96 



BELL SYSTEM TECHNICAL JOURNAL 



is settled, and the triumphant side is that of which Meitner was the 

 protagonist. Evidently we must conceive that the electron departing 

 from a nucleus leaves it in a very unstable state, from which it speedily 

 passes over into a comparatively though not absolutely stable state by 

 one or a series of transitions, of which the gamma-rays are the manifes- 

 tations. 



We have still the gamma-rays to consider. Throughout this article 

 I have taken it for granted that the gamma-rays are electromagnetic 

 waves of definite frequencies. The evidence that they are electromag- 

 netic waves has been known so long 

 that it need not be rehearsed. The 

 classical way of determining the fre- 

 quency of such a wave is to measure its 

 wave-length. With ordinary light-waves 

 this is effected by dispersing them with 

 a prism or diffracting them with a ruled 

 grating. It used to be thought that 

 these methods do not avail with X-rays, 

 because of the shortness of their waves; 

 natural crystal gratings, in which close- 

 ordered files of atoms play the role of the 

 rulings in artificial gratings, are used to 

 diffract these. Applying the crystals to 

 gamma-rays, one meets the same diffi- 

 culty as the discoverers of X-ray met 

 when they applied prisms and ruled grat- 

 ings; the waves are mostly too short to be diffracted appreciably by 

 natural crystals. The gamma-rays are spread out into a spectrum, and 

 sometimes lines are discernible in the spectrum (Fig. 9) ; but the line of 

 shortest wave-length thus far measured (so far as I know) is at 0.052 

 Angstrom units or 52 X-units, and there are certainly many others at 

 much shorter wave-lengths which the crystal spectroscope does^not dif- 

 fract far enough outward to be located. Recently people have renewed 

 the attempt to measure wave-lengths of X-rays by the methods appro- 

 priate to visible light, and have attained values of astonishing accu- 

 racy; perhaps it is not too much to hope that a comparable advance in 

 technique will bring the shortest gamma-rays into the scope of crystal 

 gratings. 



The usual method for estimating the frequencies of gamma-rays 

 consists in guessing the class to which the electrons forming a beta-ray 

 line originally belonged; taking its binding-energy from the tables; 

 measuring the kinetic energy of the electrons forming the line; adding 



Fig. 9. Gamma-ray spectrum 

 of radiothorium and thorium B 



(Wavelengths of the lines 52X, 

 145X, 168X from right to left. 

 After J. Thibaud, /. c. footnote 



26.) 



