48 Prof. Bunsen on the Preparation of 



X 1/ 



of A and B employed : if we put £= -p and tj= j-, and if we 

 suppose that the error (SA) made in the determination of A and 



* B 



A 



B is in both cases the same, then we have £= —, .and the 



o — a 



probable error of (• 



"-^■V^STr • • • ' 2 » 



If in the preceding example the error in the determination of A 

 and B in each case amounted to 0*001 grm., the most probable 

 inaccuracy in the determination of the chlorides of rubidium 

 and caesium in 1 gramme of both does not exceed 0*0055 grm. 



(c) I have also employed an indirect method for the estima- 

 tion of the chloride of lithium, as the direct separation of this 

 metal from those of the other alkalies by extraction of the chlo- 

 ride or sulphate with a mixture of alcohol and ether, is as inac- 

 curate as the determination by means of phosphate of sodium. 



8*3300 grms. of the anhydrous substance was extracted four 

 consecutive times with a mixture of ether and alcohol, the residue 

 obtained by evaporating the alcoholic liquid was again treated 

 with ether and alcohol, and the residue obtained by evaporating 

 this second extract was heated until fusion commenced. This 

 mass weighed 0*0628 grm., and, when dissolved in water and 

 treated with nitrate of silver, yielded 0*1650 grm. of chloride of 

 silver. The filtrate from the precipitate of chloride of silver was 

 carefully freed from silver by hydrochloric acid, and the potas- 

 sium precipitated as the double platinum-salt; 0*0278 grm. of 

 this double salt was thus obtained, corresponding to 0*0088 grm. 

 chloride of potassium and 0*0163 chloride of silver. The salt 

 weighing 0*0628 grm. contains y of chloride of sodium, x of 

 chloride of lithium, and 0*0088 of chloride of potassium ; if we 

 subtract the weight of the latter salt, we have A= 0*054 grm. 

 for the sum of the weights of x and y ; and for the weight of 

 chloride of silver which these two chlorides yield, we have 

 B=01487. 



If the corresponding values of a and b for the chlorides of 

 lithium and sodium are substituted in formula (1), we have 



x= 1*0823 B- 2-6525 A, 



and hence x — 0*0177 ; so that 89*08 parts of the dried sub- 

 stance contain 0*190 part of chloride of lithium. 



If the chloride of sodium be estimated by difference, we obtain 

 the following as the percentage composition of the saline mass : — 



