MEASUREMENT OF PERMEABILITY 279 



ments by Baily are to be noted.* Swinburne had pointed out that 

 if E wing's model, described in Chapter XV. gives a right idea of the 

 physical nature of the magnetisation of iron, then the hysteresis of 

 iron rotating in a constant magnetic field should show a falling off* 

 as saturation is approached. For if all the molecular magnets are 

 parallel to the imposed force they will simply rotate as the field 

 rotates relative to the iron, without dissipating energy. Baily 

 found that up to a value of B in the neighbourhood of 15,000 the 

 hysteresis increased, and for both soft iron and hard steel it is not 

 very different in a rotating field from its value in a field reversed 

 in the ordinary way. If the rotating field is further increased, 

 then hysteresis falls off and apparently would vanish for a value of 

 B about 21,000, though that value was not actually reached. 



Mechanical model to illustrate hysteresis. It may 

 be interesting to describe a mechanical model which, if it could be 

 constructed, would give relations between force P and displacement 

 d very nearly corresponding to the relations between magnetic 

 intensity H and magnetisation I. Let AB, Fig. 21 3, be a cylinder of 

 air with unit-area cross-section and of length 2D. Let CDEF be a 

 short cylinder moving within it with friction F. Let G be a piston 



FIG. 213. 



moving without friction in CDEF, but against springs SS attached 

 to CDEF. Let the piston-rod R pass out through an air-tight 

 hole at B, and let any desired force P be applied at the end of R. 

 Let us suppose that G has area practically equal to the unit-area 

 of cross-section of AB. Let us start with the piston in the centre 

 and with equal pressures of air p on the two sides. It is easily seen 

 that if the piston is displaced d to the right, then, as far as the air 

 pressure alone is concerned, R must be pulled with a force 



and the curve giving the relation between P a and d will be 

 somewhat as in Fig. 214. 



But the springs will also introduce a force P 2 , which may be 

 considered separately, on the supposition that the ends of A and B 

 are open. At first the spring force, and therefore P 2 , is pro- 

 portional to the displacement of G, and will continue so till that 

 displacement has come to a certain value d v at which the spring 

 * Phil. Trans., A, '.vol. 77 (1896), p. 715. 



