OF ARTS AND SCIENCES. 



179 



Fresenius * found on examining the method that several repetitions 

 of the treatment by evaporation and extraction were required to com- 

 plete the recovery of all litliium phosi)hate, and advised that the 

 operation be continued until residual lithium phosphate fails to appear. 

 The results of Fresenius's experiments with lithium carbonate, recal- 

 culated with the use of the number 7 as the atomic weight of lithium, 

 are given in the table appended. 



Thus it will be seen that in the nine experiments of Mayer the error 

 ranges from 0.01 33-[- grm. to 0.0089 — grm., and that of the deter- 

 minations of Fresenius from 0.0001-]- grm. to 0.0041-(- grm. for the 

 dried precipitate, and from 0.0006-]- grm. to 0.0036 — grm. for the 

 ignited precipitate. 



If the tendency of lithium carbonate to fall in company with the 

 phosphate were not to assert itself during the evaporations of solutions 

 of salts of lithium in presence of sodium hydrate and in contact with 

 ordinary atmospheric air, it would surely be strange, and this point 

 may be fairly set down as one of the weak ones of the method ; but 

 the gravest source of error, and that indicated most unmistakably 

 throughout the whole history of the process, — which has been re- 

 counted at some length for the purpose of emphasizing this very 

 matter, — is the impossibility of preparing the lithium phosphate in 

 anything like a condition of freedom from other alkaline phosphates 

 without a careful and prolonged washing which is sure to result in 

 loss of the lithium salt by solution. When it is remembered that 

 according to Mayer's determinations tri-lithium phosphate requires 

 for solution only 2539 parts of water, or 3920 parts of a mixture of 

 ammonia and water in equal portions, it is plain that the success of 

 the method depends upon the ability of the analyst to wash to a con- 

 dition of purity, and without loss of that which it is the purpose of the 



* Zeit. fur Anal. Chem., i. 42. 



