390 Preston—Radiation Phenomena in a Strong Magnetic Field. 
appearances are really due to reversal or not.* Up tothe present, however, I have 
not been able to obtain any decisive information on this important point; but on 
the whole I am inclined to the opinion that all the variations of the normal triplet 
appearance which I have observed may be sufficiently accounted for by absorption. 
This point however, I hope to return to, as soon as I obtain a stronger magnetic 
field, and for the present it is best to regard it as not absolutely decided. 
Indeed, if we consider this question from a theoretical standpoint, there seems 
to be no good reason why we should not have doublets as well as triplets, or triplets 
with weak middles, or other modifications, when the source of light is viewed 
across the lines of force. For, if the influence of the magnetic field on the vibrat- 
ing systems (be they what they may), be such as to set the plane of any one 
vibration at right angles to the lines of force, then the component of this vibration 
in the direction of the lines of force will vanish, that is, the constituent of the 
unmodified period which gives rise to the central line of the triplet, will vanish, 
and we shall be left with a doublet ; or on the other hand, if the vibration became 
directed wholly along the lines of force, the period would remain unaltered, and 
the line should be quite unaffected by the magnetic field. 
In fact, in order that a line should be converted into a normal triplet having 
the middle line equal in intensity to the two side lines combined, it is necessary to 
assume that the freedom of vibration is the same in every direction in the system. 
But if this be not the case, then modifications of the type of normal triplet will 
occur and should be searched for. Ii these are proved to exist, our knowledge of 
the difference in character of the different spectral lines will be considerably 
increased ; but if not, and if all the observed modifications are proved to be 
due to absorption, still the examination of the spectra by means of the influence 
exerted by a magnetic field retains a high value ; for, by it, we may be led to the 
grouping of the spectral lines in such a way that the general laws of their recurrence 
may be discovered and hence inferences may be made as to the nature of the 
events taking place in the vibrating systems which give rise to them. 
Another cause which may be operative in producing deviations from the normal 
triplet produced by the magnetic field may lie in the constitution of the line itself 
Thus, if a spectral line be not a simple line, but a close doublet,t+ then each consti- 
tuent will be converted into a triplet by the magnetic field ; and if the correspond- 
ing lines of these triplets, which overlap, differ in phase by the same or different 
amounts, decided modifications will be introduced into the appearance of the 
*It should be remarked that in none of these cases does the line exhibit reversal when the magnetic 
field is unexcited. 
} Michelson has already found that some of the spectral lines possess complexity of this nature: 
A. A. Michelson, Phil. Mag. vol. xxxtv., p. 280, 1892. 
