Vol. 8, 1922 PHYSICS: DUANE A ND PA TTERSON 87 
frequency than and 72 (a fact since verified by Coster), the idea of 
absorption Hmits being "Hmiting frequencies" for groups of emission Hues 
was confronted with an anomaly. We proposed an explanation of this 
apparent anomaly that was in accord with the Bohr theory. Since then, 
however, Hoyt^ and Dauvillier^ have pointed out that the W emission 
line at XI. 212 should be taken as 185 instead of the one at X1.203 as before. 
This suggestion has been confirmed by the more accurate measurements 
of the corresponding line for the other heavy elements. When the square 
root of the frequency is plotted against atomic number for the lines, 
the points all lie on a smooth curve. Furthermore, in place of 72, the work 
of Dauvillier and Coster indicates that a new 7 line, 76, at X = 1.072, 
is the result of an electron falling into the L2 orbit, and therefore belongs 
to the L2 group. Both the old (3^ line for W and the 72 lines have been 
assigned to the L3 group. 
Consequently, we now take (3b, Jq and 74 to be the shortest emission 
wave-lengths in the Li, L2 and L3 groups, respectively. 
Coster has compared his new values for the wave-lengths of each of these 
emission lines with our values of the corresponding absorption limits, 
and has concluded that the absorption limits are in general somewhat 
shorter in wave-length than the emission lines associated with them. 
Table II gives this comparison using our new values for L2 and L3. 
TABLE II 
ELEMBNT L2 76 Ls yi 
Pt .9318 =t 3 .9317 .8918 =t 9 .8950 
Au .9008 =t 4 .90125 .8610 ± 9 .8663 
Bi .7871 =t 4 .7874 .7562 =b 9 .7610 
Within experimental error the wave-lengths of the absorption limits 
are shorter than the wave-lengths of the shortest emission lines associated 
with them. 
Coster^ now places the number of L doublets at four, -q-l, ^2- ol^, 75 - jSe, 
76 - /Ss. The 7i - j82 doublet has been eliminated which is in accord with 
our suggestion that the electrons responsible for these two lines do not 
appear to start from the same orbit but different ones. He has also found 
and measured the two M absorption limits predicted by us for U and Th 
in order to explain the existence of the r] - 1 doublet. His values are in 
close agreement with our predicted ones. 
According to Smekal,* the My line is due to the passage of an electron 
from an Nz orbit to an Mz orbit. If this is true, the following relation 
between frequencies should hold: 
(L3 - Li) - M7 = Ldz - L^2 
Coster has tested this relation using his new values for ^2 and ^z, Sten- 
strom's value for Af 7 and our values of Li and L3. He finds it confirmed 
by the available experimental data except in the cases of Pt and Bi. Our 
