XII.] THE MAGNETIC PERTURBATION OF LINES. 113 



at right angles to these, their vibrations being at right angles to the 

 field of force. 



Exceptions to this rule, however, were found in the lines having 

 wave-lengths 3587-13, 3733*47, and 3865'67, which behaved in exactly 

 the reverse manner. Two other lines at wave-lengths 3722*72 and 

 3872-64, were quadruplets, the central component, which had its vibra- 

 tions along the line of force, being a close double. Some of the lines 

 which showed no modifications whatever, were those at AA 3746 '06, 

 3767-34, 3850-12, and 3888-67. 



These observers further noticed that the separation of the side com- 

 ponents of the triplets seemed to be irregular ; they found that there 

 were certain lines in which the separation was nearly the same, but 

 much greater than that of other lines where separations seemed to be 

 quite closely alike. On this basis they separated the lines in the iron 

 spectrum into two classes, in each of which the " magnetic separation " 

 was the same, but in the one set much greater than in the other. 



The lines belonging to these two sets were found to be practically 

 identical with those sets of lines into which the iron spectrum breaks 

 up when studied with reference to the shift produced by pressure, but 

 this conclusion is not accepted by Dr. Preston. 



I have already stated that on the simple theory we should get 

 triplets only, as on the simple theory of thirty years ago we should have 

 got motion of a solar vapour indicated by all the lines in a spectrum. 

 The facts are equally against the simple theory in both cases. 



The magneticians can now, however, by extending their theory,, 

 embrace and explain all the new, and at first sight extraordinary, 

 phenomena. To show how they have done it, I cannot do better than 

 quote from a lecture recently given by Dr. Preston, who is among the 

 most successful investigators of this new branch of science.* 



" According to the simple theory, every spectral line, when viewed 

 across the lines of force, should become a triplet in the magnetic field, 

 and the difference of the vibration frequency between the side lines of 

 the triplet should be the same for all the spectral lines of a given sub- 

 stance. In other words, the precessional frequency should be the same 

 for all the ionic orbits, or the difference of wave-length 8 A between 

 the lateral components of the magnetic triplet should vary inversely 

 as the square of the wave-length of the spectral line under consider- 

 ation. Now, when we examine this point by experiment, we find 

 that this simple law is very far from being fulfilled. In fact, a very 

 casual survey of the spectrum of any substance shows that the law 

 does not hold even as a rough approximation ; for, while some spectral 



* Nature, vol. 60, p. 178. 



