*30< ©KbMposiTioif br light. 



tuggtsfc to me the idea, that perhaps these three kinds of 

 Tayswere all that really existed ; a proposition that required 

 to be examined with care proportioned to its importance. 

 Accordingly I inquired into the probabilities that might be 

 brought to support it, and compared it with all the pheno- 

 mena of colour that occurred to mc, and lastly I verified it 

 by direct experiments. 



The details of these I shall reserve for the last place, be- 

 ginning with an account of the others. 

 This hypothe- 'I have already mentioned, that the supposition of three 



^is not incon- colours was not inconsistent with the formation of all the 



sistentvuth . __ , ' „ 



known pheno- tints of the spectrum. Neither is it in contradiction with 



7Ti '- fta the unchangeableness of each tint by a second refraction : 



for if a red ray of a certain degree, for example, be found 

 in the spectrum at the same place as a green of a certain 

 degree, their combination will give a yellow of a particular 

 tint ; and as these two rays have the same refrangibility, a 

 similar refractive power cannot again separate them. Ac- 

 cordingly, to have a spectrum in all points similar to that 

 which really occurs, nothing more is necessary than to con- 

 ceive it composed of three spectrums partly overlaying each 

 Other; one formed of red rays, differently refrangible, and 

 of different tints ; a second, trenching a little upon the first, 

 and having only green rays, but a similar gradation of tints 

 corresponding to- their refrangibility; and lastly a third, 

 exhibiting an analogous series of violet rays, and in like 

 "Explains the manne r trenching upon the green. On this hypothesis, 

 contiguity and there will be no disruption of the whole image, whatever 

 che ^okmrs extent be given to it by refraction : besides, it accounts for 

 seven colours separated by lines of demarcation, which no, 

 one yet has explained. 



To comprehend this, let us look at-Fig. 1. Pi. II. Which 

 is constructed in the following manner: A right line is di- 

 vided into seven parts, proportioned to the spaces of (he 

 seven colours in the spectrum, and marked by the initials 

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



that formerly adopted; for the red, yeliow and blue, have hitherto 

 been 30 considered ; while here they are the red, green, and violet, 

 the exclusive existence of which is proved by the analysis of white 

 light in several rays, as will be seen farther on 



erected 



