488 EXPERIMENTS ON LIGHT-SPECTRA. 



50 square. The sheet is fixed with a few tacks to the 

 upper part of a window, the slit being horizontal. A 

 retort-stand is placed on the table so that an eye situated 

 near and above the arm of the stand may see the light 

 of the sky (and the sun) through the slit ; the prism is 

 then supported by the arm of the stand, its refracting 

 edge downwards, and the eye applied to it. The 

 window and the sheet of pasteboard will appear to be 

 much lower than they really are, and in the place of the 

 slit a beautiful spectrum will appear, red at the upper- 

 most end, violet at the lowest. The deviation caused 

 by the carbon disulphide prism is so great that the eye 

 must be directed downwards in looking through the 

 prism, in order to see the slit and the sheet of paste- 

 board. In the spectrum on the wall the violet was 

 uppermost, but a little consideration will show why in 

 the present case the spectrum presents the colours in 

 an inverted order. We have already seen in the last 

 article that objects appear displaced upwards if the 

 prism deflects the rays of light in a downward direction ; 

 it follows that in the present position of the prism, in 

 which the light is refracted upwards, the slit must ap- 

 pear displaced downwards, and the more so the greater 

 the amount of refraction. Now, if red light alone were 

 to pass through the slit, we should see a red image of 

 the slit instead of a spectrum ; if only blue light passed 

 through the slit, we should see a blue image, displaced 

 more downwards than the red. This difference in the 

 relative positions of the various colours can be studied 

 by covering the slit with red or blue glass, though 

 this method is only imperfect, for coloured glass does 



