22 



REPORT — 1877. 



Table XII. 



It appears therefore that Tatlock's process is applicable to the analysis of 

 sulphates or nitrates, provided that there is sufficient chloride present for 

 the formation of th.G chloroplatinate of potassium. If not, it must be added 

 in the form of sodium chloride, or, in the case of sulphates, hydrochloric acid 

 may be used. When much sulphate is present, the quantity of platinum 

 solution used for washing the precipitate must he somewhat increased, or the 

 results will be too high, owing to the insolubility of the sulphates in alcohol. 

 Magnesium appears to cause no difficulty, the result 99-999 having been 

 obtained in its presence. 



When it is remembered that none of tho foregoing experiments were mado 

 on a larger quantity than -7 of a gramme (about 10 grains), it will be seen 

 that the determination of potassium as chloroplatinate is, when due care is 

 taken, as accurate as the estimation of most elements, and, when heavy metals 

 arc absent, quite as easily effected. 



In practice it is rarely required to determine potassium very accurately in 

 presence of large proportions of foreign metals, but in the accurate assay of 

 the better class products is becoming daily more important. If the propor- 

 tion of sodium salts present in a sample exceed 3 per cent, the product is 

 unfit for certain purposes ; and as tho determination of the sodium is strictly 

 dependent on that of the potassium, any error in the latter is reproduced. 



Although the results obtained by Tatlock's method show a decided loss 

 when a very large proportion of chloride of sodium is present, this error 

 nearly disappears with smaller amounts ; and as the method is available in 

 presence of svdphates, nitrates, and magnesium, and is • very readily con- 

 ducted, it seems the best suited for the general assay of commercial potassium 

 salts. 



From a general consideration of the foregoing researches on the deter- 

 mination of potassium as chloroplatinate it appears that : — 



1. Potassium in the form of pure chloride can be determined with great 

 accuracy by precipitation as chloroplatinate. If a large excess of platinum 

 solution be employed, and alcohol alone used for washing the precipitate, the 

 results have a tendency to exceed the truth. By avoiding tho use of a large 

 excess of platinum solution more accurate results arc obtained. If a small 

 volume of platinum solution be employed in the first instance for washing 

 the precipitate (as recommended by Tatlock), and the washing be then com- 

 pleted with alcohol in the usual way, the results are very accurate. Potas- 

 sium chloroplatinate appears to be practically insoluble in a concentrated 

 solution of platinic chloride. 

 ^ 2. In presence of a considerable proportion of sodium, washing the pre- 

 cipitate with alcohol alone tends to give results in excess of the tnith. If 

 the precipitate be first treated with platinum solution the results are some- 

 what low, apparently owing to the solubility of the precipitate in a solution 



