484 
PHYSICS: DUANE AND STENSTROM Proc. N. A. S. 
chemical elements, we can estimate approximately where the critical 
absorption wave-lengths lie in the M series of tungsten. Making this 
estimate, and calculating the three critical absorption wave-numbers 
we obtain the following values, each multiplied by 10^, 
Mai = 0.152 Ma2 = 0.157 Mas = 0.182. 
vSubtr acting these from the K absorption wave-number, and comparing 
the differences with the K|8 emission wave-number we get 
Ka - Mai = 5.464 
Ka - Ma2 = 5.459 K,S = 5.429 
Ka — Mas = 5.434 
One of the differences between absorption wave-numbers, Ka — Mas 
lies very close to that of the KjS line. This suggests that the K/? line is 
due chiefly to electrons falling from the third M orbit to the K orbit. 
If electrons also fell from the second M orbit to the K orbit they would 
produce a line on the short wave-length side of the principle Kj8 line. 
Faint lines on the short wave-length sides of the Kj8 lines have been ob- 
served in the spectra of a few of the chemical elements of low atomic 
number by Mr. Hjalmar working in Professor Siegbahn's laboratory. ^ 
The separation of these components of the K(3 line from each other amounts 
to a small fraction of 1%. 
According to Rubinowicz' "Principle of Selection" transfers of elec- 
trons from the first M orbit to the K orbit ought not to occur. ^ 
The faint component that has been observed by Siegbahn,^ de Broglie^ 
and Hjalmar on the long wave-length side of the K/3 line in the spectra of 
a few of the chemical elements was predicted by Sommerfeld^ by reasoning 
from his theory of elliptic orbits. 
These components of the line may account for some of the irregu- 
larities in the j(3 peaks on our curves (figs. 2 and 3). 
Some years ago Kossel suggested that the difference between the fre- 
quencies (or wave-numbers) of the KjS and Ka lines should equal that of 
a line in the L series. It is well known that this relation, although ap- 
proximately true, is not quite exact. An explanation for this discrepancy 
may be found in the supposition that the electrons producing the KjS 
line may not come from the same M orbits as those producing the La and 
L/3 lines."^ Dr. R. A. Patterson and one of us recently presented to the 
American Physical Society an account of some experiments which show 
that the wave-numbers of the Lai, La2 and LjSi lines of uranium and thor- 
ium equal the differences between certain of the L and M critical absorp- 
tion wave-numbers. These relations may be expressed by the following 
equations : 
Lai — Mai = Lai, Lai — Ma2 = La2 and La2 — Ma2 = LjSi. 
In terms of the transfer of electrons between orbits these equations mean 
