On definite Proportions. 89 
fo separate the earth from alum by adding ammonia in excess. 
But m this I could not succeed; for, when I burnt the alumina 
thus obtained, it always afforded first water, and then sulphurous 
acid and oxygen gas, Consequently alumina forms with the 
sulphuric acid an insoluble subsalt, which is but imperfectly de- 
composed by ammonia. 
I therefore took the alumina which had heen freed from sul- 
phurie acid by complete ignition, dissolved it by long digestion 
in nitric acid, and precipitated it by adding to the solution caustie 
ammonia in excess. The gelatinous earth was well washed on a 
filter, and slowly dried in the sun: when dry, it was rubbed to 
a fine powder, again digested with water, in order to separate the 
_ nitrate of ammonia, once more dried, and ignited in a small glass 
retort, At first pure water was evolved, but it was followed by 
a quantity of nitrous acid vapours, which could only be com- 
pletely expelled by a white heat. Consequently the nitric acid 
possesses the same property as the sulphuric, of affording with 
alumina a subsalt, which is not completely decomposed by am- 
monia. The subnitrate of alumina has a considerable resem- 
blance to the gelatinous silica, or to a stiff decoction of starch, 
and is easily obtained by rubbing unburnt alumina with a little 
nitric acid: the mixture exhibits after a few moments an inflated 
starch-like mass. 
I now dissolved alumina in muriatic acid, precipitated it with 
ammonia in great excess, and digested the precipitate for six 
hours with the solution, which remained strongly alkaline. The 
earth, when taken out of the filter, and well washed, was: dried 
in thesun, finely powdered, and again exposed to the sun for a 
day. When ignited in a small retort, it afforded nothing but 
water, although a small portion of the earth itself was carried up 
with the water, and was collected like a fine dust in the receiver. 
The loss was somewhat increased by this circumstance. 
This compound of waler and alumina left 64-932 per cent. 
of earth, which, being dissolved in nitric acid, showed no signs 
of sulphuric acid when examined by the test of a salt of baryta. 
Consequently 100 parts of dry alumina had been united with 54 
of water: and this water contains 47:65 parts of oxygen: the 
alumina, on the other hand, as we have seen, contains only 
46°726. I cannot insist on the perfect accuracy of either of 
these determinations ; but both of them are sufficiently near to 
the truth, to show that alumina, like the bases of salts already 
mentioned, is capable of combining with a quantity of water of 
which the oxygen is equal to that of the earth. 
Alumina, after ignition, attracts moisture very rapidly from 
the air, but retains it with a slight force; and the quantity de- 
pends on the hygrometrical state of the air, The warmth fe 
} the 
