200 
])Jv’. W. J. PtUSSELL ON THE FOPMAITON OF 
so that the hanging screen at this height exercises but little influence on the figure 
formed below it. The same hanging screen was now allowed to rest on the j^late, 
but at a distance of 15 millims. from the edge, and a pin was placed against the edge 
of tlie plate. Fig. 40 shows well the different actions Avhich came into plav, that of 
the pin, of the corners of the plate, of the screen, and of the broken corner producing 
a forked ray. The hanging screen was now raised to 3 millims. above the plate, the 
cuiTent now passed under it, and gave the curious picture (fig. 41) with di.stinct side 
wings. 
Another way of offering obstruction to these dust currents was to jjlace above the 
plate on which the deposit is to take place, other plates at different height.s, and of 
varying sizes, some larger and .some smaller than the ])late which is to receive the 
figure. First taking the case of placing a plate larger than the one on which the 
deposit is to form above it, the large one was 7-i- Inches by 4| inches and the smaller 
was 3 inches square. The upper part was supported on vulcanite pillars, so far from 
FiA ii'. Fig. 43. 
the plate as not to influence the figures foinied. It was found that when the 
distance between the.se plates was 1 millim. no deposit took place; but when this 
di.stance was 5 millims. a deposit did take place, covering mo.st of the curved outline, 
and passing into each of the corners, but le.ss deposit occurred in the middle of 
the figure (fig. 42), and was apparently an early stage of the cross. If the distance 
between the two plates be 10 millims., the amount of even deposit is less, and when 
the distance is increased to 15 millims. a considerable chano-e has occurred, and fio-. 43 
is formed, the cross still further developed, and when the distance between the plates 
is 20 millims., then a perfect cross forms. 
If the upper plate, in place of being larger, is of the same size as the lower one. 
different results take place. Plates 3^ inches scjuare were used, and the upper one 
was suspended above the lower one. When the distance between the plates was 
1 millim., again no dust entered, but when it as 2 millims. there was a small amount 
of deposit at each of the corners, and when 3 millims. a considerable increase of deposit 
