88 On the Separation of Iron 
be added, the liquor poured again through the filter, so as to 
wash the precipitate, and the neutralization performed anew. 
The precipitation of iron above described seems at first sight 
to result from a double decomposition. Were it so, the princi- 
ple of the method would be merely a difference of solubility in 
the carbonates of iron and the other metals, and as such would 
have no claim to be regarded as rigorous. Such however is not 
the case. The iron is not separated in the state of a carbonate, 
but of a sub-salt, or a simple peroxide, the whole of the carbonie 
acid escaping with effervescence at each addition of the alkali. 
The phenomenon turns on a peculiarity in the peroxide of this 
metal, in virtue of which it is incapable of existing in a neutral 
solution at the boiling temperature. If we add an alkaline, 
earthy, or metallic carbonate by little and little to a cold solution 
of peroxide of iron, the precipitate formed is redissolved with 
effervescence, readily at first, but gradually more and more slowly, 
till at length many hours, or even days, elapse before the liquid 
becomes quite clear. Meanwhile it deepens in colour till (unless 
much diluted) it becomes dark brown or red. If the addition of 
the carbonate be carried as far as possible without producing a 
permanent precipitate, the solution is perfectly neutral, and con- 
tinues clear at a low temperature for any length of time: In this 
state it may be evaporated to dryness in vacuo, and the residue 
(which does not effervesce with acids) is still soluble in water 
without letting any iron fall, and so on as often as we please. 
The compound thus formed is however far from permanent. 
It is in fact in a state of tottering equilibrium, which a very 
slight cause is sufficient to overset. Supposing the point of sa- 
turation to have been exactly attained, the addition of an ex- 
tremely smal! quantity more of the alkaline solution is sufficient 
to determine the separation of the whole, or nearly the whole, 
metallic contents ; and if the solution operated on be pretty con-' 
centrated, it fixes after a longer or shorter time into a stiff and 
almost solid coagulum. Again, if to the coagulum so formed, a 
quantity equally inappreciable of the original ferruginous solution 
be added, it gradually liquefies, and after some time is com- 
pletely redissolved (forming no inapt representation of the cele- 
brated imposture of St. Januarius’s blood )*. 
* The phenomenon described in the text appears to me to differ from or- 
dinary precipitations and solutions, in the small proportion between the 
precipitant and the precipitate, the solvent and the matter dissolved. I 
can call to mind but one instance of so small a quantity of matter operating 
a chemical change on so large a mass, viz. the decomposition of oxygenated 
water by fibrin and other animal sul bstances. The action seems to be pro- 
pagated from particle to particle. Whether the superabundant oxide of 
iron be retained in solution in a state at all analogous to that of the oxygen 
in Thenard’s experiments, might possibly deserve considezation. 
A similar 
