OF FLAMES CONTAIXING SALT VAPOUES. 
511 
This conductivity is, however, only a minute fraction of that which is observed 
when the electrodes are both hot enough to glow. If one electrode only is hot, then 
the current is much greater than Avhen both are cold, but is still small compared 
with that obtained when both are hot. The explanation of this, is considered in 
Section 7. 
If a piece of platinum foil is jDut in the flame midway between the two electrodes 
when both are just outside the flame and not visibly hot, then the current is greatly 
increased, showing that the presence of the glowing platinum enables ionisation to 
take place. The following currents were observed with an E.M.F. of 45 volts and 
3^ Rb, CO3 
( 1 .) Both electrodes not visibly hot. Current 12 divisions. 
(2.) With a piece of platinum foil in the flame between the electrodes and 
insulated. Current 40 divisions. 
(3.) Foil connected to earth. Current 400 divisions. 
The great increase in the current on connecting the foil to earth appears to be due 
to the much greater ease with which an electrode in the flame loses negative electrifi¬ 
cation than positive. This causes the foil when insulated to l^e positively charged to 
nearly the same potential as the positive electrode, which diminishes the current. 
The explanation of this will be considered in Section 7. 
The effect of putting a small bead of salt near the electrodes was also tried, 
the flame being otherwise free from salt. It was found that if the salt vapour only 
came in contact with the positive electrode, then tlie increase in the current due to 
the salt was very small, whereas if the salt vapour came in contact with the negative 
electrode the current was greatly increased. If the salt vapour passed between the 
electrodes without coming in contact with either, then the current was not increased 
at all. The following numbers were obtained with a head of potassium carbonate :— 
( 1 .) Flame without salt. Current 20 divisions. 
(2.) Salt vapour on positive electrode. Current 60 divisions. 
(3.) Salt vapour on negative electrode. Current 720 divisions. 
It is clear from this that unless the salt vapour actually comes into contact with 
the glowing electrodes, the conductivity of the flame is not affected by its presence. 
Giese (‘Wied. Ann.,’ vol. 17, p. 517, 1882) showed that when two pairs of 
electrodes were placed one above the other in a flame free from salt, or rather in the 
gases immediately above the flame, then applying an E.M.F. to the lower pair 
diminishes the conductivity between the upper pair. This effect was evidently due 
to the removal of the ions from the stream of gas by the lower pair of electrodes, 
and should, therefore, not happen in the case of the conductivity of the salt vapour 
if there is no ionisation of the vapour throughout the volume of the flame. Two 
pairs of electrodes were arranged as shown in fig. 3. The upper pair was supported 
by a glass tube passing through the tin plate above the flame, and could be fixed at 
