498 



Sir J. A. Ewing on a Nei 



j 



which admit of adjustment for direction, height, and nearness 

 to the centre. When equally strong fixed magnets are 

 chosen and the clearances are made equal, the model always 

 shows some magnetic stability, for the distribution of 

 magnetism in the wires is slightly altered by induction, with 

 the effect of bringing the unlike poles at one end nearer 

 together, and driving the like poles at the other end further 

 apart. When the pivoted magnet is made to. turn round 

 from one pair of fixed magnets to another it exhibits this 

 stability in each successive position. 



Consider next how the parts which make up the model 

 may be thought of as having equivalents in the structure of' 

 the atom. It would seem that the electrons of each atom, 

 which possess magnetic quality through rotating in orbits, 

 or perhaps through being rings, furnish two types of magnetic 

 element, a turning element near the nucleus and other elements 

 at a greater distance from it, which are held fixed — more or 

 less completely — in consequence of the forces between the 

 atom and its ' neighbours in the crystal. A. W. Hull 

 has shown by his X-ray analysis that the space-lattice of 

 iron is the centred cube, and has suggested as a probable 

 structure for the iron atom one in which two of the 

 electrons form a doublet, with the nucleus in the middle, 

 and the remaining twenty-four electrons are grouped in 

 octets, further out along the diagonals of a cube *. On 

 the basis of this suggestion, we may think of these octets 

 as supplying the four pairs of magnetic elements which 

 correspond to the fixed magnets A, B, 0, D, etc., of the 

 model, with axes that lie along the cube diagonals of the 

 atom (presumably coinciding with the trigonal axes of the 

 crystal). They point radially inwards, and in any one atom 

 all their inner " poles " are of the same name. In the inner 

 space a group, possibly the doublet, which may he taken to 

 constitute the Weber element and corresponds to W of the 

 model, turns in response to an impressed field. So long- 

 as there is no impressed field the Weber element may be 

 in any one of eight positions of stability, pointing towards 

 one of the fixed elements. Its small reversible deflexion 

 before breaking away constitutes the quasi-elastic stage 

 that is found when one begins to magnetize, or to change 

 from an increasing to a diminishing field, or vice versa, lis 

 irreversible passage from one position of stability to another 

 is what makes hysteresis. In a substance such as soft iron 

 there is no more than a narrow range of stability because 

 the clearance is small ; and the control which gives stability 

 * " The Crystal Structure of Iron," Phys. Kev. ix. p. 85 (1917). 



