On {lie DecGmposition of Light. 2I7 



Another day these txperiments were repeated with some 

 lillle alterations. When the small round image fell upon 

 the cloth, we vv&nt behind to look at it; and found that it 

 passt-d through, appearing on the back of the cloth, which 

 was muslin well stiffened with starch. In this way we could 

 make our experin)ents more conveniently} as we had only to 

 cover o«r eye with a coloured substance, and look at the 

 little spot through it. When the spot was formed by blue 

 light, it appeared green, or violet, according to the sub- 

 •Slance interposed. Through a red substance no light was 

 seen : a proof, that the preceding ciTcct did not arise from 

 white light mixed with blue. If it were viewed through an 

 orange glass, the property of which is to absorb only the 

 blue and the violet rays, the spot appeared green ; a proof 

 that it was formed in reality by green and violet ravs. 



Finally, the yellow spot exhibited similar appearances ; 

 altogether nivisible through a violet substance, it showed 

 itself 2,reen or red through substances of these colours. 



Such are the results, that confirm my opinion of the 

 elementary parts of light. Now let these be combined with 

 the effects of absorption, which ultimately leaves onlv red, 

 green, or violet rays; with the simple and natural explana- 

 tion of the principal appearances of the spectrum, by 

 means of three kinds of rays; with the happy manner in 

 which these three kinds are applicable to the properties of 

 the dial of colours, and remove its complicalion ; and I 

 think the whole will support my proposition. U it do not 

 hence appear to the natural philosopher as a fact established 

 beyond all question, at least he cannot refuse to consider it 

 as already grounded on strong probabilities, and sufficiently 

 interesting to merit a thorough investigation, whicli my 

 occupations have prevented me from pursuing any lurther. 



Recapitulatmi. . 



Thus our system of colours appears to me reduced to 

 these few data: three sorts of luminou.- rays, of a piirtir 

 cular and unknown nature; red, greirn, aiid violet. Com- 

 bined by twos, the red and green produce yellow; the 



crcca 



