SOME CONTEMPORARY ADVANCES IN PHYSICS— X 133 



the atom is permitted to take either (2/+1) or 2/ distinct orienta- 

 tions in the field; the former number if it is, the latter if it is not 

 permitted to set itself quite parallel to the field. 



It will now be shown that these are by no means idle speculations; 

 they bear directly upon certain facts accessible to observation. Before 

 bringing up these facts it is necessary to abandon the policy of speak- 

 ing exclusively about the "normal" Zeeman effect. This "normal" 

 effect received its adjective because it agrees so excellently with the 

 original theory devised years before quanta were dreamt of to explain 

 the effect of magnetic field upon spectra. It is essentially because 

 of this agreement that it is possible to develop the contemporary theory 

 of the "normal" effect in a perfectly deductive fashion, using no 

 new assumptions beyond those general ones of the orientation-theory. 

 Most spectrum lines, however, are affected by a magnetic field in ways 

 not compatible with the original theory; which is a consequence of 

 the fact that the set of new Stationary States, whereby a magnetic 

 field supplants each original Stationary State, in most cases does not 

 conform to the laws previously set forth. 



The laws to which it generally does conform were read from the 

 spectra by Lande. The one feature in which the foregoing theory 

 quite agrees with these laws is its prediction of the total number of 

 Stationary States. A Stationary State for which the angular mo- 

 mentum of the atom is determined, by virtue of the theory of multi- 

 plets which filled the preceding section of this article, as being 

 2J {\h/2-w), is actually found to be supplanted, when a magnetic 

 field is impressed upon the atom, by 2/ new Stationary States. This 

 is in agreement with one of the two alternative predictions made a 

 few paragraphs supra; to wit, with the prediction derived from the 

 assumption that the atom cannot set itself quite parallel to the 

 field. This agreement between the orientation-theory of multiplets 

 and the orientation-theory of Zeeman effect considerably strengthens 

 both. 



The several Stationary States replacing a given original State 

 are always equally spaced; but the spacing differs in amount 

 from the value ullh or eHh/^irfxc exhibited when the normal 

 Zeeman effect occurs, and which we found it possible to deduce 

 from the simple orientation-theory. The difference is this, that 

 the actual spacing is a multiple of the value uHh by a factor g 

 (generally lying between | and 2) which depends upon the original 

 State : 



AU = g<^Hh (17) 



