Vol. 8, 1922 
PHYSICS: L. L. NETTLETON 
355 
may flow against a negative potential of 100 volts or more with respect to 
the filament, the grid being at a positive potential of several hundred 
volts. If the plate is simply connected to the filament through a voltmeter, 
this voltmeter may read as high as 150 volts, or if connected through an 
ammeter, the negative current to the plate may be as much as 25% of the 
total electron current from the filament. The magnitude of this negative 
current or voltage varies greatly with the temperature of the filament, 
grid voltage, and gas pressure. There is never oscillation in the Lecher 
wires without a negative current or potential on the plate, but this negative 
current may be present and yet the Lecher wires show very feeble or no os- 
cillations. The characteristics of this negative current and the associated 
oscillations have been studied as affected by changes of grid voltage, of 
electron current, and of gas pressure. 
It was found, very unexpectedly, that both the negative plate current 
and any sign of oscillation, as indicated by the Lecher wires, ceased abruptly 
a/a 
A^ 
n 
< 
/ 
^^^^ 
\ 
A 
\. A 
A \ 
\ 
-1 — 1 — 1 
1 1 r— 
1 Preset, 
re Scale t 
'or Air 
7O00/mm Hq S 
10 
fS 
50 
^40 
30 
20 
50 
100 
iSO 
200 250 
FIGURE 2 
300 
350 
400 450 
lONlZf^riON 
500 
at very low pressures. This is shown in the curves of figure 2 for plate 
voltages, and in figure 3 for the strength of oscillation in the Lecher wires. 
These observations were all made with the grid at 250 volts and the 
filament heating currents such that the electron current was 200 mil- 
amperes. The pressure of the residual gas in the tube was measured with 
an ionization manometer. The pressures of hydrogen and air were con- 
trolled by adjusting the flame under the mercury vapor pump, with liquid 
air on the trap to keep out mercury vapor. The mercury vapor pressure 
was controlled by keeping the liquid air trap (with a pool of mercury at 
