ME. W. CROOKES ON REPULSION RESULTING EROM RADIATION. 
537 
instead of swinging to the opposite side and only returning to rest after ten or a dozen 
oscillations, as is the case when the beam is set vibrating by mechanical means. This 
behaviour points to the return movement taking place under the influence of a force 
which remains active after the original radiation is cut off, and which is only gradually 
dissipated. This force is most probably from the heat which the pith has absorbed 
raising its temperature ; and the steady return to zero seems to be due to the movement 
being controlled by the radiation of heat by the pith. 
108. A series of observations taken with another apparatus, with the object of ascer- 
taining the times of oscillation to and fro, showed that the first half, or the maximum 
deviation produced, whilst under the influence of radiation, occupied about the same 
time as the second half, or the return swing, w 7 hen the source of radiation was cut off. 
The following are the observations. The source of radiation was a candle, the intensity 
of action being moderated by filtering the rays through glass screens. 
Half oscillation, 
under influence of 
radiation. 
Whole oscillation, 
radiation being cut off 
during the return, swing.- ■ 
8 seconds. 
15 seconds. 
7-5 
15 
7-5 ,, 
14-5 „ 
7-5 „ 
15-5 „ 
7-5 
14-5 „ 
7*25 „ 
15 ....... 
7*5 „ 
15 
7'5 „ 
15 „ 
7 
14 » 
7 ,i 
14 „ 
6*7 5 „ 
14 „ 
7 „ 
14 „ 
7-25 „ 
15 „ 
7 
14 „ 
7 
13*25 „ 
8 
16 
8 „ 
16 
7-5 
15 „ 
7 
15 „ 
8 
15 „ 
8*5 „ 
15*5 „ 
7’5 „ 
15 
8 
15 
8 
15 „ 
7 
14 „ 
Mean. ..7*47 „ 
] 
Mean... 14*77 ,, 
The average time of the first half oscillation is therefore 7*47 seconds*', and of the 
second half 7‘3 seconds. This small difference is not unlikely to be due to errors of 
observation. 
After a long series of experiments the zero gradually creeps up, showing that 
one side of the apparatus is becoming warmed. The conducting-power for heat and 
* % referring to paragraphs 106 and 307 it will he seen that I have put the time of the first half oscillation 
as 11-5 seconds. This was with another apparatus, having a glass thread of different torsion. 
MDCCCLXXV. 4 B 
