72 
PROF. H. A. WILSON ON THE ELECTRICAL CONDUCTIVITY AND 
rather greater than 300 volts per centimetre. I think, when any great number of 
positive ions moves down to the negative electrode an arc starts because the positive 
ions form salt vapour by recombining with negative ions, and the salt diminishes the 
drop of potential at the negative electrode and so increases the gradient between the 
electrodes. The gradient required for an arc to start on putting in a bead of salt is, 
therefore, probably that required to make a large number of positive salt ions move 
down the flame. According to the earlier experiments all salts give positive ions 
having equal velocities. This conclusion is probably correct in spite of the large error 
made in the absolute value of the velocity. The salt vapour really extended down 
the flame nearly to the lower electrode and was not confined to the part of the flame 
close to the bead as was supposed. The potential gradient which made the ions move 
down was therefore the gradient just above the lower electrode and not that just 
below the bead. This explanation of the large velocities obtained in the earlier 
experiments was suggested by Andrade. The present experiments have convinced 
me that his suggestion is correct. 
Part III.—-Relation Between Luminosity and Conductivity oe Salt 
Vapours in Flames. 
Some experiments on the changes in the luminosity and conductivity of flames 
containing sodium chloride, due to the addition of chloroform, will now be 
described. 
A mixture of air, spray of a sodium chloride solution and gasoline vapour produced 
by a sprayer was passed through a T-tube into two similar burners. The tube to 
one of the burners led the mixture through a vessel in which chloroform could be 
placed. 
The luminosities of the two flames were compared by means of a Hilgers’ spectro¬ 
photometer. 
Each flame contained two platinum electrodes separated by a horizontal distance of 
about 8 cm. A current from a battery of from 10 to 300 cells could be passed 
between these electrodes. In each flame two fine platinum wires were introduced 
about 5 cm. apart, and either pair could be connected to a quadrant electrometer. 
These wires were horizontal and perpendicular to the horizontal line joining the 
centres of the electrodes between which the current was passed. 
The current between the electrodes was measured with a galvanometer, and the 
ratio of the current to the potential difference between the fine wires was taken as 
a measure of the conductivity. When no chloroform was used the two flames 
had equal conductivities and luminosities. 
The following table contains the results obtained :— 
