24 THE DETERMINATION OF LITHIUM. 



be based on the relation between the amount of sodium on a wire 

 and the length of time required to volatilize it. 



Champion, Pellet, and Grenier,^ starting from the suggestion of 

 Janssen, developed a method for sodium where the sodium line was 

 extinguished by a blue glass made in the form of a wedge, with a 

 corresponding wedge of plain glass after the fashion of the compen- 

 sator in a saccharimeter. They calibrated the blue-glass wedge, 

 obtaining the quantities of sodium, the bright line from which would 

 just be extinguished by the prism in different positions. Working 

 on soils, they were able to get fairly accurate figures for the sodium 

 content. They determined the sodium always in a solution sat- 

 urated with potassium sulphate. The mean of three determina- 

 tions gave 0.297 gram of sodium oxid in 100 grams of water which 

 contained 0.3 gram, and they found 0.63 gram in 100 grams of water 

 containing 0.6 gram. 



In a note ^ on the work of Champion, Pellet, and Grenier, Janssen 

 referred to the value of approximate determinations at times and 

 suggested again the value of a method based on the duration of the 

 spectral lines. 



LocKYER and Roberts^ made an attempt at the quantitative 

 analysis of alloys by study of the spectra obtained by causing a 

 spark to pass between points of the alloys. As one of the constitu- 

 ents of the alloys decreased in amount its lines disappeared from the 

 spectrum. They obtained some figures as to the smallest amount of 

 a metal which might be present in an alloy and show its lines in the 

 spark spectrum. 



Truchot* made some determinations of litliium m waters by com- 

 paring the intensity and duration of the lithium Ime obtained from 

 the sample with that obtained from standards. He made standards 

 contaming from 5 to 40 mg of lithium chlorid per liter. This method 

 has been used in the investigations at the Bureau of Chemistry with 

 the modification that standards were made containing in addition to 

 the lithium different amounts of sodium chlorid, either 10 or 100 mg 

 per cubic centimeter, and the mixed chlorids which were to be 

 tested were dissolved m such an amount of water that the solution 

 contained either 10 or 100 mg per cubic centimeter. Using a 

 straight wire, 0.01 mg of lithium per cubic centimeter was about 

 the lowest limit with a solution of lithium chlorid alone. When the 

 solution of this strength contained 100 mg of sodium chlorid, the 

 lithium line was very much more difficult to see. A straight wire 

 with a rough surface will take up nearly the same amount of liquid 

 each time, but unless considerable care is exercised the amount is too 

 variable to give good quantitative results. The use of a small coil 



1 Compt. rend., 1873, 76: 707-11. a Proc. Roy. Soc. Lond., 1873, 21: 507-8. 



2 Compt. rend., 1873, 76: 711-13. « Compt. Rend., 1874, 7S: 1022-1024. 



