513 
OP FLAMES CONTAINING SALT VAPOUES. 
The results of this experiment are therefore entirely in accord witli the view that 
the ionisation of the salt vapoTir takes place only at the surface of the glowing 
electrodes. 
Another experiment very strikingly in favour of this view was also tried. The 
two horizontal electrodes of platinum gauze were fixed in position about 4 centims. 
apart in the flame. A bead of potassium carbonate or other salt was then held in 
the flame between the two electrodes so that the salt vapour only came in contact 
with the upper electrode. It was found that the current was almost independent of 
the height of the l)ead above the lower electrode, unless it was brought so near to the 
lower electrode that the salt vapour from the bead came into contact with the 
glowing platinum, when an increase in the current between the electrodes occurred, 
which was very great if the lower electrode was negatively charged. 
The following currents were obtained with a bead of lithium carbonate and 
potential difference of 380 volts :— 
(1.) Bead very near the lower electrode— 
Current with lower electrode negative 
130’0 divisions. 
,, ,, ,, positive. 
4-3 
(2.) Bead 0’2 centim. above lower electrode— 
Current with lower electrode negative 
84'0 divisions. 
,, ,, ,, positive. 
2-8 
(3.) Bead 1 centim. above lower electrode—■ 
Current with lower electrode negative 
3‘5 divisions. 
,, ,, ,, positive. 
o ,, 
To further test the view that the salt vapour is not ionised in the flame, except at 
the surface of the glowing electrodes, the conductivity of the flame alone for very 
rapidly-alternating currents was compared with that of the flame containing salt 
vapour by the method described by Professor J. J. Thomson (‘ Cambridge Phil. Soc. 
Proc.,’ vol. 8, Part V.). 
The outer coatings of two Leyden jars were connected through two coils of wire, 
each consisting of five or six tui’iis of well-insTdated wire, aiid the inner coatings 
were charged by means of a Whimshurst machine ; so that when the charges on the 
inner coatings were allowed to dischaige to each other rapid electrical oscillations 
passed through the two coils. An electrodeless discharge bulb containing bromine 
vapour placed in one of the coils served to indicate, by the intensity of the light 
from its discharge, the absorption of energy when a conductor was placed in the 
other coil. With this arrangement the conductivity of a large Bunsen flame could 
be distinctly detected, but the'introduction of salt into the flame produced little or 
no effect, although enough salt was introduced to have increased the cuii-ent between 
two electrodes in the flame by several hundred times. 
3 u 
VOL. CXCII.—A. 
