36 Electrical Behaviour of Metallic Vapours in Flames, 



pressure, whereas the number of impacts of metal atoms 

 with molecules of the flame gases increases with increasing 

 pressure. Bui whereas metal atom with metal atom is the 

 chief cause of the liberation of the electron, it is probable 

 that the impact with a gas molecule is, to a relatively very 

 small extent, capable of freeing an electron undo- favourable 

 circumstances, since with decreasing rate of vaporization of 

 the metal we reach a fixed small limiting value of the 

 deviation of the luminous metallic vapour in an electric 

 field. Onlv brief reference is made to these points here, 

 since the whole question has recently been fully discussed by 

 Lenard *. 



§ 12. Conclusions. 



1. Negatively-charged luminous metallic carriers can exist 

 in a flame, in addition to the known positive carriers. 



2. If a streak of metallic vapour stand between, but not 

 touching two electrodes in the Hame, Ohm's law holds for 

 the conductivity of the vapour within the observed range 



3. If a cooled negative electrode he employed, with a 

 given voltage a definite maximum current is obtained on 

 displacing a luminous metallic vapour si real; between the 

 electrodes towards the negative electrode in the flame. An 

 estimate of the number of free electrons produced per second 

 in the streak may he obtained from this. 



4. The conductivity and energy of light-emission of a Btreak 

 of metallic vapour are proportional within the observed range. 



5. The energy of light-emission for a given rate of vapori- 

 zation of metal is roughly independent of the pressure. 



6. The velocity of migration of the positive carriers <A' 

 the luminous streak is inversely proportional to the pressure, 

 and hence the fraction of the time during which a carrier is 

 positively charged is independent of the pressure. This 

 leads to the theoretical conclusion that the impact of metal 

 atom against metal atom is the main cause of the liberation 

 of an electron from a metal atom. 



It gives me great pleasure to thank here Prof. 1*. Lenard, 

 in whose Enstitute at Eeidelberg these experiments were 

 carried out, for his constant advice and encouragement, and 

 for his great kindness throughout the work. To Dr. Ram- 

 Bauer I am also indebted Eor much assistance in the course 

 of the work. 



* P. Lenard, SitMungsberickte dor Heidelbe r g >11," Ueberdie 



Elektricitfttaleitung and Lichtemission metailhaltiger Flammen." 



