Behaciour of Metallic Vapours in Flames. 809 



that my hypothesis is by its nature unfit to give in a future, 

 perhaps very near, a satisfactory and perfect explanation of 

 the phenomena. I may add, that the influence of the nature 

 of the gas on the magnetic rays put in evidence by them, 

 which cannot be at all understood in the light of the ordinary 

 explanation, would suffice alone to prove the insufficiency of 

 the explanation itself, and thus confirm the necessity for a 

 new theory such as that which I have proposed. 

 Bologna, September 1912. 



LXXXII. Behaviour of Metallic Vapours in Flames. 

 To the Editors of the Philosophical Magazine. 

 Gentlemen, — ■ 



IN a paper on " The Behaviour of Metallic Vapours in 

 Flames,' 1 recently published in the Philosophical Magazine 

 (July 1912, p. 15) I proposed to investigate the variation of 

 the conductivity and tl e luminosity in a salted flame de- 

 colorized by chlorine wh< n the electrodes used for measuring 

 the conductivity do not stand in the metallic vapour (p. 23), 

 anticipating that in this case the two quantities would vary 

 together, and that the result of Smithells, Dawson, and 

 Wilson on this subject * was due to the fact that their elec- 

 trodes stood in the body of the vapour. A paper by Tufts 

 in the Physikalisclie Zeitschrift has just come to my notice, 

 describing the experiment I proposed in a slightly modified 

 form, and giving as the result of it that in this case the 

 luminosity and conductivity are indeed diminished in the 

 same ratio by the introduction of chlorine. This strongly 

 confirms the view put forward by Professor Lenard f and 

 the author J, that the light-emission is essentially connected 

 with the presence of free metal atoms and electrons liberated 

 from them, it being the mutual impact of these free metallic 

 atoms which is mainly, at any rate, responsible for the pro- 

 duction of the free electrons. The conductivity in the 

 experiment just mentioned will also be due to these electrons. 

 The conductivity in the case where the electrodes stand in 

 the vapour itself must he almost entirely due to surface 

 ionization §, accounted for by the assumption that the impact 



* That chlorine can destroy the luminosity of the vapour without 

 much influencing the conductivity. 



f SitzvngslericMe cler Seiclelberger Acothmie, Dec. ]911. "Ueher die 

 Elelstricitatsleitung urd Lichtemission mttallLaltiger Flammen." 



t Loc. cit. 



§ Cf. loc. cit. p. 18, and fig. 1 . 



