280 Prof. G. Wiedemann on the Marjnetlc 



ferricyanides (tc. is afforded by tlie experiments described at 

 p. 170j and by otliers similar to them. By those experiments 

 it is proved that when a solution of these salts is mixed with 

 the solution of a salt of iron, nickelj cobalt, or manganese, and 

 thereby the potassium in them is partially or entirely replaced 

 by the magnetic metals of the latter, the molecular magnetism 

 is augmented by the magnetism of those metals — as, e. g., on 

 the replacement of the potassium in the chloride of potassium 

 by the same metals, in which the magnetism of the potassium 

 is a vanishing quantity against that of the latter. The electro- 

 lytic behaviour shows that the ferro- and ferricyanides and the 

 analogous salts of manganese and cobalt are to be regarded 

 as compounds in which the one constituent {e. g. the potassium) 

 is electrolytically opposed to the radical combined with it 

 (ferrocyanogen &c.) in like manner as the chlorine in potas- 

 sium chloride. It results from the magnetic investigation that, 

 even in the undecomposed salts, the potassium in the ferro- 

 cyanide &c. possesses altogether the same properties as in the 

 chloride of potassium &c. The constitution of both series of 

 salts is in this respect the same ; the former are therefore, like 

 the latter, to be considered, even in the undecomposed state, as 

 binary compounds. 



It is remarkable that also the combined cyanides of nickel 

 and potassium (NiCy2 + 2 KCy), which possesses a composi- 

 tion deviating from the constitution of the ferro cyan ides, is 

 feebly diamagnetic. Consequently this salt is not to be ima- 

 gined a double salt ; much rather, as results from its electro- 

 lysis, may the potassium be opposed as electropositive radical to 

 the total remaining complex of atoms as electronegative radical. 



A quite different behaviour is exhibited, according to more 

 recent experiments, by the so-called potassium cliromicyanide 

 (CrCy3 + 3KCy), although it appears to be composed after 

 the analogy of the potassium ferri cyanide &c. ; and it is just 

 the same with the potassium cliromisulplio cyanide. If we de- 

 note by Gr the weight of the magnetic metal contained in the 

 solutions of these salts (which was partly determined directly, 

 and partly calculated from the weight of the salt present in 

 the solution), by M the magnetism of the solutions after de- 

 ducting the (negative) magnetism of the glass filled with 

 water, we obtain the magnetism a for each atom of the mag- 

 netic radical for : — 



G. M. 



Potassmm chromicyanide ... 0*8273 0-3178 42*0 



„ ferricyanide 0*9734 0*1329 15*9 



Ferric nitrate 0*5286 0*4532 100*0 



