380 Mr. Brougham’s Experiments and Observations 
thus the fringes might be distended ; still no uncommon ap- 
pearance took place; nor when other bodies were used with 
one edge of crystal, nor when polished pieces of different shapes 
and sizes were employed. The same things happened by candle- 
light, and also by refracted homogeneal light. In short, I re- 
peated most of my experiments on flexion with Iceland crystal, 
and found that they were not changed at all in their results. 
Observation 4. Having great reason to doubt the accuracy 
of an experiment tried by Mr. Martin, and in which, by a 
prism of Iceland crystal, he thought six spectra were produced, 
I was not much surprised to find, that a prism made by polish- 
ing the two contiguous sides of a parallelopiped of Iceland crys- 
tal produced only two equal and parallel images, in whatever 
position the prism was held. But though, from the imperfect 
account which Martin gives of this appearance, it was impos- 
sible to discover his error from his own words, yet chance led 
me to find out what most probably had misled him ; for looking 
at a candle through the opposite sides of a specimen of Iceland 
crystal, I saw four coloured images (besides two white ones) 
of the candle. These were parallel to one another, and in the 
same line, as represented in fig. 7. where E represents the two 
regular images, G and F two others coloured very irregularly, 
and changing colours as the crystal was moved horizontally, 
sometimes appearing each two-fold, and its two parts of the 
same or different colours. A and B were regularly coloured, 
and evidently formed by refraction, and reflected back from 
the sides. On turning the crystal round, so that its position 
might be at right angles to its former position, the images 
moved round, and were in a line perpendicular to AB, as CD. 
All this happened in like manner in the sun’s rays ; and on 
