ATOMIC WEIGHTS 



457 



Hence Pd = 105.71. Bailey and Lamb's weighings are all rednced 

 to a vacuum. 



Keller and Smith/ reviewing Keiser's work, find that palladiam- 

 monium chloride, prepared as Keiser prepared it, may retain traces of 

 foreign metals, and especially of copper. Accordingly, they prepared a 

 quantity of the salt, after a thorough and elaborate process of purifica- 

 tion, dried it with extreme care, and then determined the palladium by 

 electrolysis in silver-coated platinum dishes. The precipitated palladium 

 was dried under varying conditions, concerning which the original mem- 

 oir must be consulted, and was proved to be free from occluded hydro- 

 gen. By this method two sets of experiments were made to determine 

 the atomic weight of palladium; but for present purposes the two may 

 fairly be treated as one. The data obtained are as follows, but the 

 weights do not appear to have been reduced to a vacuum : 



Mean, 50.508, ± .0014 



Hence Pd = 107.14, a result notably higher than Reiser's. 



Keller and Smith account for the difference between their determina- 

 tions and Keiser's partly by the assumption that the materials used by 

 the latter were not pure, and partly by considerations based on the proc- 

 ess. In order to clarify the latter part of the question they made three 

 sets of experiments by Keiser's method, slightly varying the conditions. 

 rirst, the chloride was not pulverized before ignition, and slight decrepi- 

 tation took place, while dark stains of palladium appeared in the reduc- 

 tion tube, indicating loss by volatilization. Secondly, the chloride was 

 prepared from crude palladium exactly as described by Keiser, but was 

 pulverized before reduction. N"o decrepitation ensued, but traces of pal- 

 ladium were volatilized. The third series, also on finely pulverized 

 material, was like the second; but the palladiammonium chloride was 

 purified by Keller and Smith's process. The three series, here treated 

 as one, are as follows : 



Anier. Cheni. .Journ., 14, 423. 1S92. 



