ON THE ESTIMATION OF POTASH AND PHOSPHORIC ACID. 13 



ten times ita weight of distilled water*. In the experiments made on the 

 weighed solution, the required quantity was approximately measured by 

 running it from a buretto into an accurately tared beaker, and tho exact 

 quantity taken was then ascertained by weighing. In this manner the 

 amount of potassium salt employed in each experiment was ascertained with 

 great accuracy. The error in the amount taken could not be more than 

 •00005 of a gramme, or about -007 per cent, of the quantity used. With 

 the new solution the following experiments were made as a check on its 

 strength : — 



By precipitation with nitrate of silver, 



Weight of =KC1 Weight of = K01 = 



solution. Agul. 1 



Al 7-7065 -70060 1-3459 -70017 99-94 



A 2 7-7455 -70414 1-3534 -70403 09-00 



By direct evaporation, 



Wt. of sola. KC1. Residue = KOI per cent. 

 B 1 7-711 -701 -7005 99-93 



In this case half a milligramme loss of chloride of potassium, probably due 

 to decrepitation on heating the residuo, caused a difference of -07 per cent. 



In the foregoing and all subsequent experiments the following atomic 

 weights and factors were employed : — 



Chlorine 35-457 Stas, 1SG5. 



Potassium 39-137 „ „ 



(Silver 107-930 „ 1866. 



AgCl x -52023 = KC1. 



The atomic weight of platinum was calculated from the original data of 

 Berzclius, obtained by the analysis of potassium chloroplatinate, but substitu- 

 ting Stas's numbers for chlorine and potassium for those employed by 

 Berzclius. This gives the result 



Pt iv = 197-1937. 

 Hence 



K 2 PtCl x -16033 = K„ 

 K,PtCl c x -19310 = ICO 

 K 2 rtCl,. x -30560 = 2KC1. 



Fresenius, in the last edition of his ' Quantitative Analysis ' (7th English), 

 adopts the number 98-59 as the atomic weight of the divalent platinum, 

 which also gives the factor -3056, for calculating the chlorojnatinatc into 

 chloride of potassium. In former editions Andrews's number 98*94 was 

 adapted for platinum, which caused a sensible difference in the percentage of 

 potassium chloride obtained. The factor -30507, resulting from the employ- 

 ment of Andrews's atomic weight for platinum, is adopted by Drs. Frank and 

 Berrand in their communication to this Committee. 



The consequence of employing the above factors in calculating the per- 

 centage of chloride of potassium corresponding to the precipitate of chloro- 

 platinate obtained is shown in the following statement : — 



* It is obvious that the subsequent calculations would hare bsen facilitated by dissolving 

 a known weight of the pure potassium chloride in exactly nine times its weight of dis- 

 tilled water instead of ten. This consideration did not present itself till it was too late 



to take advantage of it. 



