28 THE DETERMINATION OF LITHIUM. 



easily seen from lithium chlorid solution than from a solution of 

 lithium sulphate of equivalent concentration, with which statement 

 the authors can not entirely agree, for in their work they have found 

 the amount of lithium requhed to show the lithium line m the spec- 

 troscope to be about the same whether it is present in solution as the 

 sulphate or chlorid. 



GoocH and Hart ^ determined potassium by the use of the spec- 

 troscoi)e by matching the mtensity of the red line of potassium in 

 known and unknown solutions. They used a Bunsen burner and 

 brought the solution into the flame on a spiral coil holding 0.020 gram 

 of solution and found that the brilliancy of the red Ime in the potas- 

 sium spectrum was increased by the presence of sodium salts until 

 the amount of sodium was 100 times that of the potassium, after 

 which more sodium mad« the potassium line appear less distinct. 

 Therefore they made their potassium standards to have about the 

 same concentration of sodium as the unknown solution and reported 

 very close results in their test analyses. 



GoocH and Phinney ^ determined rubidium by a method similar to 

 that of Gooch and Hart for potassium. The interference of potas- 

 sium with the rubidium determination was more marked than the 

 interference of sodium with potassium. Working with 1, 2, and 1 

 mg of rubidium in rubidium chlorid with 0.1 gram of potassium as 

 chlorid, they precipitated the greater part of the potassium salt by 

 alcohol and then found 0.8, 1.7, and 0.9 mg of rubidium, respectively, 

 showing errors of from 10 to 20 per cent. 



Nasini and Anderlini ^ determmed lithium in a thermal water by 

 the method of Foehr. At another time •* in the analysis of a brine 

 they used this method on the original water, &st obtaining a rough 

 approximation as to the amount and then maldng up a solution with 

 a known amount of lithium and about the same content of other sub- 

 stances as the water under exammation. When this solution was 

 added to distilled water and tested with the spectroscope the amount 

 of lithium present when the lithium line first appeared was different 

 from the amount required when no other salts were present. They 

 noted the necessity of having the wires and drops of uniform size, of 

 keepmg the Bunsen burner flame of constant size, and of always 

 ])lacing the wire in the same part of the flame. 



VoGEL •'"' studied the different methods of producing flame spectra 

 and concluded that for general use the illuminating gas-oxygen 

 flame was the most satisfactory as the spectra were brighter than 

 with the Bunsen burner flame. He made lists of the elements best 

 studied by the Bunsen burner flame, the illuminating gas-oxygen 



1 Amer. J. Sci., 1890, (3) 42 : 448. •• Gaz. chim. ital., 1890, SO (I) : 305. 



2 Amer. J. Sci., 1S92 (3) U : 392-400. ^ zts. anorg. Chem., 1894, 5: 42-62. 



3 Giiz. chim. ital,, 1,S94, S4 (I) : 327. 



