190 MAGNETISM 



The results for cobalt and nickel rods are also given. It will 

 be observed that the greatest lengthening in iron is about two in 

 a million, while the ultimate shortening approaches seven in a 

 million. The behaviour of cobalt is just the opposite to that of 

 iron, while nickel contracts always and is much more affected than 

 either of the other metals. In Rapports du Congres International 

 de Physique, vol. ii. p. 536, there is an account of the work done 

 on this subject. It has been found that there is an increase in 

 volume on magnetisation of the order of 1 in 10 6 for a field of 

 1000 in iron and steel, and an increase somewhat less in nickel. 



Stress and magnetisation. When a body either pre- 

 viously magnetised or under the action of a magnetising force is 

 subjected to stress, the magnetisation is affected by the stress. 

 We shall here only give a short account of two cases, referring the 

 reader for a fuller treatment of this very complicated subject to 

 Ewing's Magnetic Induction in Iron and other Metals, chap, ix.* 



Matteucci discovered that if a bar of magnetised iron "is 

 pulled lengthwise " its magnetisation was increased. Villari found 

 that the nature of this effect depends upon the degree of magnetisa- 

 tion. If the bar is weakly magnetised, then the effect of a pull is 

 to increase the magnetisation as Matteucci observed, but if the bar 

 is strongly magnetised the pull decreases the magnetisation. That 

 is, after a certain point in the magnetisation the effect is reversed. 

 This is known as the Villari reversal. Comparing this with the 

 effect of magnetisation on change of length, we note that there is 

 a reciprocity. Weak magnetisation lengthens a bar. Lengthening 

 a weakly magnetised bar increases the magnetisation. Strong 

 magnetisation shortens a bar. Lengthening a strongly magnetised 

 bar decreases, while shortening increases, the magnetisation. 



It may be noted here that with nickel and cobalt bars the 

 reciprocal effects also hold. Thus nickel shortens, whatever the 

 value of the magnetising force, and reciprocally a longitudinal pull 

 leads to diminution of magnetisation. Weakly magnetised cooalt 

 shortens, and a pull diminishes the weak magnetisation. Strongly 

 magnetised cobalt lengthens, and a pull increases the strong 

 magnetisation. These reciprocal effects are illustrations of recipro- 

 cities which are continually met with in physical phenomena. They 

 are dealt with in J. J. Thomson's Applications of Dynamics to 

 Physics and Chemistry. 



Lord Kelvin,f who was the first to investigate these effects of 

 stress with requisite exactness, pointed out that we may regard a 

 pull on a bar in a weak magnetic field as increasing its permeability, 

 and a push as decreasing its permeability. In a strong field the 

 change of permeability is reversed. 



* For earlier work see Wiedemann's Galvanism.ii*, epitomised in Encyc. Erit., 

 9th ed., Magnetism, p. 253. Later work and references will be found in Ewing 



(loc. cit.), who has himself largely contributed to tho exact study of the subject, 

 f Math, and Phys. Papers, vol. ii. 



p. 332. 



