on Magnetic Rotatory Polarization in Gases. 421 



The slight correction due to the presence of 0*05 of air can 

 be calculated with great precision. For example, we have for 

 the fourth image — 



4th image. 



White. Green. 



Mixture 3i'28 39-05 



0-05 of air 0-32 Q38 



Difference = 095 defiant gas 3096 38*67 



Pure olefiant £as 32*59 4O70 



Discussion of the Results obtained. 



I. Verifications of the Method of Observation. 



The principle of the method employed by me in this work 

 consisted, as we have seen above, in amplifying the pheno- 

 menon by causing the luminous rays to pass several times 

 through the tube full of gas. It is evident that the numbers 

 obtained for each reflected image must be proportionate to the 

 number of passages made by the luminous rays through the 

 apparatus. These numbers should therefore mutually test 

 each other; and by dividing them by 3, 5, 7, and 9 we ought 

 to obtain numbers which would represent the magnetic rota- 

 tion corresponding to a single passage through the tube. 

 With the same luminous source and the same coloured screen 

 these numbers should be constant ; but we have seen above 

 that the composition of the light which reaches the eye in 

 order to form the different images varies with them; we 

 ought, then, to find numbers regularly varying with the order 

 of the images. 



Strictly speaking, all the measurements ought to be effected 

 at the distance which separates the mirrors; and when any 

 other gas than air is in question, the column studied has not 

 exactly the same length as the distance from the mirrors. But 

 we have seen that, in the interval of 0*06 metre which sepa- 

 rates the end of the tube from the mirrors, the mao-netic in- 

 tensity is very feeble ; and the effect due to this slight differ- 

 ence in length must be regarded as absolutely unimportant, 

 so that the verifying calculations can be applied even when 

 the tube is full of any gas whatever. The slight differences 

 between numbers which, for the same gas, correspond to a 

 single passage of a pencil of rays of the same colour, are partly 

 due to variations in the colour of the images, rather than to 

 errors in the measurements. 



