under the Operation of Feeble Magnetic Forces. 233 



" magnetization-function/' introduced by Dr. Stoletow, rather 

 suggests a different conclusion. 



The curves given by Stoletow and by Rowland in their 

 celebrated researches are not exactly magnetization-curves in 

 the more natural sense ; that is to say, they do not exhibit 

 fully the behaviour of a piece of iron when subjected to 

 a given sequence of magnetic forces. But a number of such 

 curves have been drawn by Ewing which afford all necessary 

 general information. Among these we may especially dis- 

 tinguish the course followed by the iron in passing from 

 strong positive to strong negative magnetization and vice 

 versa, and that by which iron starting from a neutral con- 

 dition first acquires magnetization under the action of a force 

 constantly increasing. 



Attention is called by Ewing to the loops which are formed 

 when ihe forces are carried round a (not very small) cycle of 

 any kind. " Every loop in the diagram shows that when we 

 reverse the change of magnetizing force from increment to 

 decrement, or vice versa, the magnetism begins to change 

 very gradually relatively to the change of ,jj (the force), no 

 matter how fast it may have been changing in the opposite 

 direction before. So much is this the case that the curves, 

 when drawn to a scale such as that of the figure, appear in all 

 cases to start off tangent to the line parallel to the axis on 

 which ,jr) is measured whenever the change of .0 is reversed in 



The question here raised as to the direction of the curve, 

 after the force has passed a maximum or minimum, is one of 

 great importance. If it were strictly true that this direction 

 were parallel to the axis, it would follow generally that iron in 

 any condition of magnetization would be uninfluenced by 

 small periodic variations of magnetic force ; for example, that 

 in many telephone experiments iron would show no magnetic 

 properties. The experiments already detailed prove that when 

 the w T hole force and magnetization are small (they were not 

 actually evanescent) very sensible proportional changes of 

 magnetization accompany small changes of force, the ratio 

 being such as to give a permeability not much inferior to 

 100. ."Nothing is easier than to show that this conclusion is 

 not limited to very small mean forces and magnetizations. 

 As regards the latter, we may apply and remove a force 

 (say) of 5H. By this operation the iron is left in a different 

 magnetic condition, and the zero-reading of the magnetometer 

 is altered, probably to the extent of driving the spot of light 

 off the scale. But if we bring the needle back with the aid of 

 external magnets, we can examine, as before, the effect of 



