the Theory of Luminous Flames. 93 



the luminous flame slightly to one side^ the deposited soot 

 may be caused to burn in the air. 



By allowing a somewhat thick layer of soot to be formed, 

 and by then simultaneously removing both flames, it is pos- 

 sible to obtain a considerable quantity of unburned soot when 

 the basin has become cold. 



As confirmation of the statement that the cooling action 

 of a substance held in the flame is only of secondary im- 

 portance so far as regards the deposition of carbon upon 

 that substance, and that the main action consists in intercept- 

 ing the upgoing particles of carbon which have been sepa- 

 rated in the lower parts of the flame, the experiment with the 

 porcelain rod may be repeated. If the rod be carefully 

 watched, it is seen that soot is at first deposited only upon 

 the under surfaces of the rod (that is, upon those surfaces 

 which are opposed to the uprising current of gas), and that 

 the deposit spreads to the upper surfaces only after some time 

 has elapsed, and even then that this latter deposit is much 

 thinner than the former. 



These facts furnish further and most weighty arguments in 

 favour of the generally adopted view that the soot actually 

 exists in the flame in the form of a solid body, and not in the 

 gaseous state as Frankland supposes. For if the latter sup- 

 position were true, deposition of soot must be a consequence of 

 withdrawal of heat, and condensation of the thick light-giving 

 hydrocarbons would take place upon the upper surfaces of cold 

 bodies placed in the flame. 



Aided by knowledge gained from these experiments, we 

 shall be in a position the better to understand the action of 

 the burner as a means whereby heat is withdrawn and the 

 flame rendered less luminous, and also the action of the cold 

 inrushing gas. Just as a metallic rod placed near to a burner 

 whence a flame is issuing causes a somewhat extended dark 

 spot in the latter, so it must be admitted that the cold top of 

 the burner itself will exercise a similar action upon the non- 

 luminous part of the flame which always immediately sur- 

 mounts it. In this case a second factor, viz. the action of the 

 cold inrushing gas, also comes into play. 



The action of the burner tending to decrease luminosity is 

 naturally more marked in the case of small than of large 

 flames, because the whole area of the flame is less, and also 

 because the temperature throughout that area is less, in the 

 former than in the latter cases. By causing the flame of a 

 bat's-wing burner to become very small, the cooling action 

 of the burner may succeed in entirely extinguishing the flame. 

 The cooling action of the burner becomes very marked when 



