406 O. N. Rood—Force involved in Crookes’ Radiometer. 
blackened, each on one side, with lampblack to which a minute 
portion of spirit varnish had been added. Lach dise was folded 
so as to be double, the two leaves not 
1. being in contact. Di ied i 
front of it a plate of mica equal in size 
hoa, with itself, and distant from it about’ 
5mm. The system was arranged so 
2. single fiber of silk, and was provided 
A B with a small directing magnet. Fig. 
wana oes ae 2 gives a view of the arrangement 
from above. 
The dises thus arranged were enclosed in a clear glass flask, 
which was exhausted to a pressure of ‘25 of a millimeter and 
sealed up. 
readily be estimated, the circle itself being divided into half 
degrees. The small magnet connected with the discs was ret: 
dered nearly astatic by an external magnet: it consumed thirty- 
two seconds in making a single oscillation. : 
Experiments.—The light of a luminous gas-flame at a disiance 
of twelve inches was allowed to fall on the blackened dise not 
perited with a plate of mica, its companion being protec 
y a triple screen of sheet brass from the action of the flame. 
Under these circumstances the exposed disc moved away from 
the light: after it had come to rest ten readings were made; 
‘3°23 mean deviation away from light. 
Next, the vane provided with the mica shield was exposed, 
the other being screened. After a slight agitation It sere 
rest, and ten readings were made as before; the results o 
experiments are given below: 
0°26 away from light. 
0°-06 toward the light. 
0°°10 mean. sa : 
It will be seen that the interposition of the mica plate ere i 
to the vane actually did prevent any reaction from is Si 
place between the disc and the walls of the flask, so that pre 
