hillebrand.] SILICA, SEPARATION FROM ALUMINA, ETC. 47 
tightly fitting glass stopper and kept over sulphuric acid. It must 
not be kept long without reheating, because of being hygroscopic. 
Another form of tube used by Jannaseh for special purposes is 
shown in fig. 11. Minerals, such as topaz, which is not fully decom- 
posed by the borax method and which contains a large amount of 
fluorine, are fused at h with about six times their weight of lead oxide. 
A layer of lead oxide between c d serves to retain any fluorine escap- 
ing from the fusion. 
VI. SILICA, SEPARATION FROM ALUMINA, ETC. 
ALTERNATIVE METHODS OF DECOMPOSITION. 
PKELI M I X A K V REM AKK S. 
The practice of separating alumina, etc., by the usual methods, after 
first attacking the rock powder by hydrofluoric and sulphuric acids — 
silica being estimated in a separate portion — while attractive in prin- 
ciple, was abandoned by the writer after fair trial, owing to the disturb- 
ance sometimes occasioned }yj incomplete expulsion of fluorine and to 
a less degree by the presence of sulphates instead of chlorides. With 
exception of the comparatively few analyses made thus, the sodium- 
carbonate method has always been employed. In the case of rocks 
rich in fluorine strict accuracy would require the separation of silica 
to be made as in the Berzelian method for fluorine estimation (see 
footnote, p. 50), but in practice it is not often necessary to resort to this 
tedious procedure, since the amount of fluorine is usually small and 
it can by no possibility cause a loss of much more than three-fourths 
its own weight of silica by volatilization as silicon fluoride when the 
sodium-carbonate fusion is evaporated directly with hydrochloric acid. 
Probably the loss is less, since some fluorine perhaps escapes as hydro- 
fluoric acid. However this may be, the error is of comparatively 
slight importance, since it attaches to the constituent always present 
in greatest amount. 
Various fluxes other than alkali carbonates have been recommended 
for breaking up silicates insoluble in ordinary acids, such as lead and 
bismuth oxides, lead carbonate, borax, and boric oxide. Professor 
Jannaseh and his pupils have been especially active of recent years in 
this line of work, as evidenced by their numerous published papers. 
One of the advantages these fluxes possess over the alkali carbonates 
is their removability after serving their purpose, thus allowing the 
various separations to be made more perfectly and without the annoy- 
ing interference of several grams of foreign fixed salts, which are 
most troublesome in that part of the analysis devoted to the separation 
of silica, alumina, iron, lime, and magnesia. 
