132 MR. W. CROOKES ON REPULSION RESULTING FROM RADIATION. 
regarded as fair reasons for suspecting that the repulsions at high and at low 
pressures are not entirely due to the same cause. 
MEASUREMENT OP THE FORCE OF REPULSION. 
479. The following experiments were undertaken with the view of getting measure¬ 
ments of the actual amounts of force exerted by radiation in causing repulsion. 
A horizontal torsion balance was employed similar in construction to the one 
described in a previous paper'"' (209, 210, 211). A somewhat stiffer torsion fibre was 
employed, and the part corresponding to the pan of the balance was a clear mica disk 
16 millims. in diameter. A similar disk was fastened to the tube in which the beam 
oscillated, in such a position that when in equilibrium the pan should be 1 millim. above 
the fixed disk. This fixed disk was lampblacked on the upper surface, and had beneath 
a thin platinum spiral connected with terminals sealed in and passing through the 
glass, as in apparatus fig. 31 (461). When this spiral is heated by an electric current, 
the blacked mica disk fixed above it becomes heated, and the molecular pressure 
thereby generated between it and the mica pan causes the latter to rise. The glass 
thread attached to the beam is now twisted by means of the graduated circle, and 
the number of degrees through which the thread has to be twisted, hi order to bring 
the beam back to equilibrium, is noted. This gives a measurement of the pressure 
exerted in torsional degrees. To convert these degrees into grains it is only necessary 
to ascertain through how many degrees the glass thread has to be twisted, in order to 
balance a known weight placed on the pan. A piece of iron weighing O'Ol grain was 
used, and it was found that to restore the beam to equilibrium with this weight on 
the pan, 3 complete revolutions and 34°5 additional, or 1114'5 degrees of torsion, were 
required. The force with which the glass fibre tends to untwist itself, being directly 
proportional to the number of degrees through which it has been twisted, the value in 
grains of any number of torsional degrees is readily calculated. A ray of fight 
reflected from a mirror in the centre of the beam is used as an index, being brought 
to a definite mark on a scale for zero. 
The battery power was kept constant by the means already described (359, 461). 
480. In the following table the degrees of exhaustion are given in millionths of an 
atmosphere, and the force of repulsion is given in fractions of a grain, calculated from 
the degrees of torsion. 
* Phil. Trans., 1870, Vol. 100, part II., p. 37L 
