WHY A FLAME EMTT3 LIGHT K >>'> 



meter as well. Pieces of glass were coated with carbon, 

 and the tissue paper was not scorched except in a small hole 

 where the explosion of the fulminate had burst through. 

 This experiment showed the formation of carbon, the emis- 

 sion of a brilliant light and the localization of the heat 

 liberated. But as the decomposition in a flame can hardly 

 be as rapid as in this experiment, and as hydrogen and 

 oxygen also give a feeble light when exploded, he sought 

 to detect the rise in temperature at the moment of decom- 

 position when this is caused by heat. He arranged a 

 thermo-couple in a small tube so that only the turn of 

 wires was exposed, and after sweeping out the air passed 

 a slow current of acetylene through the tube, the arrange- 



ment being as shown in Fig. 2. The heat was raised 

 throughout the tube at a rate of about 10 C. per minute, 

 and almost as soon as the temperature of area a passed 

 800 C. it took a sudden leap to 1000 C., the gas burst into 

 a lurid flame and streams of carbon passed on through the 

 tube. Although the temperature of area & was made con- 

 siderably higher than a the carbon passing through it was 

 not luminous. This experiment would seem to leave no 

 doubt that the incandescence is caused by latent heat, yet 

 further evidence was produced. In another experiment 

 in which diluted acetylene was used it required a higher 

 heat to cause the decomposition and luminosity. This 

 latter is the condition existing in a flame, and the tempera- 

 ture there found is above that required. In other experi- 



