CH. XX. TffE DISPERSION OF LIGHT. 163 



could be the reason of the rays falling into this long shape ? 

 At first he thought that it might be because some of them 

 passed through a thinner part of the prism, and so were less 

 refracted ; but when he tested this by sending one ray through 

 a thin part of the prism, and another through a thick part, 

 he found that they were both equally spread out into a 

 spectrum. Then he thought that there might be some flaw 

 in the glass, and he took another prism ; still, however, the 

 spectrum remained long, as before. Next he considered 

 whether the different angles at which the rays of the sun fell 

 upon the prism had anything to do with it, but after calcu- 

 lating this mathematically he found the difference was too 

 small to have any effect Finally, he tried whether it was 

 possible that the rays had been bent into curves in passing 

 through the prism, but he found by measurement that this 

 again was not the reason. 



At last, after carefully proving that none of these expla- 

 nations was the true one, he began to suspect that it must 

 be something peculiar in the different coloured rays them- 

 selves which caused them to divide one from the other. 

 To prove this he made the following experiment: He made 

 a hole F, in the shutter, as before, and passed the light 

 through the prism, ABC, throwing the spectrum upon a 

 screen, M N. He then pierced a tiny hole through the screen 

 at the point -, Fig. 31; the hole in this board was so small 

 that the rays of only one colour could pass through at a 

 time. Newton first let a red ray pass through, so that it was 

 bent by the prism, H I K, on the other side of the screen, 

 and made a shaded red spot on the wall, o P, at R ; here he 

 put a mark. He now moved the first prism, A B c, a little, 

 so as to let the second, or orange ray pass through the hole 

 g. This i^y fell upon exactly the same spot of the second 

 prism, H i K, as the red ray had done, but // did not go to 



