154 
NATURE 
[ FEBRUARY 3, 1923 

exciting these lines. Thus it is found that there is a 
critical radiation frequency v-, which is very nearly 
equal to, but just greater than, Vey? and unless the 
incident radiation stream contains components the 
frequencies of which are at least as great as ve, the K 
series will not be excited. There isa precisely analogous 
limitation on the electron energies which cause the 
generation of the characteristic radiations. Thus there 
is a critical electron energy eV,, where V, denotes the 
critical potential difference through which the electron 
of charge e has to fall in order to gain this energy, 
which is connected with the critical frequency v, by 
the quantum relation eV.=hv,, and if the energy of 
the impinging electrons is equal to, or greater than, 
eV,, the characteristic radiations will be excited, other- 
wise they will not. Furthermore, if we measure the 
absorption of radiations of different frequencies by the 
element under consideration, we find that, correspond- 
K Levers or Light ELemenTs 
X Fricke Grilical Absorplion Frequencies by crystal methods. 
a= ¥ 
BD Holweck. Critical Alsorpiion by eV = hy. V= exciting voltage 
© andx. Crifical exciting volfage and photoelectric defection. 
Oxgen Kurth. Nifrogen Foole and Mohler. Boron, Hughes 
Carton. \. Kurth. 2. Richardson and Bazgoni 
3. Foote and Mohler. 4. Hughes 
B& Hydrogen. Convergence limit of Lyman Series 
| 
a ae ee ae 
ATOMIC NUMBER —= 
FIG. 4. 
ing to the excitation of the characteristic rays, there is 
a sudden increase in absorption at the critical frequency 
ve. There is also a discontinuity in the ionisation of 
the element at the same frequency. 
There is definite evidence from X-ray phenomena 
that the critical energy eV, measures the work which 
has to be done in removing an electron from its position 
in the normal atom to a point outside the atom. The 
characteristic rays are emitted when the gap thus 
created is subsequently filled up, the different lines 
arising according to the origin of the electron 
which fills the gap. If, measured in terms of 
energy, it is from a near location, we get a low- 
frequency line such as Ke; if it is from a location 
near the surface of the atom, a high-frequency line 
such as K, arises. 
Thus critical energies such as eV, give a direct meas- 
ure, in terms of energy, of the levels of the different 
electrons in the atom. Alternatively, the correspond- 
ing critical frequencies v, are the limits of the relevant 
X-ray spectra. If we can determine these limits we 
shall have found the high-frequency ends of the various 
spectra. While these ends are not, strictly speaking, 
NO. 2779, VOL, 111 | 





spectral lines, for the heavier elements at any rate, 
they are very close to the highest-frequency emission 
lines in the spectra. Furthermore, according to 
modern spectroscopic theory, they give us the funda- 
mental data on which the formule for the spectral 
series are based. 
It is a curious fact that evidence of the existence of 
such levels in the gap between what are ordinarily 
termed the X-ray and the ultra-violet spectra should 
have been produced independently and almost simul- 
taneously by a number of investigators scattered all 
over the world. These include Foote and Mohler in 
Washington, Holtsmark in Christiania, Holweck in 
Paris, Hughes in Kingston, Ontario, Kurth in Princeton, 
and myself and Bazzoni in London. While the details 
of the apparatus used by the different workers vary 
considerably, the principle involved in most of them can 
be made clear by reference to Fig. 2 (p. 119). Let 
the element under test forming the 
anode A, be bombarded by a power- 
ful electron current from the hot 
cathode F. Then anyradiation gene- 
rated at A, can pass between the 
charged plate condenser P, where 
any ions present will be removed 
from it, into the chamber on the 
left. If the rather complicated 
apparatus shown in the left-hand 
chamber is removed and replaced 
by a plate on which the radiation 
falls and by a second electrode, the 
radiation can be detected by the 
photoelectric electron emission it 
produces at the plate and measured 
by the current which flows between 
the two electrodes, the plate being 
negatively charged. Let this cur- 
rent be measured and divided by 
the thermionic bombarding current 
for a series of different potentials 
applied between F and A,; then 
if there is a sudden generation of 
characteristic rays from A, at some 
critical potential V. we should expect an increased rate 
of rise of the photoelectric current with applied potential 
to set inat V,. Thus, briefly stated, the experimental 
method is to plot photoelectric current per unit thermi- 
onic current against primary bombarding potential and 
to look for discontinuities in the resulting diagram. 
These discontinuities should occur at the critical 
potential differences V, corresponding to the energy 
levels eV, and to the frequency limits v, equal to eV,/h. 
This general type of method leaves much to be 
desired, but it seems the most practicable procedure 
at the present stage of the subject. It is open to the 
general objection that discontinuities in functional 
diagrams are often merely indications of faulty experi- 
menting, and the evidence that such discontinuities 
as are observed are really due to the excitation of X- 
rays is quite indirect and inferential. It is hoped later, 
however, to make good this deficiency by supplying a 
direct test of the frequencies of the radiations gene- 
rated ; for example, by using the magnetic spectro- 
scope which was used for determining the end of the 
helium spectrum, and by other methods. 
Fig. 4 shows the square roots of the critical fre- 
