348 On Chemical Analysis by Spectrum-observations. 



carbonate of rubidium and caesium appears to be formed which 

 is not perfectly insoluble in alcohol. On this account, we prefer 

 to prepare the caustic alkalies by the action of baryta-water on 

 the sulphates, and then to convert about one-fifth part of the 

 mixed caustic alkalies into carbonate by evaporation with carbon- 

 ate of ammonium in a silver basin. The hydrated oxide of caesium 

 is then dissolved out from such a mixture by alcohol, leaving a 

 residue consisting of carbonate of rubidium containing caesium. 

 If this mode of separation be repeated five or six times, each 

 time taking care to use as little alcohol as possible, the hydrated 

 oxide of csesium is obtained quite free from rubidium, as may be 

 proved by examination in the spectrum-apparatus. It is scarcely 

 necessary to remark that the numerous residues which accumu- 

 late in the course of the several operations must again be worked 

 up in the same way as the original substance, and that the pla- 

 tinum can be used over and over again without any great loss. 



The following experiment served as a preliminary determi- 

 nation of the atomic weight of csesium. Chloride of caesium was 

 prepared from the mixed chlorides freed from all potassium salt, 

 by the methods just described, and the contained chlorine esti- 

 mated as chloride of silver. 



0-5219 grm. of the chloride of caesium yielded 04995 grm. of 

 chloride of silver. 



The chloride of caesium of the first preparation was submit- 

 ted a second time to the purifying treatment above described. 

 1-7690 grin, of the third preparation yielded 1-6518 grm. of 

 chloride of silver. This substance was again purified in like 

 manner, and 0-3727 grm. of the third preparation yielded 0-3402 

 grm. chloride of silver. After the process of purification had been 

 repeated for a fourth time, 1*3860 grm. of the substance yielded 

 1-2518 grm. of chloride of silver. After the fifth time of treat- 

 ment, T0124 of the salt gave 0-9144 grm. chloride of silver; 

 and lastly, after the sixth purification, 0*4126 grm. of chloride 

 of silver was obtained from 0*4572 grm. of the salt. 



Hence 100 parts of the substance under examination gave — 



After the first purification 95-708 grms. of chloride of silver. 



second 



)) 



93*486 



third 



)) 



91-280 



fourth 



a 



90-318 



fifth 



a 



90-320 



sixth 



)> 



90*245 



It is evident from these numbers that, after extracting the salt 

 four times with alcohol, a substance was obtained which on a 

 repetition of the process did not undergo any alteration in com- 

 position. 



