362 Proceedings of Royal Society of Edinburgh. [sess. 
An ideal material would consist of little long-shaped grains hav- 
ing a conducting interior within an insulating shell, such as metal 
filings covered with a thick coating of insulating varnish, hut the 
author has not yet been able to experiment on these lines. He 
first used fine sawdust, and better results have been obtained by its 
use than from any other materials he has since tried. Among 
these are iron filings, magnetic sand (powdered Ee 3 0 4 ), tea (as 
purchased), oatmeal, and hoar’s bristles cut into very short pieces. 
Tea seems a good material, hut the curliness of the leaves spoils the 
appearance of the curves, while oatmeal is rather coarse, although 
it does show something. The bristles gave no result at all, while 
the filings and sand were thrown off without forming curves. 
Probably the former conducts too feebly and the latter too well. 
All the photographs are from curves taken with mahogany saw- 
dust coloured with ink except three, for which tea was employed. 
As a first trial, one side of each of three glass plates was 
varnished with shellac, and pieces of tinfoil were stuck on it. One 
plate had a single circular disc of tinfoil placed at its centre, a 
second had two discs, and the third had two parallel strips. 
The first plate was then supported horizontally by two wooden 
blocks at its ends with the varnished side downwards, so that a 
spring of thin wire would make contact with the disc. This spring 
rested on a glass plate, and was connected to one pole of an ordinary 
Wimshurst machine. Sawdust was sprinkled on the plate from a 
muslin hag whilst the machine was running ; but, instead of giving 
pictures of the lines of force, it jumped off again as soon as it 
touched the glass. The other plates were tried, one piece of tinfoil 
being connected to each pole of the machine, but with like results. 
The author next tried putting the dust on before charging the 
foil, and then tapping the plate when the machine was started. 
This proved more satisfactory, for, although the powder flew off 
as formerly, just before doing so it formed itself partly into definite 
lines. By putting on larger pieces of tinfoil and stopping the 
machine the instant the powder began to get thrown outwards, 
good curves were obtained on all three plates. 
The plates were then cleaned and heated, and a thin coating of 
a mixture of paraffin wax and vaseline was applied to that side of 
the plate on which there was no tinfoil, i.e., to the side on which 
