Behaviour of Cliemical Compounds. 171 



for the investioation of the behaviour of two salt-solutions 

 with double decomposition. An aliquot part of the salt- 

 solution the magnetism of which was investigated in the flask 

 of the torsion-a})paratus was put into the flask with a quantity 

 of potash or soda lye, or liquid ammonia, sufticient to produce 

 precipitation ; Avith the latter the flask was filled up to the 

 mark, and after agitation, before deposition, the magnetism of 

 the hydrate determined. The calculation of the magnetism of 

 the precipitated oxide took place as above, except only that 

 from the magnetism of the precipitated solution that of the 

 glass and the water, or of the soda or potash lye or the am- 

 monia (which difl'ered but very slightly from that of the 

 water), was deducted. 



If fis denotes the molecular magnetism of the dissolved 

 salts of the magnetic metals, then the molecular magnetism 

 /i.^ of the corresponding hydrates is as follows : — 



Manganous hydrate /i,/4 = 0'85/a^ 



Ferrous hydrate „ 1*12 



Cobaltous hydrate „ 1*12 



Nickelous hydrate „ 1*00 



Cupric hydrate „ 0'74 



Chromic hydrate „ 0'95 



Ferric hydrate „ 0'69-l-18 



The molecular magnetism of the ferric hydrate changes, 

 after the addition of ammonia to a solution of ferric chloride, 

 very rapidly from the smaller value to the greater — of course 

 in consequence of the oxide passing out of the colloid into the 

 ordinary state {vide i7ifra). 



According to this the 7)iolecular magnetisms of inost of tlie 

 hydrates of the peroxides are partly but little less or greater, 

 partly very approximately the same as those of the correspoiid- 

 ing salts in their solutions. Hence, taking into consideration 

 the altered density-ratios, we may perhaps assume that these 

 values would, under otherwise similar circumstances, be equal 

 to one another ; and accordingly the magnetic groups of atoms 

 would also remain unchanged the same in both series of 

 compounds. 



10. Quite different from this is the magnetism of colloid 

 solutions of the oxides. A solution of colloid ferric oxide, ob- 

 tained by dialysis of a solution of ferric chloride digested with 

 ferric hydrate, shows, in proportion to its content of iron, a 

 much feebler magnetism than a neutral and concentrated 

 solution of ferric chloride or one mixed with a large quantity 

 of acid. The molecular magnetism of the colloid solution of 

 ferric oxide is accordingly only about 0*21 of that of the 

 ferric salts. 



