on Magnetic Rotatory Polarization in Gases. 403 



inside the first coil; and taking into account the relative 

 thickness of the glasses and mirrors, it is clear that, in order 

 to ascertain the magnetic rotation to which they give rise, 

 the number obtained for the glasses inside the tube must be 

 multiplied by 0-040. 



In this way the following Table, which contains the neces- 

 sary corrections to be made to the direct measurements, has 

 been drawn up : — 



Corrections for the Glasses and the Glass of the Mirrors. 



Red. 



4th image. 3rd image. 2nd image. 1st image. 



Glasses 



Mirrors 



3-49 

 0-25 



294 

 0-25 



2-10 

 0-25 



1-26 

 0-25 



3-74 



3-19 



2-35 



1-51 



White (incandescent lime) ; yellow. 



4th image (D). 3rd image. 2nd image 1st image. 



Glasses 4'30 3"50 2-53 1-52 



Mirrors 028 029 030 030 



4-58 3-79 2-83 1-82 



Green. 



4th image. 3rd image. 2nd image. 1st image. 



Glasses 



575 4-47 



3-10 

 0-37 



3-47 



1-85 



Mirrors 



0-38 0-38 



037 





613 4-85 



2-22 





Blue. 









4th image. 3rd image. 



2nd image. 



1st image. 



Glasses 







2-16 



Mirrors 





0-42 



2-58 



These numbers are obtained from a large number of obser- 

 vations ; although the tenths of a minute may be relied on, 

 the hundredths are given as they resulted from the calcu- 

 lations. In concluding the examination of this question, I 

 may remark that, the rotations to be measured in the gases 

 being very small, all secondary considerations had to be 

 sacrificed to perfection in the arrangement of the optical 

 system, without which perfection the measurements would 

 have been impossible. 



IY. Correction for the Temperature. — I stated above that, in 



