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



unglazed porcelain crucible in a platinum crucible and bringing 

 about the ignition of the salt in this double-walled chamber. The 

 coloration of the surface of the oxide was extremely slight. Experi- 

 ence, however, eventually showed that the best course to pursue con- 

 sisted in digesting the ammonium paratungstate directly in a 

 porcelain casserole with pure nitric acid and a few cubic centi- 

 meters of hydrochloric acid until it was completely decomposed, and 

 the ammonia and hydrochloric acid were destroyed. When the 

 tungstic acid was evaporated to complete dryness, it showed a rich 

 orange-yellow color. It was transferred to an unglazed porcelain 

 crucible and there ignited gently for half an hour. This may be 

 done over a direct flame, the crucible being covered with an 

 inverted porcelain lid. Any enclosed nitric acid was expelled by 

 the gentle heat, and the weight soon became constant. The result- 

 ing tungstic oxide had a uniform yellow color. Green was abso- 

 lutely absent. This procedure eliminated the reduction caused by 

 the ammonia, and it may be added that by its use glazed crucibles 

 were employed every day in similar ignitions for several weeks with- 

 out showing the least etching or corrosion of the surface. 



Having at last gotten pure salt and pure oxide, the question arose 

 as to what method should be adopted in the determination of the 

 atomic weight of the metal. The method proposed by Taylor 

 (p. 130) was new. The results he obtained were with material not 

 especially purified, yet their fair agreement pointed to the possibil- 

 ity of arriving at a definite value with the pure substance, such as 

 was now available. Preliminary trials were executed according to 

 Taylor's suggestions, using glass apparatus just as he had done, and 

 drying at 400° to constant weight. The weighings were all made 

 on the same day and under uniform conditions. The main pur- 

 pose was to ascertain whether concordance in results could be 

 realized. The results in the subjoined table show the opposite ; 



Na^CO^ WO^ CO^ At. Weight 



1 5-9 gm. 2.45645 gm. .46775 gm. 183.07 



2 5.6 " 2.72292 « -51785 *' 183.36 



3 5-5 " 3-32953 *' -63288 «' 183.48 



4 4-7 *' 3-96720 " .75473 " 183.29 



5 4-0 " 3.44944 " -65489 •' 183.75 



6 4-1 " 3-41273 " .64796 « 183.74 



7 4.8 *' 6.10309 " 1. 16087 « 183.32 



8 3-9 " 6.39735 " 1. 21644 « 183.39 



9 3-5 " 2.17450 « -41332 ** 183.48 



10 3.1 « 1.57903 " .29966 ." 183.85 



