on Magnetic Rotatory Polarization in Gases, 315 



tion of the analyzer which corresponds to equality of tint 

 should remain nearly the same for all the images. This con- 

 clusion is verified by experiment ; besides, it is easy to see 

 that, the reflections on the mirrors being nearly normal, the 

 displacement of the planes of polarization must be inappre- 

 ciable. The mirrors are made of St.-Gobain glass wrought 

 with the greatest care and silvered on the surface. As it is 

 necessary that they should be absolutely plane in order to 

 avoid any deformation of the images, we have employed rather 

 large surfaces of glass, of which we have used only the central 

 part. The mirrors are rectangular, and are about 0*16 metre 

 long by 0*06 metre wide and 7*9 millim. thick ; they are only 

 silvered on 0*08 metreof the middle portion (PI. VII. fig. 3), so 

 that the luminous rays, on coming out of the polariscope and 

 before being received on the analyzer, pass once through the 

 thickness of the glass of the mirrors M x and M. There is thus 

 introduced into the magnetic rotations to be measured a slight 

 correction due to the influence of magnetism on this mass of 

 glass. The mirrors are placed at 0*195 metre from the bob- 

 bins. It is easy to see that under these conditions the correction 

 which corresponds to the passage of the luminous ray through 

 the substance of the mirrors is very slight. It would have 

 been possible to silver the mirrors to the ends and cause the 

 luminous rays to pass over their edges without passing through 

 the glass ; but I have preferred to allow this slight correction, 

 which was easily estimated, to remain, in order to profit by 

 the greater perfection of the workmanship and silvering of the 

 central portion. 



Analyzer. — The luminous rays, after their successive reflec- 

 tions, are received on an analyzer formed of a Foucault's prism 

 set on a graduated circle, the vernier of which reads to a minute. 

 The diaphragm of this prism is 0*017 metre. Behind the 

 analyzer is a microscope of a rather strong magnifying-power 

 (about 19 diameters), and which is independent of the move- 

 ment of the analyzer on the graduated circle, so as not to in- 

 crease the mass to be carried round in taking the measurements, 

 and to leave to the analyzer all the freedom of movement pos- 

 sible (PI. VII. fig. 4). This precaution is far from being use- 

 less in ensuring accuracy of observation. The analyzer and 

 microscope, however, are associated in the other movements, 

 forming a system which can be displaced laterally and inclined 

 at a slight angle by turning on a vertical axis, so as to bring to 

 the centre of the microscope the reflected image to be studied, 

 and to cause the optical axis of the system to coincide with 

 the more or less oblique direction of the reflected ray. 



The various parts of this apparatus can be regulated with 



