August 13, 1670.] 
THE PHARMACEUTICAL JOURNAL AND TRANSACTIONS. 
127 
set aside the solution to crystallize with or without the 
previous addition of alcohol. This method is not very 
satisfactory. 
To the methods given by Rose and Deville I have 
to add two others quite different. The first and most 
important one is the last method given for preparing the 
normal carbonate. A concentrated solution of commer¬ 
cial carbonate in warm water is made in a loosely-closed 
vessel, and left to crystallize; the operation is repeated 
again and again with the mother-liquor and more of the 
carbonate of commerce, until in place of the compact 
crop of crystals at first obtained, a bulky and loosely- 
arranged crop of thin plates is produced; this and those 
succeeding it are crops of the half-acid carbonate. 
By using not too great a charge of commercial car¬ 
bonate, and allowing the crystallization to go on, in a 
place not too cold, crystals form and grow slowly to a 
considerable size; or large crystals may generally be ob¬ 
tained by decanting the mother-liquor, after twenty-four 
hours’ standing, off the first or second crop of half-acid 
carbonate, and leaving it undisturbed in a closed vessel 
for some weeks. 
A modification of this method is to heat some commer¬ 
cial carbonate in a retort with just enough water to 
cover it, at a temperature not exceeding 60°, until all is 
dissolved, then cooling the solution, and setting it aside 
to crystallize, when it yields the half-acid carbonate 
mixed with acid carbonate. 
Pelouze and Fremy state, in then’ ‘ Traite de Chimie,’* 
that the half-acid carbonate, identical with that obtained 
by Rose, may be procured by cooling a solution of com¬ 
mercial carbonate to about 0°, and this statement reap¬ 
pears in one of our standard English works on chemistry. 
It is utterly at variance with my own experiments on 
the subject. When the solution has been strong enough 
to crystallize, the crystals obtained at 0° have always 
proved to be the acid carbonate. And further, if a solu¬ 
tion which has deposited crystals at 0° is heated up with 
more of the commercial carbonate, and again cooled, the 
crystals which form are still those of the acid carbonate. 
Pelouze and Fremy are besides in this statement in 
direct opposition to H. St. Claire Deville, who, in his 
paper “ Sur la Forme et la Composition des Carbonates 
Ammoniacaux,”t says that prismatic flattened crystals of 
the acid carbonate are produced by greatly cooling a 
solution of the commercial carbonate. 
The other new process is the distillation of the am- 
monio-magnesian carbonate, the products being a fluid 
distillate, giving crystals of the half-acid carbonate, and 
a solid directly deposited in the neck of the retort of 
which the more remote parts are also this carbonate. 
Further details of the distillation wall be afterwards 
given. 
To ensure success in preparing the half-acid carbonate 
for analysis, similar precautions to those recommended 
for preparing the normal carbonate should be taken, and, 
as in the case of this salt, when the crystals are small a 
feeling of dampness in the drying-paper must be disre¬ 
garded, and the operation arrested as soon as the crystals 
cease to make wet spots on bibulous paper pressed against 
them for a few moments. 
Sensible Properties. —Half-acid ammonium carbonate 
has a very pungent ammoniacal taste and smell, but the 
pungency is less intense and persistent than that of the 
normal carbonate. 
Form .—It occurs in the form of thin, elongated, six- 
sided plates, or, when these plates are left to grow 
in their mother-liquor, of flattened, right rectangular 
prisms, terminated by the faces of a rhombic octahe¬ 
dron. 
Measurements of the angles of the crystals are given 
in Deville’s paper in the ‘ Annales de Chimie.’ 
Chemical Composition .—Results of the analyses of the 
half-acid carbonate have been published by Rose and 
* Vol. ii. p. 483. 
f jinn, de Chimie [3], vol. xl. p. 87. 
Deville in their respective memoirs already referred to. 
Rose deduced from his analysis the formula— 
(C0 2 ) 3 (0H 2 ) 5 (NH 3 ) 4 , 
and Deville has adopted the same formula. My own 
experiments have led me to adopt a different formula, 
and the results of my analyses point unmistakably to 
the formula having four atoms of water. It must be 
taken into consideration that in no case can the samples 
analysed be quite dry, as the decomposition by which 
ammonia is liberated also sets free water; because the 
presence of this water must lower the numbers found for 
both the carbonic anhydride and the ammonia, thus 
making the numbers for the former correspond closely 
to the calculated number, in spite of the decomposition 
which the salt has undergone, and those for the latter 
show a greater deviation from the calculated number 
than the actual loss of the ammonia itself would cause. 
The results of Rose’s analyses correspond closely with 
the numbers calculated for a salt containing five atoms 
of water:— 
Pose's Analyses. 
I. II. III. IV. 
Carbonic anhydride . — 45’35 44-61 44-69 
Ammonia. 23-69 — — — 
As I have already said, when speaking of the normal 
carbonate, the accuracy of the results of an analysis of 
such changeable salts depends very much upon the suc¬ 
cess in preparing them for analysis. Rose makes no 
mention of the method he adopted for getting the salt in 
the dry state undecomposed, or of his success in doing 
so; yet it will be seen, on examining his results, that the 
ammonia is slightly in excess of, while the carbonic an¬ 
hydride is actually less than, the calculated quantity,—a 
somewhat anomalous result, admitting the accuracy of 
his analyses. 
If it be assumed that the salt analysed, which was in 
very thin plates, retained some of the mother-liquor, this 
difficulty respecting the results will be removed; for the 
mother-liquor from the crystals was itself a solution of 
normal carbonate, according to Rose. This assumption 
we are justified in making; for it has been shown by 
Deville’s results that other analyses recorded in Rose’s 
paper prove that the crystals employed must have con¬ 
tained mother-liquor in their interstices. And in further 
support of this assumption, we have also Deville’s testi¬ 
mony, that the half-acid carbonate can hardly be ob¬ 
tained at once dry and undecomposed. 
Turning now to Deville’s deduction of the same for¬ 
mula as Rose’s, from the results he himself obtained, we 
shall not find it, I think, more satisfactory than that of 
the German chemist. He gives the annexed table of his 
analyses, the samples—none of them dry—being drier in 
the order of their numbers. (The water was determined 
directly.) 
Deville's Analyses. 
I. 
II. 
III. 
IV. 
Mean. 
Carbonic anhydride 
. 41-4 
46-8 
46-6 
47-8 
4o"65 
Ammonia.... 
. 21-1 
23-7 
24-2 
23-6 
23-15 
Water. 
. 37-5 
29-5 
29-2 
28-6 
31-20 
The practical accuracy of these analyses, as such, can¬ 
not be impugned ; they were performed with great pre¬ 
caution to avoid sources of error, and the three consti¬ 
tuents were all determined by a single operation in each 
case. Now, their mean results correspond closely enough 
with the percentage numbers for the five-atoms-of-water 
formula. But then we are certainly not right in taking 
these mean results as a true indication of the composition 
of the salt. We can only properly take the mean results 
of a series of analyses as likely to be more accurate than 
the results of a single analysis, after we have satisfied 
ourselves that the unavoidable and undetected imperfec¬ 
tions in the performance of the operations are as likely 
to have influenced each result in one direction as in the 
opposite. With the present series of analyses, we know 
h 3 
