252 Mr B. Martin's Experiments on Island Crystal . 
C m, or to rise above it if viewed in the line C Z, a phenomenon 
never seen or heard of in a prism before. 
46 It matters not in what part of the prism the beam AB is 
incident ; for the effect will always be invariably the same. Nor 
does it signify whether the prism be large or small, foul or 
clear, if you can but see the light through it, the refraction 
thereof will ever be the same, as I have found by experiments 
with many of them made from the same piece of spar. 
44 There is yet another property in this same piece of crystal 
equally strange and astonishing, viz. That if it be formed into a 
prism another way, it will discover a twenty-fold power of re- 
fraction, or it will refract a beam of the solar light into twenty 
different rays ; and of course it will exhibit twenty different 
images of one object only, as the sun, a candle, &c. 
44 As these are such singular phenomena, I shall give the me- 
thod of forming the prisms to exhibit them, as represented in 
Fig. 6., where AE is a piece of crystal of the usual rhomboid 
form. It is easy to see, that every such rhombus consists of, 
and may be resolved into, two equal prisms, viz. ABDCEH 
and EFGHAC ; or, upon the other side of the obtuse angle C, 
you have the two prisms CDEFHB and ABHGFC. 
44 In the first case, if the second prism be ground away, it 
will leave the first prism ABCDEH, which, when polished, will 
exhibit the last-mentioned phenomena of the prism DEF in 
Fig. 5. of four images. 
44 But, in the second case, if the prism AGHBCF be ground 
off, there will remain the prism CDEFHB, that, when polished, 
will produce the twenty rays or images above mentioned. Such 
wonderful and different effects are occasioned by the same 
forms given to the same piece of Island crystal on its different 
parts. 
44 This system of images appears in the form of a rhomboides 
ABCD as represented in Fig. 7. with five rows of images, of 
four in a row, and all parallel to one another. It is with some 
difficulty that the two images of a candle in the acute angles B 
and D can be seen ; but by use, and a dextrous management 
of the prisms, they may all be seen. 
44 If in so short a time a multiple refraction from two tq twen- 
ty times has been discovered, who can pretend to guess at the 
