316 PROCEEDINGS OF THE ACADEMY OF [1880. 



tional titanic acid was separated upon boiling the filtrate from SiO,„ 

 after reduction with H-S. In one analysis titanic acid was sepa- 

 rated from SiOoby volatilizing the latter with HF, dissolving the 

 residue in HsSOi, diluting and boiling. Ferrous oxide Avas deter- 

 mined in the air-dried mineral as in analyses I and II. Iron and 

 alumina were estimated by precipitation b}' boiling with sodic 

 acetate in a neutral solution, dissolving in HCl, reprecipitating 

 with NH4HO, igniting and weighing together. In the filtrate 

 MnO was precipitated by bromine and ignited. 



The following method was employed for the detection of vana- 

 dium. 80 grammes of impure mineral were mixed with 90 grammes 

 of sodic carbonate and 100 grammes of sulphur, and the whole 

 heated slowly in a Hessian crucible covered by charcoal until 

 partially fused. It was then digested in warm water, filtered, and 

 to the filtrate dilute IICl was added, precipitating a copious 

 heavy flocculent brown mass of the sulphides of vanadium, copper, 

 cobalt and nickel. The precipitate was washed, ignited and evap- 

 orated with nitric acid, when it gave a red residue. This was fused 

 with a mixture of sodic carbonate and sodic nitrate, and extracted 

 with water in order to separate the oxides of copper, cobalt and 

 nickel. Solid ammonic chloride was now added to the aqueous 

 solution, when vanadate of ammonia was precipitated. Upon 

 ignition it was changed to vanadic oxide, and was found to be 

 pure, giving all the characteristic reactions. 



For the estimation of vanadium the following method was em- 

 ployed. 4^ grammes of the pulverized ignited mica were fused 

 with a mixture of 3 parts NaCOj and 1 part NaNOs, the mass ex- 

 tracted with H.O, filtered, and the filtrate digested with HiS. 

 Traces of CuS and FeS were filtered off, and the silica eliminated 

 by evaporation to dryness and addition of dilute H2SO4. II2S was 

 again added, giving a blue solution. After driving off" the H,,S by 

 heat, the vanadic acid present was estimated volumetrically by the 

 addition of a measured portion of a standard solution of per- 

 manganate of potash. 



Magnesia was determined as pyrophosphate, and the alkalies b}' 

 means of Smith's method. Phosphoric acid was precipitated as 

 phosplioraolybdate of ammonia, and weiglied as pyrophosphate of 



magnesia. 



On account of the remarkable h3'groscopic powers of Philadel- 

 phite, great difficult}^ was experienced in the estimation of the com- 

 bined water. Nearly one-half of the water in the air-dried mineral 



