90 On definite Proportions. 
the sun is sufficient to expel the greatest part of this water; and 
if we heat the alumina on a sand-bath, the whole of it is driven 
off ; while the water in the true hydrate requires a red heat in 
order to be expelled. After some days, 100 parts of burnt 
alumina had absorbed 34°5 of water in an atmosphere saturated 
with moisture, hygr. 100°. Of this water it lost in a few days 
18:5 parts, when the hygrometer stood at 7°, and the thermo- 
meter from 22° to 25° [72° to 77°]; and its weight then re- 
mained stationary. It may be questioned in what state of com- 
bination this water was retained: certainly not in the same as in 
the hydrate, 
8. Silica. 
The silica, which is separated by acids from the liquor silicum, 
contains, after being dried, as is well known, a considerable 
quantity of water. I found that it made no difference whether 
I employed silica that was precipitated at once, or that was first 
separated in a gelatinous form. 
Three portions of silica, which | had cbtained by different 
analytical operations, were dried in the same saucer on a sand- 
bath. When I ignited them after some hours’ exposure to this 
heat, they all lost a quantity of water, which varied from 11-2 to 
11-3 per cent. 
The experiments upon the ignited alumina, and others upon 
the oxide of tin, which I shall hereafter relate, determined me 
to repeat these experiments at a subsequent time. I found that 
the oxide of tin, when dried in different temperatures, retained 
different proportions of water. J therefore weighed a portion of 
silica dried in the open air, and dried it again in a sand heat: 
the loss amounted to 26°8 per cent. Being left in the scale, its 
weight increased by degrees. I then dried it again very thoroughly, 
and ignited it; and I found the loss 14°2 per cent. It formed 
little transparent grains, which lost nothing of their transparency 
nor of their form by ignition. It seems, therefore, as if the water 
retained by the silica were precisely in the same state as that 
which ignited alumina absorbs from the air. 
_ | had entertained hopes of being able to calculate the quan- 
tity of oxygen in silica, from that of the water retained in 
eorilatiation with it in a dry form; and the agreement of the 
three first experiments with each other made me suppose, that 
silica probably contains four times as much oxygen as the water 
combined with it ; and in this case the quantity would be 45 per 
cent. But I no longer place any dependence on this mode of 
determination. It may perhaps be possible, at a future time, to 
ascertain the composition of silica from its combinations with the 
fluoric acid, or with the alkalis and the earths. 
4, Oxide 
