418 
DR. HAROLD A. WILSOX OX THE 
Thus even when a very large P.D. is applied the observed curves show that the 
electric intensity between the plates is very small, almost zero in some cases, so that 
there is no reason to suppose that there is no recombination occurring between the 
electrodes when this form of curve is obtained ; in fact, a saturating P.D. has not really 
been applied, so that Marx’s argument breaks down. The presence of surface ioniza¬ 
tion has much the same effect as an increase in the distance between the electrodes, 
and consequently protects the intervening region from the applied P.D., so that it 
does not get saturated. 
When the current is reversed so that the upper electrode is negative the supply of 
negative ions is greatly diminished owing to the comparatively low temperature 
of the upjDer electrode. Consequently, since the current is mainly carried bv the 
negative ions it becomes easier to saturate the intervening space and the potential 
fall curve changes accordingly. 
I think, therefore, that the observed potential fall curves are not inconsistent with 
the view that most of the ionization occurs close to the surfaces of the efowing^ 
electrodes. 
The remainder of the present paper is divided into the following sections :— 
(1) Description of the apparatus used. 
(2) Variation of the current with the E.M.F. 
(3) Variation of the current through air with the temperature. 
(4) Variation of the current through salt vapours with the temperature. 
(5) Summary of results., 
1 . Description of the Apparatus used. 
The apparatus used is shown in fig. 1. 
It consisted of a platinum tube TT, 25 centims. long and 075 centim. in diameter, 
having a narrow tube T' joined on at one end, and a flange FF, 6 centims. in 
diameter, joined on at the other. This tube was supjDorted horizontally in a 
Fletcher’s tul)e furnace, the fire-clay Iflocks of which are shown by the dotted lines. 
The flange served to keep the furnace gases from the open end of the tube. An 
electrode, FF, consisting of a platinum tube, 12 centims. long and 0'3 centim. in 
diameter, was suj^ported on an adjustable insulated stand, along the axis of the 
tube TT. The end of this electrode was closed by a conical platinum cap which was 
about 9 centims. down the tube TT. 
At T' the platinum tulje was sealed on to a glass tube through 'which the air charged 
with spray entered. The spray was produced by a Gouy sprayer, S, which projected 
the spray into a glass bulb, G, about 8 centims. in diameter, from which the air and 
spray were led through an inverted U-tube in which the coarser spray settled. 
The solution was contained in a reservoir R. The level of the surface of the solu¬ 
tion being 30 centims. above the nozzle of the sprayer. The greater part of the 
