Vol.. 8, 1922 PHYSICS: J. A. ELDRIDGE 269 
experimental methods which have hitherto been employed for the deter- 
mination of the resonating and ionizing potentials. It seemed desirable 
to use a new type of tube, with which the actual distribution of velocities of 
the electrons after their collisions with the vapor molecules could be 
determined. 
Instead of the grid employed in the usual form of tube two diaphragms, 
placed one behind the other and each pierced with a single small hole, 
were used. These diaphragms divided the tube into two regions: an 
upper one in which any desired pressure of mercury vapor could be main- 
tained and in which the collisions to be studied took place; and a lower 
region from which mercury was removed by liquid air and in which the 
electron velocities were measured. In the upper region the electrons were 
emitted from an oxide coated cathode and accelerated by a grid which 
0 Z 4 6 8 10 12 14 16 
Refardin9 Pote n tial-\/olfs 
was maintained at any desired potential. Passing through the grid, the 
electrons entered a constant potential region and after a large number 
of collisions with the mercury vapor, some passed through the two small 
holes into the lower, mercury free region and to the receiving electrode. 
To this electrode was applied a variable retarding potential and by measur- 
ing the current as a function of this retarding voltage the velocity dis- 
tribution of the electrons was determined. 
In the curve shown the current is plotted against the retarding poten- 
tial for electrons which had, before collision, velocities ranging between 
15.8 and 16.8 volts. We find in the curves drops beginning at 0, 3, 9.1, 
10.9 and 15.8 volts, respectively. The drop at 15.8 volts represents those 
electrons which after collision retained all of their energy. The other 
drops in the current are due to groups of electrons which have lost re- 
spectively all, 13.4 (i.e. 2 X 6.7), 6.7 and 4.9 volts of energy which they had 
