418 
Proceedings of the Royal Society 
far as possible the same pressure each time, and moving with the 
same velocity. Variation in velocity produces very considerable 
variation in the amount of electricity produced, or at least left after 
the operation. It is also necessary that the brush be spread out 
to an equal extent in the successive whisks. I found that by first 
placing the brush in position at the side at a constant angle to 
the disc, and then moving it across with due regard to pressure 
and velocity, that a set of successive readings has very considerable 
constancy. Tor example, the following for copper, taken at random 
from my note-book 
AVhisk.. 
heading. 
Difference from 
Ho, 1. 
no 
+ 2-5 
2, 
100 
- 7‘5 
3. 
110 
-p 2*5 
L 
90 
--17-5 
5, 
125 
+ 17-5 
6 , 
110 
-f 2-5 
7, 
90 
- 7-5 
8. 
no 
-f- 2-5 
9. 
115 
+ 7-5 
10. 
115 
4- 7-5 
Mean, 107 '5 
The differences are fairly alternate in sign, and the mean of the 
readings may with reason be assumed as giving the effect of the rub 
aimed at in the ten trials. However, I found later on that the 
readings tend to increase as the succession of whisks goes om 
It may be asked, Does the brush change in power after making 
a considerable number of whisksT De la Eive in his memoir 
cautions us when rubbing with a stick of wood to use a new stick 
each time, or else to scrape the surface used with a bit of glass. 
Were it necessary to use a new brush each time, then to make the 
thousand and half observations which I have made would be a 
matter of expense. When the brush is heated slightly before the fire, 
it produces a larger deflection than when it is not heated. For example, 
after I had finished the series of readings for copper, given above, 
and similar series for zinc, tiu, iron, and lead, I held the brush 
