OF FLAMES COXTAINING SALT VAPOURS. 
527 
C = I + ^’lE, 
I is the maximum current carried by the negative ions coming from the negative 
electrode, and is the current due to the positive ions, which is still far from its 
saturation value. 
If E is made sufficiently great, the current due to the positive ions should also 
become saturated, and with very high E.M.F’.s it is found that the current increases 
less rapidly with the E.M.F. than the formula C = I + /jjE represents. 
The equation I — ^ = /r, — was originally obtained for the conductivity of 
Eontgenised gases in which most of the ionisation takes place throughout the volume 
ciP . . . fZN 
of the gas. In this case , where a is the constant in the equation — =aN“ 
representating the rate of recombination of the ions, I is the distance between the 
electrodes, e the charge carried by each ion, and U the sum of the velocities of the 
positive and negative ions under an electromotive intensity of one volt per centim. 
When aU the ionisation takes place at the surfaces of the electrodes, the above 
signification of 1:2 no longer holds, but the fact that the equation can still represent 
the observed relation between the current and E.M.F. shoAvs that the general nature 
of the conductivity is the same in both cases. 
In our paper referred to above, we showed that the conductivity of the halogen 
salts is approximately proportional to the square root of the concentration of the salt 
vapour in the flame. The conductivity of the oxysalts followed the same law when 
the concentration was very small, hut with greater concentrations was nearly pro¬ 
portional to the concentration of the salt vapour. Now when the concentration of the 
salt vapour in the flame is very small, the amount of salt in an extremely tliin layer 
at the surface of an electrode in the flame will be proportional to 0% where C is the 
concentration of the salt vapour, that is, supposing that the mean distances between 
the molecules of salt is larger than the thickness of the layer in which ionisation takes 
place. Consequently the conductivity should vary as C* if it is supposed that the 
amount of ionisation is proportional to the amount of salt in this layer. If the 
concentration of the salt vapour in the flame is so great that the mean distance 
between the molecules of salt is small compared with the thickness of this layer, then 
the amount of salt in the layer will be j^roportional to the concentration of the salt in 
the flame, so that at such concentrations the conductivity should be j^roportional to 
the concentration. 
In the folloAAung table the observed conductivity of KCL is compared with the 
calculated by assuming it proportional to (1) C*, (2) C':—• 
