0. N. Rood—Force involved in Crookes’ Radiometer. 407 
tically the disc remained immovable. In the next experiment 
the brass screen was entirely removed, and the light allowed to 
fall on both discs simultaneously, a plate of mica, identical in 
substance with that attached to the vane, being placed outside 
of the flask and opposite the unprotected vane, so that both 
Vanes received the same amount of radiation, the only difference 
ing that one of them was deprived of its direct communica- 
tion with the walls of the flask. Under these circumstances the 
disc without the attached mica screen moved instantly away 
from the light, as was shown by two careful experiments: 
en t away from light. 
2°38 mean. 
These results I regard as confirmatory of the theory advanced 
by Stoney, and as fatal to the idea that the motion is directly 
dependent on the impact of light or heat, for both discs received 
the same amount of heat and light. 
According to the theory just referred to in these experiments 
repulsion took place between the blackened disc and the mica 
attached to it, but being firmly fastened together, no motion of 
either could result, Assuming this repulsion to exist I devised 
va as for making it evident and for measuring its 
moun 
Description of Apparatus.—A disc of aluminium, folded double 
and blackened on one side, was provided with a small magnet, 
and suspended by a single fiber of silk in the interior of a glass 
flask, Tn front of this was similarly suspended a dise of glass 
such as is used for covering microscopic objects; it also was 
Provided with a small magnet and both discs were properly 
Oho poised. These discs were each seventeen millimeters 
su and fourteen millimeters high: the weight of each was 0°3 
stam, and the minute magnets attached to them were of nearly 
“qual strength. The distance of the points of suspension apart 
"as five millimeters, Figure 3 furnishes a view of the system 
seen from above, 
oh - le disc, G that 3. 
“s, ¥ and P’ are the points of ae 4 
suspension, C and CO’ the bade An Pg 
ia Which being on different a 
i “4 readily passed each other ; 2 i 
in S and N’ S’ are the controll: 
of ghets. The exhaustion of the flask was carried to 0-24 
4 millimeter, 
"ments.—-It will be observed that the magnets were so 
wmanged as to tend to cause the discs to touch each other, but 
