56 Py^cerih'iiQS nt the Roi/nl Irish Acaihmy. 



IV. DlSST-MMETRICAL SePAKATIONS AXD COMPARISON WITH OTHER SCBSTAXCES. 



Gmeliu* and other observers have shown by various methods that the 

 centre component of the yellow mercuiy line (X, = 5790) is displaced towards 

 the red, and the displacement is proportional to the square of the magnetic 

 field. Dufoui-f has obtained for several lines of chromium displacement of the 

 outside components without displacement of the middle one. In a recent 

 publication* I showed that there was apparently a connexion between the 

 rotation of the plane of polarization and the types of dissymmetry observed 

 with the apparatus used by me. It was pointed out that the variation in the 

 amount of rotation coiTesponded to an alteration in the relative intensity of 

 the middle component of a triplet to the outer components. It was found 

 that with the apparatus used for tungsten, molybdenum, and also for niobium, 

 a minimum intensity of the middle component of the triplets was reached near 

 X = 2900, and that corresponding t-o this turning-point for intensities, there 

 was a change in the type of dissymmetry observed from normal to abnormal. 

 There were a few exceptions, but in general the turning-points in intensity 

 corresponded to turning-points in the tj^e of dissymmetry. The following 

 Table (p. 57] contains the examples of dissymmetrical separations found in the 

 study of niobium. 



The letters a and n denote respectively " abnormal" and " nonual" dissym- 

 metry. It will be observed that here again there is a confirmation of what had 

 been observed before. The connexion between the rotation of the plane of 

 polarization and the tj^pe of dissymmetry is more firmly established. It is 

 only the tjjpc-s of dissymmetry which conform to the rule, since the actual 

 amoimts of the dissymmetry appear to be ii'regiilar, and the rule even does 

 not hold in all cases. The intensities of the red and violet components 

 are in each case the same, and this also corresponds to what was pre\aously 

 observed. 



It is iuterestiag to show, as was done for trmgsteu and molybdenum, that 

 when a verj' small displa<;ement is given to one of the kinds of components, 

 the intervals can then Ije expressed as midtiples, ^vith the factors 0, 1, 2, 3, 4 

 . . . instead of by multiples with factors showing no regularity. This, of 

 coiu-se, is a variation of the Pdtz law, and also of the Eiuige one ; and two of 

 the points to be noted ai'e that the superposed displacement is in each case 

 small, and that it is not necessary to suppose that so many components 

 have disappeared through lack of intensity. [Table XI.] 



* Phy3. Zeit , is, 212, 190S. 

 t Journal de Physique, is, 277, 1910. 



t " Dissymmetiiral Separations in the Zeem.an Ett'eet in Tungsten and ilolybilenum.' 

 Roy. Soc, Edin., vol. xxix, Pt. ii, No. 5. 



