104 METHODS OP PETROGRAPHIC-MICROSCOPIC RESEARCH. 
nicol divided by 180 gives directly the phase difference of the waves trans- 
mitted by the crystal, this difference being exactly compensated by the 
mica plate and the lower nicol. The accuracy of this method is dependent 
on (i) the homogeneity of the monochromatic light; (2) the accuracy with 
which the quarter wave plate has been cut; (3) the accuracy with which its 
ellipsoidal axes are placed parallel with the principal planes of the nicols; 
(4) the accuracy of the setting of the rotating lower nicol. Friedel claims 
an accuracy of about of the wave-length used under favorable condi- 
180 
tions, or about 3 nn for sodium light. This accuracy is not great and the 
method has not been generally adopted by petrologists. 
W. Nikitin* has recently described a compensator which is intended 
especially for measuring birefringence on thin plates showing low inter- 
ference colors. The compensation is effected by a quartz plate (0.07 mm. 
thick) whose normal includes an angle of 25 with the axis. The path- 
difference is increased or decreased by rotating the quartz plate, the angle 
of rotation being determined directly on a graduated circle. Nikitin claims 
an accuracy of at least 4 nn for this plate. The birefringence of the plate for 
the different positions can be calculated by formulas developed by Nikitin. f 
The accuracy of this device does not exceed that of the Babinet compensator 
or even that of the graduated quartz wedges, and is, moreover, limited in its 
application. 
In the determination of the birefringence of a given crystal plate in the 
thin section two distinct measurements are necessary : ( i ) that of the thick- 
ness of the crystal plate, and (2) that of the path-difference between the 
emergent light- waves (in monochromatic light). The thickness of the 
plate can be measured either by direct contact by use of the micrometer- 
screw or of the spherometer, or by means of the fine adjustment screw of 
the microscope, or indirectly by means of the interference-color or path- 
difference (in nn} of an adjacent mineral properly cut and of known bire- 
fringence, t Of these different methods, the second (with fine adjustment 
screw of microscope) is most convenient, although possibly less accurate. 
The usual method consists in bringing to sharp focus || the upper surface of 
the plate and then the lower surface as seen through the plate itself, or, if 
the plate be resting free on the object glass, to focus on the upper surface 
of the glass at the side of the mineral plate. In the first case the apparent 
thickness must be multiplied by the refractive index of the mineral to 
obtain the real thickness. As the average thickness of minerals in the thin 
section is 0.03 to 0.02 mm. an error of o.ooi mm. in setting the micro- 
meter-screw will produce an error of 3 to 5 per cent in the thickness deter- 
mination. In ordinary microscopes this error may frequently amount to 
0.002 or 0.003 nun. and the resulting error in thickness to 10 per cent. In 
making this measurement well-defined particles should be selected on the 
upper and lower surfaces of the plate and the thickness determined in the 
ZeiUchr. Kryst.. 47, 378-388. 1910. 
17.<r. Kryst 3J, 1.^.1-146. IOOO. 
IB. Mallard. Bull. Soc. Min. Fr. 6, 130-133. 1883. 
For the accurate determination of the focus on a point, the method of Michel-LeVy (Bull. Soc. Min. Fr. 
*t 133-154. 1883) of illuminating the object by a narrow beam of strong monochromatic light is satisfactory. 
If the poin tjbe not in exact focus sharply defined diffraction rings appear. An accuracy of - - mm . is claimed 
for this method. 
