378 PHILOSOPHICAL TRANSACTIONS. [aNNO 1794. 



and the other yellow, by interposing a sheet of yellow glass before one of them; 

 I now tried what the effect would be when blue glass was made use of instead of 

 yellow, and I found it to be the same; the shadows were still coloured, the one 

 blue, and the other yellow, with the difference however, that the colours of the 

 shadows were reversed, that which, with the yellow glass, was before yellow 

 being now blue, and that which was blue being yellow. I afterwards tried a 

 glass of a bright amethyst colour, and was surprized to find that the shadows still 

 continued to be coloured blue and yellow. The yellow, it is true, had a dirty 

 purple cast; but the blue, though a little inclining to green, was still a clean, 

 bright, decided colour. 



Having no other coloured glass at hand to push these particular inquiries fur- 

 ther, I now removed the candles, and opening 1 holes in the upper parts of the 

 window-shutters of 1 neighbouring windows, I let into the room from above, 1 

 beams of light from different parts of the heavens, and placing the instFument 

 in such a manner that 2 distinct shadows were projected by the cylinder on the 

 paper, I was entertained by a succession of very amusing appearances. The 

 shadows were tinged with an infinite variety of the most unexpected, and often 

 most beautiful colours, which continually varying, sometimes slowly, and some- 

 times with inconceivable rapidity, absolutely fascinated the eyes, and command- 

 ing the most eager attention, afforded an enjoyment as new as it was bewitching. 

 It was a windy day, with flying clouds, and it seemed as if every cloud that 

 passed brought with it another complete succession of varying hues, and most 

 harmonious tints. If any colours could be said to predominate, it was purples, 

 but all the varieties of browns, and almost all the other colours I ever remem- 

 bered to have seen, appeared in their turns, and there were even colours which 

 seemed to be perfectly new. 



Reflecting on the great variety of colours observed in these last experiments, 

 many of which did not appear to have the least relation to the apparent colours 

 of the light by which they were produced, I began to suspect that the colours of 

 the shadows might, in many cases, notwithstanding their apparent brilliancy, be 

 merely an optical deception, owing to contrast, or to some effect of the other 

 neighbouring colours on the eye. To determine this fact by a direct experiment, 

 I proceeded in the following manner. Having, by making use of a flat ruler in- 

 stead of the cylinder, contrived to render the shadows much broader, I shut out 

 of the room every ray of day light, and prepared to make the experiment with 2 

 Argand's lamps, well trimmed, and which were both made to burn with the 

 greatest possible brilliancy ; and having assured myself that the light they emitted 

 was precisely of the same colour, by the shadows being perfectly colourless which 

 were projected on the white paper, I directed a tube about 12 inches long, and 

 near an inch in diameter, lined with black paper, against the centre of one of the 



