550 
PROFESSOR J. A. EWING ON EXPERIMENTAL 
In the present case 33 and £ have the same direction. For any indefinitely small 
change of the field 
dE= — ^(£d33+33d£). 
When the change is cyclic, J£<i33 and |33c/£ are equal, and 
= -+-§• 
This is the energy expended in performing the cycle, per unit of volume of the 
substance magnetised, and it has a positive value for all actual cycles of magnetism in 
iron or steel in consequence of the hysteresis which exists in the relation of 3 to £• 
The energy so expended is a thing quite apart from the further dissipation which 
occurs when, on account of the changes of magnetism taking place not indefinitely 
slowly, currents are induced in the substance of the magnet and in neighbouring con¬ 
ductors, and also from that which occurs in consequence of the tliermomagnetic 
properties of the metal.* These latter sources of dissipation are functions of the time, 
and are absent when the magnetic change occurs indefinitely slowdy: their effects do 
not appear in the cyclic curves of 3 and £ when we leave a sufficient interval after 
each change of £ before taking a reading of 3- The former, however, is independent 
of the rate of magnetisation, the hysteresis to which it is due being a static pheno¬ 
menon; and if we accept the idea that the magnetic molecules of Weber, have their 
rotation opposed by a species of static friction, the dissipation of energy we are now 
considering is the work done in forcing them to rotate back and forth against this 
friction. 
§ 33. Numerical Values of J2k7£ for various Cycles .—Measurements have been 
made of the values of J33d£ in a number of the cycles of magnetisation in iron and 
steel which have been already described and figured. These, divided by 4v, give the 
corresponding values of J3>c££, or the energy expended, in c.g.s. units of work or 
ergs., in carrying one cubic centimetre of the material through the cycle in question. 
Ergs. 
(1) For the large cycle of fig. 4 (§ 20), produced by the double reversal of 
magnetisation in a piece of very soft annealed iron wire from 
33= 13,450, £=17-26, to 33= —13,190, £= — 17-26, and back, the 
value of pd£ is .. 9,300 
(2) For the small loop of the same figure, produced by removing and 
restoring the magnetising force of +17"26, the value of J2>d£ is. . 490 
(3) For the large cycle of fig. 5 (§ 21), produced by the double reversal of 
magnetisation in a less soft annealed iron wire, from 33 — 15,710, 
£=22-27, to 33= —15,550, £=—23"08, and back, the value of 
J3d£ is. 16,300 
* Sir W. Thomson, Pliil. Mag., vol. v., 1878, pp. 24, 25. 
