424 
Proceedings of the Royal Society 
sist in this, that the brush sweeps away abraded particles which have 
been left by rubbing with emery paper and flannel cloth, or with 
flannel cloth alone, that so long as any of these particles are in its 
course the deflection is more or less altered in the positive direction; 
and that after they have all been swept off by a few whisks we get 
the true unimpaired reading ? I believe that that is the proper ex- 
planation. 
On 12th December the surfaces were prepared by rubbing with 
dry bath brick and flannel, Zinc, iron, and lead were positive at 
first ; the zinc most, the iron next, and the lead least, and the zinc 
required the greatest amount of whisking to change it to negative. 
In case of 13th December the discs were merely brushed with a 
large soft brush before beginning. The zinc was at first slightly 
positive. The readings were not so satisfactory as usual on account 
of imperfect insulation due to very moist weather. 
On 14th December the discs were rubbed with warm flannel. 
Only the zinc and lead gave positive deflections, the former more 
persistently than the latter. The copper disc was then rubbed with 
sand paper, and the following series of observations taken, the first 
before the disc had been brushed in any way : — 
Eepeated Brushing. 
60 
no 
50 
115 
115 
120 
Mean 55 
Mean 115 
Eepeated Brushing. 
Repeated Brushing. 
105 
125 
105 
116 
no 
118 
106 
125 
Mean 106 
Mean 121 
The brushing was done by the brush used to make the readings. 
The first brushing has much more effect than any of the subsequent 
ones ; and this favours the idea that free particles of the metal are 
