128 MR. W. CROOKES OK REPULSION RESULTING FROM RADIATION. 
REPULSION AT ATMOSPHERIC PRESSURE. 
461. The attraction or repulsion exerted between surfaces at atmospheric pressure 
or low exhaustions has always exhibited contradictions, the air currents caused by 
the heating of the glass envelope, and those generated inside by the heating of the 
movable indicator, being sufficient to neutralise, wholly or partly, any repulsion which 
was occasioned by true molecular pressure. It was thought that by special arrange¬ 
ments the effect of air currents might be separated from that of molecular pressure, 
and the following apparatus was accordingly fitted up. 
. It consists of a torsion balance, similar to the one last used (455, fig. 27), but 
instead of the adjustable mica screen and blacked disk, the experimental end of the 
Fig. 31. 
torsion beam was arranged as shown in fig. 31. a is a clear mica disk suspended on 
the end of the torsion beam ; b is a shallow cup of platinum, lampblacked on the 
convex side ; c is a platinum wire spiral, connected with thicker platinum terminals 
passing hermetically through the glass. By heating the spiral the platinum cup is 
warmed, and the approach or retreat of the disk a is measured by an index ray of 
light reflected on to a screen in the usual manner. 
The strength of the voltaic current used to heat the spiral is kept constant by the 
system of resistance coils and galvanometer described in a previous paper (359). Two 
Leclanche cells, giving a tolerably constant current, were used to heat the spiral; 
the heat never got up to redness. The scale was kept 3 feet from the mirror. 
462. The apparatus was exhausted to dry it, and then filled with dry air which had 
passed through a tube packed with phosphoric anhydride. The disk was adjusted so 
as just not to touch the platinum cup. 
When the spiral was heated, the disk was immediately attracted to the cup. The 
mica disk was now put at different distances from the cup. On heating the spiral, 
strong attraction always took place, the disk rushing across to the cup, rebounding 
from it, and rapidly settling in close contact. A wet finger* placed on the opposite 
side to the cup, when it had got cold, drew the disk away. 
463. Exhaustion was now proceeded with, observations being taken at intervals. 
The control magnet was moved so as to let the disk lightly touch the cup ; the scale was 
moved till the index ray of light stood at —140. The magnet then brought the 
torsion beam back till the index ray of light stood in the centre of the scale, at zero. 
As now adjusted, the mica disk is several millimetres from the cup, and can move 
* Wetted to avoid electrification by friction, 
