534 
ME, W. CEOOKES ON EEPULSION EESULTING EEOM EADIATION. 
103. The preparation of the suspending thread of glass requires some care. It 
should be drawn from flint glass, as this gives much tougher threads than foreign 
glass. The diameter varies with the amount of torsion required ; it may be O'OOl inch 
or less. I select the piece best adapted for the special experiment in the following 
way : — Several threads of glass are first drawn out before the blowpipe, and a certain 
number selected as being likely to answer the purpose. These are then suspended, side 
by side, to a horizontal rod and equalized as to length. A piece of glass rod, about 
2 inches long, which is always kept for this purpose, is then cemented by shellac on to 
the end of one of the threads. Air-currents are then cut off by a glass screen, and the 
thread being set in movement by a slight twist, the torsion is measured by timing the 
oscillations. This having been done with each thread in succession, one is selected and 
mounted in the apparatus. If it works properly, well and good ; if not, it is easy to 
select a thread having the requisite amount of torsion, more or less, and substitute it for 
the one first used. 
In fitting up one of these apparatus, threads were drawn out which were found to 
require respectively : — 
44 seconds, 
30 „ 
28 „ 
11 „ 
and 
3! „ 
for a half oscillation when the glass weight was hung on to their ends. The one oscil- 
lating in 30 seconds was first used, but was found to give insufficient torsion. The one 
making half an oscillation in 11 seconds was then used, and was found to answer well. 
Before I adopted this plan days were frequently wasted in the attempt to hit upon a 
glass thread of the requisite degree of fineness. 
104. In taking accurate observations with an apparatus of this description, it is neces- 
sary to support it on a stand firmly fastened to a main wall. When resting on a bench, 
or connected in any other way to the floor, there is a constant oscillation which keeps 
the index from zero. 
The apparatus being fastened firmly to its stand, accurately levelled, and sealed on to 
the pump, a divided scale, a b (fig. 8), is placed four feet from the small mirror ; and 
immediately beneath the scale is a narrow brass slit, c, illuminated by a lamp, d. In 
front is a lens, e, which throws the image of the slit on to the mirror, where it is reflected 
back again on to the divided scale. Here the angular movement of the bright line of 
light shows the minutest attractive or repulsive force acting on the pith at the extre- 
mity of the movable index. 
In order to keep the luminous index accurately at zero, except when experiments are 
being tried, extreme precautions must be taken to keep all extraneous radiation from 
acting on the apparatus. A slightly conical paper tube, f, about 6 inches long, and 
as narrow as the angular movement of the ray of light will admit of, is cemented on to 
the glass window in front of the mirror ; and a similar tube, g , is cemented on to the 
