790 The Mobility of the Positive Ion in Flames. 



of the molecules. An alternative method of ionization that 

 may be suggested is that when the metallic salts are intro- 

 duced into the flame they expel electrons which ionize by 

 impact. If, however, these electrons are in equilibrium with 

 the molecules of the flame gases, ionization by impact does 

 not appear to be possible. The equivalent potential fall of an 

 electron in temperature equilibrium may be deduced from 

 the energy relation Ye = J mv 2 . At 2000° we find that V is 

 only 0*25 volt. At ordinary temperatures a potential drop of 

 at least 5 volts is necessary for ionization by impact. Thus it 

 appears that in a flame the electrons cannot ionize by impact 

 unless either their velocity exceeds the equilibrium value, or 

 else the energy required for ionization is much less than at 

 normal temperatures. These two points can only be settled 

 by further experiments. 



The simplest conceivable type of ionization would be 

 the separation of the metal salt into two ions, a positive 

 metal ion and a negative acid radicle. That positively 

 charged metal atoms exist in the flame seems fairly certain ; 

 for Lenard has shown that the salt vapour which exhibits the 

 characteristic coloration in the flame is positively charged. 

 It has, however, been established that the coloration of a 

 flame is quite independent of its ionization*. For instance, 

 if chloroform be sprayed into a salted flame, the coloration 

 disappears, but the conductivity is scarcely affected. These 

 two facts suggest very strongly that besides the metal ions 

 (which cause the characteristic colour), some smaller ions are 

 present which produce the conductivity of the flame: all 

 observers agree that the salts of all alkali metals produce an 

 ion of the same mobility; lithium with an atomic mass of 7, 

 for example, produces a similar ion to caesium whose atomic 

 mass is 1&5. Professor Wilson f has recently made an 

 attempt to explain this by supposing that the ions are really 

 metal atoms, but that the charged life of the ion increases 

 with the atomic mass of the metal. To get the apparent con- 

 stancy of mobility that is observed, one must assume that 

 the charged life of the atom varies as the square root of its 

 mas^. The experimental test that Professor Wilson has 

 deduced for this law leads to values of the " true " mobility 

 which are three times greater than those deduced from the 

 equation k = \e/mv ; such a discrepancy seems too large to 

 justify the law suggested. Franck and PringsheimJ have 

 recently examined a flame consisting of hydrogen and chlorine. 



* Smithells, Wilson, and Dawson, Proc. Roy. Soc. Lxiv. 1898. 



t H. A. Wilson, Phil. Mag. June 1911. 



J Franck and Piingsheim, Deutsch. Phys. Ges. xiii. p. 8 (1911). 



