Mr. 0. D. Allen on Casium and Rubidium, 195 



composition of the salt, according to the formula 



H Cs/ U • 



Calculated. Found. 



1 ^ ii. s 



C« . . 48-00 17-62 16-99 17-02 



H 5 . . 5-00 1-83 1-85 1*88 



O 11 . . 88-00 32-31 



CsO . . 131-35 48-24 .. .. 4870 



272-35 100-00 



The discrepancy between the composition as calculated and 

 found is perhaps due to a slight admixture of the neutral tar- 

 trate, which might possibly have been present owing to the use 

 of insufficient tartaric acid. 



The solubility of bitartrate of caesium was determined for the 

 same temperature and by the same methods as were employed in 

 the case of the rubidium-salt. 



I. 2*998; grms. of solution, saturated at the boiling-point, 

 gave a residue of 1*483 grm. 



One part of the salt requires 1*0.2 part of boiling water for 

 solution. 



II. 11*931 grms. of solution, saturated at 25° C, gave a 

 residue of 1*054 grm. 



III. 8-7625 grms. of solution, saturated at 25° C, gave a 

 residue of 0*7727 grm. 



One part of the salt accordingly requires 10*32 parts of water 

 at 25° C. for solution. 



The fact that bitrartrate of rubidium requires about eight 

 times as much water for solution as bitartrate of caesium, explains 

 the facility with which these salts can be separated from each 

 other by crystallization. 



In these experiments I have received the advice and assistance 

 of Professors Johnson and Brush, for which I here take pleasure 

 in expressing my most grateful acknowledgments*. 

 New Haven, August 12, 1862. 



* Since the above was written I have recovered, from the mother-liquors 

 and washings of the 1 32 grammes platinochlorides of caesium and rubidium 

 first obtained, an additional quantity of potash-free platinochlorides (chiefly 

 of rubidium) amounting to 40 grammes, making the total yield ] 72 grammes 

 from 10i kilogrammes of lepidolite. Most of this remained in solution from 

 the use of insufficient bichloride of platinum in some of the precipitations. 

 The amount of rubidium in the Hebron lepidolite thus appears to be not 

 less than in that from Rozena. 



02 



