158 Mr B. Martin's Experiments on Island Crystal. 
fore refract the rays of light without sensible colour. But this 
general law does not hold in Island crystal, as we shall now de- 
monstrate by experiment. 
In Fig. 3, let PO be a beam of the solar rays, incident upon 
the first surface CG of the crystal in a dark room, then will it 
be divided into two, viz. OL, OM, as has been said ; at the op- 
posite and parallel surface FH, there will be another double but 
very different sort cf r fraction of each of the said two rays at 
L and M ; for part of the ray OL will be equally refracted in- 
to LQ, parallel to the incident ray OP, after the usual manner 
in glass, &c. ; but the other part of the said ray OL will be un- 
equally refracted from L towards T, as they are in a prism, ac- 
cording^ their different degrees of refrangibility ; and therefore 
all the beam LT will appear of various coloured light in the 
darkened room. After the same manner, on the other part, the 
beam OM will be equally and unequally refracted into the two 
parts MR, parallel to PO ; and MS of coloured light diverging 
from the point M. 
u By these three beams refracted from the second surface of 
the crystal, there will be three different images of the hole at P, 
through which the first beam enters the room ; for by the rays 
refracted in the usual manner in LQ and MR, the hole will 
have a double image at Q and R on the screen, but entirely co- 
lourless, at the distance equal to LM, in the corresponding 
parts, from each other. 
“ But the other two rays LT, MS, paint each of them a co- 
loured image of the hole in the shutter, at such a distance on 
either side, as makes the angles TLQ, RMS, of about five or 
six degrees ; and this different refrangibility of rays is strictly 
agreeable to that in glass, but in a greater degree, as before ob- 
served. By variously inclining the surface to the incident ray, 
the angle of refraction may, on each side, be varied from two or 
three degrees to sixty or seventy. 
“ As these pieces of crystal make three images, in a right 
line in the plane of perpendicular refraction FCGH, and that 
in the middle is double ; so other pieces are found, by which 
this line of three images is refracted on each side into another 
line of three images, and the middle line of images is doubled ; 
so that, upon the whole, there are no less than twelve distinct 
