180 APPENDIX III. 



substance, and to treat the ash. with sulphuric acid, 

 to exhaust the resulting sulphates with absolute 

 alcohol and evaporate the alcohol extract to diyness, 

 and to test the residue for lithia. 



Kirchhoff and Bunsen state that less than liUO ,' J;000 of 

 a milligramme of carbonate of lithia = to about ^,^. co 

 of a grain can be detected by the spectrum analysis. 



To .determine the delicacy of the test for the 

 chloride of lithium, the following experiment was 

 made : One grain of chloride of lithium was dis- 

 solved in one litre of water. Of this solution 

 100 cub. centims. were taken and again diluted to 

 one litre, this latter solution containing O'l grain of 

 chloride of lithium to the litre. 



When further diluted to five times its bulk, the 

 lithium reaction was still seen faintly on a wire 

 taking up 0*06 grain of solution. The line is most 

 distinctly visible in the evening, in a somewhat 

 dark room. 



This dilution is equal to 0*1 grain of chloride of 

 lithium in 5 litres of water, or 1 grain in 50 litres. 

 One litre = 15,440 grains, or 50 litres = 772,000 

 grains. In 0'06 grain of this solution there are 

 therefore 0-00000008 grain chloride of lithium, or 

 about i 2i(J oo,uooth of a grain of chloride of lithium. 

 This contains only - part of lithium, so that the 

 72,ooo,oooth of a grain of metallic lithium, when pure, 

 gives the spectrum reaction. 



When the chloride of lithium was dissolved in 

 urine, the test was from twice to six times less 

 delicate than in distilled water. 



