510 PHARMACEUTICAL MEETING. 
The percentage composition of the Pharmacopoeia salt containing— 
3 NH 4 0 = 40*20 
r0 5 = 36-59 
5 HO = 23*21 
100-00 
it will be seen therefore, as before stated, that the ammonia is a trifle lower, 
and the acid higher, in a proportionate degree ; moreover, it would have taken 
a great deal of trouble to have prevented this loss, which the importance of the 
subject scarcely warranted, even if it could have been done; otherwise these 
percentage compositions would without doubt have absolutely agreed. Only 
in one case does the difference amount to more than 1^- per cent., and a loss 
of only one milligramme (*001) in the substance operated on becomes increased 
to nearly one per cent, when multiplied so many times in order to obtain the 
percentage. It is obvious then from these analyses that the phosphate of the 
Pharmacopoeia is decidedly a triammonic salt, or one which contains 3 atoms of 
ammonia united with 1 atom of phosphoric acid; and it is also a curious fact 
that no analysis of such salt has been published, possibly from the errors which 
arise from the loss of ammonia. Mitsclierlick, who examined these phosphates, 
published results only of the di- and monammonic varieties; and Gmelin, quoting 
Berzelius as his authority, mentions concerning the triammonic, merely that it 
is precipitated from a solution of the disalt, by mixing it with excess of am¬ 
monia, without saying anything about the possibility of its being redissolved 
and crystallized. Prom the words he uses, I conclude he thought it impracticable. 
In analysing these phosphates it is impossible to ascertain whether the sample 
be tri- or diammonic by estimating merely the ammonia, since both salts con¬ 
tain it in almost the same proportion as regards weight, owing to the fact that 
one contains 5 atoms of water of crystallization, the other none; by estimating the 
phosphoric acid instead, we of course immediately see which sample we have in 
hand. It was mentioned the other evening as being rather an unlikely fact, 
that the salt would contain an uneven number of atoms of water of crystalliza¬ 
tion. Now I did not undertake so much to establish a formula for the salt, as to 
ascertain whether, when made according as the Pharmacopoeia directed, it con¬ 
tained 3 atoms of ammonia or no. Moreover, the analysis would have to be con¬ 
ducted differently to speak definitively upon this latter point, as the water 
must be directly determined ; still the reasons advanced are, I think, scarcely 
tenable, inasmuch as we have a very great number of salts which do contain 
uneven atoms, and as regards the possibility of writing the formula with the 
double oxygen atom, it would stand 2 [(NH 4 ) 3 , P 0 4 ] -f-5 II 2 0,—a formula truly 
somewhat complex, yet nevertheless correct. I next analysed some commercial 
specimens of the phosphate, to see the composition of the salt in general use. 
I confess I did not expect to find 3 atoms of base, nor was I disappointed, for 
the mean of several experiments gave 53-29 per cent, of P0 5 , the percentage 
composition of the diammonic salt containing 53*79, P0 5 ; the formula of 
these salts exactly coincided. As the specimens examined were purchased since 
the British Pharmacopoeia came out, it is obvious that either the salt was not 
made according to its directions, or else that it was old stock in hand previous 
to its appearance. It is scarcely likely, I think, that a salt made with 3 atoms 
of NII 4 0, unless very carelessly kept, would become so entirely converted into 
that containing only 2 atoms, as to give a percentage agreeing so exactly with 
the calculated one. 
Seeing therefore the difference in composition between the commercial and 
officinal varieties, the next point to be considered is which is the best salt for 
medicinal use, and which can be most readily and economically prepared. The 
