496 M. Verdet on the Optical Properties developed in 



superieurc, was a rectangular parallelepiped polished on its six 

 faces, whose edges were respectively equal to 37"2 millims., 26*0 

 millims., and 12-5 millims. T describe the two samples as No. 1 

 and No. 2. The specimen of flint was a parallelopiped of square 

 base 43*3 millims. long by 14'5 millims. on the side, two of its 

 bases, and a couple of its lateral faces, being polished. These 

 three pieces were not entirely free from defect ; but by elevating 

 or lowering the support on which they were placed, and by 

 making them slide on this support, the ray of light could always 

 be made to traverse a space where the defect had no perceptible 

 influence, so that by a convenient position of the analyser the 

 incident light could be completely extinguished when the electro- 

 magnet was not magnetized. The bisulphide of carbon was con- 

 tained in small glass cylinders, closed by plates of ordinary glass j 

 one of these cylinders was 44 millims., and another 31 millims. 

 long. I satisfied myself beforehand that the rotation due to the 

 terminal plates of glass was entirely imperceptible. 



The whole of the experiments establish a very simple law, — 

 " The rotation of the plane of polarization is proportional to the 

 magnetic action." The following Tables, which only contain a 

 part of my experiments, give the proof of this law. In these 

 tables F is the magnetic action measured by the deviation directly 

 observed in the galvanometer, R the double of the rotation, and 



T> 



Q the value of the quotient r= when we suppose R expressed in 



minutes. In each table the words white light or homogeneous 

 light indicate the manner in which the optical observation is 

 made ; the thickness of the substance traversed by the luminous 

 rays is also indicated. 



Experiments on heavy Glass, No. 1. 



Mean . . 3-86 



* The hundredths of division and the seconds that we shad see in these 

 tables result from the calculation of the mean. 



t In these two experiments, the distance from the divided rule to the 

 mirror was not the same, so that the magnetic actions given in the Tables I . 

 and II. are not measured with the same unity. 



