382 Experiments and Observations, &c. 
there is more loss of acid than alkali, in all 
probability for the reason already assigned, 
namely, the affinity of water for the alkali 
but not for the acid. 
I observed in one experiment made to as- 
certain whether subcarbonate of ammonia in 
the dry state attracts carbonic acid, that the 
‘common opinion on this head is erroneous. 
A quantity of subcarbonate of ammonia was 
exposed over mercury to carbonic acid gas. 
It seemed to have no effect at first; but in 
the course of a few hours the gas was all ab- — 
sorbed by the subcarbonate.* 
* The common subcarbonate of ammonia of the shops, 
when taken from the middle of a transparent hard cake, is, 
I find, tolerably uniform in its composition; it consists of, 
59 carb. acid © 
24.5 ammonia | 
16.5 water 
100 
consequently it may be accounted a compound of 1 atom 
subcarbonate with 2 atoms of carbonate and water. 
The simple carbonate consists of 1 atom acid, 1 ammo- 
nia (6), and 1 water: or, 
58 acid 
18 ammonia 
24 water 
100 
but it generally contains rather less acid and more alkali. 
Erratum. Page 23, line 22, for 7.4 read 8.5. 
