t 



SPECTROSOOPTC DETERMINATION OF LITIIIITM. 27 



stcite was strongly ignited to give the spectrum, lie found that as the 

 sohition was dihited the spectra lost their brilliancy so slowly that it 

 was very difficult to fix the point at wliich they disa])peared. He con- 

 sidered it possible to estimate rougiily the amount of ]>otassium in a 

 solution by this method, tlie error allowed being 10 times the quan- 

 tity of metal to be determined. By trying, like Truchot, to match the 

 intensity of lines he had no better success. He mentioned as the 

 limit for potassium 0.0006 nig of ])otassium in the form of a salt. It 

 seems possible that the error of 10 times the quantity sought should 

 have read "0.1 times the quantity sought." Owing to the strain 

 upon the eyes and the inaccuracy of the results, he gave up the attempt 

 to use a spectroscopic method. 



LiVEiNG and Dewar^ made some experiments on quantitative 

 analysis by obtaining absorption bands due to small quantities of 

 sodium vapor in iron tubes which were kept very hot. The results 

 were not of any practical value. 



Bell 2 published a description of a spectroscopic method for the 

 quantitative determmation of lithium which was identical with the 

 method of Ballmann, as noted by Hofmann.^ He used a small loop 

 for bringing the solution into the flame and called attention to the 

 necessity of keeping the flame of constant mtensity and always 

 bringmg the loop mto the same part of the flame. He developed the 

 use of the method for lithium determinations, but stated that it 

 might be used for other substances, especially thallium. Calcium, 

 strontium, barium, and potassium, he said, do not give lines suffi- 

 ciently persistent, except in strong solutions, which is true also of 

 caesium and rubidium, while sodium is so ubiquitous as to give very 

 uncertain results. 



FoEHR,* commentmg on Bell's ^ paper, stated that he himself had 

 made many experiments on quantitative analysis by the spectro- 

 scope, but did not obtain as good results as those reported by Bell. 

 Instead of diluting until the spectrum failed to appear, he added a 

 solution of the substance a little at a time to pure water, testing the 

 mixture by means of the spectroscope after each addition. The end 

 point was the concentration at wliich the spectruin ffist appeared and 

 was found to be much sharper by this method. He determined potas- 

 sium by adding the solution of the potassium salt to a saturated solu- 

 tion of ammonium clilorid until the spectrum of potassium appeared, 

 and also made determinations by adding an excess of ammonium 

 clilorid solution to the unknown })otassium solution, then adding 

 solution of known potassium content until the potassium Ime ap- 

 peared again. Foehr stated that the lithium line is much more 



1 Proc. Roy. Soc, Lond., 1879, Z9: 482^89. * Chem. Ztg., 1885, 9 : 1013. 



2 Amer. Chem. J., 1885, 7 : 35. o Amer. Chem. J., 1885, 7 : 35. 



3 Ber. (1. ehcm. Gas., 188.5, IS : 2897. 



