Hl 



345 



efFecl nol only aü'ord a confirmation of the validity of tlie considerations in § 5, 

 but also, in many respects, seem to give indications of the way in wliich these 

 considerations may be extended. 



Let us now discuss more closely each of the lines considered in the tables. 



As regards //„ it is seen that the intensity of the undisplaced perpendicular 

 component has been found to be the strongest of all components observed, in agree- 

 ment with what should be expected from the large values of the corresponding 

 (juantities R' - and /?"-. Further it is seen that the increase of tiie intensities of the 

 parallel components in the direction of increasing J is very well illustrated by the 

 values of the corresponding R^'s. In addition to the components observed, the 

 theory predicts the existence of weak parallel components at J = 4; 8 and of weak 

 perpendicular components at J == ^ 5 and J = ±6. 



For H,3 Stark records parallel as well as perpendicular components corres- 

 ponding to J = 2, 4, 6, 8, 10 and 12, but, according to Stark's own statement, it 

 was very difficult to obtain good photographs of the Stark effect of this line, and 

 a long exposure was necessary in order to obtain all components on the plate. 

 These difficulties may account for the small discrepancies which seem to exist 

 between the different observations on one hand and between these observations 

 and the theory on the other hand, since during the long exposure any unreal 

 component ("Geist") due to the imperfection of the grating would have special 

 opportunity to appear. Thus according to the theory no parallel components at 

 J = 4 (and J = 12) and no perpendicular components at J = and J == 8 

 should appear, while Stark's photographs would indicate the existence of such 

 components. (It must, however, be remarked that, as seen from the table, the 

 perpendicular component at J = 8 was not recorded in Stark's publications before 

 1915). Further the intensity of the perpendicular component J = 6 would according 

 to the theory be stronger than the perpendicular component J = 4, in agreement 

 with the photograph reproduced in fig. 5 on Plate II, but in disagreement with the 

 values 9,7 and 12,6 for the relative intensities of these components appearing in the 10"' 

 column of Table II. The possibility for the appearance of a parallel component at 

 J = has been discussed on page 58 and 59. On the whole it will l)e seen that the 

 agreement between the theory and the observations is satisfactory, and it seems 

 probable that this agreement will be improved by further experiments. 



In case of the Stark eflect of Hy it is seen that the agreement between the 

 measurements and the estimate afforded by the theory is rather distinct for most 

 components; but the perpendicular component at J = (113 — 002), and also the 

 parallel component at J 5 (212 — 002), appear undoubtedly stronger than we 

 would expect from the corresponding values of /?' -'. This may ha\e connection with 

 the fact that, for the corresponding transitions, the final stales correspond to a 

 circular orbit of the electron (compare page 60). 



In case of the Stark etfect of //,; it is especially salisiactory lliat it has been 

 possible to explain (he non-appearance of a component corresponding to the trans- 



