RADIANT HEAT AND ITS RELATIONS. 219 



visible if we could burn the gas in a clean atmosphere. 

 But the atmosphere, even at the summit of Mont Blanc, 

 is dirty ; in London it is more than dirty ; and the burning 

 dirt gives to this flame the greater portion of its present 

 light. But the heat of the flame is enormous. Cast-iron 

 fuses at a temperature of 2,000 Fahr. A piece of platinum 

 is heated to vivid redness at a distance of two inches be- 

 yond the visible termination of the flame. The vapor 

 which produces incandescence is here absolutely dark. In 

 the flame itself the platinum is raised to dazzling white- 

 ness, and is finally pierced by the flame. "When this flame 

 impinges on a piece of lime, we have the dazzling Drum- 

 mond light. But the light is here due to the fact that 

 when it impinges upon the solid body, the vibrations ex- 

 cited in that body by the flame are of periods different from 

 its own. 



Thus far we have fixed our attention on atoms and 

 molecules in a state of vibration, and surrounded by a 

 medium which accepts their vibrations, and transmits them 

 through space. But suppose the waves generated by one 

 system of molecules to impinge upon another system, how 

 will the waves be affected ? Will they be stopped, or will 

 they be permitted to pass ? Will they transfer their mo- 

 tion to the molecules on which they impinge, or will they 

 glide round the molecules, through the intermolecular 

 spaces, and thus escape ? 



The answer to this question depends upon a condition 

 which may be beautifully exemplified by an experiment on 

 sound. These two tuning-forks are tuned absolutely alike. 

 They vibrate with the same rapidity, and mounted thus 

 upon their resonant stands, you hear them loudly sounding 

 the same musical note. I stop one of the forks, and throw 

 the other into strong vibration. I now bring that other 

 near the silent fork, but not into contact with it. Allow- 

 ing them to continue in this position for four or five seconds, 



