18 BULLETIN 600, U. S. DEPARTMENT OF AGRICULTURE. 
After a number of trials it became evident that the method which 
calls for the separation of calcium as sulphate by alcohol, after the 
removal of silica, was the best available. Test analyses by two 
analysts on-syuthetic mixtures showed this method to be accurate. 
It is outlined in Mellor’s, “A Treatise on Quantitative Inorganic 
Analysis,” p. 606, J. B. Lippincott Co., Philadelphia (1913). 
This method was found satisfactory after modification; that is, 
the removal of the iron precipitated with the calcium sulphate and 
the recovery of magnesium from the iron group after the precipi- 
tation of phosphoric acid, by the use of citric acid and ammonium 
phosphate. : 
SILICA. 
It is impossible to prepare plants for analysis in such a manner 
as to have them free from contaminating soil, which is made up 
largely of silica. An approximate way of determining silica is car- 
ried out by digesting the dehydrated silica and sand with a solution 
of sodium carbonate and hydroxide as outlined in Bul. 107, of the 
Bureau of Chemistry. The finely divided silica of constitution 
dissolves, and the soil grains are supposed to resist attack. It is 
difficult to obtain results by this method which agree well, particu- 
larly when working with plants high in silica. For the separation 
of silica, two evaporations with an intervening filtration were in- 
variably made. 
THE CALCIUM PRECIPITATION. 
The calcium sulphate precipitated by alcohol frequently contains 
iron but never phosphorus, provided, of course, the washing has 
been thorough. It has been our practice to dissolve the calcium 
sulphate in hydrochloric acid and precipitate the iron, if any, with 
ammonia and add it to the main precipitate. When the final volume 
of the alcoholic solution in which the calcium-sulphate precipitation 
was made is from 150 «. ¢. to 200 ¢. c. from one to two milligrams of 
calcium sulphate remain in solution. This was shown by the test 
analysis of synthetic test solutions. 
THE IRON AND ALUMINUM. GROUP. 
The skill of the analyst is put to a severer test in making this 
group separation than at any other stage of the analysis. The pre- — 
cipitation must be quickly and completely made and the filtration 
hastened with all possible dispatch. This is to avoid the coprecipita- 
tion of magnesium phosphate. To this end a small but known excess 
of ammonia is quickly added to the hot mixture. The solution is 
then boiled a minute or so and the filtration made by suction in a 
