192 
DR. AV. J. RUSSELL ON THE FORMATION OF 
cross, and over tlie rest of the plate, except at the corners, there is an even deposit 
of dust (fig. 15). 
Other curious results are juoduced if the cojDper cylinder, heated to about 130°, be 
placed on the upper side of the heated plate, instead of the under one. Then a cross 
is formed, but it is very much broadened out, and a deposit of dust has formed round 
tlie base of the cylinder (fig. 16). If the plate be not heated, but the hot cylinder 
put upon it, then a modified effect, shown in fig. 17, is produced, and lastly, again 
reversing the heating, putting a cold cylinder on a heated plate, the cross is well 
formed, and a curious deposit, square in shape, is found round the base of the 
cylinder (fig. 18). All these forms are readily and constantly produced when the 
centre of the plate is heated or cooled as above desciibed. It will now be obvious 
why three wires form the liest kind of support for the plates on which a .symmetrical 
figure is to be formed. If a large solid support be required, a cork is probably better 
tlmn anything else, hut a cork heated to 100° (J. caused, wlien supporting a square 
plate, a uniform deposit to take place over very nearly the whole surface. 
Ihere is still another condition which affects the formation of these ffo-ures, and 
that to a very considerable extent : it is whether the plate on which the deposit is 
forming be horizontal or not. If not horizontal, the figure always has a tendencv, 
as it were, to slide down the plate. The smoothness of the glass is not essential to 
this effect, for if a cop])er plate be painted over with lampblack and a little shellac 
in alcohol, which gives it a rough surface, identical figures are formed. Fig. 19 sliows 
the deposit formed if the })late is placed on a slope of only 2 degrees, but if the slope 
be increased to 5 degrees, then the deposit assumes the form shown in hg. 20, and if the 
slope be 15 degrees, then the deposit has the form shown in fig. 21. These three figures 
show in an interesting Avay the great effect which the slope of the plate produces. 
Ihere is another way by winch the formation of these figures uiay be controlled and 
