1898-99.] Mr Davicl Eobertson on Lines of Force. 363 
the carves were made. When this had set, new curves were made 
and were fixed by heating them until the wax melted. 
Photo. No. 11 is from part of one of these first curves, but 
with this exception, all the photographs have been taken from 
curves specially made on clear unvarnished glass, and not fixed. 
Glasses from photographic plates were used, three of them being 
whole-plate size and the remainder half-plate. The first illustration 
shows the apparatus itself, as arranged for getting the later curves. 
The glass plate lies horizontally in a wooden frame held in a retort 
stand, while two wires or electrodes, standing on a glass plate, 
make contact with the tinfoil. Each of these wires is about 24 
centimetres long, and carries a little sliding piece at the top, which 
is pressed upwards by a spring so as to ensure contact. It was 
found that this method of supporting the plate loosely caused the 
least confusion from vibratidn figures. 
In order to get a perfect control of the action, the electrodes 
are connected to the poles of the machine through a pair of 
thick bare wires lying on the glass table. Another wire is laid 
across these so as to short circuit the machine until all is ready. 
The sawdust, or other material, is spread as uniformly as possible 
over the plate by means of a sieve or muslin bag, and the Wims- 
hurst is set agoing. The cross wire is lifted, and, while the foil is 
being charged, the wooden frame (or the arm holding it) should be 
briskly tapped. The short circuit wire is dropped again whenever 
the dust begins to move outwards from the tin. This motion is 
caused by the attracted charge being deposited on the glass, 
leaving only the repelled charge on the dust. It is so great 
immediately over the tinfoil that that part is very often cleared of 
dust altogether. The secret of getting good results lies in stopping 
the action at the right moment. 
When there is only one piece of tinfoil on the plate, only one 
electrode is used, and when two pieces have to receive like charges, 
the two electrodes are connected together and to one of the 
horizontal wires. In both these cases the other horizontal wire is 
still left connected to its pole of the machine, but is, of course, 
not allowed to touch the electrodes. 
The curves obtained in this manner were photographed in a 
reducing camera fixed to a wall, so that its axis was vertical and 
