PHYSICS: WEBSTER AND CLARK 
183 
Results of Our Experiments. — ^We have partially tested these predic- 
tions for a platinum target in a Coolidge tube furnished through the 
kindness of Dr. Coolidge. The potential was obtained from Professor 
Trowbridge's 40 kv storage battery, and measured by a Chaffee elec- 
trostatic voltmeter calibrated by a manganin resistance of 0.8940 
megohms. The spectrometer was similar in general plan to that used 
in the previous work on rhodium, but with many improvements in 
construction and accuracy. 
The principal results of the work to date are shown in figure 1, which 
contains graphs of intensity as a function of potential for two lines, 
ai and ^2, of the Li series and three lines, jSi, and 71, of the L2 series. 
In each case the graph leaves the axis at the potential required to give 
an electron an energy quantum of the proper frequency, and each graph 
is of the form of one for general radiation only, from the start to a 
definite potential at which its curvature suddenly changes. This 
change denotes the first appearance of the spectrum line. For both 
the Li lines, this change occurs at 11.45 kv., or the quantum potential 
of the jSs line, as predicted, within the limits of error caused by the un- 
certainty in h. For 184 and jSi also the critical points are identical, but 
quite different from that of the Li lines. In none of these four cases can 
there be any doubt as to the existence of critical points, although the 
intensities are somewhat inaccurate, since even /Si is so weak there 
that a fall of the electroscope over the shortest scale division requires 
about half a minute. 
In the interpretation of the 71 curve, difficulty arises from the fact 
that the critical point is so close to the quantum potential that the 
general radiation itself has not acquired a uniform rate of increase. 
This curve has therefore been repeated in figure 2, with curves for gen- 
eral radiation of wave lengths just above and below 71. From these 
the general radiation with 71, itself may be estimated. It is repre- 
sented by the dotted line, which crosses each of the other curves at 
about the proper distances and would lie midway between them at very 
high potentials. From these graphs it appears that 71 also has a crit- 
ical potential, identical with that of /3i and ,84, which appears to be 
13.20± 0.05 kv. The wave length corresponding to this potential 
is 0.93 7 A, not that of 74, the predicted head of the series, but even 
longer than 72. This result, which was quite unexpected, raises a ques- 
tion as to the nature and behavior of the three very faint lines 72, 73, 
and 74, on which we hope to obtain more evidence in the course of this 
work. 
The calculations of the critical wave lengths of the series is dependent 
