POTASSIUM. 99 



M. H. Klaproth ; F. H. Wollaston : 39.517 (O = 

 16). 



Klaproth found that 441 CI = 591 potassium oxide. 

 Hence "Wollaston deduced the value 491 (0 = 100) for 

 K. [If CI = 35.457, this relation gives K = 39.517.] {Phil 

 Trans., m, 1814, 20.) 



J. J. Berzelius : 39.193 (O = 16) ; 244.958 (O = 

 100). 



Berzelius found that 100 KCl = 192.4 Ag CI. If Ag = 

 1351 .607 ; CI = 442.65 ; the above value follows. {Poggend. 

 Ann., 8, 1826, 190.) 



F. Penny : 39.073 (0 = 16). 



Penny made six experiments on the conversion of the 

 chlorate into the chloride. Potassic chlorate was dried at 

 about 105°, dissolved in a flask with HCl, evaporated, dried 

 and weighed. The cake contained some free HCl. It was 

 broken up, pulverized, and a known quantity heated to dull 

 redness but not to fusion, and the HCl expelled allowed for. 

 The mean result was that 100 KCl O3 = 60.823 KCl ; ex- 

 treme difference, 0.015. This relation gives KCl = 74.527 

 and if CI = 35.454 (Penny,) the value for K follows, 

 i^umerous experiments were also made on the introconver- 

 sion of the nitrate, the chloride and the chlorate, which 

 established the difference between a chloride and a nitrate, 

 besides confirming the value of K. The salts were purified 

 by recrystallization and were carefully tested for impurities. 

 The weighings are all for vacuum. (Phil. Trans., 129, 1839, 

 18.) 



C. Marignac : 39.2 (O = 16); 245 (0 = 100). 



By six experiments on the decomposition of the chlorate 

 by heat, 100 chlorate were found to lose 39.161 oxygen ; 

 extreme difference 0.012 ; hence KCl = 932.14. By com- 

 paring this value with the molecular weight and the compo- 

 sition of argentic chloride, CI was calculated at 442.13, leav- 

 ing for K the number 490. Confirmatory experiments were 

 made on potassic perchloride. The chlorate was purified by 

 recrystallization. The weighings are for vacuum. [Liebig's 

 Ann., U, 1842, 23.) 



