[ 601 ] 



LVTII. Precision-measurements in the X-Ray Spectra. 

 By Manne Siegbahn *. 



1. Introduction. 

 rrJHE systematical measurements o£ the X-ray spectra, 

 JL discovered by G. Gr. Barkla, with the help of crystal- 

 lattices, taken up by W. H. Bragg, H. G. J. Moseley, 

 de Broglie, and others f, have given a great many resulis 

 of the utmost interest. The theories of the constitution of 

 the atom and its radiation have received empirical basis which 

 has greatly furthered this branch of physical knowledge. 

 Through the theoretical investigations of Debye, Kroo, 

 Sommerfeld and others, the empirical material has, however, 

 in some respects been distanced. It will be seen that the 

 spectral formulae of the authors just mentioned require more 

 precise measurements of the wave-lengths than those which 

 have hitherto been attained, to allows a desirable control. 



For this purpose the author has worked out methods for 

 more precise measurements of wave-lengths in the X-ray 

 spectra. The results of this investigation, which extends 

 over the longer waves (greater than 1 A.E.) show that the 

 new values have about a hundred-fold accuracy compared 

 with the older. In a following paper a method for accurate 

 measurements of the short waves will be given. 



2. Method and Apparatus. 



In all the older methods of wave-length measurements the 

 reflexion angle (cf>) in the Bragg formula 



n\=2d sin <£> 

 was in some way or other geometrically determined, which 

 highly limited the attainable accuracy. In the following 

 method the angle of the reflected rays is determined as the 

 angle through which the same photographic plate has to be 

 turned in order to receive impressions of the nth order reflexion 

 on both sides of the direct ray. These angles were simply read 

 off on an accurate circle-scale. When the plate was turned 

 accurately 4(/> (and the crystal, nearly 2</> -(- 180°), the 

 spectral lines on both sides would cover one another. When 

 the angle was nearly 4c£ the lines on the plate fell at some 

 little distance apart. By measuring this distance the small 

 correction to 4(/> could readily be determined with desirable 

 accuracy. 



The method requires that plate and slit have the same 

 distance from the rotating axis of the reflecting face of 



* Communicated by the Author. 

 t Siegbahn, Jahrb. d. llad.und Elektronik, 1916. 

 Phil. Mag. S. 6. Vol. 37. No. 222. June 1919. 2 T 



