30 THE DETERMINATION OF LITHIUM. 



solution was made up according to a former analysis of the water 

 wluch gave 0.047 gram per liter of lithium. A rough comparison 

 showed that this solution contained about twelve times as muchlithium 

 as the sample for analysis; therefore, a new solution was prepared 

 with the same amounts of sodium, potassium, and calcium chlorids, 

 and one-twelfth as much lithium. When 147 cc of this solution 

 were added to 25 cc of distilled water the line litliium could not be 

 seen. With 148 cc one trial gave the litliium line, while with 149 cc 

 the majority of the drops taken showed the line, and when 150 cc 

 were added every drop tried gave the line. One hundred and forty- 

 seven cubic centimeters of the water mixed with 25 cc of distilled 

 water failed to show the lithium line, with 148 cc added the line 

 appeared in a majority of the trials, and with 149 cc added it was 

 shown by every drop. These readings would indicate an error of 

 less than 1 per cent in the determination which is a greater accu- 

 racy than we have regularly obtained when working vnih this method. 



MoTTRAM ^ determined the amounts of sodium and potassium in 

 tissues by a spectroscopic method depending on the relation between 

 the weight of the substance and the time it required to completely 

 volatilize in the flame as shown in the spectroscope. He found that 

 it took 6.5 minutes to volatilize 0.001 mg of sodium as sulphate and 

 that the presence of other substances and the acid radical combined 

 with the metal aft'ected the time required for volatilization. This 

 is an extension of the method of Truchot,^ and the suggestion of 

 Janssen ^ in regard to the use of the duration of the spectrum as a 

 measure of the amount of material. Using a straight wire we made 

 some rough tests of this suggestion when using Truchot's method. 

 The results were not entirely satisfactory. 



Hempel and Klemperer ^ devised a complicated method for obtain- 

 ing quantitative results by spectroscopic analysis of soils. It is 

 much like one of the methods used by Mitscherlich.^ They used a 

 spark to volatilize water over which was passed hydrogen which was 

 burned with oxygen before the slit of the spectroscope. A solution 

 of known content of lithium or potassium was let into the vessel 

 until the lithium or potassium line was shown in the spectroscope, 

 and a solution of the soil then added to the water with the passage 

 of the spark, and when the oxyhydrogen flame gave the spectrum of 

 lithium or potassium the quantity present could be calculated. 

 From 0.0082 to 0.035 per cent of litliium oxid was found in various 

 ignited soils, with an average value of 0.0148 per cent for 10 soils. 



WliUe Xasini and Anderlini, Ranzoli and Abati all gave most 

 excellent results obtained bv the method used bv Nasini and Ander- 



' Arch. Middlesex Hosp., 1909, 15: lOfi-HT. i Zts. angew Chem., 1910, 23: 1756-1759. 



2 Compt. rend.. 1874, 7S: 1022. 6 Pogg. Arm., 1864, 121 : 459-488. 



3 Compt. rend., 1873, 76: 711-713. 



