the Theory of Luminous Flames. 99 



tained at a red heat, a marked increase in the luminosity of 

 the flame produced by burning the gas would, I believe, be 

 noticed. 



I have already remarked that the cooling effect of the 

 burner may be neutralized by heating the burner, and that 

 the entering gas is hereby itself unavoidably raised in tempe- 

 rature. Inasmuch, then, as it is impossible to pass a heated 

 stream of gas through a cold burner without raising the tempe- 

 rature of the latter, or to pass gas through a hot burner with- 

 out increasing the temperature of the former, it is evident that 

 in an experiment in which the gas is passed through a hot 

 burner the effect obtained will be equal to the sum of these 

 two causes. If the results of experiments are to be depended 

 upon, every precaution must be taken, and especial care must 

 be given to a determination of the quantity of gas con- 

 sumed, so that, if a difference in luminosity be really ob- 

 served when the tube of the burner is heated, it may be 

 certainly known whether or not any change in the consump- 

 tion of gas has occurred. 



Blochmann has shown that the diminution noticed in the 

 consumption of gas after the issuing stream is ignited is due 

 to the action of the heated burner ; hence one would infer 

 that if the gas be passed through a strongly heated tube, a 

 considerable diminution in the quantity consumed will be 

 observable. It was therefore most important that determina- 

 tions of the quantity of gas consumed should be carried out 

 simultaneously with determinations of the intensity of the 

 light obtained by burning the gas. Such determinations have 

 been carefully carried out by me, the gas being passed from a 

 meter through a platinum tube which could be heated by 

 means of a Bunsen's lamp. 



Observations taken before heating the tube and after it had 

 become cold gave the same result. Two platinum tubes were 

 used, each 10 centims. in length, one being 4 millims., the 

 other 8 millims. in width. These were heated so that they 

 became red-hot to the very point whence the gas issued. The 

 light radiated from the hot tubes was proved to be so small in 

 amount as not to influence the results of the experiments. 

 Inasmuch as the relation existing between intensity of light 

 and consumption of gas is not as yet accurately known (the 

 results of different observers differ very much), I have as- 

 sumed that the calculated diminution in intensity of light is 

 proportional to the observed diminution in consumption of gas. 

 Thus, as 16-litre gas gave light equal to 0'4 candle, I have 

 supposed that 15'5-litre gas would give light equal to 0*38 

 candle. The numbers thus obtained are given in column 7. 



H2 



