Vol. 8, 1922 
PHYSICS: R. C. WILLIAMSON 
257 
by the positive ions was not due to the full applied accelerating voltage 
(Va), since they started from a somewhat extended region between the 
electrodes. A shding shutter over the vapor nozzle made it possible to 
cut off the vapor jet. 
Holding the retarding potential constant and varying the accelerating 
field, observations were made and a curve plotted showing current to the 
electrometer versus accelerating voltage, when the jet was illuminated 
IS 
so 
^+25 
-so 
-75 
-WO 
Wei 
// 
= SV 
= (0V. 
( 
f 
] ^ 
1 
? 
/ 
or ^ 
Jgr 
p 
5" /O /5 
Acce/erat//7<7 T^otentici/ m Vb/ts 
20 
FIG. 1 
ahc, ghi, klm, were taken with radiation unscreened, mini- 
mum wave-length about 1850 A.; a'h'c' , g'h'i' , k'l'm' 
were taken with screen, minimum wave-length about 
3100 A. I'j gives change when screen is removed. 
cd-de is the change when vapor jet shutter is closed 
and reopened, and ef is due to inserting screen. 
by unscreened radiation from the arc. This procedure was repeated with 
a screen interposed which cut out quite sharply wave-lengths less than 
3100 A. Pairs of such curves were obtained for retarding potentials of 
1, 5, and 10 volts, as shown in figure 1. 
The curves taken with the radiation unscreened show negative currents 
with marked breaks toward positive values at accelerating voltages which 
depend upon the retarding potentials. The negative currents are due to 
