on Magnetic Rotatory Polarization in Gases. 319 



It is known that, if the pile is suitably placed, a slight rota- 

 tion of the incident plane of polarization can be considerably 

 amplified. 



I had two piles of glasses made, constructed exactly according 

 to the directions given by M. Fizeau * ; but I did not find, in 

 the present case, that the use of them gave us all the results 

 that we had a right to expect. 



In the first place, it is clear that, if a polariscope of the kind 

 described is employed, it is necessary to have a prism the two 

 planes of which only make a very slight angle ; for this angle 

 is itself increased by passing through the piles of glasses, and 

 it may happen, as in the case of the polariscope described 

 above, that all sensitiveness disappears. In fact, the piles of 

 glasses gave an amplification of nine times a slight rotation 

 of the incident plane of polarization ; and the principal planes 

 of the two halves of the image from the polariscope were con- 

 siderably thrown out. I attempted to make use of a large 

 Nicol's prism furnished with a half-wave-length plate, so as to 

 form a polariscope of variable sensitiveness ; but this appa- 

 ratus necessitates the employment of a monochromatic source of 

 light, and did not give good results. Even with monochromatic 

 light the two halves of the image are not in the same physical 

 conditions ; the one is polarized elliptically, and the other recti- 

 linearly. For such delicate measurements as those here spoken 

 of, I very much prefer the Nicol cut in half. If a suitable 

 polariscope had been sufficient, I could have introduced this 

 arrangement with the same precision as the method of ampli- 

 fication by the mirrors. But the use of piles of glasses has, in 

 the present case, a much more serious objection. I had 

 intended to use them concurrently with the mirrors, so as to 

 obtain rotations of several degrees ; but unfortunately the 

 slightest variation in the angle made by the incident rays with 

 the normal of the piles of glasses changes the amplification 

 considerably, so that it is not the same for the different images. 

 It would then be necessary, after each measurement, to deter- 

 mine experimentally the amplification due to the piles of 

 glasses, which would greatly complicate the observations. 

 Besides, this amplification is not the same for rays of different 

 colours ; it is thus necessary to make use of a source of light 

 almost perfectly monochromatic and, if the rotations of rays 

 having different wave-lengths are studied, to measure experi- 

 mentally in each case the amplification which corresponds to 

 them. It will hereafter be seen what difficulties arose from 

 the want of homogeneity in the source of light we were obliged 



* Annates de Chimie et dc Physique, viii. p, 153 (1859). 



