334 
ME. W. CEOOKES ON EEPULSION EESULTING- EEOM EADIATION. 
as great, and has the advantage of being capable of being increased or diminished within 
very wide limits. 
A glass tube ( a b), 16 inches long and 1 inch diameter, has a 4-inch bulb blown on 
to the lower end. c d is a bar of pith f- inch wide, 3J inches long, and yg- inch thick. 
One half ( d ) is coated with lampblack, and the other half left white, as shown in the 
figure. The pith bar is suspended in the bulb by a very fine cocoon fibre. Through 
the centre of the pith bar, at e, and at right angles to it, is passed a magnet (101) about 
i inch long, made out of a fine steel sewing-needle. To the ends of this magnet are 
attached cocoon fibres, which support a small square of silvered glass (/), hanging freely 
below the pith bar and at right angles to it ; a reference to the end view will show the 
arrangement. At the upper part of the tube (a b) are seen the tube filled with cocoa- 
nut-shell charcoal, and the spiral glass tube for connexion with the mercury-pump. The 
arrangement for an experiment is shown in fig. 3 , plan. A ray of light (gfh) from a 
slit in front of a lamp falls on the mirror (/), and is thence reflected on to the scale (h). 
The apparatus is so placed that the index ray falls near zero when the magnet ( e ) has 
assumed its normal north-south position. It may be brought accurately to zero, and the 
sensitiveness increased or diminished at will, even during an experiment, by means of a 
control-magnet on a cork sliding up and down the tube, as shown at i, either close to 
the bulb or at some distance off, and acting with or contrary to the earth’s magnetism, 
according to the sensitiveness required. 
The instrument was exhausted and reexhausted, with repeated heatings of the char- 
coal-tube, in the manner already described (131). It was finally sealed off from the 
pump, the charcoal still remaining attached to it, and it was set aside for some months. 
The following experiments were tried with it after it had arrived at its maximum sen- 
sitiveness : — 
136. The bulb was placed in a box lined with black velvet, apertures being cut to 
allow the index ray to pass in and out and the experimental light to fall on it. The 
index ray was passed through diaphragms in cards and a cell of water, to keep the heat 
from the lamp from acting on the pith. The face of the pith was also protected by 
black screens from all side radiation, and the path of the experimental ray of light was 
guarded on each side by a double row of bottles filled with water, and packed at the 
top and bottom with cotton-wool. Without this precaution I was unable to go suffi- 
ciently near the apparatus to observe the movement, without introducing irregularities 
from the heat of my body. The light was only allowed to shine on the black surface 
of the pith, a screen shading it from the white half. 
The scale (h) on which the index ray of light (gfh) fell was 5 feet 6 inches from the 
hanging mirror (/). It was divided into millimetres ; the measurements given below are 
the actual movements of the index ray of light along this millimetre-scale. The candle (k) 
was a “ parliamentary standard” (109). It was surrounded on three sides with black velvet 
screens, and an assistant, standing close to it, raised or depressed a black shade in front 
of it, as I called “ light ” or “ dark,” watching the index ray of light at the same time. 
