114 BELL SYSTEM TECHNICAL JOURNAL 



servations on paramagnetic substances, except for two ions which 

 yield values not reducible to agreement with the new theory. 8 The 

 new method of calculating magnetic moments thus leads to values 

 which confirm the contemporary atom-model. It would not be desir- 

 able to dismiss the old method and the old theory too hastily, 

 considering that they lead to values which are claimed to be integer 

 multiples of an apparently fundamental constant; but this constant 

 has proved so intractable to theory that it would be gratifying to be 

 able to discard it. 



The arrangement of atoms in two samples of Heusler alloys was 

 investigated with the X-ray method by J. F. T. Young at Toronto. 

 These alloys are mixtures of the metals, copper, manganese, and 

 aluminium in certain proportions; they are strongly ferromagnetic 

 while the component metals are not ferromagnetic at all. Of the 

 two samples, one had a much higher permeability than the other; 

 the atoms of the former sample were arranged in a body-centered 

 cubic lattice, with no trace of the characteristic lattices of the com- 

 ponent metals. The atoms of the latter sample were arranged in a 

 face-centred-cubic lattice. Thus these alloys furnish an additional 

 instance of the frequent, though not by any means universal, correla- 

 tion between body-centred-cubic lattice and strong ferromagnetism. 

 L. W. McKeehan of the Western Electric used the same method to 

 investigate palladium containing great quantities of occluded hydro- 

 gen. The space-lattice of the hydrogen-free metal was distended by 

 a certain fixed percentage by saturating it with hydrogen; and it 

 appeared that when the palladium contained a lesser quantity of 

 hydrogen than the maximum or saturation amount, some parts of it 

 were quite saturated and others contained no hydrogen at all, instead 

 of the whole lattice being equally enlarged; it is probable that the 

 individual crystals of the metal are saturated one by one as the hydro- 

 gen creeps in. 



8 The value for beta-iron as quoted by Gerlach does not agree with the theory. 

 As for the values assigned by Weiss to the three ferromagnetic metals iron, nickel 

 and cobalt, obtained from direct measurements of the saturation-intensity at the 

 temperature of boiling hydrogen, the first two do not agree with the theory, the last 

 agrees very well (assuming the electron to be in a one-quantum orbit). Of course, 

 it is likely enough that the theory should not be applied to ferromagnetics. It 

 seems fitting to quote a remark of Andrade about theories of magnetism in general 

 "... the substances selected for verification of theories are of a very limited 

 class, called of normal behavior rather because they agree with the theories than 

 because they represent a numerical majority." 



REFERENCES 

 D. M. Bose and S. K. Ghosh: Nature 111, pp. 463-464; 1923. 



A. H. Compton: Physical Review, 2d ser., 21, pp. 483-502 and 715; 1923. (X-ray 

 scattering.) 



