428 
Froceedings of the Royal Society 
times with the brush before any readings were noted, as the deflec- 
tions appeared too small. A new brush was used. Two series of 
observations were taken; in the former case the copper was observed 
before the electro-plates, in the latter after the electro-plates. 
On 21st each disc was rubbed with chamois leather beforehand. 
The bismuth was null at the first whisk, the antimony was negative. 
On 27 th December an elaborate system of readings was taken. 
The entries in Table I. are deduced from the following four series 
of means : — 
1st Series, 
after being 
rubbed with 
chamois 
leather. 
2nd Series, 
after brush 
was heated 
slightly 
before the 
fire. 
3rd Series, 
after being 
rubbed with 
warm 
flannel. 
4 th Series, 
after being 
heated before 
the fire. 
Copper (disc), 
101 
309 
319 
Gold, . 
79 
228 
192 
212 
Silver, 
36-6 
83 
129 
184 
Nickel, 
3 
39 
50 
56 
Antimony, . 
+ 38 
70 
Bismuth, 
+ 20 
56-5 
+ 2 
48 
Aluminium, 
35-3 
119 
Magnesium, 
German Silver, 
49-6 
+ 77 
28 
109 
The electro-plates had been lying in a cold room since the 21st, and 
they were very moist when first taken up. This moisture on the 
surface probably explains why the ratio of silver to copper and of 
nickel to copper, changes so considerably. The other discs were 
brought from a warm room. I consider that the positive values 
entered are due to the existence of small particles of some kind on 
the metal. 
The observations taken on the 28th are given in full in the 
accompanying table (Table III.). After the first five readings the 
disc was commonly whisked half a dozen times or so with the brush 
used. I have marked when this wms done. It had the efifect of 
increasing the reading. All the discs were brought from home, and 
a new brush was used. It will be observed that the mean for the 
copper electro-plate is less than for the copper disc ; the result 
deduced from the different ratios is 2 to 3. In the second series 
only bismuth gave positive deflections, and these were soon changed 
to negative. Zinc was negative from tlm first, and repeated whisk- 
