542 NEWTON'S DISC EXAMINED BY A PRISM. 



seven strips together must be given out by the red 

 alone if it is to appear white, and as the spectrum in 

 our disc is given twice, it follows that the quantity of 

 light emitted by the whole fourteen strips would have 

 to be ejnitted by two, in order to make the disc appear 

 white. Thus, although the colours are actually by the 

 duration of the impression recomposed to white, there 

 remains still a . deficiency of light, which renders the 

 disc by far less bright than a really white disc upon 

 which the same amount of light falls. If more light is 

 thrown upon the rotating disc than upon the surround- 

 ing objects, the disc becomes really nearly white. In the 

 second place we must consider that it is impossible 

 to obtain colours that is, pigments which are as pure 

 as the colours of the spectrum, each pigment being, in 

 fact, a combination of various colours ; nor can the 

 coloured sectors be arranged in the exact proportions 

 side by side in which they appear in the spectrum. 



A separate impression of fig. III. is added to the frontispiece ; it 

 will enable the student to experiment with it, after cutting it away 

 and making a hole in the middle. The black edge should not be 

 removed, as the grey, by contrast with the black, appears more 

 nearly white than without the black edge. Additional light may be 

 thrown upon the coloured disc in the following manner : Place the 

 whirling-table vertically upon a table near a window, and a retort- 

 stand upon the window-sill. Clamp a small mirror in the fork of 

 the retort-stand, and move it until it reflects direct sunlight upon 

 the coloured disc. Only a portion of the disc, in the form of a 

 parallelogram, will be thus more bright than the vicinity, but the 

 effect will be obvious at once. 



If a prism be applied to the examination of each of 

 the seven colours in the disc, each may be proved to be 

 not a simple colour, but a mixture of various colours. 

 In order to avoid in this experiment that any one of 



