ON THE ESTIMATION OF POTASH AND PHOSPHORIC ACID. 11 



the precipitate thoroughly dried at 100° C. to a temperature of 140° C. for 

 one hour. This loss represents only -02 for 100 parts of potassium chloride. 



In order to obtain a satisfactory basis of investigation, it was necessary, in 

 the first place, to obtain perfectly pure potassium salts ; and as a necessary 

 condition of the requisite purity was complete freedom from sodium com- 

 pounds, their preparation was found less easy than might be expected *. In 

 the first place an attempt was made to obtain pure potassium chloride by 

 repeatedly evaporating pure nitre with hydrochloric acid. The result showed 

 that the reaction took place with far less facility than was expected, and the 

 process was abandoned. Chloride of potassium was next obtained by dis- 

 solving the purest commercial acid potassium carbonate in hydrochloric acid, 

 filtering, and repeatedly crystallizing the product. Ignition of the crystals 

 on platinum wire in the Bunsen fiame showed the presence of sodium in 

 abundance, and two determinations of the real chloride of potassium as 

 platinum salt gave 98-93 and 98-85 per cent, respectively. 



A highly satisfactory product, was at length obtained by the _ following 

 process : — Cream of tartar was dissolved in boiling water, the liquid filtered, 

 and the acid tartrate of potassium obtained by cooling the solution. The 

 product was recrystallized, and then tested for sulphates and sodium, neither 

 of which was found. The dried crystals were ignited, the mass dissolved in 

 water, the liquid filtered, nearly neutralized with hydrochloric acid, a few 

 drops of ammonium oxalate added, the solution again filtered, and then eva- 

 porated to dryness. The resultant chloride of potassium was heated to 

 fusion and reduced to powder. The product was absolutely free from sul- 

 phates, completely soluble in water, and the solution was perfectly neutral. 

 The salt showed no trace of sodium when heated on platinum wire in the 

 Bunsen flame. 



The hydrochloric acid used in the experiments was prepared by acting on 

 common salt by non-arsenical sulphuric acid and passing the washed gas into 

 distilled water f. 



The platinic chloride was obtained by reducing the commercial chloride 

 (which contained iron and other impurities), by boiling with caustic soda and 

 alcohol, thoroughly washing (first by decantation and afterwards on the 

 filter) the resultant platinum black, boiling it for some time with hydro- 

 chloric acid, and again thoroughly washing with hot water and igniting in a 

 muffle. The metallic platinum was boiled with nitric acid, rewashed and 

 reignited, and then weighed and dissolved in aqua regia. 



The platinic-chloiide solution thus obtained was evaporated nearly to 

 dryness, first with hydrochloric acid and then several times with water, in 

 order to get rid of the free acid as much as possible* Ultimately the solu- 

 tion was diluted, filtered from any insoluble residue (which was ignited and 

 weighed, and the weight deducted from the original), and the filtrate further 

 diluted until 100 cub. ccntims. contained about 6 grammes of metallic 

 platinum. 



* A shorter method than those tried would probably have been to ignite pure 

 potassium chlorate. — A. 31. Allen. 



t For some years all the hydrochloric acid used in my laboratory lias been prepared 

 by this process. It is more convenient, and furnishes a far superior product to that 

 obtained by distilling the impure liquid acid. — A. It. Alle.v 



+ This method of preparing pure chloride of platinum is practically identical with 



that recommended by Messrs. Chalmers and Tatlock in a paper read before the Chemical 



lie Philosophical Society of Glasgow, April 20th, 1868. The Committee has 



adopted I process for recovering the platinum from the precipitates and filtrates 



obtained in the experiments, 



