222 Dr. L. J. Blake on the electrical Neutrality of 
surface remained very constant. The striking-distance varied 
in different experiments from 3 to 8 millim. The ball and con- 
denser-plate B in metallic connexion with it were insulated 
just previous to the electrification of the liquid ; and the 
presence of any electricity conveyed to the ball was deter- 
mined by the method described on page 218. It was necessary 
that the ball be highly polished and the liquid carefully cleaned 
in order to avoid dust-particles. Their influence is shown by 
the deflections in scale-divisions given in the following tables 
of experiments with sea-water. 
45 scale-divisions - 
= 1 Daniell 
Positive Ch 
arge. 
Experi- 
ment. 
Liquid at 
room-temp. 
Liquid at 
100° C. 
Electrified brass 
disk replacing 
liquid surface 
under the ball. 
Without special 
precaution 
against dust- 
particles 
With precaution... 
1. 
2. 
3. 
4. 
-36-25 
-37 
+2025 
+25-75 
-14 
-91-75 
-4375 
+16-25 
+30-5 
+21 
+38-5 
] 
Negative C 
large. 
Without special f 
precaution 
against dust- | 
particles |_ 
With precaution -J 
1. 
2. 
3. 
4. 
5. 
6. 
+40-5 
+25-5 
-10 
+ 8-5 
-25-25 
-19-5 
+98-75 
+26-5 
- 6 
- 4-5 
-15 
- 8-5 
+ 3-75 
+81-25 
+12 
-25-25 
-16-75 
It was found by direct experiments that the loss of elec- 
tricity upon B, and consequently upon the ball, during three 
minutes amounted to 39-44 scale-divisions. Between the 
action of cold and copiously evaporating sea-water was a dif- 
ference of only 4 to 13 scale-divisions (0*088 to 0'288 Daniell) 
in those experiments made with precaution against dust- 
particles. The kind of electricity upon the ball was the oppo- 
site to that upon the sea-water. Had convection occurred, the 
reverse must have evidently taken place; and from the high 
potential of the sea-water, a strong convection of electricity 
was to be expected. 
(e) To experiment with evaporating mercury, the following 
