304 OPTICAL PROJECTION 



Glass coloured with copper oxide is pretty pure red. 

 Ammoniated copper gives a fair blue ; but a much better plan 

 is that discovered by Mr. H. G. Madan, of superposing 

 Chance's signal green, which stops all red, on a rich cobalt- 

 blue glass, which transmits little but red and blue. The two 

 only transmit rays between p and G. Green is obtained by 

 superposing blue and yellow glass, or by the copper and 

 bichromate as just described. 



In all colour experiments, it has to be remembered that 

 deeper tints are needed in proportion to the brilliancy of the 

 light used in the lantern. 



CHAPTER XIX 



THE SPECTRUM 



175. Continuous Spectra. The continuous spectrum is 

 easily projected in the manner already described in Chapter 

 XVIII. But when the object is to illustrate the principles of 

 spectrum analysis, it is advisable to pay special attention to 

 several points, which it has not been hitherto necessary to 

 enforce. Always with the arc -light, and generally with the 

 oxy-hydrogen, it is advisable to use two prism-bottles, or else 

 the cinnamic ether direct-prism of high dispersion. A 

 convenient plan is to have the two bottles set, either in a box 

 with open ends, or on a piece of black board, in shallow 

 cavities retaining both at the proper angle of medium devia- 

 tion. 



In most typical experiments, it is permissible to converge 

 more light upon the slit, by adjusting the latter at the focus 

 of either an ordinary or cylindrical lens. 



And thirdly, special care must be taken that the spectrum 



