356 
PHYSICS: L. L. NETTLETON 
Proc. N. a. S. 
the bottom) at different temperatures, the mercury vapor pump operating 
at its maximum rate to reduce the pressure of any gas to the minimum. 
The curves are plotted with the readings of the ionization manometer as 
abscissae. This was done because it is probable that the essential thing 
is not the presence of the gas itself, but of gaseous ions. Since different 
gases at the same pressure give different amounts of ionization when under 
the same ionizing agent, it seemed better to plot their ionization instead of 
pressure. The pressure corresponding to a given ionization current is 
I 
-J 
05 
'ure ^ca/e 
for Air 
If j 
— 1 ) — 
A 
.© p) o — 
-A . 
mm Hq 
50 
100 
ISO 200 
FIGURE 3 
250 300 
IONIZATION 
for mercury about half, and for hydrogen about twice that for air. The 
pressure scale for air is given on the curves, so by multiplying and dividing 
by 2 the approximate pressures of hydrogen and mercury, respectively, 
will be obtained. 
These curves apparently indicate that a small amount of ionization is 
necessary for this type of oscillation. The kind of gas does not seem im- 
portant. Since the gas pressure at which the oscillation stops is so low, be- 
ing less than .00005 mm. there can be no connection with the mean free 
path of the ions, for at these pressures the path is many times longer than 
the dimensions of the tube. It is not likely that the commercial tubes used 
by other investigators are sufficiently exhausted to give any indication of 
the dropping off of the curves at the low pressure side. So far, no staisf ac- 
