xvin THE CLASSIFICATION OF THE ELEMENTS 487 



the agreement between the reduced atomic weights and 

 the atomic numbers was closer and the number of vacant 

 spaces was reduced to thirty-seven. 



Experimental determination of atomic numbers. 

 Atomic numbers have acquired great importance in recent 

 years, owing to the discovery that several properties of the 

 atom, which are not directly related to the irregularly-distri- 

 buted atomic weights, are related in a very simple way to 

 integral ATOMIC NUMBERS closely analogous with those put 

 forward by Nevvlands in 1878, but showing seven places less 

 between hydrogen and gold. Amongst these properties 

 are 



(1) The scattering of a-particles by gases. 



(2) The absorption of X-rays by different elements. 



(3) The high-frequency spectra of the elements. 



Of these methods, the last is the simplest to explain, and 

 appears to give the most exact results : it may therefore be 

 described as a type of the new methods of determining 

 atomic numbers experimentally. 



Moseley (1914) on the high-frequency spectra of the 

 elements. When X-rays fall upon a crystal, a diffraction- 

 spectrum is produced, in just the same way as when ordinary 

 light falls upon a ruled diffraction-grating. The two phe- 

 nomena are very similar to one another, but in the case of 

 X-rays the linear dimensions are about 10,000 times smaller. 

 Thus while the yellow light of sodium consists of two radia- 

 tions of wave length 5890 and 5896 x io~ 8 cm., the radiation 

 from a rhodium target in an X-ray bulb gives radiations of 

 wave-length 0*54 and o'6i x io~ 8 cm., the latter in its turn 

 being a doublet of wave-lengths 0*614 and 0-619 x io~ 8 cm. 

 (Bragg, Royal Institution Lecture, June 5, 1914). The 

 nature of the X-ray spectrum depends on the nature of the 

 elements composing the target (Moseley, Phil. Mag., Dec. 

 1913, 26, 1024-1034; April, 1914, 27, 703-713).. The 



