42 



Prof. J. A. Ewing. On Effects of [May 25, 



weight was put on and taken off. By making several steps, instead 

 of only one, in the application and removal of the load, the existence 

 of hysteresis may easily be demonstrated by this method • but I have 

 preferred the direct magnetometric method, which has the immense 

 advantage of exhibiting the actual magnetic state of the stretched 

 wire at any time. 



Each wire was hung vertically with its upper end on a level with a 

 mirror magnetometer. It was then annealed by heating to bright 

 redness with a spirit-lamp, and after it had become cool, weights were 

 progressively applied. 



During the earliest part of the first loading certain very interesting 

 apparently anomalous effects occur, which will be described in the 

 detailed account. Apart from these, which are easily distinguished, 

 the following is the normal action : — 



If to the annealed wire any load not exceeding the elastic limit is 

 successively applied and removed (without shock), its application 

 causes a decrease and its removal an increase of magnetisation. The 

 "on" and " off " curves of stress and magnetism are widely different, 

 and afford an excellent instance of hysteresis. 



Next, let the wire be stretched beyond its limit of elasticity. The 

 stretching is accompanied by a decrease of magnetisation, which 

 continues so long as the wire keeps "running down." When the 

 load is removed it is found that a great diminution of magnetisation 

 has taken place ; but besides this, the wire has undergone a very 

 remarkable change with respect to its subsequent behaviour under 

 stress. 



For let weights now be gradually applied : they cause at first an 

 increase* of magnetisation, but this passes a maximum and falls off 

 slightly as that value of the load is approached which previously was 

 applied to produce the permanent set. Let the load then be gradually 

 removed : the magnetisation at first increases, passes a maximum (at 

 a considerably lower value of the load than that which gave the 

 maximum during application), and finally diminishes rapidly to its 

 previous value with no load. 



These effects will be clearly seen by reference to fig. 1, which shows 

 the results of a small part of one set of observations. The ordinates 

 are proportional to the total magnetisation, and the abscissas are 

 the loads in kilogrammes. In this case the wire was of moderately 

 soft iron, 079 millim. in diameter, and had a well-defined limit of 

 elasticity at about 10 kilos. 



The upper part of the diagram shows the effect of gradually apply- 



* This agrees with. Sir W. Thomson's observation that with low magnetising 

 forces the effect of "on" is to increase, and "off" to diminish magnetism. The 

 description of the wire examined by him shows that it was in fact in the state de- 

 scribed in the text. (See " Phil. Trans.," 1879, p. 56.) 



