180 PROCEEDINGS OF THE AMERICAN ACADEMY 



process to save, a precipitate peculiarly prone to retain foreign matter 

 and soluble in the washing mixture in the proportion of ten milligrams 

 to every 40 cm.^ of the latter. Of course washings will never be 

 entirely saturated, nor will the precipitate be as soluble at the begin- 

 ning of the operation as at the end, when the precipitant no longer 

 exerts an action which tends to lessen solubility ; but in view of the 

 difficulties which present themselves, it is sufficiently obvious that exact 

 results obtained by Mayer's process owe their apparent accuracy to a 

 fortuitous balance of errors. The difference of 0.0222 erm. between 

 the extremes of Mayer's experimental results should not be surprising ; 

 and, at the best, the process is tedious and not entirely trustworthy, — 

 Jacts of which its author was not unmindful. 



In Rammelsberg's method of separating lithium chloride from the 

 chlorides of sodium and potassium the sources of error are, in brief, 

 the solubility of sodium chloride and potassium chloride in the ether- 

 alcohol mixture, the influence which the presence of small amounts of 

 water exerts upon the solubility of these same salts, the difficulty of 

 bringing the chlorides to the anhydrous condition without decomposing 

 the lithium chloride to a greater or less extent, and the mechanical 

 difficulties of transferring the fused or crusted chlorides to a suitable 

 receptacle for digestion and agitation in the solvent, and of extracting 

 perfectly the soluble constituents of closely compacted matter. Of the 

 last two items nothing need be said in explanation beyond simply 

 noting them. The third is particularly important, inasmuch as the 

 tendency of lithium chloride, first noted I believe by Mayer, to ex- 

 change chlorine for oxygen when ignited in presence of water, results 

 in the formation of lithium hydrate or, in contact with products of 

 combustion, lithium carbonate, both of which are insoluble in the 

 mixture of ether and alcohol, and remain with the sodium and potas- 

 sium chlorides. As to the effect of water in the mixture, an experi- 

 ment of Mayer, in which it was found that 100 cm.^ of a mixture of 

 alcohol of 9G% and ether of 98% dissolved 0.1100 grm. of sodium 

 chloride, is instructive. In regard to the solubility of the chlorides 

 of sodium and potassium in the mixture of anhydrous ether and 

 alcohol, Rammelsberg's statement, that from 0.9770 grm. of jjure 

 strongly heated sodium chloride with an undetermined amount of 

 lithium chloride the mixture extracted 0.0130 grm., is unfortunately 

 meaningless in the absence of information concerning the amount of 

 solvent employed. J. Lawrence Smith * found, in making an exami- 



* Am. Jour. Sci. [2], xvi. 56. 



