26 Mr. Tomlinson on [July 



library, except my own, which is necessarily very limited, 

 I could not examine one or two eminent and recent treatises 

 on optics ; I have, however, carefully examined seven or 

 eight works on chromatics, and have found nothing at all 

 bearing on the nature of these experiments, except in Sir 

 David Brewster's Treatise on Optics, to whom I feel it due to 

 state that the adoption of stained glass by me was suggested 

 by the following passage in his work, p. 309. ** Accidental 

 colours may also be seen by looking at the image of a candle, 

 or any white object seen by reflection from a plate, or sur- 

 face of coloured glass, sufficiently thin to throw back its 

 colour from the second surface. In this case the reflected 

 image will always have the complementary colour of the 

 glass. The same effect may be seen in looking at the edge 

 of a candle reflected from the water in a blue finger-glass ; 

 the image of the candle is yellowish, but the effect is not so 

 decided in this case, as the retina is not sufficiently impressed 

 with the blue light of the glass." 



22. I should be sorry to be supposed to question the 

 statement of so distinguished a philosopher and optician as 

 Sir David Brewster, but I may be allowed to state that with 

 coloured glass alone I am seldom successful in obtaining a 

 sight of the complementary colours of the glasses, possibly 

 because my disks are not sufficiently thin : Whereas, with 

 the plane mirror reflector behind the disks, the complemen- 

 tary colours are immediately observed in all their well 

 defined beauty. 



23. With regard to the reflected image of a candle in 

 water contained in a blue glass, the experiment is, for the 

 reason stated in the quotation, imperfect. I have found it 

 to succeed tolerably well with a strong homogeneous white 

 light. It succeeds also with the light obtained from the 

 mixed gases on lime, but this light is not homogeneous. I 

 have decomposed it by a prism, and have obtained the 

 prismatic spectrum from its rays. 



Note,— In the foregoing paper I have omitted to state, 

 that the sulphate of chrome affords a remarkable and 

 beautiful illustration of complementary colour, its solution 

 being intensely green by reflected, and of a brilliant ruby 

 by transmitted light, and yet this solution yields purple 

 crystals ! Still, I find that there are circumstances under 



