52 Prof. W. F. Barrett on the Molecular Changes that 



its precise analogue is to be found in the conduction of electri- 

 city in magnetized iron and nickel. 



5th. A bar of wrought iron is more easily magnetized in the 

 direction of its fibre ; and steel once magnetized in a given di- 

 rection and then demagnetized, is more readily magnetized in 

 its first direction than in any other — a fact first pointed out by 

 M. Marianni, and recently again observed by M. Jamin. 



Lastly, it is well known that mechanical blows aid the assump- 

 tion of magnetic power in steel, but tend to lessen and can even 

 destroy it when assumed ; and the same also is true of heat, 

 which no doubt acts in a similar way, viz. by lessening the co- 

 hesion of the particles of steel. 



All these facts may be embraced under the assumption first 

 made by M. De la Rive, that magnetization is expressed by a 

 definite movement, or a marshalling of the molecules of iron — 

 the placing, as Dr. Tyndall puts it, of their longest dimensions end 

 to end. Now iron is not the only magnetic body. Nickel and 

 cobalt share the magnetic properties of iron to a very high 

 degree; and to a much less extent the metals chromium and 

 manganese are also magnetic. If, then, magnetization is an act 

 associated with an altered structure of iron, we should expect 

 to find a certain correspondence to iron in the properties of the 

 other magnetic metals. That this is the case, is I think shown 

 by the paper which has preceded this, " On the Relationship of 

 the Magnetic Metals" * 



I was anxious to try further whether the molecular disturb- 

 ances found on magnetizing iron were also exhibited by nickel 

 and cobalt. One would of course expect to find analogous 

 changes in these bodies; but I am unaware that they have 

 hitherto been examined. 



Messrs. Johnson and Matthey very kindly lent me an extremely 

 fine bar of nickel and one of cobalt. Both bars are cylindrical, 

 a little over 9 inches long and 1 inch in diameter. Though as 

 pure specimens as they can be rendered commercially, the co- 

 balt I find contains a very appreciable amount of iron; the 

 removal of which body, as chemists well know, is a matter of 

 the utmost difficulty. 



The relative magnetic powers of these two bars deserves a 

 moment's consideration. Nickel is invariably ranked above 

 cobalt in the scale of magnetic metals, Faraday and others 

 placing it next to soft iron. But the bar of nickel I have used, 

 when submitted to the same magnetizing current as the cobalt 

 bar, exhibits far less portative force than the cobalt. It is re- 

 markable that the iron impurity contained in the cobalt is able to 

 produce so powerful an influence. The nickel, like other speci- 

 * Phil. Mag. Dec. 1873. 



