ON THE ESTIMATION or POTASH AND TIIOSTHORIC ACID. 



15 



was but slightly in excess of the amount required to convert the whole of 

 the potassium into chloroplatinate. 



Table II. — ltesults of experiments on Pure Chloride of Potassium by Pro- 

 cesses I. and III., employing only a slight excess of Platinum solution. 



These results showed a great improvement, and indicated pretty clearly 

 the importance of avoiding a large excess of platinum solution when alcohol 

 only was employed for washing the chloroplatinate. 



The following Table shows the relative accuracy and limits of variation 

 obtained in experiments on pure chloride of potassium by methods L, III., 

 and IV. 



Table III. — Analysis of the Results obtained in the estimation of Potassium 

 when in the form of Pure Chloride. 



From these experiments, therefore, it was concluded that the method of 

 estimating potassium by precipitation as chloroplatinate was very accurate 

 when proper precautions were taken. 



This conclusion is generally accepted, the chief discrepancies arising when 

 mixed alkaline chlorides are analyzed, the different methods then giving results 

 which sometimes exhibit wide variations. 



The following Table shows the results obtained by the analysis of various 

 mixtures of the pure chlorides of potassium and sodium : — 



