Prof. Bunsen on Ccesium. 243 



chlorides are converted into carbonates : rather more tartaric acid 

 is then added to the solution of the carbonates than is sufficient 

 to convert the caesium-compound into the neutral, and the rubi- 

 dium-compound into the acid tartrate. The finely divided solid 

 mass obtained by evaporation and trituration is brought on to a 

 funnel containing a small filter, and then left for some time 

 exposed to a moist atmosphere. The deliquescent csesium-salt 

 drops slowly down, whilst the acid tartrate of rubidium remains 

 unaltered in the funnel. In this way a substance was obtained 

 which, when converted into chloride, gave upon analysis the fol- 

 lowing numbers : — 



1*3033 grm. of this chloride of caesium yielded 1*1422 grm. 

 chloride of silver. This corresponds to 



Chlorine. Caesium. 



21-709 .... 78-291 



This salt exhibited in the spectroscope but a faint rubidium 

 reaction, but it had taken up a small quantity of lithium- and 

 potassium-compounds from the tartaric acid. In order to sepa- 

 rate these impurities, a boiling and somewhat dilute solution of 

 the salt was precipitated by chloride of platinum, and the preci- 

 pitate, after having been washed by decantation with hot water 

 and dried, was reduced in a stream of hydrogen. This mode of 

 purification was repeated until the composition of the chlorine- 

 compound remained unaltered. In this way the following results 

 were obtained :— 



After the second purification with PtCl 2 , 

 1-8987 grm. of substance yielded 1*6342 grm. chloride of silver. 



After the fourth purification with PtCl 2 , 

 1*3835 grm. of substance yielded 1*1781 grm. chloride of silver. 



After the fifth purification with PtCl 2 , 

 1-3682 grm. of substance yielded 1 1644 grm. chloride of silver. 



After the sixth purification with PtCl 2 , 



1*2478 grm. of substance yielded 1*0623 grm. chloride of silver. 



These numbers correspond respectively to the following com- 

 position of the chloride :— 



Chlorine. Caesium. 



21-293 78-707 



21-057. ..... 78-943 



21-045 ...... 78-955 



21-052 78*948 



The mean of the three last experiments, in which the compo- 

 sition may be assumed to have become constant, shows that 



E2 



