CONTEMPORARY ADVANCES IN PHYSICS 235 



to the law tluU X varies iinersely as (7^ — 6). The coiislaiU A is one 

 which we adjust so as to get the empirical value of the constant d. 

 It is pleasant to be able to say that this constant /I —that is to say, the 

 hypothetical extra held — does not meddle at all with the multiplying 

 factor: the theory allows us to write: 



_ bNfjr 6 depending ov\ A 



k(T — 6) b depending on w ^ ^ 



and (if it is the right theory, of course!) we can go on estimating atomic 

 moments for substances of this category just as well as we can for the 

 substances for which 6 is zero. Most published values of fx correspond 

 to such cases. 



I am going to say very little about the extra field, or "Weiss field" 

 as it is often called, because it is still one of the mysteries of physics. 

 One realizes readily, of course, that if all the little atomic magnets turn 

 themselves partially or totally Into alignment, each one of them ex- 

 periences a magnetic torque which is due to all the rest. It may be 

 shown that this is proportional to the magnetization /, which looks 

 very promising indeed; but alas, when it is calculated its magnitude 

 turns out to be thousands of times too small. People used to say that 

 .47 must be a field of non-magnetic origin, which is just another way of 

 saying the same thing. At present it is commonly believed that the 

 force in question is what is called an "exchange" force, that is to say, 

 an electrostatic force among electrons, of which the modus operandi can 

 be discovered only by quantum mechanics. I am told that this 

 quantum-mechanical theory has not yet been persuaded to deliver a 

 really satisfactory result; but probably we shall be obliged to accept it 

 in default of any other. 



Now I call your attention to the fact that if the temperature should 

 be made equal to or lower than 6, this last equation would predict 

 something very wild and strange: an infinite, or a negative, suscepti- 

 bility. This is a curious situation, and there are several cases in which 

 we can appeal to experiment to resolve it. Take the elementary metal 

 nickel, for example; if one measures the susceptibility over the range 

 between 400° and 900° C. one gets a gorgeous curve of just this charac- 

 ter, for which the value of 6 is around 370°; now if one investigates 

 nickel at temperatures below 370°, say around room-temperature, one 

 learns that it \s ferromagnetic. The same holds true for iron, for cobalt, 

 for a diversity of alloys, except that 6 varies from one case to another.* 



* There are however cases in which the substance does not display the distinguish- 

 ing marks of ferromagnetism (notabh- remanence) when I < 6; ant! incidentally 

 there are cases in which 6 is negative; all of these are knotty problems for theor\-. 



