CONTEMPORARY ADVANCES IN PHYSICS 89 



I will pass over the theory first propounded for these data, which I 

 have quoted elsewhere, and over the paper in which Compton ad- 

 vanced practically the present theory; and reproduce one of the 

 spectrum-curves of scattered radiation from the work of Duane and 

 Clark. This curve represents the scattering of high-frequency X-rays 

 by molybdenum. The primary light was composed of several lines, 

 the various lines of the i^T-spectrum of tungsten. In the scattered 

 spectrum, these lines appear unshifted as narrow sharp peaks. We 

 consider the two at the extreme left, and then the two humps which 

 rise from the points marked B and A . These points correspond to the 

 wave-lengths of quanta which originally belonged to the spectrum-lines 

 at the extreme left, and have lost exactly the energy necessary to 

 extract a i^-electron from the molybdenum atom. The humps in all 

 probability are composed of quanta, which have extracted such elec- 

 trons and in addition have given them greater or smaller amounts of 

 extra kinetic energy. 



Years after the work of Clark and Duane had been discontinued, 

 Bergen Davis and D. P. Mitchell undertook to study what they 

 designated as the "fine structure " of the lines in the spectra of scattered 

 X-rays. They had many improvements of technique at their disposi- 

 tion, improvements many of which were due to Davis himself; for 

 instance they had a spectrometer of Davis' design, by which it was 

 possible to appreciate the true narrowness of a very narrow X-ray line, 

 instead of having it spread out by defects of the apparatus into a simu- 

 lation of a wide band. They irradiated graphite with the spectrum-line 

 known as Ka\ of molybdenum, of which the wave-length is 0.721 A.; 

 and in the spectrum of the rays scattered at 90° they found not only 

 this line, but four others of slightly greater wave-lengths of which 

 three are shown in Fig. 10. The outermost, beyond the right-hand 

 limit of the picture, comprises quanta which collided with free electrons 

 — it is the Compton shifted line. The outermost of the other three 

 consists of incident quanta which have given up just the amount of 

 energy required to extract one of the K electrons — one of the pair 

 which are by far the most tightly bound of all the six which belong to 

 the carbon atom. More precisely, the extraction energy of these 

 electrons is evaluated from the X-ray spectrum of carbon at 287 

 equivalent volts, while the loss of energy suffered by the quanta is 

 estimated by Davis and Mitchell at 279; the difference of less than 

 4 per cent is within the uncertainty of experiment. The two innermost 

 of the shifted lines, composed as they are of quanta which have yielded 

 up 29 and 50 equivalent volts respectively, are presumably tokens of 

 collisions in which superficial electrons were torn away from carbon 



