SPECTROSCOPIC DETKRMINATION OK JJTHIUM. 29 



flame, the spark, ami by absorption, reporting lithium as best shown 

 in the gus-oxygen flame. 



Ranzoli,* eomparing the spectroscopic and gravimetric methods 

 for litliium after obtaining the results quoted above (p. 14), 

 detei-mined the lithium by Foehr's method and found 0.52078, 

 0.51688, and 0.5151 gram of litliium clilorid instead of 0.54 gram, 

 the average being 4.15 per cent low. Then he used the method as 

 modified by Nasini and Anderlini - and found 0.53612 and 0.58914 

 instead of 0.5400 gram. On some solutions of unknown lithium con- 

 tent he obtained 0.2708 gram when the solution contained 0.2700 

 and obtained 0.18953 gram from a solution containing 0.189 gram. 

 The average error on the three solutions was 0.365 per cent. Ran- 

 zoli advocated the use of tliis method in preference to the gravi- 

 metric method for all waters, as the determination is made on the 

 original water added to distilled water and there is no chance for 

 loss or gain of lithium in manipulation. 



Hermann,^ in some investigations on the presence of lithium in 

 the human bod}", determined the presence of lithium by the spectro- 

 scope. With the apparatus used, he could detect 0.0002 mg per 

 cubic centimeter. He used a drop of 0.01 cc volume. 



Nutting,^ working with the spectra of alloys, confirmed to a cer- 

 tain extent Lockyer's ^ supposition that arc spectra might be used 

 for quantitative analysis of allo3"s. He said (par. 4, p. 137): 



4. Spectroscopic quantitative analysis to within an error of perhaps 5 per cent 

 appears to be practicable, provided: (a) Selected lines of similar character are uped 

 for comparison; (b) spectra are taken with an arc or spark with capacity and induct- 

 ance, with Sufficient current to produce plenty of metallic vapor in proportion to the 

 ambient gas: (c) allowance is made for difference in the natural intensity and hard- 

 ness of the spectra of various pure metals taken under the same conditions; (d) allow- 

 ance is made for differences in the atomic weiglits of components, pro\-ided these 

 differ by a considerable amount. 



Abati ® determined the lithium in a mineral water by the method 

 of Nasini and Anderlini. He used a bunsen burner with the center 

 of the flame 4 cm from the slit of the spectroscope, which had a 

 width of 0.2 mm. He used a loop made from wire 0.17 mm in 

 diameter wound in six turns, making a spiral 1.82 mm in diameter 

 and 1.1 nmi liigh. The loop was filled with solution, the water gently 

 evaporated and the dry loop brought into the flame. The water to 

 be examined was diluted until the lithium line was barely visible. 

 This diluted solution was added, a little at a time to 25 cc of water 

 until the line appeared in the spectroscope. At first the standard 



1 Oaz. chim. ital., 1901, 31 (I) : 40-48. 



= Gaz. cMm. ital., 1890, 30 (I) : 305. 



» tjberdas Vorkommen ties Llthions im inr-nsohlidion Organisniiis, iKOo. 



* Astrophys J., 1905, St- 131-KJ7. 



6 I^^e. cit. 



6 Gaz. chim. ital., 1900, 3b' (II) : 85o-8t)0. 



