92 BOTANICAL GAZETTE [February 



same quantity of the same add was added to the blank determina- 

 tion, so that subtracting the blank subtracted the sulphur added 

 in the sulphuric acid as well as that present in the reagents. In 

 every case the lo cc. was measured between the lo and 20 marks 

 on the burette. According to Koch, barium sulphate does not 

 precipitate readily when the concentration of the SO4 ion is low. 

 The addition of the sulphuric acid is then necessary to bring the 

 concentration of the SO4 ion up to the point where precipitation 

 takes place readily. The method as finally adopted is as follows: 

 The equivalent of 10 gm. of oven dry soil was weighed into a 

 nickel crucible, moistened with a few drops of distilled water, and 

 part of a weighed 20 gm. of sodium peroxide stirred in a little at 

 a time with a nickel rod. (If the moisture was just right, reaction 

 took place immediately without the application of heat, and the 

 charge was fairly dry by the time most of the sodium peroxide had 

 been stirred in. If too little water had been added, it was neces- 

 sary to heat with an alcohol lamp to start the reaction. If too 

 much water was added, it was necessary to heat with the alcohol 

 lamp to bring to the desired degree qf dryness before adding the 

 last of the sodium peroxide.) After the charge was fairly dry, the 

 rest of the sodium peroxide was placed over the charge, the crucible 

 covered, and heated over a bunsen burner, raising the temperature 

 gradually to a fairly high temperature which was maintained for an 

 hour. After cooling, the fused mass was removed with hot dis- 

 tilled water to a 600 cc. beaker, neutralized with concentrated 

 HCl, and then 10 cc. additional concentrated HCl added. The 

 beaker was then heated for five or six hours on the steam bath with 

 occasional stirring. It was then transferred to a 500 cc. flask, 

 covered, and made up to the mark. The solution was shaken 

 frequently for several hours and the 250 cc. filtered off. The 

 250 cc. of filtrate was transferred to a 600 cc. beaker, heated on 

 the steam bath, and the iron, aluminium, and silica precipitated 

 with ammonium hydroxide, allowed to stand a few minutes, and 

 then filtered into a one liter beaker. The precipitate was washed 

 with hot distilled water until the combined filtrate and washings 

 had a volume of approximately 600 cc. Exactly 10 cc. of approx- 

 imately N/io H2SO4 was then added, heated to boiUng, and 10 cc. 



