30 DECOMPOSITION OF LIGHT. 



action on the two simple kinds do not give a marked pre* 



ponderance to one of them, it will be a mixed combination, 



Absorb others that will first disappear. The mass of the body being 



till bvincreasing gradually increased, the destruction of the rays will go 



rheir thickness by nevv mixtures, still progressively ; the kind least acted 



they transmit ... , ./ ... ., 



none, upon will remain the last, and it will necessarily be one 



of these three, red, green, or violet; after which no more 

 light will be transmitted. Such are the phenomena of 

 absorption, and its different gradations. 



I shall give here one of the last results of my experi- 

 ments, which might have created some confusion, had it 

 been mixed with the preceding considerations, yet tends to 

 confirm their principles. 

 White pro- I had an inclination to try, whether the light from a 



binTn' 1 o^ Se given pa,rt of the sin S le B l )cctru w> combined with that of 

 points of the another part chosen for the purpose, would produce 

 spectrum. ^; te> 



To carry this into escculion, I placed before the image 

 of the spectrum, received on the cloth a screen, by which 

 one portion was concealed, and another left open. This 

 screen, however, was perforated by a small hole, through 

 which passed a ray of coloured light belonging to that part 

 of the spectrum which was concealed. Lastly, this little 

 coloured ray was received on a metallic speculum, placed 

 between the screen and the cloth, and inclined so as to 

 throw it on a given point of that part of the spectrum, 

 which arrived freely on the cloth. Thus the colour arising 

 from the mixture of two rays of light was observed. 



I varied the trial of this apparatus on different points, 



the corresponding tints of which were diametrically op- 



Cicare^t white posite on the dial of colours. In several instances I did 



from the ex- not Q^^ain a white free from all tint of colour, because, a 



tremity of the ' ' 



red, and boun- certain proportion in the quantity as well as quality of the 



ad bl grCen e ' eraents * s necessary : but having carried the little image 



almost to the limit of the green and blue, it gave a decided 



and bright white, when thrown on the extremity of the 



red. 



This very remarkable fact adds fresh support to my 



proposition respecting the compound state of the colour in 



certain parts of the spectrum, simplified to the utmost 



For, 



