ON THE ESTIMATION OF POTASH AND PHOSPHORIC ACID. 27 



into the beaker previously used, treat with water, add some tincture of litmus, 

 heat to boiling, and add normal alkali solution till the fluid is justbluo, and 

 remains so after continued boiling." 



With a view of preparing pure chloride of potassium for these researches, 

 the Committee abandoned the method employed last year, in favour of a 

 simpler process suggested in a note appended to the last report. 



Commercial chlorate of potassium was rccrystallized and heated until no 

 more oxygen was evolved, and the product was dissolved in water, filtered, a 

 few drops of hydrochloric acid added to the clear solution, and the whole 

 evaporated to dryness and ignited in a muffle at a low red heat. The product 

 was powdered and bottled. When heated on clean platinum wire in a 

 Bunsen flame it gave no trace of yellow colour to the flame. 



Thirty-five grammes of this pure salt were dissolved in 315 grammes of 

 pure water, in order to make a solution containing exactly one tenth of its 

 weight of pure chloride of potassium. 



In the following investigation a roughly measured quantity of this solution 

 was run into a beaker and the exact weight taken. This method was 

 adopted during the experiments of last year in preference to pipette measure- 

 ments, which were not considered satisfactory. 



The solution was first analyzed by precipitation by nitrate of silver. The 

 following tablo shows the quantities taken, and the number of parts of KC1 

 found for one hundred taken * : — 



The six following experiments were made by adhering strictly to the 

 description of the process already quoted. A quantity of potassium chloride 

 solution containing about -7 gramme or 10 grains of tho salt was employed 

 in each case. The standard alkali employed was very carefully prepared and 

 was strictly normal. 



It was not found practicable to wash the precipitate till the washings were 

 no longer acid to litmus. The washing was therefore arrested when the 

 filtrate gave no reaction with silver nitrate. 



8 C 



Expt. Wt. of Soln. C. c. normal alkali = KOI - KCl found per 100 



used. parts taken. 



1 7-1155 18-80 S -70110 98-54 



2 7-1100 18-83 P -70236 98-80 



3 7-0200 18-75 S -69937 99-90 



4 7-0190 18-75 S -69932 99-62 



5 7-1365 19-10 S -71237 99-50 



6 7-0210 18-70 P -69745 99-34 



Although it was not to be expected that there could be any advantage in 

 employing caustic potash in the titration instead of caustic soda, it was 

 considered that the case was one in which it was just possible that there 

 might be a choice, and therefore both alkalies -were tried. The alkali em- 

 ployed in each experiment is distinguished by the letters P and S placed 

 after the number of centimetres of normal alkali required. 



* The following are the atomic weights employed in the investigation : — 

 K = 39-137; Si=28; Ag=107-93; C1-35--457; F = 1896. 



