332 
ME. W. CEOOKES ON EEPULSION EESULTING- EEOM EADIATION. 
along the axis of the tube a 1). The whole is supported on a stand furnished with finely 
cut levelling-screws, and, according to the principle of the horizontal pendulum, the 
sensitiveness of the beam to any force applied at the pith end can be increased or dimi- 
nished at pleasure by tilting the end a of the apparatus up or down ; this can be easily 
effected by turning the milled head of the screw f. A ray of light from a lamp is reflected 
from the mirror to a graduated scale, and appropriate screens are used to cut off 
from the pith disk all radiation, except that being experimented on. The apparatus is 
connected to the pump by means of the glass spiral shown at the upper part of the 
tube c. On lowering the end a of the horizontal tube, by means of the screw/', the 
oscillation of the beam becomes very rapid, and its sensitiveness diminishes. On raising 
the end a the time of oscillation can be increased to any desired amount, with corre- 
sponding increase in sensitiveness. The other levelling-screws are for the purpose of 
bringing the beam into the centre of the horizontal tube. If the tube is too much 
tilted up, the centre of gravity gets too high, and the pith falls to one side or the other 
of the tube. The most convenient degree of sensitiveness I found to be that accom- 
panying an oscillation at the rate of one per minute. 
The ray of light used as an index of movement; was reflected to a graduated scale 
4 feet off. The instrument, mounted and adjusted as above described and highly 
exhausted, was found to be very sensitive. A ray of light from a candle 10 feet off, 
falling on the pith, would cause the index ray to move through 15 divisions ; when 5 feet 
off the index moved about 60 divisions, and when 20 feet off the index moved between 
3 and 4 divisions. These movements were sufficient to show that the motion of the 
pith was in inverse proportion to the square of the distance the candle was from it. 
I tried numerous experiments with this apparatus, and verified the law perfectly ; but 
there were difficulties connected with working with it which induced me to devise 
another instrument, free from the objections attending the use of the horizontal 
pendulum, and at the same time simpler to make and equally sensitive to faint 
radiation. 
134. The objections to the use of the horizontal pendulum were the following: — 
When sufficiently sensitive to indicate readily the action of faint light, I found it 
almost impossible to bring the index to zero ; the oscillations were so slow, and, taking 
place in a vacuum, kept on for so many minutes that my patience became exhausted 
with waiting for the next observation. But if I ventured to move away, or, when stand- 
ing close to the apparatus, even to shift the weight of the body from one leg to the 
other, that was sufficient to alter the level of the floor, and therefore of the horizontal 
beam ; the spot of light would suddenly fly ten or twelve degrees in another direction, 
and all the tedious waiting had to be gone over again, and possibly another zero had to 
be taken. A person running up stairs, a child playing in the adjoining room, a passing 
carriage, or a rail way- train, all had their influence on the level of the laboratory floor. 
I tried fixing the apparatus to a main wall of the house, but this did little good. When 
the instrument was brought to its highest pitch of sensitiveness, to watch the move- 
