352 PROF. . Gr. BARKLA ON X-RAYS AND THE THEORY OF RADIATION. 
pass through hydrogen (the direct ionization in which is negligible), and this indirect 
ionization was measured. Though the ionization produced in hydrogen by the 
secondary corpuscular radiation is not necessarily quite the same as that produced in 
aluminium by the absorption of its own corpuscles, the factor giving aluminium 
ionization from that in hydrogen is constant or nearly so. Consequently changes in 
the relative values of the ionization in aluminium and air may be observed equally 
well by this indirect method. 
The ionization in a chamber with its back face of aluminium and filled with 
hydrogen was compared with that produced by the same beam of X-rays in air. In 
the first case the ionization was ionization by the electrons ejected from hydrogen 
plus that by the net gain of electrons from aluminium at the end. In the second case 
the ionization was produced b} r electrons from the air itself plus that produced by 
electrons from aluminium ends. Approximately these are (a) ionization in hydrogen 
by electrons from the aluminium plates, and ( b) ionization in air by its own 
electrons plus a comparatively small ionization in air by a net gain of electrons 
from the aluminium plates at the end of the chamber. (The latter of these might 
have been approximately eliminated from the results of experiments on the 
corpuscular radiation from aluminium. This was unnecessary, however; an approxi¬ 
mate correction was obtained by extrapolation.) 
Homogeneous beams of various wave-lengths were employed, and the relative 
ionizations in air and in aluminium were obtained. A correction had, however, to be 
made to the relative ionizations obtained in this way, for the air ionization was due 
to the electrons ejected from a given mass of air, whereas the aluminium ionization 
was produced by the electrons from a mass which increased with the frequency of the 
primary radiation. This is due to the fact that the shorter the wave-length of 
the primary radiation, the greater is the velocity of electrons which this radiation 
ejects, consequently the greater is the depth of material from which they emerge. 
An approximate correction for this may, however, be found by extrapolation. This 
is explained below in its application to the determination of the relative ionizations 
in air and in copper. The final results of the experiments on air and aluminium are 
shown graphically in fig. 6, curve 2. The rise in the air ionization at a wave-length 
about 0'56 x l(r 8 cm. (\/p in A1 — 2'5) is well marked, and the subsequent rise in 
aluminium ionization is shown in the change in relative ionization in the opposite 
direction at a wave-length of about 0'52x 10~ 8 cm. (\/p in A1 = T9). It will be seen 
that the curve so obtained is very like the relative ionization curves for S0 2 and air 
obtained directly. This, of course, is as it should be, the atomic weights of aluminium 
and sulphur being close together, and their characteristic radiations consequently 
being of neighbouring wave-length. The difference between the two indicates that 
the radiation characteristic of sulphur is more penetrating than that of aluminium, 
for the relative ionization begins to approach the normal value with more penetrating 
primary radiation (of shorter wave-length) in the case of sulphur than that of 
