426 BELL SYSTEM TECHNICAL JOURNAL 



its faces being parallel to the f 100) strata. If one has a monochromatic 

 beam of X-rays incident on that face and revolves the crystal so that 

 6 varies steadily from zero to 90°, then for the several angles given by 

 equation (15) with various values of n diffraction-beams spring forth. 

 One may set up a photographic plate beside the crystal, and find the 

 imprints of all the beams upon it after the rotation is completed; or 

 alternatively one may revolve an ionization-chamber at double the 

 angular speed of the crystal, so that whenever a beam shoots out the 

 chamber is in the right place to capture it, and then the curve of 

 ionization-current versus angle d shows a peak for every value of 6 

 corresponding to an integer value of ti. If the incident beam comprises 

 many wave-lengths, one finds their spectrum spread out in the ioniza- 

 tion-current curve; three examples are shown in Fig. 14, where each 

 of the sharp tall peaks is due to the first-order diffraction-beam of a 

 monochromatic wave very intensely represented in the primary wave- 

 mixture. Or one may hold the crystal and the collector still and vary 

 the wave-length, obtaining a peak wherever X is such that for some 

 integer value of n the equation (15) is satisfied; the curve of Fig. 16 

 was obtained in this way, using waves of negative electricity. 



The process of measuring wave-lengths of X-rays is usually con- 

 ducted by this method, using a crystal such as rocksalt for which the 

 density is very accurately known. For if we know the density of the 

 crystal we know how many atoms it contains in a given volume; and 

 if we know in addition how many atoms constitute the atom-group 

 which is repeated over and over again to form the crystal, we can 

 compute by simple division what is the volume of the unit cell, and 

 what therefore is the spacing from one atom-group to the next — the 

 edge of the elementary cube. If we then set up the crystal so as to get 

 reflections from the 100 face we know that the edge of the cube is the 

 quantity d which figures in the equation (15) ; and measuring then the 

 values of Q corresponding to several diffraction beams we can identify 

 the corresponding values of n, and so evaluate X. Diffraction of 

 X-rays by ruled gratings can now be called upon in confirmation — or in 

 correction, as certain recent data indicate. 



The determination of the number of atoms in the atom-group is the 

 delicate point of this computation ; and perhaps it is to be accounted a 

 piece of luck that with the first crystals used in the spectroscopy of 

 X-rays the guess was easy and was rightly mad€. The routine of 

 determining it is but a part of the general process of learning from 

 the diffraction as much as possible about the atom-group; and this 

 deserves an article to itself, or many. Two examples of this process 

 however are interesting, useful and very simple. 



