on Magnetic Rotatory Polarization in Gases. 429 



These numbers correspond to a single passage of the lumi- 

 nous rays through the tube ; we see with what precision they 

 mutually test each other. By dividing these numbers by the 

 rotation obtained for carbon bisulphide under the same condi- 

 tions, and which was found to be equal to 4520', the following 

 table is obtained: — 



Magnetic Rotations of Gases compared with liquid Carbon 

 Bisulphide (yellow light D). 



Gas. 



Rotation for one 

 passage. 



Rotations com- 

 pared with carbon 

 bisulphide. 



Oxygen 



0-663 

 0-720 

 0-729 

 1-365 



1-780 

 3300 

 3-625 



0000146 

 000159 

 0000161 

 000302 

 0-000393 

 0000730 

 0-000802 



Air 



Nitrogen 



Carbonic acid 



Nitrous oxide 



Sulphurous acid 







In the course of the experiments an occasion offered of veri- 

 fying the preceding numbers. 



We have seen how the determination of the magnetic rota- 

 tion of the glasses in the tube inside the first coil was effected. 

 This rotation may be compared with that of the gases. Now 

 it happens that for air the comparison is immediate ; the rota- 

 tion of the glasses (say 1 centim. of crown glass) is equal 

 to ten times the rotation of 3 metres of air (say to 3000 cen- 

 tim. of air submitted to the same magnetic influence). The 

 relation of the rotations is then 



3^0 =,>00033 - 



Now in my former researches I found that the magnetic rota- 

 tory power of crown glass for the yellow light was 0*48 of 

 that of carbon bisulphide. By adopting this number, which 

 is independent of our present experiments, we should find for 

 the rotatory power of air 0*000158*. 



* Since the time when I began these researches, MM. Kundt and 

 Routgen have published numbers relative to several gases. The only 

 ones common to our researches are oxygen and air. The numbers 

 given by these authors in their first memoir (Wiedemann's Annalen, 

 t. viii. p. 278) are too large, because they have compared the rotations 

 obtained directly with white light for gases, and with yellow light D for 

 carbon bisulphide. For oxygen the number that I find is identical with 

 that which they have given for high pressures. The agreement is, no 

 doubt, owing to the fact that oxygen has no sensible rotatory dispersion, 

 and that the errors which they committed respecting the refraugibility of 

 light became then unnoticed. In a second memoir the same authors have 



Phil. Mag. S. 5. Vol. 12. No. 77. Dec. 1881. 



2K 



