92 On definite Preportions. ’ 
ammonia. All these solytions afforded however a mixture of 
hydrate and subsalt, from which heat expelled first some water, 
and then some of the acid. The ignited hydrate lost in eis 
experiments from 27 to 15°5 per cent. 
I next digested the precipitate obtained from the nitrate with 
a great excess of caustic ammonia. The hydrate, when washed, 
and dried in the sunshine, now lost 22-15 per cent.: the fluid 
which escaped, was not pure water, but a strongly ammoniacal 
fluid. I had not therefore yet obtained a pure hydrate. 
Two years and a half before, when I was examining the iron 
which contained silicium, I had left about 20 gramines of this 
iron, moistened with water, to oxidate spontaneously. But the 
mass assumed by degrees a solid consisteney, and even after the 
expiration of this time, was not completely oxidated throughout 
its substance. Not being able therefore to employ it for the 
analysis of silica, as I originally intended, I collected and dried 
a quantity of the yellow ochre that was formed on it: but acci- 
dentally this substance was placed, with others that were to be 
dried, ina sand heat, and here in all probability lost a part of 
the water that was chemically combined with it ; for the hydrates 
of iron support but little heat, without assuming a darker colour, 
and losing part of their water. By ignition, its weight was 
further diminished 10 per cent. The red oxide that was left 
contained 8*2 per cent. of silica. I was therefore working again 
upen a triple combination of silica, oxide of iron, and water, 
The quantity, which I had to employ for this experiment, was 
too small; and was already consumed, so that I could not ex- 
amine this preparation of iron with greater accuracy: but it was 
evidently an artificial combination resembling the foreign ore 
which I have already mentioned. This ove will undoubtedly be- 
come the subject of further investigation, and the comparative 
analysis of the artificial triple combination, obtained by the oxi- 
dation of iron containing silicium, will perhaps be indispensable ; 
for the ore is mixed in various and inconstant proportions with 
dust, and with different earthy substances, so that it can afford 
no certain results. 3 
{t is also obvious from what has been related, that it it diffi- 
cult to obtain a perfectly pure hydrate of iron, since the oxide 
combines, at the moment of precipitation, as well with acids as 
with ammonia, according to the degree of excess in which this 
alkali is present. 1 ther efore allowed some iron filings to be oxi- 
dated in pure water, which I changed daily, and collected the 
hydrate. After some weeks I had obtained enough for an ex- 
periment on a small scale. I dried it in the sunshine for several 
days, and then ignited it in a platina crucible. It left 85:2 per 
cent. of red oxide, which was partially magnetic ; a circumstance 
which 
