120 SIR DAVID BREWSTER ON THE DECOMPOSITION AND DISPERSION OF LIGHT 
the edges of the primitive or secondary forms of the mineral. In tabasheer, 
where the vacuities contain air, which we can expel and send back at pleasure, 
a fine blue light is dispersed, depending, no doubt, on the size of the vacuities. 
In a very remarkable specimen of calcareous spar, crowded with hemitrope veins, 
I have observed a copious internal dispersion produced by the reflection of light 
at the different surfaces, which, though in optical contact, have different degrees 
of extraordinary refraction. 
All these phenomena, however, are essentially different from those which 
form the subject of this paper, with the exception of the phenomena of fluor-spar, 
in so far, at least, as they are the result of imperfect crystallization. The epipo- 
lism which Sir Joun Herscuer ascribes to this mineral, or its internal dispersion, 
according to my experiments, does not belong to the species, but only to particu- 
lar varieties, and not even to the variety, but merely to particular parts of it. 
It is therefore the result of inequal or imperfect crystallization. The nucleus is 
perfect, a coating supervenes, having a different tint by transmitted light, and 
dispersing a fine blue light, and so on through a succession of strata, dispersing dif- 
ferently coloured lights, and separated by non-dispersing spaces. An extraneous 
element, therefore, depending on the state of the solution, has been successively 
introduced into the crystal, and if it had the same refractive and dispersive 
power as the fiuor-spar, it could not reflect any portion of the intromitted beam : 
But if there is any difference in the mean refraction, or in the dispersive power, 
or if the difference consists merely in the unequal length of certain portions of the 
two spectra, then, in all these cases, light will be dispersed by the extraneous 
element. If, for example, we place a film of oil of cassia between two prisms of 
flint glass, the light reflected from the film will be b/ue. The index of refraction 
for certain of the red rays is the same in the glass and in the oil, and consequently 
none of these rays enter into the reflected pencil, which must therefore be blue, 
whatever be the inclination of the incident rays. If we now suppose this film of 
oil to be solidified, and disseminated in infinitely small atoms throught flint glass, 
or a fluid that has the same action as the glass upon light, we should have the 
phenomenon of a blue dispersion.* 
A beam of blue light thus produced should be polarised at the polarising 
angle, and partially polarised at other angles; and if this is not its character, we 
must look for some cause by which it has been counteracted. We have already 
seen that, in the Bohemian yellow glass, none of the light is polarised by reflec- 
tion, and that in the quiniferous solution only a part of it is so polarised, the 
whole pencil in the one case, and the residual pencil in the other, having a qua- 
quaversus polarisation. This effect cannot be the result of an opposite polarisation 
* In the experiment with Prussian blue, which is a very splendid one, the particles are mechani- 
cally suspended in the water; so that we have here an ocular demonstration that the particles are 
the cause of the dispersion and the quaquaversus polarisation. 

