on the Affections and Properties of Light . 379 
was observable in the image, farther than what happens in re- 
flexion from any other polished body. Some pieces, indeed, 
doubled and tripled the image, but only such as were rough 
on the surface, and consequently presented several surfaces to 
the rays. When smooth and well polished, a single image was 
all that they formed. The same happened if I viewed a 
candle, the letters of a book, &c. by reflexion from the Iceland 
crystal. 
Observation 2. I ground a small piece of Iceland crystal 
round at the edge, and gave it a tolerable polish here and there 
by rubbing it on looking-glass, and sometimes by a burnisher 
(it would have been next to impossible to polish it completely). 
I then placed the polisaed part in the rays near the hole in the 
window-shut, and saw the chart illuminated with a great va- 
riety of colours by reflexion, irregularly scattered, as described 
above; * I therefore held the edge in the smoke of a candle and 
blackened it all over, then rubbed off a very little of the soot, 
and exposed it again in the rays. I now got a pretty good 
streak of images by reflexion, in no respect differing from 
those made in the common way. Nor could I ever produce a 
double set, or a single set of double images, by any specimen 
properly prepared, either on a chart by the rays of the sun, or 
on my eye by those of a candle. 
Observation 3. I ground to an even and pretty sharp edge 
two pieces of Iceland crystal, and placed one in the sun’s rays. 
At some feet distance I viewed the fringes with which its sha- 
dow was surrounded, and saw the usual number in the usual 
order. I then applied the other edge so near that their spheres 
of flexion might interfere in the manner before described, -f and 
* Phil. Trans, for 1796, p. 270. . f Ibid. p. 256. 
MDCCXCVII, 3 D 
