M. B. Riggs — Composition of Tourmaline. 39 



pleteness of the disintegration. In the case of the tourmalines, 

 with their low silica percentages, the solution is usually com- 

 plete if the decomposition has been complete. After the 

 alkalies have been freed from the other bases by the usual 

 well-known methods, though the greater part of the boric 

 acid will have been driven off by the repeated evapora- 

 tions, more or less may remain. This residual is removed 

 by two or three evaporations with methyl alcohol. The sep- 

 aration of lithia from the other alkalies was made by the G-ooch 

 method, i. e., by boiling the mixed chlorides in amyl alcohol, 

 and the further separation of the potash and soda effected after 

 the usual manner. 



Silica. — On the strength of the belief that tourmaline some- 

 times contained fluorine in considerable amounts, the silica 

 was at first estimated by precipitating it with carbonate of am- 

 monia. "When the bases, thrown out by this reagent, are so 

 in excess of the silica as they are in tourmaline, its precipitation 

 is quite easy and quantitative. Nevertheless, when the method 

 was used of duplicate determinations, the higher was taken in 

 the summation of the analysis. (Save where the silica is 

 greatly in excess of the bases thrown down by ammonium car- 

 bonate, the addition of zinc or like compounds, as is commonly 

 recommended, is wholly unnecessary. In fact, even in dealing 

 with such minerals as the lepidolites, which contain about fifty 

 per cent of silica and less than thirty of alumina, the addition 

 of zinc compounds gives no better results than can be obtained 

 by simple evaporation with carbonate of ammonia, the evaj> 

 oration being repeated several times.) 



So soon as it became evident that the amount of fluorine, 

 if present at all, was so small that its influence in carrying off 

 silica (the amount of silica carried off by fluorine on evap- 

 orating a soda fusion with hydrochloric acid is but a small part 

 of the tetrafluoride equivalent) could be neglected, the ordi- 

 nary method of separation was employed. In all cases the 

 silica was corrected by evaporation with hydrofluoric acid. 



Alumina. — The only point worthy of mention in this con- 

 nection is the necessity of testing the alumina for the silica 

 which it frequently contains. This is usually done by fusing 

 the ignited oxides with bisulphate of potash. But when they 

 amount to as much as they do in the tourmaline a carbonate of 

 soda fusion works more satisfactorily. It can be continued 

 longer and at a higher temperature. This fusion is readily con- 

 verted into a sulphate fusion and the desired silica separation 

 thus accomplished. As regards other determinations nothing 

 in particular need be said. 



