PROF. C. G. BARKLA ON X-RAYS AND THE THEORY OF RADIATION. 
343 
It should, however, be pointed out that in elements of lower atomic weight a 
considerably smaller fraction of the energy of primary radiation appears to be 
transformed into fluorescent (characteristic) X-radiation. Measurements of the 
energy of corpuscular radiation, too, indicate a smaller fraction transformed into 
radiation of this type than is the case in bromine. Thus experiments point to a 
marked loss of energy in such substances. # 
With elements of higher atomic weight, however, there appears to be a close 
approximation of the experimental values with those given by the above simple 
theory. 
We shall, in what follows, deal principally with the results obtained with ethyl 
bromide, the behaviour of which is certainly typical of many substances. If in light 
elements, as seems probable, a considerable portion of the energy absorbed is 
unaccounted for, this is a matter for future investigation. It does not, however, 
affect the conclusions to be obtained with other substances, but indicates that the 
conditions are not the simplest possible. 
Process of Radiation. 
It is inconceivable that the processes resulting in the emission of fluorescent 
X-radiation and corpuscular radiation are either independent or even that they are 
alternative conclusions (dependent on some critical condition in the atom) to similar 
initial processes. The above experimental relations are inexplicable on any such 
assumption. The emission of each electron of the corpuscular radiation is associated 
with the emission of a quantum of the corresponding fluorescent X-radiation. But 
we have already shown that the electron after emission takes no part in the 
fluorescence, so are left with the conclusion that the atom which emits an electron 
emits also a fluorescent radiation characteristic of the atom itself. We shall now 
consider the process in greater detail. 
The above approximate relations show that the energy of primary radiation 
absorbed per K electron expelled is greater than that per lower frequency electron 
expelled by a quantum of K radiation, t In the measurements upon which this 
conclusion is based we are unable to distinguish between L, M, and N electrons, but 
as L electrons constitute quite 80 per cent, of the whole, and for the others the energy 
* Further experimental work on the light elements is desirable, for it is remarkable that such a 
variation should occur with a comparatively small variation in the atomic weight of the radiating 
substance (fig. 4), for the line spectra of these elements are similar, and in addition the absorption curves 
(as in fig. 2) are exactly similar. Similar absorption curves indicate similar processes of absorption. 
f If we ignore the first approximation altogether and examine the experimentally determined values of 
Energy absorption per K electron - energy absorption per L electron , we gee that the possihle error is 
energy of quantum of K radiation 
necessarily very great. The experimental values 1 • 4, 115, O'9, O'7 and 0'9 are of the order of 
magnitude of unity. 
