250 Mr B. Martin’s Experiments on Island Crystal. 
“ For though I had there shewn that if a beam of light 
AB (Fig. 4.) were let into a dark room through a hole in the 
window-shut at A, and were to fall perpendicular upon the side 
ED of a parallelepiped DF of Island crystal, it would pass di- 
rectly through it to the opposite side at C, and there, in many 
pieces of crystal, it will be divided into three parts, viz. CK, 
CH, Cl ; of which the first being unrefracted will shew an image 
of the hole A without colours ; but the other two parts CH, Cl, 
are refracted on each side from the perpendicular CK, and there- 
fore make two coloured images of the hole at H and I. 
(( Even in those rare pieces (of the same form) that separate 
the beam AB into twelve parts, each part is uniformly refracted 
from the perpendicular ray in the middle on every side, as I 
there observed. And the same thing was thought to be the 
case in all refractions through prisms , let the number of parts 
into which the beam was divided, be less or more. 
“ But a lump of this spar or crystal (from the Peak in Der- 
byshire) ^ has lately convinced me that we are not yet capable of 
discerning the limits or laws of the refracting powers of bodies ; 
because in many prisms made from this large piece of Island 
crystal, it is proved to be fact, that the same beam of light pas- 
sing through the prism will in part be refracted from the per- 
pendicular , and in part refracted towards it. 
“ As this is so novel and singular a case, I shall illustrate it 
by Fig. 5., where AB is a beam of light from the hole A in the 
shutter, and incident upon the prism DEF in the point B, from 
whence it is refracted to C in the other side ; and there it will 
suffer a strange kind of refraction into different sorts of rays as 
follows, viz. (1.) There will be two parcels of rays CG and CH, 
refracted from the perpendicular CP, and each of them gives a 
coloured image of the hole at G and H. (2.) Another part of 
Che beam is much more refracted from C to I, and there ex- 
hibits an image of the hole A, in much stronger and more vivid 
colours. (3.) But there is yet a small part of the original beam 
AB remaining, and which is refracted (contrary to the others) 
into the rays CK towards the perpendicular CP, and makes an 
image at K much less luminous than any of the other, yet has 
all the different sorts of colour as they have ; but (4.) The co- 
lours of this last image at K are quite in an inverse order to the 
