Diopside — Forsterite — Silica. 251 



It is evident that in some composition between these two, 

 the value of the optic axial angle passes through zero. In 

 order to study more closely this change in the plane of the 

 optic axes, it was desirable to prepare crystals of known com- 

 position, in which the width of the individual twinning 

 lamellae was greater than in material prepared by crystallizing 

 glass of the pyroxene composition. It has already been 

 shown that, when the three-phase-boundaries for the ternary 

 system have been determined by the method of finding the 

 temperature of beginning of melting of various mixtures 

 (temperature method), a mixture can be chosen which, when 

 held at a certain temperature, will give liquid, pyroxene of a 

 certain composition and another crystalline phase. Thus a 

 mixture of known composition can be chosen, held at a 

 measured temperature and the composition of the crystals of 

 pyroxene to be obtained can be predicted. It is not necessary, 

 however, to depend on the prediction, for the composition of 

 the pyroxene can be determined by measuring symmetrical 

 extinctions on 010. A check is thereby obtained on the predic- 

 tion and therefore, ultimately, on the determination of the 

 three-phase-boundaries by the temperature method. The 

 prediction is, therefore, an unnecessary part of the process, 

 though very desirable since it removes the necessity of a num- 

 ber of trials. 



When pyroxene crystals, embedded in glass, are prepared by 

 this method it is found that the width of the individual twin- 

 ning lamellae is greater than in the same pyroxene prepared by 

 the other method, It is, therefore, much less of a task to find 

 a crystal which will give a good interference figure and a 

 dependable value for the optic axial angle. In practice it is 

 necessary to start with a mixture of suitable composition, in 

 the form of glass, and to raise the glass quickly to the desired 

 temperature. Mixtures containing a. moderate excess of silica 

 above pyroxene composition are more suitable than others, for, 

 during the time necessary to raise the temperature to the 

 desired point (2 or 3 minutes), no crystallization takes place 

 in these somewhat viscous mixtures. All crystallization takes 

 place, therefore, at the desired point, during the time the 

 charge is held there, and it is thereby insured that all the 

 pyroxene has the composition appropriate to this temperature. 

 With less viscous, less siliceous mixtures, the charge cannot be 

 raised to the desired temperature quickly enough to prevent 

 some crystallization en route. Any pyroxene formed at a lower 

 temperature will not have the desired composition and, once 

 formed, it is very difficult to get rid of, for later crystallization 

 is likely to take place about these early crystals. It is, of 

 course, even more undesirable to start with material already 



