1014 Prof. R. W. Wood and M. G. Ribaud on the 



The absorption lines which we have investigated are 

 numbered, 2, 3, 4, 4', 5 and 6. These numbers conform to 



Fie-. 3. 



4- 



77\ P : : itNi|,ffl r ••• 



/- — "~ I r :::::riJ^IF77^ 



'+ -' 



E F I 



ItE3» 



A. Mercury arc. 



B. Lens forming image of arc on bulb at D. 



C. Polarizing prism. 



E. Half-wave plate. 



F. Cell of bichromate of potash and neodymium to remove 



yellow and violet lines. 



G. Lens forming image of E 3 mm. inside the slit of 



collimator I. 

 H. Analysing nicol. 



those used in previous papers on the resonance spectra. The 

 line 4' is much weaker than the others, and did not record 

 itself on the earlier photographs made with vapour less 

 dense than that used in the present case. 



The curves of rotation for these lines are shown by fig. 4 : 

 they are only roughly quantitative. 



IWL- 



The direction of the rotation by the absorption line No. 4 is the 

 same as for the T> lines of sodium. 



In this group of lines the direction of the rotation changes 

 as we pass from line to line, which explains perfectly why 

 the angular magnitude of the rotation is so small in comparison 

 with that exhibited by the D lines. The -}- rotation due to 

 a given line is nearly neutralized by the — rotations of its 

 two neighbours. We also see at once why the rotation on 

 one side of a line may be much greater than on the other. 

 For example, we have very strong rotation to the right of 

 line 4 since the opposed rotation due to the faint line 4' is 

 very small, and the effect contributed by the next line 5 is 

 of the same sign. The rotation to the left of 4 is, however, 

 very small since Hue 3, with its opposite rotation, lies very 

 close to 4. In the case of line 3 the strong rotation is to 



