Mr. 0. D. Allen on Casium and Rubidium. 193 



zation,the solution from which the first crystals had been removed 

 was concentrated to nearly one-half its original volume, when, 

 by cooling, a very small quantity of salts of the two alkalies was 

 deposited. This operation was repeated three times, when a 

 portion of the solution, evaporated to dryness and examined by 

 the spectroscope, gave only the lines belonging to csesium. The 

 several intermediate products containing both alkalies were then 

 united, and another portion of each salt separated from them in 

 the same manner. JBy repeating this process of fractional cry- 

 stallization four times with about 40 grammes of the mixed salts, 

 23*77 grammes of bitartrate of ca?sium, and 12*511 grammes of 

 bitartrate of rubidium were obtained, while 3*74 grammes re- 

 mained unseparated. It was found that the csesium-salt thus 

 obtained, although exhibiting no impurity when tested by the 

 spectroscope directly, i. e. after conversion by ignition into car- 

 bonate, was still mixed with a trace of rubidium, as on converting 

 it into chloride a faint line characteristic of the latter metal was 

 perceptible. The separation of two or three more small crops of 

 crystals sufficed to render the residual solution perfectly free 

 from any admixture that could be detected by a spectroscope of 

 ordinary power*. The rubidium-salt was also more carefully 

 tested in the same manner, but was found to be entirely pure. 



The process above described thus furnishes a simple and easy 

 method of separating in a perfectly pure state a large share (in 

 these trials about 90 per cent.) of a mixture of the two alkalies. 

 It requires no great expenditure of time, since the solutions em- 

 ployed can be concentrated at high temperatures, and, on cooling, 

 immediately deposit well-formed crystals. 



Composition and solubility of the Bitartrates of Ccesium and 

 Rubidium. — Bitartrate of rubidium crystallizes from hot solutions 

 in colourless, transparent, flattened prisms, which are often half 

 an inch or more in length, even when formed rapidly from small 

 quantities of solution. They remain unaltered in the air, and 

 also are unchanged at a temperature of 100° C. The pulverized 

 salt, dried at 100° C, was burned with chromate of lead in the 

 usual manner. 



I. 0*4681 grm. gave 0*0902 grm. water, and 0*354 grm. 

 carbonic acid. 



To determine the base, the salt was heated to a temperature a 

 little below redness, the resulting carbonate extracted with water 

 from a small residue of carbon which could not be burned away 

 without volatilizing rubidium. The carbonate was then con- 



* The instrument used was a modification of Bunsen and Kirchhoff's 

 spectroscope, devised by Prof. J. P. Cooke, and manufactured by Messrs. 

 A Ivan Clark and Sons, of Cambridgeport. 



Phil. Mag. S. 4. Vol. 25. No. 167. March 1863. O 



