340 FKAGMENTS OF SCIENCE. 



and the metal dish that holds them is heated by a 

 spirit-lamp. The vapour from the mixture ignites and 

 we have a monochromatic flame. Through this flame 

 the beam from the lamp is now passing; and observe 

 the result upon the spectrum. You see a shady band 

 cut out of the yellow, not very dark, but sufficiently 

 so to be seen by everybody present. 



But let me exalt this effect. Placing in front of 

 the electric lamp the intense flame of a large Bunsen's 

 burner, a platinum capsule containing a bit of sodium 

 less than a pea in magnitude is plunged into the flame. 

 The sodium soon volatilises and burns with brilliant 

 incandescence. The beam crosses the flame, and at the 

 same time the yellow band of the spectrum is clearly 

 and sharply cut out, a band of intense darkness occupy- 

 ing its place. On withdrawing the sodium, the bril- 

 liant yellow of the spectrum takes its proper place, 

 while the reintroduction of the flame causes the band 

 to reappear. 



Let me be more precise: The yellow colour of the 

 spectrum extends over a sensible space, blending on 

 one side with the orange and on the other with the 

 green. The term ' yellow band ' is therefore somewhat 

 indefinite. This vagueness may be entirely removed. 

 By dipping the carbon-point used for the positive elec- 

 trode into a solution of common salt, and replacing it 

 in the lamp, the bright yellow band produced by the 

 sodium vapour stands out from the spectrum. When 

 the sodium flame is caused to act upon the beam it is 

 that particular yellow band that is obliterated, an in- 

 tensely black streak occupying its place. 



An additional step of reasoning leads to the con- 

 clusion that if, instead of the flame of sodium alone, we 

 were to introduce jnto the path of the beam a flame in 

 which lithium, strontium, magnesium, calcium, &c., 



