Electricity from a Nernst Filament. 233 
to about the middle. The slits were just wide enough to 
aliow the cylinder being slipped round the filament and the 
electrodes when the former was in 
position. In fig. 1 the wires of the 
thermal junction are shown passing 
through the cylinder D. The cylinder, 
however, was only employed in the ex- 
periments on the magnetic deflexion of 
the ions. In those experiments the use 
ot the thermoconple was dispensed with. 
The filament was placed mid-way be- 
tween two circular disks of platinum foil, 
2°8 cms. in diameter and 7 mms. apart. 
A plan of the disks is shown in fig. 3. 
The black part of the disk represents a 
gap about 1 mm. wide formed by cutting 
a strip out of the platinum. To the 
points O, O’, thick platinum wires were 
welded, and these were fastened to the 
copper leads M, M (fig. 2). Only one 
lead for each diskis shown in the diagram. 
By means of a current of about 8 
amperes the central U-shaped part of 
_ the disks could be raised to a bright red 
heat. By this means it was possible 
to “light”? the filament. When the 
filament had been got to glow, the 
current through the disks was cut off 
and the disks were connected to earth 
through a galvanometer. ‘The disks 
thus served the double purpose of acting 
as the “‘ heater” and as the electrodes to 
collect the electricity discharged from the 
filament. The shaded portions in figs. 1 and 2 represent 
sealing-wax joints which were made air-tight by coating with 
shellac varnish. The wires forming the ‘thermocouple were 
pure platinum and 10 per cent. platinum-rhodium, 0:0035 cm. 
in diameter. The junction was made by fusing the wires 
together in the oxyhydrogen flame. In order to standardize 
the thermocouple, a junction of the same wires as those used 
during the experiments was attached to a piece of stout 
platinum wire, which could be heated by means of an electric 
current. Very small grains of potassium sulphate were 
placed on the stout wire near the junction, and the current 
through the wire was increased until the grains of salt 
(observed through a microscope) just melted. The thermo- 
