Mr B. Martin’s Experiments on Island Crystal. 251 
colours of the other three at G, H, I ; for in them (as in com- 
mon prisms), the red rays are next the perpendicular, but in 
the rays CK they are farthest from it. 
66 From the phenomena of this experiment, we are assured of 
a peculiar property and idiomatic difference in the constitution 
and nature of this species of Island crystal. For here we dis- 
cover a power residing in every part of this substance, that acts 
upon the particles of light in two different and directly contrary 
modes , a thing perhaps never observed in the operations of na- 
ture before. The same needle will be made to move in two 
different and contrary directions by magnetism, but then this 
magnetic power is found to reside in two different parts of the 
magnet. There are also experiments to shew, that light bodies 
will be moved in contrary directions by the action of electrical 
fluids ; but these fluids possess not the same body, for one is 
found in glass , the other in amber. But, in the case before us, 
if we say the refraction of the rays CG is occasioned by attrac- 
tion in the prism, we may, with equal propriety, say, the re- 
fraction of the ray CK is produced by repulsion , or that there 
is a power of a positive and negative nature residing in the 
same, and every point of the prism, which is a matter very dif- 
ficult to conceive. 
66 The position of the prism in the Figure is such as makes 
the rays CG, GH, Cl, CK, all stationary , that is, when the 
rays BC go through it parallel to the base DE ; and then all 
those refracted rays are as near to one another as they can be ; 
but upon moving the prism ever so little out of that position, 
you will observe a very quick motion in the two extreme rays 
Cl, CK, towards contrary parts to i and k, while the two mean 
rays CG, CH, move very slowly to g and h. 
“ And it is further to be observed, that the ray CK is not on- 
ly refracted in a direction (contrary to the rest) towards the per- 
pendicular CP, but even so far that the image of an object shall 
appear considerably below the direction of the incident ray. 
Thus, let Cmbe parallel to AB, then will the ray C be refract- 
ed below it to C l; and, therefore, if you view the flame of a 
candle through the prism (posited as in the Figure), you may 
make the image of the flame appear either below the candle 
seen from K, or to coincide with it as viewed in the direction 
