346 



SCIENCE 



[N. S. Vol. LIV. No. 1398 



roeyanide is diamagnetic. No cataclysm of 

 the atom lias occurred in. these chemical 

 changes. On the other hand if we turn to 

 magnetostriction for help in interpreting the 

 work of the Comptons and explain magneto- 

 striction as due to the orientation of the 

 elementary magnets it would appear that 

 their negative results may be due to the fact 

 that they worked at only one field strength, 

 whose value is not given in their papers, and 

 at that field strength the orientation had not 

 proceeded far enough to give measurable 

 effects. For instance, in the case of an iron 

 rod, as the magnetic field strength is increased 

 from zero upwards, the rod first elongates 

 and then shortens, becoming shorter at high 

 field strengths than in its virgin state. At 

 that field strength where the length once 

 more becomes equal to the original length, 

 at that point one would expect negative results 

 in the work of the Comptons. In iron this 

 field strength is about at the point where 

 saturation occurs. From the magnetostric- 

 tive viewpoint the Comptons should find 

 maximum effects at those field strengths 

 where maximum changes in length occur. The 

 Comptons used magnetite which is quite dif- 

 ferent from iron in the manner in which its 

 length changes in a magnetic field. Yamada 

 found that at several hundred Gauss field 

 strength, it was still increasing its length and 

 no maximum attained. The question may 

 legitimately be raised as to whether the orien- 

 tation of the elementary magnets had been 

 carried on sufficiently to give the Comptons 

 the effects they were looking for. A further 

 study of the Joule effect in magnetite is 

 being started to throw more light on this 

 subject. 



2. Would negative electrons revolving in 

 orbits or negative electrons rotating, a la 

 Parson, alone suffice as a picture of the ele- 

 mentary magnet? The theories we have so 

 far discussed seem to convey the idea that 

 they would. Why not attribute magnetic 

 phenomena to a positive nucleus spinning on 

 its axis? Barnett's work indicates the nega- 

 tive charge as the portion of the elementary 

 magnet which is in motion. This does not. 



however, debar the positive nucleus from con- 

 tributing some part of that property which 

 we know as susceptibility and which we have 

 been discussing. In other words induction 

 may be a part of the property of the nucleus 

 and we shift at least a part of that property 

 from the mass to the elementary magnet.^* 

 What is it that gives magnetic characteris- 

 tics? These are questions which our general 

 subject of susceptibility raises. There are a 

 number of items which, as it seems, bear upon 

 these queries. Maurain^" deposited thin films 

 of iron and nickel and found he had to have 

 a certain thickness of film before he obtained 

 definite magnetic ijroperties. For iron this 

 was 8.3X10-^ cm. and for nickel, 20X10"^ 

 cm. Wilson-^ in measuring the magnetic 

 fields in a rotating iron cylinder arrives at 

 the size of a magnetic particle as 10 X 10"^ cm. 

 which checks fairly well. Hull,-'' working on 

 the X-ray analysis of iron and nickel finds 

 the distance of 2.47 and 2.50 X 10'* cm. 

 respectively as the distance between nearest 

 atoms. These values seem to be commensu- 

 rate. As already pointed out the spacing of 

 the atoms seems to play a very important 

 part in magnetic phenomena. Hull calls 

 attention to the fact that' it might be antici- 

 pated that ferromagnetic substances would 

 have the same crystal structure. This is not 

 true for iron and nickel are different accord- 

 ing to Hull's observations. It is evident that 

 ferromagnetism does not depend upon any 

 particular arrangement of atoms but more 

 probably upon distance between atoms which 

 would explain the fact that this property is 

 lost when the temperature is increased beyond 

 a definite value. A center cubic arrange- 

 ment may be more favorable to ferromagnet- 

 ism, but is not a principle or essential factor. 

 Arnold and Hicks'* state: 



The elements giving iron higt permeability and 



^iPhys. nev., abstract, Feb., 1911. Fhys. Bev., 

 34, 40, 1912. 



25 Maurain, Jour, de Fhys., 1, 151, 1902. 

 20 Wilson, Froc. Boy. Soo., 69, 435, 1902. 

 27 Hull, Fhys. Bev., 14, 540, 1919. 

 2s Arnold and Hieks, Nature, Apr. 17, 1902. 



