Vol. 6, 1920 
PHYSICS: DUANE AND PATTERSON 
523 
The positions of /^s as determined from the measurements of Overn^ 
and Siegbahn^^ are marked above the curve in figure 1. According to 
their measurements this faint Hne has a wave-length a small fraction of 
one per cent shorter than our value, which would bring it still further 
from the critical absorption in the direction of shorter wave-lengths. 
The curve in figure 2 represents the spectrum of tungsten in the neigh- 
borhood of the critical absorption La2. The drop corresponding to the 
absorption and the peaks corresponding to the 7 emission lines appear 
on the curve. The critical absorption lies between the emission lines 71 
and 72, and therefore has a wave-length longer than that of 72. Overn's 
and Siegbahn's values for 72 are in exact agreement with our values. 
According to the conception of radiation held by many scientists the 
critical absorption wave-length corresponds to the short wave-length limit 
of the group or series of emission lines with which it is associated. If 
this be true, or, to speak more accurately, if the term "group of lines" is 
defined in such a way that this is true, then 185 cannot belong to the first 
group in the L series of X-rays, and 70 cannot belong to the second group. 
The strongest evidence in favor of believing that belongs to the same 
atomic mechanism that produces the lines in the first group appears to 
be that the square root of its frequency increases from atom to atom 
nearly as a linear function of the atomic number, corresponding in this 
respect to all the lines in the first group but not to those in the second and 
third groups. In terms of the theory of atomic orbits this means that 
electrons falling into the Li orbit produce the line /Ss, as they do all the 
lines in group 1. Webster and Clark^ found that /Ss in the platinum 
spectrum appeared at a lower voltage than that required to produce the 
lines in the second group. This proves conclusively that jSs cannot belong 
to the second group. The experiments on critical potentials, however, 
are not sufficiently accurate to decide whether 185 appears at exactly the 
same voltage as the other lines in group 1, for /Ss is a weak line, and the 
difference between its wave-length and that of the critical absorption Li 
amounts to only 0.7% for tungsten. We are not, therefore, compelled 
to assume that can be produced by electrons in the X-ray tube having 
quantities of energy less than that given by the quantum equation, 
Ve = hv. 
The argument in favor of supposing that the line 72 belongs to the 
mechanism that produces the lines in the second group rests largely upon 
the fact that for the various chemical elements the difference in frequency 
between 72 and /Js equals the frequency interval between the lines in the 
other pairs belonging respectively to the two groups. It also equals the 
difference in frequency between the tw^o critical absorptions La2 and 
Lau and is given quite accurately by Sommerfeld's L doublet formula. 
Hence, according to the theory of electron orbits, 72, in common with the 
