234 PHILOSOPHICAL TRANSACTIONS. [aNNO 171 6. 



N. B. When the candle enlightens the painted paper, set an opaque body as 

 B, between the candle and lens ; lest the image of the candle, being also pro- 

 jected, should disturb the experiment. 



Exper. IV, — Having made a hole of -^ inch diameter in the window-shutter 

 of the darkened room, I suffered a sun-beam to come into the room, which I 

 intercepted with a prism at the distance of 5 inches from the hole; and after its 

 refraction in passing through the prism, I received it on a sheet of white paper, 

 where it was coloured, making an oblong image of the sun, or spectrum, of 

 about 9 inches in length, and 2 in breadth, which breadth was nearly equal to 

 the diameter of the round image of the sun received on a paper at the same 

 distance from the hole, which here was 18 feet. Or if the sun be too high, a 

 looking-glass being put instead of the prism, will throw a white round spectrum 

 on the paper, which held at the said distance of 18 feet, will have its diameter 

 equal to the breadth of the coloured spectrum. 



The colours of the spectrum, as in fig. Q, were these ; red, orange, yellow, 

 green, blue, purple and violet, though the violet was so faint in this as to be 

 scarcely perceivable. 



N. B. The axis of the prism in this, and all the other experiments hereafter 

 mentioned, must be perpendicular to the ray that falls on it ; and the plane into 

 which the ray enters, must be held in such a position, that the angle which such 

 a ray makes with that plane when it enters, may be equal to the angle made by 

 the middle line of those rays which emerge after refraction, on the other side of 

 the refracting angle of the prism, with the plane out of which they emerge. 

 That is, Zbdg = zaeh. 



If the plane ac, on which the sun-beam falls, be turned nearer to a perpen- 

 dicular to the sun-beam than before, the spectrum will be much longer : if it 

 be more inclined to the said beam, the spectrum will be shorter ; and in both 

 cases less distinct. See the spectrum de and the spectrum de, in fig. 10 and 11, 

 where nh represents the hole in the window-shutter in each case; ac, ac the 

 plane of the prism on which the rays enter; bc, be that out of which they 

 emerge ; p, p the perpendicular, and c, c the refracting angle. 



If the plane ac be still more oblique to hp, all the light will be reflected, and 

 there will be no coloured image or spectrum made by refraction at all ; as in 

 fig. 12. 



But if it be held so as to be more nearly perpendicular to the sun-beam than 

 in fig. 10, the whole beam will indeed enter the prism; but meeting with bc 

 the lower surface of the prism, or rather the surface of the air contiguous to it, 

 some of the light will by the plane bc be reflected to de, passing almost per- 



