Vol. 6, 1920 
PHYSICS: D. L. WEBSTER 
27 
In the present work, to increase the intensity even at a sacrifice of re- 
solving power, I used a long radius, either 40 or 80 cm., with the tube 
75 cm. from the mica and the plate as shown in figure 1. In practice, 
even with the best mica, there are faults in its structure that make it neces- 
sary to take several photographs at each voltage with different positions 
of the mica so as to insure clear reflection for each line in one case or 
another. To avoid scattered rays the mica was held on the edges of two 
FIG. 1 
brass plates, far enough apart so that no direct rays struck them, and the 
direct beam through the mica was caught in a sheet of lead. The lines 
were identified from a chart made from the wave-length measurements 
of Siegbahn and Friman,^ and the voltmeter (Chaffee type, as in the pre- 
vious work) was calibrated by the wave-length \miny of the high frequency 
end of the third order spectrum as determined from this chart. In all 
exposures near any critical voltages the voltage was regulated to a mean 
deviation of to Vs of 1% from the average value. 
An important improvement in the voltmeter, introduced in the work 
at Harvard, was the use of a single insulator only, directly under the sus- 
pended system, with fairly good electrostatic screening. This prevents 
disturbances from indeterminate charges on the insulator which may 
otherwise make serious errors. 
Bach spectrogram was taken on one or two Eastman Dupli-tized X-ray 
films backed by a Seed Graflex plate. For locating very faint lines this 
combination was useful because with the plate and film, or two films, 
in^contact after development any relative motion of the two would make 
