VOL. LXXXVII.] PHILOSOPHICAL TRANSACTIONS. ] Q7 



rays was broadest, and that the other parts diminished in breadth regularly towards 

 the violet. I now delineated 1 or 2, at about 3 inches from the shadow; and 

 though, from the pin's irregularities, the sides were by no means smooth, yet the 

 general shape was in every pin, and with every prism used, nearly as represented in 

 fig. 5, pi. 4, divided in the direction ra, according to the colours of the spectrum 

 in which they were formed; rob a was red, and the broadest; that is, ra was 

 broader than ob, the confines of the red and orange; and gdev was the violet, 

 narrowest of all. 



Observ. 2. Between the pin and the prism, -^ of an inch from the pin, was 

 placed a screen, through a small hole in which, of twice the pin's diameter, the 

 rays of the spectrum passed, and were reflected into images by the pin; these were 

 pretty distinct and well defined, when received on a chart 4- a foot from the pin. 

 They were oblong, having parallel sides and confused ends; they were wholly of 

 the colour whose rays fell on the pin, unless when the white, mixed with those at 

 the confines of the yellow and green, caused the images to be of all the colours. 

 When the prism was turned round on its axis, so that different rays fell on the pin, 

 the images changed their sizes as well as their positions; they were largest when 

 red, and least when violet. 



Observ. 3. In case it may be thought that the sides of the hole, through which 

 the rays passed in observation 2, by inflecting, might dispose them, before inci- 

 dence, into beams of different sizes, I removed the screen, and placed the pin 

 horizontally, the axis of the shadow being now at right angles to that of the pris- 

 matic spectrum ; and moving the prism on its axis, again I observed the contrac- 

 tion and dilatation of the images by reflexion, though now they were rather less 

 distinct, from the greater size of the incident beam; and to show that there was 

 both a change of size and of place, without any manner of deception, I placed 

 one leg of a pair of compasses in a fixed point of the spectrum, and the other in 

 the middle point of an image formed by the violet -making rays. The prism being 

 then moved till the image became red, I again bisected it, and found its centre 

 considerably beyond the point of the compasses, which was indeed evidently much 

 nearer one end of the image than the other; besides, that the red image, when 

 measured, was longer than the rest ; and this satisfied me that there were 2 changes, 

 one of place with respect to the fixed point, the other of size, with respect to the 

 centre of the image. Lastly, as far as I could judge, the dilatation and con- 

 traction appeared even and uniform. 



Observ. 4. I remarked that the fringes or images, by flexion, were always in- 

 creased in size when formed out of red-making rays, and were less in every other 

 colour, and least in violet, besides being moved farther from the edge of the 

 shadow in the former rays than in the latter; and this agrees with an observation 

 of Sir Isaac Newton, as far as he tried it, which was with respect to deflexion. In 

 making several experiments with prisms, I hit on a very remarkable confirmation of 

 this. I observed on each side of the spectrum 4 or 5 distinct fringes, like the 

 images by reflexion, coloured in the order of the spectrum, but quite well defined 



