On the Decomposition of Light. 1 69 



CCvcn colours separated by lines of demarcation, wliich no 

 one has vet explained. 



To cf)raprehend this, let us look at fig. 1. PI. IV. which 

 js constructed in the follow ing manner: A right line is di- 

 vided into seven parts, prop -rtioned to the spaces of the 

 seven colours in the spectrum, and marked by tlie initials 

 of those colours. On each of the points of division 1 have 

 erected an ordinatCj and afierwas-d drawn the arbitrary in- 

 clined line ad, then I g cutting the former in c, and lastly 

 € h cuttmg the preceding inf. I suppose that the modi- 

 fications of the red rays, on which their different refrangi- 

 bility depends, are represented by the ordinates correspond- 

 ing to the line ad: these quantities express nothing relative 

 either to the velocitv of the rays, or the magnitude of their 

 particles ; perhaps thev may have a relation to their density, 

 or to any other quality whatever that constitutes their dif- 

 ference. In Newton's system of seven classes of prin)itive 

 colours, there are likewise red ravs differently refrangible; 

 this therefore is not a difficulty peculiar to the state of things 

 I am examining. In like manner the ordinates of the line 

 g h will be the modifications of the green ; and those of the 

 line li e the modifications of the violet. This being under- 

 stood, it becomes evident that the first division of colours 

 from a to i will be red alone ; that it will be followed by a 

 mixture or combination of green and red from h to c, n\ 

 which the quantity of the latter will predominate more and 

 jDore, and consequently give orange ; after which another 

 mixture of red and green will proceed from c to d, in which 

 the green will predominate more and more, forming yel- 

 low J then from (i to e will be green alone ; from e toy" the 

 mixture of green and violet that produces blue; i'xQmfiag 

 the mixture producing indigo; and lastly from ^ to It pure 

 violet. 



But another very striking property of the spectrtui), 

 which has not hitherto been explained, is th.e greater bright- 

 ness of the yellow compared with all the others. This pro* 

 ceeds evidently in my figure from being the sum of the red 

 iind of the green in the same space. In the blue loo there 



is 



