204 M. Stas on the Determination of Atomic Weights, 



ture warmed with continual agitation as long as the precipitate 

 increased. The salt was put into a cylinder which stood on 

 a flask in connexion with the air-pump. By this means all the 

 mother-liquor could be drained away. It was washed in the same 

 way with ammoniacal water. From the wash- waters, by eva- 

 poration and ignition and treatment in a similar manner, afresh 

 portion of carbonate of lithium was obtained. The salt was now 

 dissolved in nitric acid, again precipitated with carbonate of am- 

 monium, and the same operations repeated five times. In this 

 manner 55 grammes of pure salt were obtained from 1200 

 grammes of crude carbonate of lithium. The salt obtained from 

 the mother-liquor was diffused in water and brought into solu- 

 tion by means of carbonic acid gas (prepared by igniting bicar- 

 bonate of sodium). The clear solution boiled in platinum vessels 

 separated a pure carbonate of lithium, which was once more sub- 

 jected to a similar treatment. The salt was dissolved in a platinum 

 boat heated to 160°-175°, and a current of dry hydrochloric 

 acid passed over it, whereby all spirting was avoided. When the 

 decomposition was finished, the chloride of lithium formed was 

 fused in the atmosphere of hydrochloric acid gas, and cooled in 

 a current of nitrogen. The platinum boats were directly put 

 into cylindrical vessels, which were provided with metal caps and 

 stopcocks. The apparatus were pumped empty and weighed. To 

 dry these cylinders, the author introduces small iron boats con- 

 taining oxide of copper and soda. For this purpose the author ig- 

 nites in a porcelain tube the iron boats, which are half-filled with 

 a mixture of 1 part of saltpetre and 2\ finely divided copper. This 

 mass dehydrates more rapidly than anhydrous phosphoric acid. 



For the purpose of an accurate investigation of the chloride of 

 lithium, the contents of one of the boats was dissolved in water. 

 The solution slowly but distinctly turned litmus-paper blue. 

 The alkaline reaction remained even after part of the chloride of 

 lithium had been volatilized in a current of hydrochloric acid, it 

 being allowed to cool in a current of nitrogen. Entirely vola- 

 tilized in a current of hydrochloric acid, chloride of lithium left 

 no trace of a residue. 



If this chloride of lithium was mixed with solution of silver in 

 the ratios of the atomic weights CI = 35*5, Li = 7, Ag=108, the 

 liquid always contained an appreciable trace of silver, which 

 could be estimated by means of normal solution of sal-ammoniac. 

 Here, as well as in the case of Na CI and K CI, the phenomenon 

 was observed that an almost precipitated solution of silver, which 

 still contains 1-2 milligrammes of silver in solution, is precipi- 

 tated both by chloride of sodium and by solution of silver. 

 The more acid the solution of silver, the more marked is the 

 phenomenon. 



