162 
METHODS OF PETROGRAPHIC-MICROSCOPIC RESEARCH. 
and the narrow fringe between them is of about the same degree of darkness. 
There is, in short, no point of absolute extinction on the bar. The width of 
the bar increases as the margin of the field is approached (Plate 2, Fig. 3). 
As the positions of extinction of the two possible waves emerging in a 
given direction in air do not coincide precisely and are not exactly 90 apart, 
there is evidently a range of weak illumination between the positions of 
uniradial total extinction. For each of the two possible waves, however, 
the positions of extinction are precisely 90 apart, and if the crossed nicols 
be rotated through 90 there will be only a slight change, due to surface film 
effects, in the position of the axial bars in the interference figure from an 
unmounted plate. On mounted crystal plates the rotary effects of the sur- 
faces of the glass mount enter the problem and there a rotation of the crossed 
nicols through 90 often produces a small though perceptible shift in the 
FIG. 956. 
position of the dark brushes in the interference figure, as is evident from the 
series of measurements on the interference figures represented by Figs. 94, 
95. o b, 96, a, b. The distortion due to the analyzer may be responsible in 
part for the observed shift in the positions of the dark brushes in the inter- 
ference figure on rotation of the nicols. 
OBSERVATIONS IN CONVERGENT POLARIZED LIGHT.* 
To test the above conclusions, a series of measurements was made in strong 
sodium light on clear mounted and unmounted cleavage flakes of muscovite 
and anhydrite. The petrographic microscope (Plate i, Fig. 3) was first 
accurately adjusted, a cap nicol being used whose vernier divisions read 
directly to 3', which was also the interval of the vernier of the lower nicol ; 
with this arrangement both nicols were situated outside the optical system, 
Amer. Jour. Set. (4), 21, 202. 1911. 
