POLARIZED LIGHT AT THE SURFACE OF A UNIAXAL CRYSTAL. 
597 
ducted in a somewhat different manner. Unless certain rather elaborate conditions are 
fulfilled, the angle through which a Nicol’s prism is turned is not the angle through 
which the plane of polarization of the emergent light moves, and the difference 
between the two may, as I have shown (Phil. Mag., Nov., 1880, “ On Nicol’s Prism”), 
be very considerable. Besides, when I began the experiments the only circle available 
was one graduated to S' of arc, which Was not sufficiently accurate for my purpose. 
I allowed, therefore, plane polarized light to fall on my spar prism, and by turning it 
varied the angle of incidence until only one ray emerged. I observed the angle of 
incidence and the deviation, and from them calculated the azimuth of the plane df 
polarization of the incident light on the electro-magnetic thedry. I then took a small 
cell with plane parallel glass sides and filled it with a weak solution of sugar; this 
I placed iri the path of the incident light and thus produced a change in the position 
of its plane of polarization which brought the extraordinary ray into view again. 
This, by adjusting the angle of incidence On the spar prism, can again be made to 
vanish, and the angle of incidence and deviation being observed we can obtain a 
second value for the azimuth of the plane of polarization of the incident light. The 
angle between these two azimuths is the angle on the electro-magnetic theory through 
which the plane of polarization has been turned by the sugar cell. But this can be 
directly observed and the theory thus tested. Then I removed the sugar cell, altered 
somewhat the original plane of polarization, and again made the same observations, 
thus obtaining a series of values for the rotation produced by the cell corresponding 
to different angles of incidence on the spar prism. 
The same observations were made using only the extraordinary ray. Thus the 
uncertainty arising from the want of adjustment and bad graduations of the polarizer 
circle was avoided. 
The second difficulty was perhaps more serious. It was impossible to estimate with 
anything like sufficient accuracy the position of the spar prism for which the light of 
either ordinary or extraordinary ray was just quenched. To obviate this the apparatus 
was arranged as follows. The spar prism was mounted on the table of a spectrometer, 
kindly lent me by Professor Stokes, with a circle on silver and verniers reading to 1 0". 
The instrument, and the method of adjusting the prism and focussing the telescope 
and collimator, have been described at length in my paper “ Oil Plane Waves in a 
Biaxal Crystal” (Phil. Trans, 1879, p. 293). The sodium light was replaced by a 
strong source of white light, a powerful paraffin lamp, or the oxyhydrogen lime-light. 
A biquartz with the line of separation horizontal was placed between the polarizer and 
the collimator slit, and carefully adjusted by set screws, so that the light fell on it 
normally. Between the biquartz and the slit was placed a convex lens of about 
20 eentims. focal length, arranged so as to form an image of the biquartz on the slit 
of the collimator. The light from the slit fell on the spar prism, and two spectra, an 
ordinary and extraordinary, were formed and viewed by the telescope. Each of these 
spectra was divided horizontally into two parts, corresponding to the two parts of the 
