316 M. H. Becquerel's Experimental Investigations 



great nicety ; and we thus see that the two mirrors present a 

 series of images which are reflected without any deformation. 

 Including the direct image from the polariscope, they corre- 

 spond to columns of air represented by one, three, five, seven, 

 nine times the distance of the mirrors. A tolerably large 

 number of reflected images can be obtained ; but, in order to 

 preserve a sufficient luminous intensity, distinctness, and an 

 apparent diameter of the image suitable for the polariscope, 

 I have been obliged to stop at the fourth rejected image, 

 which corresponds to nine passages of the luminous ray 

 through the bobbins. With the magnifying-power just 

 indicated, this fourth image is as clear as the direct image of 

 ordinary polarimeters or saccharimeters. 



Suppose, now, that the electric current is caused to pass in 

 one direction : one of the images, the fourth for instance, 

 must be reduced to uniformity of tint, as in ordinary polari- 

 meters ; then, if the electric current is reversed, one half of this 

 image will be seen darker than the other. If the analyzer is 

 then turned so as again to obtain uniformity of tint in the 

 two halves of the image, the angle by which it has been 

 turned will measure twice the magnetic rotation sought. 



The method supposes, in theory, that the light studied is 

 monochromatic ; but in the experiment itself, as the rotations 

 are very small, it is not necessary that the incident rays should 

 have the same wave-length. However, we shall return later 

 to this important question. 



Under the most favourable conditions the rotations could 

 be measured to nearly 30 // . 



Glasses which close the Tube. — In the experiments that we 

 have just described the luminous rays only pass through air, 

 and the rotation obtained represents, with the exception of a 

 slight correction, the magnetic rotation of that gas. When 

 the experiment is tried with other gases, it becomes necessary 

 to close the tube by glasses, as we have above indicated. 



The luminous rays pass through the glasses G, G 2 as many 

 times as through the tube itself ; and as at each passage the 

 latter presents four surfaces, it is clear that the luminous rays 

 corresponding to the fourth reflected image will have traversed 

 thirty-six surfaces of glass. If to them are added the four 

 surfaces of the mirrors, it will make forty surfaces of glass 

 during the passage of the light. From this it is seen what 

 must be the perfection of the workmanship of these surfaces 

 in order that the images should not experience any appreciable 

 distortion. In addition to this, if it happened that the glasses 

 were even very slightly unannealed, the depolarization of the 

 light, which would have been almost imperceptible during a 

 single passage, would become considerable, 



