October M, 1921] 



SCIENCE 



347 



low coercive force are those which cause it to 

 crystallize in large crystals. 



Aston-" also says: 



It seems true, other things being equal, that 

 the heat treatment which will give to pure iron a 

 coarseness of crystallization, and above all a uni- 

 formity and regularity of such structure will be 

 accompanied by a low coercive force, and the 

 effect of heat treatment is augmented by the ad- 

 dition of silicon or analogous elements, as arsenic 

 or tin, all of which increase the coarseness of crys- 

 tallization of the material. 



It seems to be generally conceded that 

 manganese is the essential constituent in the 

 Heusler alloy. We don't know the mag- 

 netic properties of manganese any too well but 

 its being associated so closely with iron, 

 cobalt and nickel in the periodic system indi- 

 cates the possibility of its possessing latent 

 magnetism which under favorable conditions 

 becomes active. Ross suggests that the pres- 

 ence of the other metals beside manganese 

 exerts a helpful influence in making the 

 manganese elementary magnets farther apart 

 and so increasing its magnetic activity by the 

 removal of the intense intermolecular forces 

 which are supposed to act in the metal man- 

 ganese. This point of view is further cor- 

 roborated by the fact that the susceptibility 

 of copper containing minute quantities of 

 iron is far greater than that calculated from 

 the amount of iron present. One of the most 

 thorough researches undertaken on a phase of 

 this subject was by Perrier and Onnes^" who 

 stiidied the susceptibility of a liquid mixture 

 of oxygen and nitrogen and the influence of 

 the mutual distt.-'c; if the molecules of oxy- 

 gen upon paramagi. ism. In this work the 

 oxygen at the lo cemperature is paramag- 

 netic and inasmuch as the nitrogen did not 

 enter into chemical combination with the 

 oxygen it was possible to separate the oxygen 

 molecules as much as desired by making the 

 percentage of nitrogen larger. Their general 

 results may be summed up by saying: 



29 Aston, Trans. Faraday Soc, Vol. 9, July, 1913. 



30 Perrier and K. Onnes, Proc. Boy. Acad. Am- 

 sterdam, 16, 901, 1914. 



The specific magnetization coefScient of oxygen 

 becomes considerably greater, in proportion as the 

 concentration diminishes. 



There is much to be investigated along this 

 line. 



This discussion leads inevitably to the 

 question as to where we shall locate the origin 

 of the property of susceptibility? Will a 

 group of electronic orbits account for mag- 

 netic phenomena or must we have added to 

 their effect that which arises from the positive 

 nucleus ? Could we have a group of small 

 coils to replace the group of little magnets 

 with which Ewing once worked and obtain 

 results such as he did? I have been working 

 on this problem the past two years and so far 

 have not been able to realize experimentally 

 what Ewing did. It must be emphasized 

 again that Ewing in his classical experiments 

 worked with elementary magnets in which 

 each elementary magnet itself showed all the 

 properties which the group did. An attempt 

 to explain the magnetostrictive effects on a 

 molecular hypothesis makes it look very much 

 as though one needed another factor to add 

 to the electronic orbit to explain that particu- 

 lar field of magnetic phenomena. 



Space forbids to give all the reasons why 

 one is led to think of the atom as the seat 

 of the phenomena we meet with in magnet- 

 ism, or that the atom is the elementary mag- 

 net. The classical argument against this 

 point of view is that the iron atom is fer- 

 romagnetic, ferrous sulphate is paramagnetic 

 and potassium ferrocyanide is diamagnetic. 

 Iron is a constituent of all three. Why this 

 wide divergence of property? From preced- 

 ing arguments it would appear that inter- 

 stitial relations might answer the question. 

 Oxley^'- put it another way by saying, in 

 speaking of diamagnetism, that the molecular 

 structure is distorted by the near approach of 

 the other molecular structures so that the 

 self induction of the electi'onic orbits is af- 

 fected. The magnetic theories of Langevin 

 and Weiss are essentially atomic theories and 

 that the susceptibilities of the elements is 



siOxley, PMl. Trans., 214 {A), 109, 1914.— 215, 

 A, 79, 1915. 



