172 On definite Proportions. 
haps from the same causes of error which affected Berzelius’s first 
experiment. Gilbert.| MM. Thenard and Reard employed 489 
grammes of alum, from which they obtained 61 or 62 of alumina, 
a quantity which, when moist, would occupy the bulk of ten 
pounds of water, and which must have been received either on a 
very large filter or on several small ones : in both cases, the wash- 
ing it, and its separation from the paper, must have been sub- 
jected to great difficulties. It is also probable, that the separa- 
tion of the sulphate of potass from the fluid obtained, by the ad- 
dition of lime, could give no very correct result, Probably there- 
fore the difference of our determinations is rather to be attributed 
to the more or less appropriate methods employed than to the 
experimenters. Besides, it is certain that an analysis on too. 
large a scale can never afford a very correct result; nor do these 
gentlemen appear to have been in pursuit of very minute ac-. 
curacy in their experiments. 
A double Subsalt. 
There exist also some double subsalts; but I have hitherto 
examined only one, that is, the combination known in pharmacy 
under the name of cuprum ammoniatum. In order to prepare 
this salt, I dissolved some sulphate of the oxide of copper in 
caustic ammonia, precipitated the double salt with alcohol, washed 
it again with alcohol, and dried it in the air: It is very difficult 
to observe with accuracy the moment of the attainment of per- 
fect dryness; for the salt is decomposed on the surface, before 
the alcohol has been expelled from4ts internal parts; it then be- 
comes by degrees of a sky-blue colour, and at the edge green. 
Hence it is impossible to obtain a very correct analysis of this 
salt; but it will not be difficult, with the assistance of the laws 
of combination which have been here developed, to diseover its 
true composition, since the result of the analysis cannot deviate 
far from the truth. 
I drove off the superfluous ammonia from a part ef this salt, 
on a sand-bath, until it became quite gray; it had lost 20-33 
per cent. in weight. When I repeated the experiment in a smalh 
retort, I found that a little water escaped at the same time. 
The gray powder, when water was poured on it, became first of 
a light blue, and then, as I gently heated it, of a dark-brown 
colour. This unquestionably depended on the decomposition of 
the salt, and the formation of a portion of neutral sulphate of 
ammonia and copper, while some hydrate of copper was first 
produced, and then decomposed by the application of heat. The 
solution had a faint blue colour, and left 40 per cent. of sulphate 
of ammonia, mixed with a small quantity of the double salt. 
The part undissolved, being a mixture of black and green oxide, 
amounted 
é 
