286 
DR. FRANK HORTON ON THE ORIGIN OF THE 
alternating current and in this way the temperature of the filament was raised, and 
its electron emission was investigated at different temperatures. The thermionic 
currents were plotted against the corresponding temperatures and a curve was drawn 
through the points so obtained. Some readings from this curve are given in the 
following table which also contains, for comparison, the values of the thermionic 
current from the filament at the same temperatures when heated by an alternating 
current passing through it. In both cases the potential difference applied between 
the terminals of the discharge tube was 208 volts, and the pressure of the residual 
gas in the apparatus was 0'0002 mm. 
Temperature, 
degrees Centigrade. 
Thermionic currents 
: 1 = 10 7 ampere. 
Filament heated 
by alternating current. 
Filament heated 
on tantalum wire. 
1754 
585 
480 
1690 
319 
256 
1637 
163 
146 
1590 
78 
78-5 
1546 
35-2 
42-8 
1503 
22-6 
23-4 
1466 
8-6 
9-7 
An inspection of the above table shows that the electron emission is about the 
same whichever method of heating the filament be employed. The greatest difference 
is about 20 per cent., which is not large considering the experimental difficulties. 
From the table it appears that the emission from the filament when heated by 
conducting a current increases more rapidly with the temperature than it does when 
heated upon a tantalum wire. This was not always the case. Several filaments 
were tested and it was found that the emission was sometimes greater with one 
method of heating and sometimes with the other. The rate of increase with 
temperature of the thermionic emission from a Nernst filament heated by an 
alternating current in a vacuum depends very largely upon the previous heating of 
the filament. With filaments which had been heated for many hours at a high 
temperature it was generally found that the thermionic current above about 1500° C. 
increased with the temperature much less rapidly than with a new filament. 
It is clear that in these experiments there must be very much more electrolytic 
decomposition of the Nernst filament, followed by re-combination of the constituents, 
when the filament is itself conducting the heating current, than when it is heated on 
the tantalum wire. In the latter case only a very small portion of the heating 
current flows through the filament, for the resistance of this, even at the highest 
temperature, is very large compared with that of the tantalum wire. 
In comparing the electron emission in the two cases it is necessary to consider the 
