Diopside — Forsterite — Silica. 243 



break must disappear. In MgSi0 3 mixtures, therefore, even 

 if crystallized in such a manner that they contain free silica, 

 the break should not occur, since there can be no calcic 

 pyroxene present. 



In the earlier work it was noted that when the intermediate 

 mixtures were crystallized from glass at a low temperature 

 and a heating curve run on such material the lower break was 

 not obtained. Preparations crystallized in this manner con- 

 sist of homogeneous xyyroxene. There is no silica and, there- 

 fore, the lower break is absent. It is clear that in the absence 

 of either silica or medium-calcic pyroxene this lower break 

 disappears, in fact, it must be absent in all mixtures of 

 CaMgSi 2 6 and MgSiO s when the preparation started with is 

 an equilibrium mixture, i. e. homogeneous pyroxene. 



Figure 20 shows that not only should the lower of these 

 two breaks be absent in all pyroxene mixtures but the higher 

 one, termed the eutectic in the earlier work, should also dis- 

 appear in the more magnesian mixtures, 58-100 per cent MgSi0 3 , 

 when they are crystallized in such a manner that they consist 

 of homogeneous pyroxene. The iinding of the higher break 

 in the more magnesian mixtures is the direct result of the 

 fact that the preparations were crystallized in such a manner 

 that they consisted of forsterite, pyroxene of various composi- 

 tions and free silica. The melting-together of pyroxene and 

 forsterite gave the break. 



In order to compare directly the behavior of homogeneous 

 material and of inhomogeneous material of the above kind I 

 have made quenchings of the two side by side, taking the mix- 

 ture diopside 63 per cent. The quite different behavior of 

 the two is displayed in the table below : 



Inhomogeneous 

 stringers of glass formed 

 increased amount of glass 



30-40 per cent glass 



I have also run heating-curves on the mixture En 81 Di 19 . 

 Homogeneous material, made by crystallizing the glass, gave 

 no breaks at these lower points, whereas material crystallized 

 by cooling the liquid gave a distinct break at 1372° owing to 

 the presence of free silica. 



In the earlier work a series of breaks recorded as the " upper 

 points" (Table II, p. 8) were obtained in the more magnesian 

 mixtures. These temperatures correspond very closely with 

 the temperatures, located by quenching, at which the last of 

 the pyroxene disappears (dissolves), leaving liquid and forsterite. 

 The degree of correspondence is shown in Table Y, columns 

 II and III. 



Temp. 



Homogeneous 



1382° 



no chano-e 



1390° 



no change 



1398° 



beginning of melting- 



minute amount of glass 



