272 Messrs. Rowney and How's Analysis of 



The same quantity of ash served to determine the potash 

 and soda, sulphuric and phosphoric acids, perphosphate of 

 iron, lime and magnesia, silicic acid and accidental sand and 

 charcoal. For this purpose, Jthe hydrochloric acid solution 

 was evaporated to dryness, gently ignited and extracted with 

 hydrochloric acid. The solution thus obtained was divided 

 into different parts. The first portion served for the determi- 

 nation of the potash and soda. 



For this purpose the acids, lime, magnesia, &c. were re- 

 moved by baryta, the excess of baryta by carbonate of am- 

 monia, and the ammoniacal salts by gentle ignition. The 

 residue, potash and soda, were estimated partly by separating 

 them by means of bichloride of platinum (analyses of the 

 ashes of the root and seed) and partly by the indirect method, 

 namely, by converting the mixed chlorides into sulphates, 

 weighing these and ascertaining the amount of sulphuric acid 

 by means of chloride of barium (analyses of the stem, leaves 

 and fruit). 



In the second portion, sulphuric and phosphoric acids were 

 determined, the former as sulphate of baryta, the latter by 

 neutralizing the filtrate from the former with ammonia and 

 precipitating the phosphoric acid by means of sesquichloride 

 of iron and acetate of potash. This precipitate was dissolved 

 in hydrochloric acid, a sufficient quantity of tartaric acid was 

 added, and the phosphoric acid estimated in the form of py- 

 rophosphate of magnesia, by precipitating with ammonia, 

 chloride of ammonium and sulphate of magnesia. The latter 

 precipitate, frequently containing a small quantity of iron, 

 was redissolved in hydrochloric acid, and after the addition 

 of some tartaric acid reprecipitated by ammonia. A third 

 portion served for the estimation of perphosphate of iron, lime 

 and magnesia. For this purpose the liquid was neutralized 

 with ammonia, some acetate of potash was added, and the 

 solution strongly acidulated with acetic acid, in order to keep 

 the phosphate of lime, which might be precipitated, in solu- 

 tion; on heating perphosphate of iron subsides, from which 

 the sesquioxide of iron was calculated according to the for- 

 mula 2Fe2 0g + 3P05. From the filtrate the lime was preci- 

 pitated by means of oxalate of ammonia, and after the sepa- 

 ration of the lime, the magnesia by means of phosphate of 

 soda. Chlorine and carbonic acid were determined in sepa- 

 rate portions of the ash. In this manner the following expe- 

 rimental numbers were obtained : — 



