on the Affections and Properties of Light. 383 
the effect taking place when the rays are perpendicular. Sus- 
pecting, therefore, that it might be owing to flexion, I made 
the following experiment, which undeceived me. 
Observation 3. I covered one side of a specimen of Iceland 
crystal, three inches deep, with black paper, all but a small 
space ~ of an inch in diameter, and placed a screen with a hole 
of the same size, six feet from the hole in the window-shut of 
my darkened chamber, so that the rays might pass through the 
screen, and fall on a prism placed behind, to refract them into 
a small and well defined spectrum, which was received on a 
chart two feet from the prism. This spectrum I viewed through 
the crystal, and of course saw it doubled ; but the two images 
were by no means parallel ; the unusual one inclined to the red, 
and its violet was considerably farther removed from the violet 
of the other, than the two reds were from one another; which 
shews, that the most refrangible or least flexible rays were 
farthest moved from their course by the unusual action, and 
proves this to be very different from flexion.* 
From all these observations this conclusion follows; that 
the remarkable phenomenon in question arises from an action 
very different from either refraction or flexion ; and whose na- 
ture well deserves to be farther considered. It may possibly 
belong to the particles of Iceland crystal, and in a degree to 
those of rock crystal, from the form and angles of the rhom- 
boidal masses, whereof these bodies are composed. Nor is this 
conjecture at all disproved by the fact that glass shaped like 
these bodies wants the property ; for we cannot mould theparticles 
of glass, we can only shape large masses of these; whereas we 
* When a candle or line is viewed through a deep specimen, the unusual image is 
tinged with colours. 
