226 BELL SYSTEM TECHNICAL JOURNAL 



were an atom-model with stationary electrons and an atom-model with 

 revolving electrons, which were in competition with each other, and 

 the latter of which has by now driven the former utterly out of the field. 

 Remember now, that the atom-model with the revolving electrons 

 triumphed over the other one not primarily because of its magnetic 

 quality, but because of the theory of spectra which Bohr and others 

 were able to derive from it. Revolving electrons in atoms were first of 

 all proved to be responsible for spectra, and then it was noticed that 

 they are capable of causing magnetism. Therefore when the physicist 

 says that magnetism is a quality of atoms, he is not making a confes- 

 sion of incompetence, but an inference from a highly-developed and 

 successful theory of spectra; and this makes all the difference in the 

 world to the value of the statement. Indeed the situation is even better 

 than I have intimated ; for there are dozens of cases in which first an 

 atom-model or a molecule-model has been constructed expressly to ex- 

 plain the spectrum of the substance in question — then, the magnetic 

 moment of its system of revolving electrons has been computed — then, 

 the magnetic moment of the atom or the molecule has been measured — 

 and the two have agreed! This is really an understatement, which 

 needs to be broadened so as to include the cases in which the spin of 

 the electron plays a part; but I pass them over, intending to defer the 

 broadening to the latter part of the article, which is to be devoted to 

 these matters of atomic structure. For the moment, let me make just 

 one more allusion to them, a very important one. Electrons revolving 

 in orbits around a nucleus obviously possess angular momentum. 

 Therefore, if an atom has a magnetic moment due to revolving electrons, 

 it has an angular momentum also. This again is an understatement, for 

 it contains a restriction which can be removed in view of the broaden- 

 ing which is later to be made. It appears to be a general rule that in 

 the atom, magnetic moment and angular momentum always go together. 

 A magnetic atom is a gyroscope — necessarily and automatically. This 

 is a fundamental principle, and from it flow some strange and striking 

 consequences. Everyone who has worked or played with the classic 

 gyroscope of our laboratories knows that it has quaint and tricky idio- 

 syncrasies. W' ell ! the atom has them too ; but it has others in addition. 

 Angular momentum, on the atomic scale, is subject to peculiar laws of 

 quantum mechanics; and the atomic magnet-gyroscope behaves in 

 extraordinary ways, of which our laboratory gyroscopes give not the 

 faintest intimation. 



To summarize my introduction then: the first step in the theory of 

 magnetism consists in referring it to the individual atom. This sounds 

 like a confession of defeat, but it is nothing of the sort; it is a claim of 



