4 RESEARCHES UPON ATOMIC WEIGHTS. 



criticize experimental work without an actual repetition of the experiments, 

 for frequently some constant source of error is so securely hidden that it may 

 be detected only by the most careful investigation, no attempt at criticism 

 is made here. 



Nevertheless, it is worth while calling especial attention to the careful re- 

 searches of Morse and Jones and Morse and Arbuckle upon the ratio Cd : CdO, 

 wliich yielded the value 1 1 2.38, and that of Bucher upon the ratios CdCl2 : zAgCl 

 and CdBr2 : '2AgBr which yielded the values 112.31 and 112.33, Our experi- 

 ments indicate that the real value for the constant in question is slightly 

 higher than any of these results. 



In a recent determination of the specific gravity of cadmium chloride,* 

 anhydrous cadmium chloride was prepared by ignition of a double chloride 

 of cadmium and ammonium in a current of hydrochloric-acid gas, in a state 

 of so great purity that it was considered worth while to make use of the salt 

 for a determination of the atomic weight of cadmium. 



PURIFICATION OF MATERIALS. 

 CADMIUM CHLORIDE. 



The general method of purification of the cadmium material was that of 

 fractionally precipitating cadmium sulphide. One kilogram of metallic cad- 

 mium was dissolved in aqua regia, and the solution, after being boiled to expel 

 chlorine and oxides of nitrogen, was filtered, and diluted to about 4 liters. 

 The solution contained traces of lead, copper, thallium, nickel, iron, and zinc. 

 When a current of hydrogen sulphide was passed through the solution, the 

 first small fraction of cadmium sulphide which was precipitated was dark- 

 colored, nearly black, owing to the presence of lead, copper, and thalHum. 

 This fraction was removed by filtration and rejected. A second larger fraction 

 of the sulphide, although it contained no appreciable amount of the latter 

 metals, also was discarded. The third fraction consisted of all that could be 

 precipitated by saturating the solution with hydrogen sulphide. However, 

 this did not contain more than one-quarter of the original material, for the 

 solution was very strongly acid, owing to the large excess of acid used in dis- 

 solving the metal and the accumulation of the acid formed during the precipi- 

 tation. The solution was separated from the precipitate by decantation, and 

 was then diluted to 16 liters. Upon saturating this solution with hydrogen 

 sulphide a fourth fraction of cadmium sulphide was obtained, and a second 

 dilution of the solution made possible the precipitation of nearly all the 

 remainder of the cadmium in still a fifth fraction. 



The third, fourth, and fifth fractions of the sulphide were separately washed 

 until free from chlorides. As the electrolytes were eliminated, the cadmium 

 sulphide showed a tendency to pass into colloidal condition, which necessitated 



' Baxter and Hines: Amer. Chetn. Jour., 31, 220 (1904). 



