Vapour from electrified Still Surfaces of Liquids. 219 
By a convection of electricity, the kind of electricity upon 
A, as mentioned above, must have been the opposite to that 
upon the surface of the electrified liquid. Now the tables not 
only do not show this, but in every case exactly the opposite. 
Upon A was always the same kind of electricity, and conse- 
quently upon B and C the opposite kind to that with which 
the liquid was charged. The following facts explain this result. 
During the experiment B and C were exposed to a strong induc- 
tion. Of the two electricities induced upon B and 0, one would 
be fixed by the surface of the liquid ; the other would be 
free and would diffuse itself upon B. If the insulation of B 
and C was not perfect, then a loss of free induced electricity 
would occur. After the liquid was swung away from C, 
there would remain upon B and C electricity equal in quantity, 
but opposite in kind, to that dissipated from defective insu- 
lation during the experiment. This electricity would induce 
upon A the same kind of electricity as that of the electrified 
liquid. The correctness of this explanation of the charge upon 
A is confirmed by the following. Direct measurements of 
the insulating power of B and C were made previous to the 
experiments. B and C were charged either directly or indi- 
rectly by means of the 480 zinc, water, copper elements; and 
the loss during three minutes, which was the duration of 
the experiments mentioned above, was measured. This loss 
amounted to, 
With positive charge, 
7 scale-div. out of 380 scale-div., or 1*8 per cent. 
4*0 „ „ 259 „ „ 1'7 „ 
2-5 „ „ 258 „ ;„ 0-9 „ 
With negative charge, 
8*5 scale-div. out of 408 scale-div., or 2*0 per cent. 
3*5 „ „ 252 „ „ 1*3 „ 
The induction upon B and C in these last experiments was 
maintained as nearly equal as possible to that in the first ex- 
periments. The close agreement of these figures with those 
of the deflections in the first tables, justify the conclusion that 
the electricity upon A in the first experiments was owing, not 
to a convection of the electricity present upon the surface of 
the liquid, but to a loss of electricity upon B and 0, occurring 
in consequence of defective insulation during the experiment. 
(c) The imperceptible variation in potential-difference be- 
tween the poles of the zinc-water-copper battery rendered 
possible a method of experiment still more sensitive, in which 
insulation did not come at all into consideration. Without 
