" and E' ; consequently A' is situated 

 in fig. 4 within the angle, which is formed by the stable part of 

 F' and the metastable part of E' . As however this angle, is divided 

 into two parts by the metastable part of (" , we cannot tell yet 

 within which of those two angles we have to draw curve A'. In 

 order to examine this, we consider the hexahedron EBDCF; we 

 find from this; 



E'F'\A'\B'C'D' (7) 



Hence it is apparent that we must find at the one side of A' the 

 curves E' and F' , at the other side the curves B' , C" and D' . 

 Consequently it follows from this that A' must be situated between 

 the metastable parts of the curves C" and E' . 



We should have been able to find the same with the aid of the 

 hexahedion EABDF; hence it follows: 



E'F' C A'B'D' (8) 



Now it appears from this that we must find at the one side of 

 C' the curves E' and F' , at the other side the curves .4', i?' and D' . 



It is apparent from fig. 4 that we may expi'ess the previous results 

 in the following way : 



when the six phases form the anglepoints of an asymmetrical 

 octohedron, then the six monovariant curves form in the P, 7-diagram 

 four onecurvical and one twocurvical bundle. 



Type III. In fig. 5 the six phases form the anglepoints of the 

 hexahedron EABDC, within which the point F is situated. 

 In order to transform this hexahedron into an octohedron, we 

 unite F with the three anglepoints of a definite side-plane of 

 the hexahedron; we find this side-plane in the following way. In 

 fig. 5 S represents the point of intersection of the diagonal CE 

 with the triangle ABD. We imagine the hexahedron to be divided 

 into six tetrahedrons, which terminate in the point aS. As the point 

 >S' is situated within the tetrahedron SB DC, we take for the side 

 plane, mentioned above, the triangle BDC and we unite therefore 

 the point /'" with the points B, C and D. 



