96 Mr, J. F, W. Herschel on the [Feb. 



like a large scale, it might hitherto be regarded as insuperable. 

 In consequence of this, and of the importance of the inquiry, 

 there is hardly a chemist of eminence who has not proposed 

 some process for the purpose but (with the exception of that 

 which depends on the insolubility of the persuccinate of the 

 obnoxious metal, which 1 have not tried, and which is too 

 expensive to be resorted to for any but the nicer purposes of 

 /analytical research), they are all of them either inadequate to 

 the end proposed, intolerably tedious, or limited in their appli- 

 cation. That which I have now to propose, on the other hand, 

 is liable to none of these objections, being mathematically rigo- 

 rous, of general application, and possessing in the highest degree 

 the advantages of facihty, celerity, and cheapness. It is briefly 

 this: 



The solution containing iron is to be brought to the maximum 

 of oxidation, which can be communicated to it by boiling with 

 nitric acid. It is then to be just neutralized while in a state of 

 ebullition^ by carbonate of ammonia. The whole of the iron to 

 the last atom, is precipitated, and the whole of the other metals 

 present (which I suppose to be manganese, cerium, nickel, and 

 cobalt), remains in solution. 



The precautions necessary to ensure success in this process 

 are few and simple. In the first place, the solution must con- 

 tain no oxide of manganese or cerium above the first degree of 

 oxidation, otherwise it will be separated with the iron. It is 

 scarcely probable in ordinary cases that any such should be pre- 

 sent, the protoxides only of these metals forming salts of any 

 stability ; but should they be suspected, a short ebullition with 

 a little sugar will reduce them to the minimum. If nitric acid 

 be now added, the iron alone is peroxidized, the other oxides 

 remaining at the minimum.* Moreover, in performing the pre- 

 cipitation, the metallic solution should not be too concentrated, 

 and must be agitated the whole time, especially towards the end 

 of the process ; and when the acid reaction is so far diminished 

 that log-wood paper is but feebly affected by it, the alkaline 

 solution must be added cautiously, in small quantities at a 

 time, and in a diluted state. If too much alkali be added, a 

 drop or two of any acid will set all right again ; but it should be 

 well observed, as upon this the whole rigoitr of the process 

 depends, that no inconvenience can arise from slightly surpass- 

 ing the point of precise neutralization, as the newly precipitated 

 carbonates of the above enumerated metals are readily soluble, to 

 a certain extent, in the solutions in tvhich they are formed (though 



• Dr. Forchhammer, in a paper recently published in Thomson's Annals of Philo. 

 sophy^ contends that the proto-salts of manganese are absolutely void of colour. To this 

 I can only say, that I have not succeeded in depriving the muriate of its pale rose colour 

 by any length of ebullition with sugar or alcohol, after which, however, not a trace of 

 deutoxide could be detected in it. I cannot help regarding the process here proposecl 

 for freeing manganese from iron as preferable to that of Dr. F. 



