514 
MR. H. A. WILSON ON THE ELECTRICAL CONDUCTIVITY 
It thus appears that nearly all the ionisation of the salt vapour, to which the 
conductivity of the flame is due, takes place at the surfaces of the glowing electrodes, 
although the electrodes are ceidainly colder than the flame gases. If we regard a 
molecule of the salt as consisting of oppositely-charged ions or electrons held together 
hy the attractions between their che.rges, then, when the molecule is very near to a 
conductor like the electrode in the flame, the induced charges on the conductor 
diminish the attraction between the ions composing the molecule which may enable 
it to be ionised, even while the molecules not near the electrodes are quite stable, 
though at a higher temperature. 
(5.) The Relative Velocities of the Ions in the Flame. 
The way in which the current through the flame depends on the temperature of 
the negative electrode indicates that the part played hy the negative ions in carrying 
the current is more important than that played by the positive ions. This fact, and 
the results obtained in investigating the fall of potential between the electrodes, 
suggested the idea that the velocity of the negative ions, due to a given potential 
gradient, is much larger than the corresponding velocity of the positive ions. 
To test this, experiments were made in which the potential difference between the 
upper and lower electrodes necessary to cause the positive or negative ions to move 
down the flame against the upward stream of gases was determined. The apparatus 
used for this purpose is shown in fig. 4. 
AA, Iiead of salt and support. P, screen above lower electrode. 
A bead of salt was put in the flame between the electrodes, and the current from 
the lower electrode measured. It was found that when the uj)per electrode was 
positive, introducing the bead caused no increase in the current unless the potential 
difference between the electrodes was greater than 100 volts, when the electrodes 
