423 M. Vcrdet on the Optical Properties developed in 



tained a strongly coloured solution whose negative rotary power 

 was very well marked. With about half the quantity of salt, I 

 obtained a solution almost entirely destitute of action on polar- 

 ized light. Alcoholic solutions gave me similar results. But 

 the best solvent which I have found is wood- spirit j this, by 

 itself, has scarcely any action on polarized light, and although it 

 dissolves a considerable quantity of the perchloride, yet remains 

 much more transparent than water, aether or alcohol containing 

 the same quantity of salt. Thus, on dissolving 55 parts of 

 the crystallized perchloride in 45 parts of wood- spirit, a liquid 

 is obtained M'hich, on account of its transparence, is suitable 

 for exact experiments, and whose action on polarized light is 

 negative in sign, and equal in absolute value to almost 

 twice that of heavy glass, or three times that of sulphide of 

 carbon*. 



Of all bodies hitherto examined, this solution is the one which, 

 under the influence of magnetism, effects the greatest deviation 

 in the plane of polarization. In my apparatus, a thickness of 

 10 millims. placed between the hexagonal armatures, which were 

 27 milhms. apart, on being subjected to the action of magnet- 

 ism, developed by twenty Buusen's elements, gave a rotation of 

 6° 31' 30" t. The greatness of this result induced me to choose 

 the solution of perchloride of iron in wood-spirit, in order to 

 determine whether the negative magneto-rotary power of iron 

 salts varied with the magnitude of the magnetic action according 

 to the same laws which occur with the positive rotary power of 

 ordinary substances. For this purpose I compared the rotation 

 produced by a thickness of 1 centim. of the solution with that 

 produced by an equal thickness of sulphide of carbon; the 

 magnitude of these rotations being made to vary either by 

 varying the strength of the electro-magnet, or by altering the 



* From experiments upon these various liquids, I obtained several values 

 for the molecular magneto-rotary power of perchloride of iron ; but my re- 

 sults were rather discordant. In fact, it is known that on treating crystallized 

 perchloride of iron with water, aether, alcohol, or wood-spirit, something 

 more than a simple solution takes place. A considerable elevation of tempe- 

 rature is always observed, which is sure evidence of the occurrence of some 

 more or less complex chemical change. With alcohol, scther and wood- 

 spirit, 51. Kuhlmann has long since shown that a definite crystallizable 

 compound is formed between the perchloride and its solveut. The hypo- 

 thesis therefore on which the calculation of the molecular powers was 

 based, ceases here to be admissible. — See Memoir of M. Kuhlmann, Me- 

 moires de la Societe des Sciences de Lille, vol. xvi. 1839. 



t This number is con-ected for the influence of the glass plates closing 

 the ease, and for the influence of the colour of the solution ujion the posi- 

 tion of the passage-tint. The rough difference between the two azimuths 

 of this tint, corresponding to the two opposite directions of the current, 

 wa8 9° 40'. 



