SPECTROSCOPIC DETEKMINATION OF LITHIUM. 31 



liiii, siicli close results have not been obtained when using this method 

 in the Bureau of Chemistry. When, therefore, lithium is present in 

 weighable amounts, and reasonable quantities of water are available 

 for examination, the amyl alcohol separation is i)referred. On the 

 other hand, the results obtained in the Bureau of Chemistry do not 

 agree with the statement of Kayser ^ that no satisfactory quantita- 

 tive analyses can be made by the spectroscope. For the estimation 

 of lithium in many waters on which work has been done, the method 

 of Nasmi and Anderlini could not be used because the waters con- 

 tained so little lithium that the spectrum could not be obtamed from 

 tlie water without considerable concentration. As in most of the 

 work complete analyses were required, the alkaline chlorids were 

 obtained pure, so that it was very little trouble to separate the lithium 

 by amyl alcohol, and when it was separated the lithium could be 

 determined by the spectroscope. The extraction with amyl alcohol 

 from the solution of the chlorids seemed the only way likely to give 

 all the litliium when the total amount was less than 1 mg. On 

 account of the fact that the lithium always remained in solution, there 

 was no serious trouble from either of the difficulties mentioned by 

 Ranzoli as objections to the gravimetric methods, 



METHOD USED IN THE BUREAU OF CHEMISTRY, WITH EXPERIMENTAL 



DATA. 



After the preparation of the sulphates of sodium, lithium, and 

 potassium for weighing, they are dissolved in a small amount of 

 water, from 1 to 10 cc according to the amounts of lithium expected 

 to be found. By trymg the solutions in the flame before the spectro- 

 scope it was quickly determined whether to dilute to the vanishing 

 point of the lithium line, as did Ballmann ^ and Bell,^ or whether to 

 make to a volume of 25 to 100 cc and add to distilled water as Foehr 

 did. In consideration of the errors of observation, it has not seemed 

 worth while to add sodium and potassium sulphates to the lithium 

 solution used for comparison. The potassium and sodium sulphates 

 are rarely present to the amount of more than 0.0005 gram each or 

 at the most 0.0010 gram each. In most cases the potassium spectrum 

 does not appear at all in the work and the sodium line is not much 

 brighter than it is in the lithium solution of known strength which 

 has been kept in a glass bottle for a short time. The spectroscope 

 used in the mvestigations is an ordinary high-grade instrument made 

 by Kriiss. The scale is illuminated by a small electric light, and 

 switches under the edge of the table enable the observer to illuminate 

 the scale or the room without moving. 



1 Handbuch der Spectroscople, 1910, B: 24. » Amor. Cheiii. J., 1885, 7: 3.5. 



« Zts. anal. Chem., 1875, 74; 297. 



