Section III., 1904 [ 47 ] Tkans. R. S. C. 



VIII. — A Revision of the Atomic Weight of Potassium. 

 I. The Analysis of Potassium Chloride. 



By E. H. Archibald, A.M., Ph.D. 



(Presented by Prof. B. J. Harrington, and read June 22, 1904). 



Although the compounds of pptassium are both plentiful and 

 numerous, and many of them have been well studied, yet the number 

 of investigations which have had for their object the determination of 

 the atomic weight of this element, is not as large as one should expect. 

 Indeed, the value at present accepted rests almost entirely upon the work 

 of Stas and Marignac; but very different interpretations are given of 

 this work, so that, while the international committee deduce the value 

 39-15, Eichards gives 39 '14, while Clarke proposes as low a value as 

 39 -ll. Some analysis of potassium iodide, which gave a very low 

 number for this constant, accounts for the number given by Clarke; 

 while the later work of Stas furnishes the data for the value given by 

 Richards. 



Quite recently, Eichards and Archibald ^ studied the decomposition 

 of potassium nitrate in the presence of pure silica, and from their 

 observations found the value 39 -141. A few determinations were also 

 made of the chlorine in a sample of potassium chloride, purified, by 

 means of the choroplatinate, and these analyses give 39.139, a number 

 in good agreement with the value found from the nitrate analysis. 



As no serious difficulties are encountered in preparing pure potas- 

 sium compounds, it was thought advisable to extend the study of 

 potassium chloride to include the analysis of a sample which had been 

 subjected to a different method of purification from that examined by 

 Eichards and Archibald; also to make an extended series of analyses of 

 potassium bromide. The following is an account of the analysis of 

 potassium chloride. 



There is no need to review here the work which has already been 

 done on the atomic weight of potassium. This has been most com- 

 pletely done by Clarke.^ For purposes of comparison, the results of 

 the analysis of potassium chloride given by Eichards and Archibald are 

 reproduced here. In calculating these values, the atomic weight of 

 silver was taken to be 107*930, and chlorine 35*455, oxygen being 

 16.000. 



* Proc. Amer. Acad., 1903, 38, 443. 



* Constants of Nature, Part V. Smithson. Misc. Coll. 1897. 



