some varieties of South American Guano. 125 



dish for twenty-four hours : this precipitate, collected, washed 

 and dried, weighed 98'07-grs., indicating an excess of -012 in 

 estimating ammonia, merely an error of experiment to this 

 trifling amount. The solution when gently heated evolved a 

 little carbonic acid, but did not deposit any carbonate of ba- 

 rytes. Satisfied by this experiment that the acid, as well as the 

 neutral carbonates of ammonia, precipitate a barytic salt equi- 

 atomically as respects the bases, and therefore that the pro- 

 posed plan of indirect estimation of ammonia was an accurate 

 one, I subjected 5\> grs. of nmriate of ammonia to distillation 

 with an excess of a solution of the ordinary carbonate of soda, 

 carrying on the distillation until about one-half of the liquid 

 had passed into the receiver, in which distilled water was of 

 course placed; long before this quantity of liquid was distilled 

 a slip of turmeric paper suspended from the tubulure of the 

 retort ceased to be reddened ; the contents of the receiver, a 

 solution of a basic carbonate of ammonia, were added to a so- 

 lution of nitrate of barytes, and then subjected to a stream of 

 carbonic acid gas until the precipitate became arenaceous and 

 the solution slightly affected litmus paper; this exposed to the 

 air for a day, and the precipitate collected, and dried, gave 97*63 

 grs. of carbonate of barytes. As in the former case, the fil- 

 tered solution when gently heated gave off a little carbonic acid 

 gas, but deposited nothing. These results being so favourable, 

 and the process simple, the trifling deficiency evidently being 

 traceable to defective manipulation alone, I have adopted this 

 process for estimating ammonia in the analyses subsequently 

 detailed, considering it susceptible of greater accuracy than 

 the hydrochloric acid method, and as accurate as, and far less 

 troublesome and expensive than, the employment of salts of 

 platinum, and of alcohol for this purpose. 



The process for separating the phosphate of lime from the 

 phosphate of magnesia depends upon the insolubility of the 

 oxalate of lime in very dilute acetic acid, and the solubility of 

 the magnesian salt in that menstruum. Thus, when a solu- 

 tion of these mixed phosphates in hydrochloric acid is treated 

 with ammonia slightly in excess to precipitate the charac- 

 teristic gelatinous precipitate, and acetic acid is gradually 

 added till such precipitate is redissolved, the solution being 

 acid to litmus paper, but only faintly so to the taste, oxalate 

 of ammonia produces the characteristic precipitate of oxalate 

 of lime, which base is completely thrown down ; the filtered 

 solution, when evaporated and treated with a little phosphate 

 of soda and ammonia, affording the ammoniacal magnesian 

 phosphate, which contains no lime. I do not give this pro- 

 cess as one which is necessarily rigidly accurate, for if the 



