[ 51 ] 



VIII. On the Molecular Changes that accompany the Magnetiza- 

 tion of Iron, Nickel, and Cobalt. By W. F. Barrett, Professor 

 of Physics in the Royal College of Science, Dublin*. 



THE magnetization of iron is accompanied by certain mole- 

 cular changes in tbe metal which are well known to phy- 

 sicists. The object of this paper is an endeavour to extend our 

 knowledge of some of these changes. Further inquiry in this di- 

 rection seems to be needed as presenting one avenue of approach 

 to a better insight of what may be termed the "molecular archi- 

 tecture " of a magnet. 



The wonderful transition of iron from an ordinary to a magne- 

 tized condition makes no alteration in the appearance, the tem- 

 perature f, the weight, or the total bulk of the iron; but it is 

 associated with the changes alluded to, which are briefly as 

 follows : — 



1st. The act of magnetization causes a slight increase in the 

 length } and a corresponding diminution of the breadth of an iron 

 bar — a fact discovered by Mr. Joule in 1842, confirming the 

 previous observations of MM. Gay-Lussac and Wertheim, that 

 there was no alteration in the total volume of the iron. This 

 elongation, however, does not occur when the iron is submitted 

 to a definite longitudinal strain ; and when the strain is still 

 greater, the iron invariably shortens when magnetized. 



2nd. A sound is emitted by the iron on magnetization and 

 again on demagnetization. This was revealed by Mr. Page in 

 1837, and studied by many physicists subsequently. In iron 

 wires the sound or clink seems composed of two distinct noises, 

 one of which intensifies by a moderate strain, but is destroyed 

 and the whole sound enfeebled by a still higher strain. 



3rd. M. Wiedemann has proved that an iron wire hung in the 

 centre of a helix and twisted is more or less untwisted when a 

 current traverses the helix and magnetizes the wire. M. Mat- 

 teucci has shown that twisting a magnet lessens its force, but 

 stretching a magnet slightly adds to its power % ; and according 

 to M. Guillemin, a strip of iron bent by its own weight is partly 

 straightened by magnetization. 



4th. The conduction of heat in magnetized iron is greater 

 across than along the magnetic axis — a fact discovered by Dr. 

 Maggi, and enlarged by Sir W. Thomson, who has shown that 



* Communicated by the Author, having been read before the British 

 Association at Bradford, September 1873. 



t Nevertheless it is stated that repeated magnetization and demagneti- 

 zation raises tbe temperature of an iron bar. 



X A refined and capital method of exhibiting this is described by Mr. 

 Gore in the ' Proceedings of the Royal Society ' for January 28, 1869, 

 vol. xvii. p. 265. 



E2 



