246 N. L. Bowen — The Ternary System: 



complete solid solution is proved by the fact that any mixture 

 between the two can be crystallized to a homogeneous crystal- 

 line mass which shows only the one kind of crystals. Com- 

 plete solid solution is again proved, though further proof is 

 unnecessary, by the character of the fusion surfaces found for 

 the ternary system, i. e. the absence of any quintuple point 

 (point at which liquid and three solid phases are in equili- 

 brium). Moreover, any member of the series can be obtained 

 in contact with liquid in the ternary system. 



In order to prepare a pyroxene of a given composition, a 

 mixture of the desired composition is converted to a glass, by 

 quenching it from the liquid state, and the glass is then crys- 

 tallized by heating at a temperature of from 1300° to 1350°. 

 The heating is best prolonged for several days in order to 

 increase the grain of the crystalline aggregate. The reasons 

 for this procedure have already been stated in detail ; suffice it 

 to repeat here that, if the mixture is crystallized simply by 

 allowing it to cool from the liquid state, the result is, for the 

 more magnesian mixtures, pyroxene of variable composition 

 (zoned pyroxene), forsterite and silica. The precaution is 

 unnecessary in the more calcic mixtures, those containing up 

 to about 35 per cent MgSi0 3 , 65 per cent CaMgSi 2 6 , for these 

 do crystallize to homogeneous pyroxene if cooled quickly from 

 the liquid state. 



In the earlier work on these mixtures the material submitted 

 to microscopic examination was crystallized simply by cooling 

 the liquid and the result stated above was found — all mixtures 

 as far as about 35 per cent MgSi0 3 , 65 per cent CaMgSi 2 O fi 

 were found to consist of homogeneous pyroxene. The properties 

 of these pyroxenes were carefully determined by Wright and 

 Larsen and found to vary continuously with composition. In 

 the present work it has been possible to show, by crystallizing 

 from glass, that this continuous variation of properties does 

 not cease with the 35 per cent MgSi0 3 mixture, but extends as 

 far as the pure MgSi0 3 itself. The properties found for the 

 pyroxenes as far as the 35 per cent MgSi0 3 composition agree 

 precisely with those found by "Wright and Larsen.* 



Tivinning. — The variation in the tendency to form twinned 

 crystals appears to be continuous, though it is, of course, a 

 property which does not admit of quantitative measurement. 

 The twinning is always after 100. In pure clino-enstatite 

 every grain when turned into the proper position shows nar- 

 row twinning lamellae. In 80 per cent MgSiO.,, 20 per cent 

 CaMgSi 2 O fi the same is true, but the twinning lamellae have a dis- 

 tinct tendency towards a greater average width. In 60 per cent 

 MgSi0 3 , 40 per cent CaMgSi 2 O c the breadth of the lamellae is 

 still greater ; many of the grains show only two twinning 

 *This Journal (4), xxvii, p. 38, 1909. 



