Transjmrent Bodies by the Action of Magnetism. 41 1 



simple proportion, it was easy to reduce all the observed rotations 

 to one constant magnetic action. 



This plan of correction necessitated two systems of measure- 

 ment, optical and magnetic. In consequence of the great num- 

 ber of determinations which I had to make, it became essential 

 to shorten the time required for each experiment ; I soon, there- 

 fore, had recourse to a method equally exact, and much more 

 expeditious than that of correction ; namely, direct comparison 

 between the optical actions of bodies and that of distilled water. 

 To effect this comparison, on some occasions the measure of the 

 rotation produced by the transparent substance was taken be- 

 tween two measures of the rotation produced by an equal thick- 

 ness of distilled water ; sometimes, when the battery had been 

 recently put together, and showed an action of almost perfect uni- 

 formity, the operation was still fm-ther abridged in the following 

 manner. Distilled water was first submitted to the action of the 

 electro-magnet, and the azimuth of the plane of polarization of the 

 emergent light determined. The water was then replaced by the 

 substance under investigation, and the azimuth of the plane of 

 polarization again determined. The direction of the current being 

 then reversed, the consequent displacement of the plane of po- 

 larization was measured. Finally, distilled water having been 

 again placed between the armatures, a fourth and last azimuth 

 \\^\s determined. Half the difference between the two azimuths 

 belonging to the water, and half the difference between the two 

 derived from the transparent substance, were considered as ex- 

 actly expressing the corresponding rotations caused by the same 

 electro-magnetic force*. 



rotations of the plane of polarization effected by a thickness of 20 millims. 

 of sulphide of carbon were 4° 16' and 3° 18'. The proportion between 

 the rotation (exjjressed in minutes) and the magnetic action was therefore 

 2'542 in the first, and 2-529 in the second case, showing that this relation 

 may be considered constant. Moreover, the difference between 1007 and 

 78'3 far exceeds the variation which the strengh of an electro-magnet can 

 undergo in a series of experiments extending over several days, if the 

 nitric acid of the battery remain unchanged. I never allowed this variation 

 to attain the tenth part of the initial power of the magnet. 



The empirical result just mentioned could not be deduced from the pro- 

 portion between the rotation and the magnetic action the existence of which 

 I i)roved in my first memoir, because we are not at liberty to assume that, 

 wlicn the electro-magnet varies in power, the action which it exercises at 

 dilFerent points of the interval between the armatures varies in the same 

 proportion. 



* The following are the complete elements of a comparison, effected by 

 the second process, between water and a solution of sulphate of nickel. 

 Tlie two liquids were cmjdoyed in thicknesses of 40 millims. ; the chstance 

 between the hexagonal armatures was 50 millims., and the battery con- 

 aistcd of twenty Bunsen's elements. 



Distilled water being placed in the apparatus, and the electro-magnet 

 2E3 



