H. E. Merwin — Per oxidized Titanium Solutions. 125 



to 22° C. the solution is compared with a 100 cc solution con- 

 taining -01 g. of Ti0 2 , tt cc of H 3 9 and 2 or 3 CC of concentrated 

 sulphuric acid. The ratio, r, of depths (or volumes) of the 

 solutions is obtained as described in the preceding section. 

 This ratio is, however, much larger than the fluorine would 

 give, owing to the alkali sulphates. The ratio that the alkali sul- 

 phates alone would give, if free from interfering substances, 

 is about 125. Different samples never give quite the same 

 ratio. Those that give a ratio much higher than 125 probably 

 contain fluorine. The safest way is to make determinations of 

 this ratio on two 8-gram portions of the carbonates used in 

 the fluorine estimation,* and to use this ratio in making the 

 correction. Having obtained this ratio — call it m — the formula 



for computing the fluorine is : = grams of fluorine. 



v * 23,000 & 



Accuracy to one '0002 of a gram may be expected. The 

 probable error is therefore not half as great as with the 

 standard gravimetric methods. f 



If the fluorine expected amounts to '0025 to *0120 grams, 

 the test solution is made acid with 12 cc of concentrated sul- 

 phuric acid, and compared as before described. The formula is 



= grams of fluorine : m is to be determined and should 



6,300 s 



not much exceed 108. Accuracy to *0005 g. may be expected. 



Thanks are due to Professors J. E. Wolff and T. W. 



Richards for suggestions and criticism. 



Petrographical Laboratory, Harvard University, May, 1909. 



* The 8 grams of carbonates dissolved in about 75 cc of water are treated 

 precisely like the filtered fluorine solution. 



f Hillebrand, W. F., The Analysis of Silicate and Carbonate Rocks : Bull. 

 U.S. Geol. Survey, No. 305, p. 158, 1907. 



