84 Penjield and Pratt — Chemical 



is about 4-1 which can be lowered by addition of KN0 3 . The 

 fused salt is a clear mobile liquid, through which the particles 

 of mineral move freely, and separations can be made in this as 

 accurately as in any of the heavy solutions. On cooling, the 

 fusion solidifies to a cake with the heavier and lighter portions 

 at the bottom and top respectively. The test tube readily 

 breaks away from the fused mass, the cake can be cut in two 

 and the minerals separated by dissolving the nitrates in water. 

 The latter can be reclaimed by evaporating the solutions to 

 dryness on a water bath and finally fusing. By eliminating 

 the heavier and lighter portions and repeating the separation 

 remarkably pure products were obtained, of nearly uniform 

 specific gravity. The manipulations are very simple and the 

 results extremely satisfactory. A preliminary experiment that 

 was made showed that staurolite does not suffer any decompo- 

 sition or loss in weight when exposed to a temperature of 

 250° 0. The separated material, when examined with the 

 microscope, was found to be homogeneous and very free from 

 visible inclusions. 



Method of analysis. — The silica and bases were determined 

 by well known methods. The evaporations were carried on in 

 platinum, the purity of the silica tested by evaporation with 

 hydrofluoric acid and account taken of the small quantity of 

 silica carried along and weighed with the sesquioxides. Espe- 

 cial pains was taken in the determination of ferrous and 

 ferric iron. The very finely pulverized mineral was treated in 

 a small platinum bottle with a mixture of strong hydrofluoric 

 and sulphuric acids and boiled vigorously for about twenty 

 minutes, the neck of the bottle being covered by a cone of 

 platinum foil. The contents of the bottle were then diluted 

 with cold, boiled water, washed into a casserole and titrated 

 with potassium permanganate. Preliminary experiments were 

 made by treating known weights of ferrous sulphate in the 

 same manner and it was found that no appreciable oxidation 

 from the air took place. As the staurolite is very slowly 

 attacked by hydrofluoric acid only a portion in each experi- 

 ment went into solution. After titration, the insoluble portion 

 was filtered off and the filtrate evaporated in a platinum dish 

 till all the hydrofluoric acid was expelled. After diluting, the 

 iron was reduced by hydrogen sulphide, the excess of the latter 

 removed by boiling and the total iron determined by means of 

 potassium permanganate. The determinations give the ratio 

 of ferrous to ferric iron in that portion which had been dis- 

 solved by the hydrofluoric acid, and the total iron in the min- 

 eral having been previously found in that portion used for 

 silica and bases, the percentages of ferrous and ferric iron 

 are readily calculated. Direct determinations of water were 



