ELECTRODYNAMIC QUALITIES OF METALS. 
67 
ten “ 011 s ” and “ offs ” were not sufficient to sliake out the effect of the powerful 
magnetizing force of the current, and allow the barrel to take the magnetization due 
to the vertical component of the terrestrial magnetic force. Indeed, they only 
diminished the residual magnetism in the curves both from left to right and from 
right to left by about 100 divisions, out of total residuals from 230 to 440. It is 
interesting to see by contrasting the right hand ends of the curves with the left hand 
ends, how much stronger the residual magnetism is when helped, than when opposed 
by the vertical magnetic force of the earth. 
220. Comparison between the results of the effects of longitudinal and transverse 
pull shows that an aeolotropic property of different magnetic inductive susceptibility in 
different directions is temporarily developed in soft iron by aeolotropic stress (that is 
to say by stress not consisting of positive or negative pressure equal in all directions). 
The results show that with low magnetizing forces, negative pressure perpendicular to 
one set of parallel planes of soft iron produces an augmentation of magnetic suscepti¬ 
bility in the direction of the pressure and diminution of the susceptibility in all 
directions at right angles to it. The effects of positive pressure have not yet been 
tested experimentally, but it is certain that they will be opposite to the effects of 
negative pressure. Independently of experiment, we may also infer that the effects 
of infinitely small positive pressure perpendicular to one set of parallel planes, and 
infinitely small negative pressure of equal amount perpendicular to a set of parallel 
planes at right augles to them, must be equal and opposite in the directions of these 
pressures, and therefore must leave the magnetic susceptibility unaltered in the 
directions inclined 45° to them. This is exactly the stress which is experienced in 
a twisted wire of circular section ; the amount of the stress being zero in the axis of 
the wire, and being elsewhere in simple proportion to distance from the axis. The 
directions of the positive and negative pressures at any point of the substance are 
two lines in the tangent plane of the cylindric surface through it, co-axal with the 
boundary of the wire, and inclined at 45° to the normal plane section. Hence, 
when the torsion is infinitely small, the magnetic susceptibility of the wire in the 
direction of its length must be unaltered, and if finite amounts of torsion produce 
any change in the magnetic susceptibility, the amount of this change must ulti¬ 
mately (for very small torsions) vary inversely as the square of the amount of torsion, 
as we see by remarking that whatever effect is produced must be independent of 
the direction of the torsion, there being nothing of helicoidal quality in longitudinal 
magnetization. 
221. In Wiedemann’s ‘ Galvanismus ’ (vol. II., §§ 476-498) an abstract is given of 
researches in this subject by Matteucci, Wertheim and Edmund Becquerel.* One 
main result of all these investigations is that torsion in either direction diminishes the 
temporary inductive longitudinal magnetization of soft iron. 
222. Nearly two years ago I instituted a series of experiments on the subject 
* Matteucci, ‘ Comptes Rendus,’ t. xxiv., 1847 ; Weetheim ‘ Comptes Rendus,’ t. xxxv., p. 702, 1852. 
K 2 
