328 BELL SYSTEM TECHNICAL JOURNAL 



change-in-length for iron, nickel, and six permalloys in which the 

 percentages of nickel are those indicated beside the curves. The 

 term "permalloy," I recall, is applied to iron-nickel alloys containing 

 more than 30 per cent of nickel, of which the initial permeability is 

 remarkably high ; the heat-treatments which these alloys had under- 

 gone conferred that quality on them, the nickel had been treated in 

 quite and the iron in nearly the same way. The abscissa is intensity 

 of magnetization, for the reason aforesaid; consequently the curves 

 terminate when this reaches its saturation-value (not, however, 

 attained in the experiments on iron and nickel). 



These curves display the gradations from a steady lengthening 

 reminding the onlooker of the initial lengthening of iron (not, however, 

 followed by a contraction) to a steady contraction approaching the 

 scale of that which nickel displays. Intermediate there lies an alloy 

 which is influenced very little, indeed up to a high stage of magnetiza- 

 tion it suffers no perceptible change at all; and this is precisely the 

 alloy having the greatest permeability and the least hysteresis in the 

 entire series. Upon this correlation McKeehan founded his theory 

 of ferromagnetism. 



This series of curves reveals other curious facts; for instance the 

 extreme ineffectiveness of the first stages of magnetization in devel- 

 oping the strain — the 46 per cent-nickel alloy had expanded, by the 

 time it' was magnetized one third of the way to saturation, by less 

 than one one-hundredth as much as it was destined to expand in 

 acquiring the remaining two thirds of its final magnetization. This 

 is significant; and more significant yet is the point, that when the 

 magnetic field is applied to one of the permalloys containing less 

 than 80 per cent of nickel and subject to a length-increasing longi- 

 tudinal tension, the magnetostriction is much reduced — that is to say, 

 the mechanical tension seems to have effected of itself a large part of 

 that task of extension which else it would have been incumbent 

 upon the magnetization to perform. It effects a great deal more, of 

 course; the extension due to even a moderate load is vastly greater 

 than the extension which even the greatest of magnetic fields could 

 by itself ever cause; the point is, that the former extension seems 

 to include the latter. Furthermore, elongation by tension is found to 

 produce just as great and no greater an increase in the electrical 

 resistance of a permalloy wire than the much smaller maximum 

 elongation attending magnetization. Yet tension by itself does not 

 magnetize; hence the change which it produces inside the wire does 

 not entirely overlap the effect of magnetic field. It is also true, as one 

 would expect, that tension acting upon a permalloy containing more 



