168 Dr. T. Preston on Radiation Phenomena 



side lines B and C show as doublets and in others as triplets: 

 and again, in some quartets the distance between the central 

 pair A is almost as great as, and it may be greater than, the 

 distance between the side lines B and C. Further, as the 

 magnetic field increases in strength, the distance between the 

 members of the central pair A (fig. 4) increases at the same 

 rate as the distance between the side lines B and C. Again, 

 when the side lines B and C are each resolved into doublets 

 (or triplets) the separation of the constituent lines of each of 

 these doublets (or triplets) increases with the magnetic field 

 like the separation of the components of the normal triplet 

 ABC shown in fig. 1 ; and in face of these facts the reversal 

 theory becomes quite untenable. 



The general phenomenon, therefore, which remains to be 

 explained is the further resolution of each constituent of the 

 normal triplet into a doublet or a triplet or some other system; 

 and, as we shall see immediately, the electromagnetic theory 

 proposed by Dr. Larmor * may be extended to embrace all 

 the phenomena yet observed. Before proceeding to consider 

 this explanation, however, it is necessary to refer to a 

 particular case which was recently announced as having been 

 observed by MM. Becquerel and Deslandresf, and subse- 

 quently by Messrs. J. S. Ames, R. F. Earhart, and H. M. 

 Reese J, and which they refer to as an example of " reversed 

 polarization." This phenomenon is represented to be as 

 follows. 



Consider the triplet shown in fig. 1 ; then in the normal 

 state of affairs the vibrations in A are parallel to the lines of 

 force, w r hile the vibrations in B and C are perpendicular to 

 the lines of force. Now in the spectrum of iron the authors 

 just named have recorded that they observed triplets in which 

 the vibrations in the middle line were perpendicular to the lines 

 of force, while the vibrations in the side lines were parallel to 

 the lines of force — the reverse of the normal case. Stated in 

 this way the phenomenon is very startling, and appears at 

 first sight to be directly contradictory to all theoretical 

 expectation. But if we return to fig. 4, it will be seen at 

 once that this phenomenon, supposing it to exist, can be 

 regarded merely as an extreme case of the quartet. For, as 

 we have already said, the horizontally vibrating lines A of 

 the quartet may be close together or widely separated. They 

 may be even more widely separated than the vertically 

 vibrating fines B and C (fig. 5), and in some particular cases 



* Dr. J. Larmor, Phil. Mag;, vol. xliv. p. 505 (1897). 

 t Comptes Rendus, t. cxxvi. p. 997, April 4th, 1898. 

 % Astro-Physical Journal, vol. viii. p. 48, June 1898. 



