132 SMITH AND EXNER — ATOMIC WEIGHT OF TUNGSTEN. [April 7, 



molybdic oxide, ferric oxide, manganese oxide, etc. Large amounts 

 of ammonium paratungstate were taken for the purification of the 

 tungsten trioxide by methods which may now be presented. 



Method I. — This first fraction was collected apart, dissolved in 

 distilled water and recrystallized ; the first crystallization was again 

 set aside and dissolved, this operation being repeated ten times. 

 A portion of the sixth crystallization was ignited in a platinum 

 crucible and the resulting oxide was digested on a water-bath with 

 a 2% solution of sodium carbonate free from iron. The oxide dis- 

 solved, but its solution was quite turbid. Upon standing, a white 

 residue settled out, which after washing with water and decomposi- 

 tion with a few drops of hydrochloric acid showed tungstic acid 

 and iron. Portions of the tenth recrystallization behaved similarly. 

 The mother liquors, including that from the tenth crystaflization, 

 assumed a dark-brown coloration upon concentration, indicating 

 thereby that not only iron, but that also the vanadico-phospho- 

 tungstate, already alluded to, continued with the ammonium tung- 

 state. Vanadium too was found in a portion of the tenth crystal- 

 lization when it was heated in an atmosphere of hydrochloric acid 

 gas. Hence it was concluded that this method was unsatisfactory 

 and it was abandoned. Chemists, who in the past were content to 

 look upon ammonium paratungstate as being very pure when its 

 color was perfectly white (and from published statements most 

 have been content with this criterion), which it is after the third 

 or fourth crystallization, cannot have had pure substance for their 

 investigations, hence the fluctuation in their results is easily com- 

 prehended. 



Method 2. — This may be called the method of Borch. In it 

 the mineral was fused with sodium carbonate, the fusion exhausted 

 with water and after filtration the liquid was precipitated with 

 calcium chloride. The resulting calcium tungstate was filtered, 

 washed and decomposed with hydrochloric acid. The liberated 

 tungstic acid was dissolved in ammonia water and the ammonium 

 salt crystallized out. The method was carried out with rigid 

 adherence to the printed instructions, but it proved entirely unsatis- 

 factory. 



Method J. — Solutions of ammonium paratungstate, decidedly 

 brown in color, were boiled with precipitated calcium carbonate. 

 Ammonia and carbon dioxide were evolved, while calcium tung- 

 state was precipitated. The solutions lost nothing in color. After 



