Zeeman Effect in Satellites of Mercury Lines. 337 



finding the position by the one and the course by the other, 

 we were finally led to the solution of the proposed problem. 



The results are by no means so simple as in the Zeeman 

 effect of most principal lines, in which they are divided 

 into branches at regular intervals, however numerous their 

 number may be. The anomalies in the Zeeman effect o£ 

 satellites mainly consist in the complexity of the separated 

 intervals, so that there is scarcely any simple proportionality 

 with the field strength. 



The reader will probably not be satisfied unless we give 3, 

 complete set of photographs which were taken into account ; 

 but as a large number of plates will be necessary for this 

 purpose, only some of the typical ones are given, for 

 uniform as well as for heterogeneous fields ; the latter 

 will serve as a guide for judging the course taken by the 

 different components. Sometimes there appeared faint 

 traces of s-components in photographs of ^-components, 

 owing to imperfect polarization, and vice versa-, but by 

 taking proper precautions, these could be easily detected. 

 Generally the definition of s-componenfs was inferior to 

 that of p-components, so that the complex separation of 

 the satellites was difficult to decipher, and conclusive results 

 for ^-components must be reserved for further investigation. 



The present experiment is limited to the violet line 4047 

 and the green line 5461. The photographs in different 

 fields are given in J?l. V. figs. 3-ll 9 in which botJi v- and 

 s-coinponents are given. For mere qualitative inspection 

 as to the course taken by tjie lilies, the photographs in 

 heterogeneous fields give a good illustration of the anoniajies 

 in the separation. 



Arranging the photographs of the type given in these 

 figures in proper order, we can graphically represent the 

 change in the position of the satellites relative to the 

 principal line by means of figs. 1 & 2, in which the abscissae 

 give the distances from the principal line, and the ordjnates 

 the strength of the field in kilogauss. The initial position 

 of the satellites in H = is taken from our measurements 

 made on crossed spectra, obtained by means of the echelon 

 grating and Lummer-Gehrcke plate. The length of the 

 line drawn underneath is proportional to the intensity. The 

 shaded portion shows the actual breadth of the line. 



Referring to the ^-components of 4047, we notice the 

 dissymmetry in the intensity of the different branches, The 

 satellites +59 and +117 are divided linearly proportional 

 to the field so long as it is weak, but the -f branch (t. e„ 

 that in the direction of increasing wave-length) gradually 



rial, Mag, 8, f Vol 27, No, 158, Feb. 1914. Z 



