114 METHODS OP PETROGRAPHIC-MICROSCOPIC RESEARCH. 
pensated for exactly by a wedge of another mineral;* in short, that the 
use of very thick wedges of quartz or selenite showing interference colors 
of the higher orders is of little value unless the observations be made in 
monochromatic light and the wedges calibrated for that light. 
H. Tertschf has recently described a method for measuring the birefrin- 
gence (y a) on a section nearly normal to an optic axis. From the two 
standard equations 
( 1- JL\ = (L L\ sin e sin * sin'F = (*- ) (JL L 
\a\ y\J Va 2 y*J \a 2 ?-/ \a l r 
and the substitution of /3 for ai and y\, which is permissible for a section 
nearly normal to an optic axis, he derives expressions for both a and y. 
These expressions involve not only /3 and the angles 8 and 61, which the 
normal to the section includes with the optic axes, but also sin 2 V, and they 
can not furnish results of a high order of accuracy. Under favorable 
conditions, however, the method may be of service in ascertaining the 
birefringence roughly. 
SUMMARY. 
Briefly stated, the most serious error in the determination of the bi- 
refringence of a mineral plate or grain is the measurement of its thickness. 
If the de Chaulnes method for measuring the thickness be used, another 
error is caused by multiplying the apparent thickness by the refractive 
index /3 instead of the correct index, but this is not serious. The error in 
the determination of the path-difference of the emergent light-waves is 
slight, especially if monochromatic light be used together with the Babinet 
compensator or one of the graduated wedges described above. If white 
light be used, the character of the white light, the relative dispersion of the 
mineral plate and of the inserted wedge, the selective absorption of the min- 
eral, are all factors which enter the problem and may seriously affect the 
accuracy of the measurement of the path-difference. As a general rule ex- 
perience has shown that the birefringence can be determined on favorable 
plates or grains in the thin section with a probable error of less than 10 per 
cent. 
A. Cornu. Bull. Soc. Min. Fr.. 6, 135. 1883. fT. M. P. M. 3, 310. 1911. 
