270 Transactions of the Royal Microscopical Society. 
referring to PI. CXXIII., Fig. 2. I have constantly had this question 
before me, but for years felt that it was a thing more to be desired 
than hoped for, and I had almost come to the conclusion that it 
could not be accomplished, when all at once I hit upon a plan which 
is even better than anything I could have wished for. 
In studying the spectra due to different kinds of interference in 
connection with the colour of feathers, insects, and minerals, I was 
led to examine the spectrum of crystals of quartz cut so that the 
light passes along the line of the principal axis. As is well known, 
the light then experiences no double refraction, but is circularly 
polarized. When viewed in a polariscope, a slice of about one- 
fourth of an inch in thickness gives fine colours, and when rotated 
no change whatever results ; but if either the polarizer or analyzer 
be rotated, the colours vary much and return to the same tint at 
each half revolution. On studying the spectrum of this light, one 
or more well-marked black bands will be seen, and on rotating the 
polarizer or analyzer these move over the spectrum, and again 
occupy the original positions on the completion of each half revolu- 
tion of the polarizer or analyzer. The number of these bands in- 
creases with the thickness of the plate of quartz, and their width 
diminishes, so that by using a very thick piece they may be made 
as numerous and narrow as may be desired. The question then is 
to choose such a thickness as may not be practically inconvenient, 
and yet give sufficiently numerous and well-defined black bands. 
The thickness which I have adopted is IJ inch. With this the 
whole visible spectrum is divided into eight spaces by seven well- 
defined bands, which in a prism spectrum are apparently at very 
uniform intervals, as shown by I of PL CXXIII., Fig. 2, but at a 
much less wave-length interval at the blue end than at the red end, 
as will be seen from the table given below. 
In order to make use of these properties of quartz it is necessary 
to have it mounted between two Nicol's prisms, in the manner shown 
by PI. CXXIII., Fig. 1. The upper can be rotated for accurate 
adjustment, and the lower is permanently fixed in a mounting with 
an ivory circle 2^ inches in diameter, each half of which is divided 
into ten parts, and these again into five smaller divisions, so that 
there is no difficulty in reading off to yi^th of a half revolution. 
After placing this circle at the zero point, the upper Nicol is rotated 
so that the centre of the second dark band from the red end of the 
spectrum exactly coincides with the sodium line, or with the solar 
line D, as shown by I of PL CXXIII., Fig. 2. All the other dark 
bands are then of course in perfectly constant and definite positions, 
depending on the action of quartz on light of various wave-length. 
On rotating the ivory circle each band gradually passes from the 
red end towards the blue, until when the circle comes to the next 
zero point, commencing the next semicircle, the series of bands is 
