THE NEW ENGLAND STATES. 17 



would at least approximately determine an amount of lithium 

 equivalent to 0.01 of a part per million or less seemed highly desir- 

 able. Naturally, the employment of some method which would take 

 advantage of the very delicate emission spectrum of lithium sug- 

 gested itself as a possible means of solving the problem. A search 

 of the literature showed that such attempts had been made. A 

 dilution method based on the disappearance of the lithium line 

 was proposed by Ballmann * and later modified by Bell. 2 A spec- 

 troscopic method has also been proposed by Fruchot. 3 Bell's 

 method was tried, but the results were far from satisfactory, and 

 it was believed that the Fruchot method gave more promise of satis- 

 fying the needs of the problem in hand. With the assistance of 

 W. D. Collins, of the Water Laboratory, the method has been finally 

 so modified and perfected that results of great accuracy may be 

 obtained in solutions of high dilution. By this method, using a 

 hydrogen flame, a quantity of lithium equal to 0.00001 mg of 

 lithium in 1 cc of solution can be detected, while variations of 

 0.00001 and 0.00002 can be determined with a fair degree of accuracy. 

 The smallest amount of lithium which it is possible to detect with 

 the spectroscope, using a hydrogen flame, is about 0.0000001 mg. 

 This method has been employed in checking most of the analyses 

 for this report. Where lithium is reported as a trace it means 

 that no weighable amount was determined by the amyl alcohol 

 separation using the amount of water noted, but by examination 

 with the spectroscope the lithium line was found. 



METHOD OF STATING RESULTS. 



The results of the chemical analysis have been first stated in the 

 so-called ionic 4 or radical form followed by a statement of the 

 hypothetical combinations, calculated from the former according to 

 the empirical method formulated by Haywood, 5 in which the sodium 

 is first combined with nitrous, nitric, and metaboric acids, potassium 

 with iodin and bromin and calcium with phosphoric acid. The 

 residual basic ions are then assigned in the following order ammo- 

 nium, lithium, potassium, sodium, magnesium, calcium, manganese, 

 iron, and aluminum; the residual acid ions in the following order 

 chlorin, sulphuric acid, bicarbonic acid, and carbonic acid. In case 



1 Zts. anal. Chem., 1875, 14: 297. 



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

 sCompt. rend., 1871, 78: 1022. 



The term "ion" employed throughout the statement of analyses is used in its most comprehensive sense 

 and to replace the less comprehensive term "radical" heretofore employed. The use of the term to include 

 both actual and potential ions is sanctioned by Ostwald, Principles of Inorganic Chemistry, p. 244. The 

 same difficulty is experienced in the statement of silica as under the older form of statement. Whether 

 silica exists in waters in combination as soluble silicate or in a colloidal form has not been definitely deter- 

 mined. Evidence collected by the author indicates that ha some cases soluble silicates do exist, and v^cn 

 necessary to balance an analysis this view has been taken throughout the bulletin. 



*U. S. Dept. Agr., Bureau of Chemistry, Bui. 91. 



84644 Bull. 13911 2 



