Traiisparent Bodies by the Action of Magnetism. 425 



numerous iron -based glasses examined by M. Matthiessen*, 

 some might be found of a suitable nature, but I have not been 

 able to obtain permission to examine them. 



No better success attended me in seeking to prepare an easily 

 fusible compound of iron of sufficient transparence in the liquid 

 state for the study of its action on polarized light, without the 

 necessity of dissolving it. The properties of the double chlo- 

 ride of aluminium and sodium induced me to imagine that the 

 corresponding chloride of iron and sodium, if it existed, might be 

 suitable. This compound actually exists, and may be easily pre- 

 pared by heating together 33 parts of perchloride of iron and 12 

 parts of chloride of sodium ; it melts easily over an alcohol lamp, 

 but in the liquid state is quite opake. 



Nickel. — All the salts of nickel which I have tried have a 

 positive magneto-rotary power, so that their solutions exert a 

 greater action on polarized light than that of the water which 

 they contain. This positive rotary power is sufficiently marked, 

 and comparable with that of zinc and tin salts. It is particularly 

 necessary, in determining the power of nickel, to take account 

 of the influence which the colour of the emergent light exerts 

 upon the position of the passage-tint. Thus, in light which has 

 passed through 40 millims. of a moderately strong solution of 

 chloride of nickel, the red, orange, violet, and indigo are almost 

 extinguished, the blue and the yellow are considerably weakened, 

 and the maximum intensity corresponds with the green rays. 

 Hence it follows that the passage-tint is much more deviated 

 than if the light passed without alteration through the chloride. 



Cobalt. — The magneto-rotary power of salts of cobalt is posi- 

 tive, but weaker than that of nickel compounds, and rather diffi- 

 cult to be shown, because no considerable quantity of cobalt salts 

 can be dissolved in water without greatly diminishing the trans- 

 parence of the liquid. The colour of the emergent light exerts 

 an opposite action on the position of the passage-tint to that 

 which occurs in the case of nickel salts, lied being the domi- 

 nant colour, the deviation of the passage-tint is diminished in 

 such a manner, that if the necessary correction were neglected, 

 the magneto-rotary power of cobalt salts would appear negative 

 and very weak. 



Manganese. — Salts of the protoxide of manganese have an in- 

 considerable positive magneto-rotary power; but, as their solu- 

 tions are quite colourless, the action is very easil\ recognized. 



Salts of the sesquioxide of manganese have so high a colouring 

 power that they cannot be employed in experiment. But in the 

 laboratory of the College de France, I found a compound pro- 

 bably corresponding to these salts, namely, the double cyanide 



* Comptes Rendus (les Seances de VAcademie des Sciences, vols. xxiv. 

 and XXV. 



Phil. Mag. S. 4. Vol. 15. No. 102. .June 1858. 2 F 



