ELECTRON EMISSION FROM GLOWING SOLIDS. 
285 
pressure so that this increase in pressure should be negligible, but the results only 
justify the statement that there was no great difference between the emissions when 
the filament was switched momentarily from the alternating circuit on to the 
continuous current circuit, the resistances in the two circuits having been adjusted so 
as to maintain the temperature of the filament as nearly constant as possible. Since 
the filament could not be left on the direct current circuit for more than a second or 
two at a time, the comparison was very difficult to make. For this reason the 
following experiments were devised. 
II. A Comparison of the Electronic Emissions from a Nernst Filament when heated 
by conducting an Electric Current and when heated without conducting the 
Heating Current. 
The ideal method of making this comparison would, at first sight, appear to be to 
heat the filament in a vacuum, firstly in the usual manner by means of an alternating 
current, and secondly by placing the discharge apparatus in some form of furnace. 
Such an experiment is, however, not practically possible, for a good vacuum cannot 
be maintained in a vessel at a high temperature ; not only do the walls of the vessel 
evolve gases, but the material which would otherwise be most suitable for the 
experiment—quartz—becomes permeable to gases at about 1000° C. 
A Nernst filament is in the form of a fine tube and the method of thermal heating 
adopted was by placing the filament upon a tantalum wire which just fitted it, and 
heating the wire by an alternating current from the transformer. Tantalum wires 
of the required diameter were prepared for me by Messrs. Siemens, and supplied in 
straight pieces about 5 cm. long. In performing the experiment the negative 
emission from a filament was investigated when the filament was heated by an 
alternating current in a vacuum obtained by the use of charcoal cooled in liquid air. 
The thermionic current as a rule decreased slightly during the first few hours 
heating of a new filament; after this it was fairly steady even when tested on 
different days. The filament was then disconnected from the platinum leads and 
its two ends were cut off, leaving the middle glowing portion only. This was 
mounted on a tantalum wire cut to such a length that about 3 mm. of wire protruded 
from each end of the filament. The ends of the tantalum wire were fitted inside two 
short pieces of platinum tube which had already been welded on to the platinum 
leads of the discharge tube. The platinum tubes were of about the same dimensions 
as the filament, and were pinched up tightly on to the tantalum, thus making good 
electrical contact. In this way none of the tantalum wire was exposed ; the central 
part was covered by the filament and the ends were encased in the platinum tubes 
through which the heating current was conducted. These leads never became 
sufficiently hot to give a thermionic emission comparable with that given by the 
filament. After exhausting the apparatus, the tantalum wire was heated by an 
