DIMENSIONS OP RINGS AND RODS OP IRON AND OTHER METALS. 215 



It appears that the retraction undergone by the iron continues to increase with 

 increasing magnetising forces until its amount is nearly twice that of the greatest 

 elongation. But towards the end of the experiment there are indications that a limit 

 is nearly reached. 



The nickel curve is of the same character as that which I have already published, 

 but it goes further, and the retraction is finally as much as 113 ten-millionths of the 

 length. The beginning of the curve is slightly different from that given in my former 

 paper. This difference was found to be accounted for by the fact that the nickel 

 there referred to was permanently magnetised. 



The curve showing the results of the cobalt experiment is of the most remarkable 

 nature. No evidence of any change of length appears until the magnetising force 

 exceeds 30 or 40 units. Then the length of the rod begins to diminish, and continues 

 to diminish until the force reaches about 400. But beyond this point the rod 

 gradually becomes longer, and the retraction with a force of 800 units is only three- 

 fifths of its maximum amount. 



[In showing some of these experiments at the conversazione of the Boyal Society 

 on May 9th, 1888, I had the privilege of taking my current from the large secondary 

 battery used for lighting the building. By means of this it was found possible to 

 carry the field up to about 1350 units, a far higher degree of intensity, I believe, than 

 had ever been previously obtained for any experimental purpose without the use of 

 electro-magnets. In this field the cobalt rod was elongated, the extent of its elonga- 

 tion being about half that of its greatest retraction in weaker fields. The cobalt 

 curve in fig. 4 would therefore, if prolonged, cut the horizontal axis, probably at 

 about 1100 (at which point there would be neither retraction nor elongation), and 

 would continue on the upper side of the axis. Iron and cobalt, therefore, behave 

 oppositely. With continually increasing magnetising force iron is at first extended 

 and afterwards contracted, while cobalt is at first contracted and afterwards extended. 

 -May 14th, 1888.] 



It is difficult to imagine what can be the physical meaning of this effect. The 

 possibility suggested itself that cobalt might acquire a maximum magnetisation with 

 a magnetising force of about 400, further increase of the force resulting in a diminu- 

 tion of the magnetisation. The importance of establishing such a property in any 

 substance* induced me to make some experiments upon the magnetisation of cobalt 

 in strong fields, t but there was no indication of any such maximum as was looked for, 

 the magnetisation being still on the increase when the experiment was stopped with 

 a force exceeding 700. 



It would be interesting to ascertain whether the retractions of iron and nickel 



* See MAXWKLL'S ' Electricity,' vol. 2, 844. 



t In ROWLAND'S experiments (' Phil. Mag.,' vol. 48, 1874) the magnetising force was not carried 

 beyond 147 C.G.S. nnits. 



