20 THE CHEMISTRY OF THE SUN. [CHAP. 



to use a spirit-lamp ; but sucli a lamp is not the best instrument 

 for such inquiries, because the temperature of its flame is too 

 low to volatilise many substances which otherwise could be 

 used to colour it. In these days of gas we are able to supersede 

 the spirit-lamp by the Bunsen burner, the flame of which is 

 hotter than that of the spirit-lamp. 



Suppose now the flame to be impregnated with some sub- 

 stance which colours the flame a bright red. If such a flame is 

 examined in the manner already indicated, we See at once 

 why the flame is red, for the light is not dispersed, it behaves 

 exactly as the red light did which Newton filched from his 

 continuous spectrum of sun light and made to pass through a 

 second prism. It contains no orange, yellow, green, blue, or 

 violet rays, and therefore its spectrum instead of containing all 

 these colours will lie wholly in the red, so that we should not 

 represent the spectrum, using the graphical notation already 

 introduced (p. 9), by 



y o m & 



as we did in the case of the candle, but simply by 



If we use a substance which colours the flame yellow we shall 

 get simply 



f 



and similarly a substance with a green flame would give us 



only. 



So much for a first general statement : now let us go a little 

 more closely into detail. Place the improvised Bunsen burner 

 and the platinum wire with the salt on it in front of the straight 

 slit and look at the slit through the prism ; it will be found that 

 there is only a yellow line visible. If the things have been 



