480 Scientific Proceedings, Royal Dublin Society. 
negative, and for the same wave-lengths the ordinary index is 
greater than the extraordinary. It follows at once that the polarized 
rays are dispersed into spectra which are considerably different in 
deviation. Hach wave-length, according as this has been polarized 
within the crystal in the ordinary or extraordinary direction, seeks 
its own focus. 
The ordinary vibrations will, for each wave-length, show a 
greater deviation than the corresponding wave-lengths of the 
extraordinary vibrations. Thus two overlapping spectra are 
formed. The violet and part of the blue are solely formed of 
polarized ordinary vibrations, and are therefore pure; the green, 
orange, and red of the ordinary are mingled with the violet, blue, 
and green of the extraordinary, and are thereby rendered impure; 
while finally, at the end of the spectrum nearest the refracting 
angle, the red of the extraordinary spectrum exists unadulterated 
with shorter wave-lengths. 
The relative positions of the two spectra may be roughly com- 
pared by a Haidinger’s dichroscope, which simultaneously shows 
two images of the crystal, the one of ordinary rays only, the other 
of extraordinary. ‘The light refracted through the lower faces of 
these images is seen to be differently coloured: a violet ordinary 
image being accompanied by a dark extraordinary; a green ordi- 
nary by a violet extraordinary, and so on. 
These spectra may be more accurately observed and compared 
on a refractometer—or a goniometer used as such—the collimator 
being fitted with a nicol over the lens. When the nicol is placed 
with its plane of polarization in the horizontal position, the optical 
axis of the erystal being also horizontal, we observe the extraordi- 
nary spectrum only, and when the nicol is now turned through a 
right angle, the ordinary spectrum only. The phenomena observed 
are very striking and beautiful, the dispersion being considerable ; 
the refracting pyramid of the octahedron lighting up with brilliant 
flashes as the nicol is rotated. Accurate measurements are compli- 
cated by the effects of the balsam, and were only attempted so far 
as was requisite to obtain assurance as to the nature of the effects 
observed. ‘The colours appearing in anatase when rays refracted 
through the pyramids are examined by a nicol, present, in short, 
a polychroism, due neither to absorption nor to interference, but to 
refraction only. 
