Electricity from a Nernst Filament. 253 
y 
positive leak with the potential at 0°018 mm. pressure is given. 
We are now ina position to explain the difference in the 
shapes of the curves A and C (diagram 8), representing the 
variation of the positive leak with the temperature in a 
vacuum. Curve A was obtained with filament II., with 
which long sets of observations extending over several days 
had been previously made. For this filament, therefore, the 
positive leak would diminish only very slowly with the time. 
The filament for which curve C was obtained was, on the con- 
trary, a new filament, and had never been kept glowing 
for a long time at atmospheric pressure. The rate of dimi- 
nution of the positive leak with the time for the latter 
filament would accordingly be much greater than in the 
former case. 
Some observations made on March 22nd with filament II. 
suggest that the positive leak diminishes much more rapidly 
with the time when the filament is kept positively charged 
than when uncharged. Further, charging the filament nega- 
tively was found to stop the diminution of the positive leak 
with the time. The observations indicating the existence of 
this effect are given in Table XII. The filament was kept at 
a constant temperature throughout. 
TABLE XII. 
| | . 
| / : Potential on 
Tontperabure Time | © otential |Filament during Dakitia 
Pressure. Filament. (approx.). | Filam Bae Beaune | Leak 
| observations. 
mm. | 2C; hm | Y 
0018 1540 7 O | +400 168 
j | +400 
7 10 +400 heal 
| 0 
| 715 | +400 | 145 
) —409 
“20 | +400 | 169 
This effect may be due to some sort of polarization of the 
filament. The author hopes to make at an early date some 
further experiments on this point, and also to see if the positive 
leak can be completely destroyed by keeping the filament 
glowing for a very long time. 
