218 EXPERIMENT STATION RECORD. 



boiling the silicic-acid mixture with a solution of sodium carbonate or of 

 sodium hj'droxid, neither of which attack the sand but dissolve the 

 amorphous silicic acid derived from the substance itself. To this end 

 the silicic-acid residue after weighing is boiled with a concentrated 

 solution of sodium carbonate, or, according toFittbogen/ with a 5 to 6 

 per cent solution of sodium hydroxid, in a platinum dish with contin- 

 uous stirring (to avoid bumping), filtered, washed, and the residue 

 ("sand") weighed. The latter is subtracted from the crude ash in order 

 to obtain the pure ash. 



The results of this troublesome operation, however, are not always 

 satisfactor3^ The difficulties of separating quartz and sand from 

 amorphous silicic acid have been well illustrated in mineral analysis.^ 

 It is consequently preferable to free the materials to be incinerated 

 from sand and earth as far as possible, in order that the determination 

 and deduction of sand ma}^ be avoided. 



ANALYSIS OF PURE ASH. 



So much has been written upon the determination of the several ash 

 elements that it will be impossible to review the entire subject here. 

 Only brief mention can be made regarding the shortest means for 

 estimating the principal constituents. 



While the whole series of mineral elements which occur in plants 

 may be present in the ash, in most cases only the more important ones 

 are considered, and among these the phosphoric acid in particular 

 gives rise to some analytical difiiculties. 



It is convenient to use 3 aliquots for the analysis, made up as 

 described above, determining the iron and alumina, lime and magnesia 

 in the first, the sulphuric acid in the second, and the potash and soda 

 in the third; a separate solution for the phosphoric acid and chlorin 

 being made up with nitric acid. 



In the first aliquot the iron and alumina are precipitated with ammo- 

 nia and acetic acid, the precipitate being mainly neutral iron phos- 

 phate, as aluminum phosphate occurs but rarely in important amounts. 

 If the above is not satisfactory, the precipitate may be dissolved and 

 the constituents determined separately — the phosphoric acid with 

 molybdic acid, the iron by titration with potassium permanganate, 

 and the aluminum by difference or by the method of Berthelot and 

 Andre* of precipitating alumina as phosphate with sodium h3^posul- 

 phite. The lime is determined in the filtrate from the iron phosphate 

 precipitate b}^ means of ammonium oxalate, and the magnesia in the 



^Landw. Vers. Stat., 13 (1871), p. 114. 



'' Michaelis, Ber. Deut. Chem. Gesell., 28 (1895), Ref. p. 1020; 29 (1896), Ref. p. 562; 

 Lunge, Ibid., 28 (1895), Eef. p. 1020; 29 (1896), Ref. p. 188. 



^Ztschr. Analyt. Chem., 35 (1896), p. 630; Ann. Chim. et Phys., 7. ser., 5 (1895), 

 p. 429. 



