242 BELL SYSTEM TECHNICAL JOURNAL 



One naturally asks about the size and the magnetic moment of the 

 domains. It is useless to remember how the latter was determined 

 for paramagnetic bodies from the features of their I-vs-II curves, 

 since the theory which made that possible is not applicable here. 

 Moreover, the super-atoms share with ordinary atoms the quality of 

 being invisible: no feature of the ordinary surface of a metal indicates 

 them, and no technique of etching the surface seems able to delineate 

 them. (It must be said, however, that ferromagnetic powders 

 sprinkled over ferromagnetic metals may distribute themselves in 

 remarkable picturesque patterns, and perhaps these sometimes simu- 

 late the pattern of the underlying domains.*) But fortunately the 

 super-atoms are not inaudible; at least, it is not a very extravagant 

 statement to say that they can be heard. Let a girdle of wire around 

 a rod of some ferromagnetic substance be connected through an 

 amplifier with a microphone, and let a gradually-increasing magnetic 

 field act lengthwise on the rod: the microphone will then emit a 

 machine-gun patter of sharp clicks (with suitable amplification it may 

 be very dramatic!) each of which corresponds to the sudden shift of 

 the magnetic moment or "big arrow" of a domain from one of its 

 possible orientations to another. Now if an electrical instead of an 

 acoustical device is attached to the girdle of wire, the magnitude of 

 the moment which thus re-orients itself at a single click may be 

 assessed. It turns out that the moments are of very various magni- 

 tudes; a mean may, however, be estimated, and this mean is some 10'^ 

 times as great as the moment of a single atom. Therefore the average 

 domain comprises a million billions of atoms, and must therefore be 

 about .002 cm in breadth ; but there is a wide range of sizes about the 

 average. As for the individual atoms of the ferromagnetic metals, 

 their moments may be derived from equating Nn to the values (ob- 

 tained by extrapolation from observations at various low temperatures, 

 to absolute zero) of that "saturation of saturations" defined above. 

 They are by no means out of the common. Iron and its congeners are 

 readily magnetizable, not because their atoms are extraordinarily 

 magnetic — which is not at all the case- — but because their atoms have 

 this curious propensity of cohering together in large groups, developed 

 to an extraordinary degree. 



To many features of ferromagnetism, of which whole monographs 

 might be or have been written, I can give only brief mention or none 

 at all. There are the "magneto-caloric effects," arising because, 

 when a ferromagnetic body is heated, the dis-alignment of the atoms 



* Cf. the article of R. M. Hozorth in the preceding number (January 1936) of this 

 Journal. 



