242 Prof. Bunsen on Ccesium. 



The residues which I obtained from the preparation of half a 

 pound of pure chloride of rubidium from lepidolite likewise gave 

 no larger a yield of caesium-salt; so that the quantity of material 

 with which I have been obliged to conduct the following experi- 

 ments has also been but very small. 



Whilst I was thus engaged with the preparation of caesium- 

 salt for the purpose of making a more exact determination of 

 the atomic weight of this metal, Messrs. Johnson and Allen* were 

 fortunate enough to obtain a specimen of American lepidolite so 

 rich in caesium that they were able to prepare about 30 grms. of 

 bitartrate of caesium from a few kilogrammes of the mineral. 

 With this material, Messrs. Allen and Johnson prepared the acid 

 tartrates of caesium and rubidium, employing the great differ- 

 ence in the solubility of these salts for the purpose of separating 

 the two metals by repeated crystallization. They thus obtained 

 a chloride of caesium which, according to four analyses made by 

 them, possessed the following composition : — 



Chlorine. Caesium. 



21-044 78-956 



21-031 78-969 



21-043 78-957 



21-063 78-937 



The purest chloride compound which I was able to obtain by 

 help of the method which I first used contained f — 



Chlorine. Caesium. 



First purification . . 22-334 77'666 



Second purification . 22'334 77-666 



Third purification . . 22*316 77'684 



From the close agreement of the above numbers, it is clear that 

 in the method of separation which I employed at the time of the 

 discovery of cassium a limit is reached beyond which the purifi- 

 cation cannot be carried. The method employed by Messrs. 

 Johnson and Allen is likewise inapplicable in the case of 

 the small quantity of material to which my experiments were 

 limited, and I have therefore endeavoured to purify the salt by 

 another method. This is based upon the fact that the acid 

 tartrate of caesium, like the corresponding rubidium-salt, is 

 unalterable in the air, whilst the neutral tartrate of caesium deli- 

 quesces very readily. This method of separation answers very 

 well even with very small quantities. Having determined, by 

 precipitation with silver, the quantity of rubidium in a mixture 

 of the chlorides free from potassium, lithium, and sodium, the 



* American Journal of Science and Arts, vol. xxxv. p. 94, Jan. 1863 ; 

 and Phil. Mag. S. 4. vol. xxv. p. 196. 

 t Phil. Mag. vol. xxii. p. 503. 



