RECENT ADVANCES IN SCIENCE 369 



only three sets of planes. The probability of any reflecting 

 position is rendered finite by the imperfect parallelism of the 

 bundle of rays and the thermal and other irregularities of the 

 atoms which give a finite reflecting range. 



A definite fraction of the energy will, therefore, be reflected 

 along each cone depending upon the number of co-operating sets 

 of planes with the same spacing, the position of the nuclei 

 with regard to the reflecting plane and on the position of the 

 scattering electrons within the atom. In elements of high 

 atomic weight, the majority of the electrons will be close to the 

 nucleus so that the position of the nuclei will be the pre- 

 dominant factor. In Hght atoms, the position of the valency 

 electrons will largely determine the relative intensity of the lines, 

 whose position and number alone can then be relied upon to 

 determine the crystal structure. 



The actual intensities of the reflections are very small even 

 under the best of conditions, and necessitate long exposures 

 to intense radiation. Exposures cited by Hull range from 20 

 to 300 milliampere-hours at constant potential. Every wave- 

 length in the beam is scattered into a different system of cones 

 which in a continuous spectrum would produce a general blacken- 

 ing of the photographic film. Use is made of the characteristic 

 radiation of the anticathode metal, whose properties have been 

 determined by the earlier crystal methods. The system of 

 lines due to these particular wave-lengths stands out from the 

 general effect and can be separately determined. 



Hull obtains a still greater homogeneity by making use of 

 the sudden increase in the absorption, by a metal, of rays whose 

 frequency exceeds that of its kr line. By using a thin filter of 

 a metal whose kr wave-length is intermediate between the k^ 

 doublet and the k^ of the anticathode, the former, which 

 fortunately is the longest and most intense of the series, is 

 practically isolated. The optimum potential for exciting the 

 tube, the material and the thickness of the filter may be cal- 

 culated for any particular combination of anticathode, filter 

 and crystal powder from previous work on intensity, wave- 

 length and absorption. 



The most generally useful combination seems to be a molyb- 

 denum Coolidge tube at 30,000 volts with a filter of 0*35 mm. 

 of powdered zircon. For metals or other heavy scatterers, 

 a tungsten tube run at 100,000 volts and an ytterbium filter 

 of 0-15 mm. is most suitable. Debye and other German 

 investigators obtained good results with unfiltered copper 

 radiation. 



The tube should be designed to give as great an output as 

 possible continuously, with as short a distance between target 

 and powder as is consistent with safety and proper definition 



