Dissociation and Electrical Conductivity. 441 
From these results it is seen that the metal is a necessary 
part in the changes causing the conduction; the changes 
might therefore be looked for at the metallic surface. 
From a zine plate that had been used the adherent mixture 
was lightly brushed off; this caused no change in the elec- 
trical conductivity. If the surface was then cleaned, either 
with glass-paper or by washing well in distilled water, it was 
found that its activity had been removed, but it could be 
restored by fresh addition of salt and iodine. 
The washing water was tested, after filtering, for free 
iodine and for iodine in combination. No free iodine was 
found, but a large amount was present in combination. 
If a zine plate that had been used was washed in dilute 
HCl instead of in water, a very large leak was obtained when 
it was replaced in the apparatus without drying and reheated. 
It was found then that a clean piece of zinc if dipped in HCl 
gave the same result, and this although it was previously 
heated to nearly 200° to drive off any adherent acid and 
moisture. The zinc plate was now replaced by one of 
platinum, on which was put some zine chloride. When 
heated, this gave a very great leak with a negative charge on 
C,asmaller one with a positive charge. For example, at 
ordinary temperatures the time taken to leak through 5 divs. 
was 5 minutes, when heated to 300° the time required was 
12 sec. only, the charge being negative. 
The outer cylinder A had been made of brass to this point, 
but after heating in it zine chloride it was found to be im- 
possible to make it insulate properly at the higher tempera- 
tures. The electroscope also was not suitable for exact 
measurements. At this point, therefore, an iron cylinder 
was obtained, the insulated electrode C was formed from a 
platinum wire, and the dish B was also platinum. No in- 
creased leak was then noticed on reheating, no matter what 
substance had been used. The electroscope was replaced by 
an electrometer, and cells were used for charging the 
apparatus. 
Ezperiments with Zine Chloride. 
A small piece of solid ZnCl, was placed on B, and the 
current between C and A measured at different temperatures. 
While an observation was being made the temperature was 
allowed to rise very slowly, the mean of the values at the 
beginning and the end of the reading being taken as the 
temperature of the experiment. 
Greater regularity could be attained in this way than by 
trying to keep the temperature quite constant. The numbers 
