36 REPORT— 18 75. 



original sample, is delivered into a small basin, diluted with 400 grms. or 

 so of water, and acidified slightly with hydrochloric acid. 



"About 500 grains of platinic-chloride solution (containing at least 25 

 grains of platinum) are added, and the fluid evaporated nearly to dryness on 

 the water-bath. A few drops of water are then added to the residue, and 

 the evaporation repeated to expel the excess of hydrochloric acid. About 

 50 grains more of the^strong platinic solution are mixed with the precipitate, 

 and the whole stirred well and set aside in a cold place for at least an hour 

 with occasional stirring. The precipitate is then thrown on a very small 

 filter (unweighed), the basin rinsed out with about 10 drops more of the 

 platinic solution, and the precipitate on the filter washed with 10 or 15 drops 

 more. The basin with the filter and contents are then washed with the small- 

 est possible quantity of alcohol of 95 per cent, strength, and dried at 100° C. 

 The dried precipitate is transferred as completely as possible to a small cap- 

 sule, in which it is further dried until it assumes a distinct orange-colour, and 

 weighed. The filter and trace of adhering precipitate is ignited on a cruci- 

 ble lid, and the residual metal with its corresponding chloride of potassium 

 calculated to potassio-platinic chloride, and the weight added to that of the 

 precipitate. The following factors are employed : — To bring the precipitate to 

 potassium -1603, to potash -1930, and to chloride of potassium -3056. The 

 figures are based on Stas's numbers for potassium and chlorine, and Berzelius's 

 equivalent for potassium." 



Professor Fresenius writes : — 



" I am quite ready to go once more closely into the question regarding the 

 estimation of potassium in commercial potash salts. However, I can only do 

 so occasionally when I am at leisure, for there are a great many experiments 

 still to be made before being able to give a satisfactory answer. As far as I am 

 concerned myself, as well as the analyses in my laboratory, these questions are 

 of no great importance. The great object I have in view is to be accurate ; 

 saving time is only a secondary consideration. I begin by entirely separating 

 from the solution the sulphuric acid, lime, and magnesia ; then I weigh the 

 pure chloride of the alliali-metals, estimate the potassium as platino-chloride 

 of potassium according to the method given in ray manual of quantitative 

 analysis, make sure that it is quite pure, and generally also estimate the 

 chloride of sodium in the washings by evaporating and heating the residue in 

 a current of hydrogen*, partly to have a check, principally, however, to make 

 sure that it contains no more potassium. This method, however, is not practi- 

 cable in works because it is too elaborate." 



In reference to the estimation of potash, Messrs. Wallace, Tatlock, and 

 Clark also write : — 



" It is a notorious fact that while the results obtained by the process we 

 follow, as compared with those got by what we may term the alcohol method, 

 agree verjf closely in the case oi potassium compounds free from sodium, wide 

 differences have jjeen observed when the potassium salts were of low strength 

 from the presence of sodium compounds. To this fact it would not be difii- 

 cult to get manufacturers and merchants to testify. This remark applies 

 specially to the case of potassium products from kelp, of which many thousand 

 tons are made in Glasgow every year, some of which contain a large propor- 



* How the proportion of cbloride of sodium present is deduced from the weight of the 

 ignited residue is not stated. Probably by washing with water and subtracting the weight 

 of the residual metallic platinum. 



