350 
PROF. C. G. BARKLA ON X-RAYS AND THE THEORY OF RADIATION. 
that is, they were based on the absorption laws on the assumption that air had 
no spectral line within this range of wave-lengths. An unobserved spectral line 
would involve greater absorption on the shorter wave-length side, and would be 
accompanied by increased ionization. The ionization was, of course, experimentally 
observed ; but this was originally thought to be that accompanying what might be 
termed normal absorption in air. A greater absorption would necessitate a correction 
of the values obtained and would account for the apparent discrepancy. 
The existence of a characteristic X-radiation may, however, be shown by various 
methods—-by experiments on (a) X-ray fluorescence (excitation by a primary 
X-radiation); (b) corpuscular radiation; (c) ionization; (cl) absorption. All these 
methods have been applied and the results have demonstrated the existence of 
a series of X-radiations previously unknown ; the absorbabilities (consequently the 
wave-lengths) of these radiations from several elements have been determined. 
[The experiments, the results of which are given under the headings (b), ( c ), and 
( d ) of this section, were undertaken in collaboration with Miss M. P. Whtte, whom 
I also wish to thank for help in the preparation of diagrams and tables for the 
lecture. The results briefly referred to in section (a) were obtained ’ during an 
investigation with Miss J. Dunlop of the scattered radiation. A more detailed 
account of these investigations will be published later.] 
Ionization Method. 
Experiments showed that with a primary radiation of decreasing wave-length the 
i p ionization coefficient in CJELBr p -i i x • , , i j 
value ot ——— :: — . . " . — tailed to increase at the expected rate beyond 
ionization coefficient in air 
a wave-length of about 0‘56 x 10~ 8 cm. It was evident, however, that the emission of 
a characteristic radiation from air would be accompanied by a rise of the air ionization 
and consequent fall of the ratio. Experiments were therefore made upon the 
ionization in sulphur dioxide and methyl iodide relative to air, in order to determine 
whether the sudden change occurred in the ethyl bromide ionization or in the air 
ionization. 
The ionization coefficient in sulphur dioxide had been shown to bear an approxi¬ 
mately constant ratio to that in air through a long range of wave-lengths.* The 
range of wave-lengths employed had, however,, not included that at which the 
observed change occurred. On extending the series of experiments it was found that 
ionization coefficient in SO., 
the ratio 
ionization coefficient in air 
- underwent a diminution at the same wave¬ 
length as had given a low value for 
ionization coefficient in C 9 ELBr 
A similar 
ionization coefficient in air 
effect was observed iti comparing the methyl iodide and air ionization coefficients. 
* Barkla, ‘ Proc. Camb. Phil. Soc.,’May, 1909; ‘Phil. Mag.,’August, 1910; Barkla and Philpot, 
‘Phil. Mag.,’ June, 1913. 
