596 
MR. R, T, GLAZEBROOK ON THE REFRACTION OF PLANE 
position ol the tace of incidence with reference to the axes of the crystal and also the 
directions of the two refracted wave normals, the theory enables us to calculate the 
angle between these two positions of the plane of polarization of the incident light. 
But this angle is capable of direct measurement by experiment, and so the truth of 
the theoretical formulae can be tested. 
The experiment in its simplest form is as follows : a plane polarized beam of sodium 
1 ight falls on a crystal of Iceland spar cut in the form of a prism, the position of whose 
faces relatively to the optic axis can be determined. The angle of incidence is 
observed and also the deviations qf the two emergent wave normals. 
From these data we can calculate the positions of the refracted wave normals in the 
crystal, and also of course the positions of the planes of polarization of the light 
travelling in these two directions respectively. 
Let us suppose the polarizer to be a Nicol’s prism, mounted in such a manner that 
the position of the plane of polarization of light emerging from it can be determined 
by means of a graduated circle attached to it; turn the Nicol until the extraordinary 
refracted wave disappears, then observe the position of the plane of polarization. But, 
there being only one ray, the ordinary, traversing the crystal prism, we can obtain from 
theory the azimuth of the plane of polarization of the incident light. Let us call this 
measured from some fixed plane 0 O . If the position of this fixed plane with reference 
to the Niool can be found with accuracy we have here a means of comparing theory 
and experiment. We can eliminate the uncertainty in our knowledge of the relative 
position of the two planes of reference by turning the Ntcol until the ordinary ray 
disappears, and reading the circle again ; the difference between the two circle readings 
is the angle through which the plane of polarization has been turned; but the theory 
gives us again in this case, when there is only one refracted ray, the extraordinary, the 
value 0 E of the azimuth, measured from the same plane as before, of the plane of 
polarization of the incident light; the difference 0 O -—0 E should be equal to the 
difference between the circle readings. Or again, having obtained as above a value for 
0 O , alter the position of the spar prism so as to change the angle of incidence, and 
proceed as before; we can thus get a series of values of 0 Q , the position of the plane 
of polarization of the incident light, for different angles of incidence as given by theory 
when the ordinary wave only traverses the crystal. But the readings of the polarizer 
circle give an experimental series of values of this same quantity and a comparison of 
these two series affords us a test of the theory. Since, in general, these two series of 
angles are not measured from the same zero point, there will, even if theory and 
experiment agree, be a constant difference between the two series depending on the 
difference of zeros. If the difference between corresponding values in the two series 
is not constant, but varies as the angle of incidence changes, w 7 e must infer that 
theory and experiment do not agree. We can test in the same manner the formula 
for the case in which the extraordinary wave only is propagated. 
In practice, mainly in consequence of two difficulties, the experiments were con- 
