Zeeman Effect in Satellites of Mercury Lines. 341 



The principal line is divided into a nonet, of which three 

 belong to the ^-components, and the remaining six to the 

 s-components. The separation takes place regularly at 

 equal intervals, and the lines are strong and broad, so that 

 the branches o£ different satellites which cross each other 

 are obscured in weak fields. Fig. 2 shows the separation 

 for p- and s-components, the shaded portion representing 

 the actual breadth of lines. 



The satellite — 242 lying farthest from the principal line 

 changes in a singular manner. It was already noticed by 

 Lunelund that in fields of a few thousand gauss, the in- 

 tensity of the + and — branches becomes equal, that of 

 the former far exceeding that of the latter. In fact, the 

 — branch becomes so faint that it is difficult to observe it 

 beyond 10 kilogauss ; the curve then begins to move in the 

 opposite direction. This, however, goes on quite slowly, so 

 that it is only in field of 21 kilogauss that the line returns 

 to its initial position in zero field. The decrease in wave- 

 length goes on steadily increasing at a constant rate, and 

 with gradual increase in intensity. The type of separation 

 in both branches is nearly parabolic, and resembles — 60 in 

 4017 ; the only difference being that the + branch, instead 

 of osculating to the principal line, approaches the — branch 

 of the satellite — 74, which runs parallel to — component 

 P_! of the principal. The s-component of — 242 is a 

 doublet of nearly equal intensity in weak fields, so that 

 it is divided into a quartet as found by different inves- 



tigators. 



The linear relation between &\ and H is satisfied in the 

 different branches of the satellite —74, which is apparently 

 divided into a nonet. The separate lines all run parallel to 

 the branches of the principal line, but are of unequal in- 

 tensity. The one parallel to P_j is very weak, and the 

 + branch which runs parallel to P +1 is not very strong, and 

 a portion, which probably passes through the principal lines, 

 cannot be followed continuously on account of obliteration 

 by the diffuse boundary of the principal line. The 

 — p branch is strong in weak fields, but gradually 

 diminishes in intensity, and ultimately becomes very faint 

 as it approaches the + branch of —242, which on the con- 

 trary becomes very intense. It is questionable whether this 

 branch of —74 ceases to exist at this stage, or takes the 

 same course as —242 after merging into it. The increase 

 in intensity can evidently be attributed to the resonance 



