SPECTROSCOPIC DETEEMINATION OF LITHIUM. 



37 



While in work of this character it would have been better to have 

 used cliemicals which were entirely free from even the most minute 

 traces of lithium, it may be well to add that blanks should alwa3^s 

 be made and that in this case the blank obtained of 0.0008 mg is the 

 largest blank determination ever found in any of the authors' work. 

 From the above data it will bo seen that 250 cc of the original mineral 

 water after subtracting the blank contained 0.0008 mg of lithium or 

 0.003 mg per liter. Several determinations of the lithium in two 

 liters of other samples of the same water gave figures of less than 

 0.005 nig of lithium per liter. The use of the combined amyl alcohol 

 and spectroscopic method on some samples of soil gave the following 

 results on the air-dried soils. 



Results on air-dried soils from the use of combined amyl alcohol and spectroscopic method. 



I Two determinations by the spectroscope of the amount of lithium in tlie same sample. 



As a regular practice m analytical work, the approximate deter- 

 mination of Uthium in the mixed clilorids obtained in water analyses 

 saves very much time by indicating whether it is worth while to 

 extract the lithium for weighing. A similar test is regularly made 

 on the solution in nitric acid of the oxids of the calcium and stron- 

 tium. By making the solution to a definite volume it is easy to tell 

 whether there is enough strontium present to separate and weigh. 

 Though potassium is regularly determined in all complete analyses 

 of water in the Bureau of Cliemistry, it is occasionally desired to 

 obtain a close approximation of the amount present in some sub- 

 stance where the information is not worth the trouble and expense 

 of a gi-avimetric determination. In many of these cases relatively 

 large percentage errors are of no consequence. It is often, however, 

 very desirable to be able to say that there is present 0.001 mg or 

 0.03 mg rather than report each amount as "trace" or "spectro- 

 scopic trace," which may be done if the lines are seen in the spectro- 

 scope without any idea as to the amount of the material required to 



