OP ARTS AND SCIENCES. 



187 



These results show very plainly that the solvent effect of anhydrous 

 amyl alcohol used for washing under the conditions described is tri- 

 fling in the extreme, and may be neglected utterly providing the amount 

 of the washing is not altogether disproportionate to the needs of the 

 case. 



We pass next to the consideration of the separation of the chlorides 

 of sodium and potassium from lithium chloride. Weighed portions of 

 the test solutions were concentrated and treated with amyl alcohol in 

 the manner described until the precipitated salt was entirely free from 

 water and the supernatant alcoholic solution of the lithium chloride 

 boiled constantly at a point not far from that of the amyl alcohol 

 employed. Then the liquid was cooled, a drop or two of strong 

 hydi'ochloric acid added in accordance with the evident suggestion of 

 the preliminary expei'iments previously mentioned, and heat again 

 applied until the boiling had continued, as before, for some minutes at 

 one point. The filtration, washing, drying, and weighing of the resi- 

 due were effected as in experiments (18j to (22). In those of the 

 experiments in which the lithium salt in solution was also determined, 

 the end was accomplished by evaporating the filtrate and washings 

 to dryness, treating the residue with sulphuric acid, and igniting and 

 weighing as lithium sulphate. In the following table the weight 

 of insoluble chloride actually found is given in one column, and this 

 weight, corrected according to the data previously determined for the 

 solubility of the chloride in the residual amyl alcohol appears in the 

 column adjoining. So also the weight is given of the lithium sulphate 

 actually found, and an adjacent column contains the result of correct- 

 ing this weight for the accompanying sodium or potassium sulphate, 

 or both, upon the hypothesis that these salts are neutral sulphates 



