136 Hartung and Rivett .- 



plished by sulphuric acid and potassium sulphate. By '■ free 

 nitrogen " is to be understood that nitrogen which may be distilled 

 (as ammonia or substituted ammonias) from a sample of earth by 

 boiling with sodium hydroxide solution for one hour. It cannot be 

 taken as giving a measure of the ammonium salts present before 

 combustion, because even after prolonged boiling the steam dis- 

 tilling was found to be alkaline. Doubtless there is a progressive 

 action between the alkali and the nitrogenous matter; to limit the 

 time of distillation to one hour is arbitrary. The estimations of 

 total chloride proved to be very tedious. Portions of each sample 

 were extracted three or four times with boiling water; the resulting 

 filtrates contained brown colloidal matter. They were evaporated to 

 small volumes, and in one case taken completely to dryness. In this 

 case, sodium chloride crystals separated along with the brown 

 matter. Evaporation did not suffice to render the latter insoluble; 

 the greater proportion again formed a colloidal suspension on the 

 addition of water. To each extract a few drops of nitric acid 

 were added, the solution boiled and silver nitrate added. The silver 

 chloride formed was fine grained and yellowish owing to colloidal 

 contamination. It could not be made to coagulate except in one 

 case (D), where much less colloidal matter had been removed in the 

 extraction. After settling, the supernatant liquid was decanted 

 through a Gooch crucible, and the residue dissolved in ammonia. 

 The deep brown colour of this solution was possibly due in part 

 to the reduction of the ammoniacal silver complex to colloidal metal, 

 by organic matter. On reprecipitation with excess nitric acid, the 

 silver chloride was white and coagulable. The liquid was poured 

 through the Gooch crucible, but on addition of water or dilute 

 nitric acid, the precipitate turned to a milky suspension, which 

 the asbestos layer in the crucible would not retain. Boiling again 

 coagulated it. The liquid was decanted away, the precipitate dis- 

 solved in ammonia (the solution again being deep brown in colour), 

 and reprecipitated with nitric acid. A repetition was necessary 

 before a satisfactory product was obtained. In the following tables, 

 all figures represent percentages : — 



TABLE IT. 



A. IS. C. P. A. B. C. D. 



Calculated on dried material. 

 83.0 - 71.1 - 82.0 - 63.4 



17.0 - 28.9 - 18.0 - 36.6 



All subsequent figures are percentages calculated upon dried 

 material. 



