19S Bowen — Birefringence of Torbernite. 



deduced from the v same diagram by neglecting the green. 

 Thus, assuming approximate equality in the retardation 

 for red and violet by a given section, the interference 

 color corresponding to a retardation of 25(W will be one 

 containing red and violet in equal amounts and as the 

 dominant colors. The color of the section will, there- 

 fore, be practically that of the sensitive-tint plate ; thin- 

 ner sections will be bluer, and thicker sections redder. 

 At a retardation of about 40(W the color will be nearly 

 a pure red and as the section thickens this will become 

 bluer until at about 53(W, a color close to the sensitive 

 tint will be observed again. This corresponds rather 

 closely with the real sensitive tint, for this is the point 

 at which green is cut out even in a mineral with normal 

 birefringence. At this point the color becomes bluer as 

 the section thickens and redder as it thins as with the 

 normal sensitive tint. It is readily seen, then, why the 

 interference colors of the grains, seen under the micro- 

 scope, are always a red, blue or purple and are quite 

 conspicuous even in thin grains that with normal bire- 

 fringence would give only inconspicuous grays. It is a 

 mistake to state that the interference color is always the 

 simple complementary color of that for which the min- 

 eral is isotropic. 



I am indebted to Dr. H. E. Merwin of the Geophysical 

 Laboratory for the use of a monochromator and for 

 information regarding the dispersion of the liquids used. 



Mineralogy Department, Queen's University, 

 Kingston, Ontario. 



