1886.] Electromagnets and the Magnetisation of Iron, 487 



obtained. The diameter of a ring was found to be increased by a 

 comparatively small magnetising current, and diminished by a strong 

 one. Now the retraction in question does not begin to occur until 

 after the stage of magnetisation, loosely called the " saturation point," 

 has been passed, when, according to the common belief, the magneti- 

 sation of the iron has practically reached a limit, and is not sensibly 

 affected by any further increase of the magnetic force; and hence 

 arises a difficulty in accounting for the phenomenon. The most 

 obvious method of explaining the retraction is to assume that under 

 the influence of increasing currents the magnetic attraction of the 

 particles of the iron towards one another is increased, and thus the 

 rod becomes compressed. But this cannot be the case if the magnetic 

 condition of the iron has become constant and independent of the 

 magnetising current. And a similar objection will apply to any 

 hypothesis which assumes (as I think all must) that some pro- 

 perty of the iron dependent upon its magnetic condition varies in a 

 sensible degree with the magnetising force after the "saturation 

 point " has been passed. 



I was led by considerations of this nature to the belief that it would 

 be desirable to make some experiments with the view of ascertaining 

 whether the lifting power and general magnetic condition of an iron 

 rod are as nearly uniform under strong magnetising forces as is com- 

 monly supposed. Two pieces of apparatus were therefore prepared. 

 The first consisted of a rod of iron hooked at each end and divided 

 transversely in the middle, together with a long solenoid, inside which 

 the divided rod could be placed. The second was an iron ring cut into 

 two equal parts each of which was encircled with a coil of insulated 

 copper wire. In both cases the construction was such that an intense 

 magnetic force could be produced with comparatively small battery 

 power. The divided ring could be used either as a semicircular elec- 

 tromagnet with a semicircular armature, or, if the current were passed 

 through both coils, as two semicircular electromagnets. 



Merely to test the hypothesis of Joule and Rowland, two or three 

 determinations were made of the weight which could be sustained 

 when the current was caused to circulate around one only of the semi- 

 circles, the other being used as an armature. With a current of 4*3 

 amperes the weight supported was 13,100 grams per square centi- 

 metre of surface ; with a current of 6' 2 amperes the weight supported 

 was 14,200 grams per square centimetre. In the latter case, there- 

 fore, the lifting power exceeded that which both Joule and Rowland 

 considered the greatest that could be imparted to a magnet by an 

 infinite current. Had it been worth while to incur the risk of 

 injury to the insulation of the coil, there is no doubt whatever that 

 by applying stronger currents the lifting power might have been 

 carried still further — for there was no indication that a limit was 



