Analysis by Spectrum-observations. 333 



does not differ from the true combining proportion more than is 

 the case with a large proportion of the atomic weights at present 

 considered as correct, and received without question. 



It is almost needless to add that the residues obtained by the 

 treatment above described, when operated upon again, yielded 

 a considerable quantity of chloride of rubidium. 



Although impossible to determine with exactitude the quantity 

 of rubidium contained in lepidolite, it appeared of interest to 

 ascertain this as accurately as possible. For this purpose a 

 specimen of lepidolite was employed, found at Eozena near 

 Hradisko, which was seen by means of spectrum-analysis to 

 contain traces of caesium as well as rubidium. The solution ob- 

 tained from 13-509 grms. of this lepidolite fused with lime, was 

 precipitated in the usual manner with chloride of platinum, and 

 the precipitated double chlorides of potassium, rubidium and 

 platinum reduced with hydrogen, yielding 2*0963 grms. of the 

 chlorides of potassium and rubidium. These salts were again 

 precipitated by chloride of platinum, and the precipitate thus 

 obtained, boiled out with small quantities of water until the 

 solution appeared of a light yellow colour. The united wash 

 waters, on evaporation and cooling, deposited a second crop of 

 crystals, which were treated [in a similar way to the first preci- 

 pitate. The platinum double salt which separated out a third 

 time, was likewise submitted to the same treatment, and the 

 operation repeated until the precipitate formed, on concentrating 

 the wash-waters, easily dissolved on boiling in a small quantity of 

 water without leaving any residue. The whole of the insoluble pla- 

 tinum double salt yielded,after reduction in a current of hydrogen, 

 00421 grm. of chloride of rubidium, corresponding to 024 

 part of oxide of rubidium in 100 parts of the lepidolite in 

 question. This determination, together with an analysis of the 

 other constituents, made by Mr. Cooper in my laboratory, gives 

 the following composition for the lepidolite from Rozena : — 



Silicic acid 50-32 



Alumina 28-54 



Peroxide of iron 073 



Lime 1-01 



Magnesia - 51 



Oxide of rubidium .... 0-24 

 Oxide of caesium .... trace. 



Lithia 0-70 



Fluoride of lithium .... 0-99 

 Fluoride of sodium . . . . 1'77 

 Fluoride of calcium . . . . 1206 



Water 312 



9999 



