MR. W. CEOOKES ON EEPULSION EESULTING- EEOM EADIATION. 
541 
have not attained the neutral point until the exhaustion was within a very small 
fraction of a millimetre (43, 47) ; whilst if the metal is in the form of thin foil the 
neutral point may easily be got lower than with pith. 
I am inclined to believe that the true action of radiation is repulsion at any pres- 
sure, and that the attraction observed when the rarefaction is below the neutral point 
is caused by some modifying circumstance connected with the surrounding gas, not 
necessarily of the nature of air-currents (80). As a proof of this I have not unfre- 
quently obtained repulsion from radiation when the apparatus was full of air at the 
normal pressure. 
113. The following experiments are too few in number, and have not been varied 
sufficiently as to conditions, to enable many inferences to be drawn from them. 
However, they afford glimpses of a law governing the position of the neutral point. 
A torsion-apparatus was fitted up similar to the one described in paragraph 102. 
The beam was of glass, and at one extremity was fitted with a spring clip, also of glass, 
so that different bodies could be experimented with. Disks of platinum foil, 1 centi- 
metre in diameter and weighing T28 grain each, were prepared, and they were fixed 
in the clip at the end of the torsion-beam, either singly or two, three, or four together, 
in such a manner that while the disk exposed was always 1 centim. in diameter, 
the weights should be in the proportion 1, 2, 3, 4. At the other end of the beam a 
movable counterpoise was arranged, so that the length of beam from the platinum 
disk to the centre was always the same. 
The neutral points were as follows : — • 
No. of disks. 
Barometer. 
Gauge. 
Diff.= 
Neutral point. 
Differences. 
1 . 
760 
682 
78 
8 
o 
760 
690 
70 
16 
3. 
760 
706 
54 
24 
4. 
760 
730 
30 
114. Two pieces of platinum, a and b , were now cut from the same sheet, each having 
1 square centim. of surface, a was left the full size, but b was carefully folded in four, 
so as to expose a surface of only a ^ of a square centimetre, the weight remaining the 
same. The neutral points were then taken. The average of several observations 
(which, however, were not quite so concordant as could have been wished) were, below 
a vacuum, 
a. b. 
136 millims. 70 millims. 
The pieces of foil were then coated with lampblack, and observations again taken. 
This time the neutral points came out — 
a. 
66 millims. 
b. 
124 millims. 
