250 JY. L. Bowen—The Ternary System : 



would have to be twice the formula weight in order to make 

 the molecular volumes nearly equal and give likelihood of solid 

 solution, but it is by no means certain that this similarity of 

 molecular volumes is necessary. At any rate, the agreement 

 seems worth recording. 



Refractive Indices. — The refractive indices of the pyroxenes 

 formed by crystallizing the glass were determined by matching 

 with liquids in sodium light and then determining the refrac- 

 tive index of the liquid on the total refractometer. Only the 

 maximum and minimum indices y and a can be determined by 

 this method. The measured values of 7 and a are given in 

 Table VII and are plotted in fig. 23. The average probable 

 error of the values is d= '002, though the accuracy is usually 

 somewhat better in the compositions with low birefringence 

 and occasionally not so good in those with higher "bire- 

 fringence. 







Table VII. 







Composit 



ion 



y 





Diopsic 



e Clino-enstatite 



a 



100 





... 



1-694 



1-664 



92 





8 



T692 



1-664 



84 





16 



1-684 



1-662 



75 





25 



1-684 



1-660 



65 





35 



1-678 



1-655 



56 





44 



1-675 



1-655 



37 





63 



1-671 



1-653 



19 





81 



1-665 



1-651 



_. 





100 



1-660 



1-651 



Optic Axial Angle. — The optic axial angle of various 

 members of the series was measured on grains under the micro- 

 scope. Sections showing the emergence of one axis were used 

 for those having a large angle, but for a considerable range of 

 compositions the angle is sufficiently small that sections show- 

 ing the emergence of both axes were available. The material 

 formed by crystallizing the glass at a low temperature was not 

 entirely satisfactory for this purpose in compositions towards 

 the magnesian end of the series, on account of the increasing 

 prevalence of poly synthetic twinning towards that end. It was 

 possible to prove, however, with this material, that the optic 

 axial angle, beginning from the diopside end, grows smaller. 

 In ]\IgSi0 3 63 per cent, diopside 37 per cent the angle is very 

 small and the plane of the optic axes, as in all the more calcic 

 members, is //010, normal to the twinning plane. In MgSi0 3 

 SI per cent, diopside 19 per cent the angle is again small, but 

 the plane of the optic axes is normal to 010 and parallel to the 

 trace of the twinning plane. 



