I 



produced in Iron Wires by Magnetization. 349 



four times as long as the shortest rod used, and the ratio of 

 the diameter to the length of the rod was only 1 to 32. 

 Besides the uncertainties introduced by such a poor arrange- 

 ment for a uniform field, great errors would be produced by 

 the strong poles created at the ends of the rod. 



Object of Experiment. 

 A careful study of the results obtained for iron by the 

 different observers will show that although the general appear- 

 ance of their curves is very similar, yet their absolute values 

 vary widely, two specimens often having maximum elonga- 

 tions that differ 20 or 30 per cent. There is, unfortunately, 

 no way of comparing these results, for it has been the 

 custom to use as coordinates the change of length and the 

 intensity of the external field. For a given apparatus this 

 intensity depends only on the current used, and not at all on 

 the specimen to be examined. The elongation is dependent 

 upon the intensity of magnetization in the wire, and this is 

 the quantity that varies with the specimen employed. For 

 that reason the relation should be found between these 

 two quantities. 



It is of course essential to have the rod uniformly magne- 

 tized throughout its length; that is, the field should be uniform 

 and there should be no free poles. These conditions may be 

 best obtained in one of two ways : either by having the metal 

 in the shape of a ring and observing the change in diameter 

 of the ring when magnetized by a solenoid wound upon 

 it, or by using long wires of the metal. In the second case 

 only the middle part of the wire should be observed, and the 

 solenoid used to magnetize it should be considerably longer 

 than the portion of the- wire experimented upon. In this 

 investigation the latter method was chosen, as rings are less 

 convenient and also because it was desirable to observe the 

 effects of tension in the metal. 



When a rod of iron is magnetized, the change in length 

 observed is due to several causes — three at least; and to 

 obtain a correct idea of the phenomena these causes and their 

 effects should be separated. There is, first, the direct action 

 of the magnetism, and this may possibly be due to the orient- 

 ing of the magnetized particles of the rod. Secondly, there 

 are indirect actions of the magnetism which tend to change 

 the length of the rod. These indirect actions are the me- 

 chanical stresses created in the rod by the magnetism. The 

 first of these mechanical stresses is the tractive force of the 

 magnet, and is measured by B 2 /87r. That this force exists, 

 tending always to contract the rod, is seen from the fact that 



