152 Magnetic Rotation of Sodium Vapour at the D Lines. 



The two constants A and B can be determined from these 

 data. Similarly we can determine A and B from the 360° 

 rotations at wave-lengths 5887*13 and 5898*57, or from the 

 540° rotations at wave-lengths 5887*70 and 5898'09. If 

 the formula expresses the experimental results, the constants 

 should be sensibly the same in each case. 



Having determined the constants we can calculate the 

 rotation at the centre (X = 5893*17), which in the above case 

 is 720°. 



The results of the calculations are given in the following 

 table : — 

 Observed d at Centre 720°. D 2 = 5890-2. D r = 5896-2. X at Centre 5893-47. 



8. 



X. 



A. 



B. 



6663 

 7055 

 6975 



A 

 B' 



d at centre 

 (ealc). 



180 

 360 

 540 



5885-83 

 5899 64 



5887-13 

 5898-57 



5887*70 

 5898-09 



12578 

 12767 



12897 



1-88 

 1*81 

 1-85 



528° 

 535° 

 536° 



A little calculation shows that the variation in the values 

 of A and B is not much greater than is to be expected from 

 experimental errors, so that we can regard the formula as 

 holding very well in the region outside of the D lines. 



Between the D lines, however, the formula does not hold 

 at all, the observed rotation at the centre being much larger 

 than the calculated, 



This has been found to be the case with all of the vapour 

 densities employed, and the same thing was observed in the 

 earlier work. 



It is clear from this that the rotation between the D lines 

 is not simply additive, at least if the individual rotations are 

 symmetrical about the lines, as indicated by the formula. 



It must be remembered that the D lines are not single 

 lines, and that neither gives rise to the normal Zeeman triplet. 

 The mo'ecular mechanism is, therefore, in all probability 

 much more complicated than the one considered in the 

 theoretical treatment. The fact that the rotation curve can 

 be computed from the dispersion curve by the aid of the 

 Becquerel formula, makes it seem worth while to determine 

 the dispersion curve between the D lines with greater ac- 

 curacy than has been done before. It seems most probable 

 that the rotatory power is not symmetrical with respect to 

 the lines, and if this is so, the same should be true for the 

 dispersion. This matter will be subsequently investigated. 



