(M'l CTIVITY AND LUMINOSITY OF FLAMES CONTAINING VAP<il:lsKI> SALTS. 123 



As the lithium salts have but a small conductivity in the flame, it was thought 

 possible that tin- ,i It.- ration of conductivity alx>vc tabulated might depend on the l.-n^c 

 excess of hydrochloric acid in the flame. The experiments were therefore extended 

 to the salts of caesium and potassium, which have so high a conductivity as to preclude 

 tin- jK>ssibility just suggested. 



In the following table the results are given in scale deflections : 



The above figures show that just as in the case of lithium salts, the conductivity of 

 flames containing the salts of caesium and potassium does not disappear on removal of 

 the colour by means of chloroform, and the influence of the E.M.F. on the changes 

 of conductivity that are noticed is the same in form as in the case of the lithium 

 salts. It appears that at a certain E.M.F. the addition of chloroform would produce no 

 effect in the conductivity. For the caesium chloride solution used it follows from the 

 table that this E.M.F would be about Tl volt. 



Conductivity of Salts vaporised in the Flame of Cyanogen. 



It has already been stated in the introduction that the equal conductivity of various 

 salts of the same metal was attributed by ARRHENIUS to the conversion of each salt 

 into hydrate by the large excess of water vapour present. 



It appeared of interest to investigate the behaviour of salts in a gaseous medium 

 containing only a small percentage of water vapour, and for this purpose we chose the 

 flame of cyanogen. To avoid the presence of water altogether is impracticable, if the 

 salt has to be sprayed ; but if we suppose that in a coal-gas flame any considerable 

 proportion of the hydrogen has been burned before it reached the neighbourhood 

 of the electrodes, the quantity of water vapour in such a flame will be very 

 great compared with that of a cyanogen flame, in which the only water is that 

 introduced by the sprayer (partly as drops and jmrtly as water vapour). As 

 we estimated this difference in the amount of water to be something like 10 to 1, we 

 thought it probable that, if the hypothesis of ARRHENIUS were correct, marked 



R 2 



