ME. W. CEOOKES ON REPULSION RESULTING FROM RADIATION. 
533 
other. When I had succeeded in suspending the needle with an equal tension on each 
silk fibre, I found their elasticity to vary ; and as soon as the vacuum was approached 
one was sure to contract more than the other, twisting the needle out of the axis of the 
tube, and sometimes causing it to touch the side. This method of suspension was there- 
fore abandoned. 
By increasing the length of the needle, and also of the fibre used to suspend it, it was 
possible to employ fibres with a considerable amount of torsion, and still preserve the 
delicacy of the apparatus. Fine platinum wire was first tried ; but this was soon aban- 
doned in favour of glass fibres, which were found to answer so perfectly that I have 
since used nothing else. 
102. Fig. 7 shows the form of apparatus which I have finally adopted, as combining 
the greatest delicacy with facility of obtaining accurate observations, and therefore of 
getting quantitative as well as qualita- 
tive results. It is a torsion-apparatus in 
which the beam moves in a horizontal 
plane, and may be called a horizontal 
torsion-balance, a b is a piece of thin 
glass tubing, sealed off at the end b 
and ground perfectly flat at the end a. 
In the centre a circular hole, c, is 
blown, and another one, c\ at the end ; 
the edges of these holes are ground 
quite flat, a, c, and c 1 can therefore 
be sealed up by cementing flat trans- 
parent pieces of plate glass, quartz, 
or rock-salt, a, d , and d' on to them 
(83). To the centre of a b an up- 
right tube, ef\ is sealed, having an arm, 
g, blown on to it for the purpose of 
attaching the apparatus to the pump. 
h i is a glass index, drawn from circular or square (22) glass tube, and as light as possible 
consistent with the needful strength. A long piece of this tube is first drawn out before 
the blowpipe ; and it is then calibrated with mercury until a piece is found having the 
same bore throughout ; the necessary length is then cut from this portion. jTc is a very 
fine glass fibre, cemented at j to a piece of glass rod, and terminating at k with a stirrup, 
cut from aluminium foil, in which the glass index, h i, rests. In front of the stirrup is a 
thin glass mirror, shown at k, silvered by Liebig’s process, and either plane or concave as 
most convenient. At the ends of the glass index (Ji i) may be cemented any substance 
with which it is desired to experiment ; for general observations I prefer to have these 
extremities of pith, as thin as possible, and exposing a surface of 10 millimetres square. 
The pith may be coated with lampblack or silver, or may retain its natural surface. 
Fig. 7. 
