366 



SMITHSONIAN MISCELLANEOUS COLLECTIONS 



VOL. 54 



experience obtained in their conduct gave a sound basis for further re- 

 searches, which were undertaken b}' Smith and Exner.' These authors 

 discuss at length the sources of error in former determinations of the 

 atomic weight of tungsten, and point out the difficulty of preparing 

 pure material, a difficulty which was at last overcome. From a pure 

 ammonium tungstate they prepared pure tungsten, the pure trioxide, and 

 pure tungsten hexchloride, free from oxychloride, and with these sub- 

 stances their atomic weight determinations were made. At this point 

 only their syntheses of the trioxide will be considered, their other series 

 being discussed later. Their figures, with vacuum weights, and the 

 usual percentage column are given below : 



.0007 



Hence W = 184.075, ±.0064. 



There are still other experiments by Eiche,^ which I have not been 

 able to get in detail. They cannot be of any value however, for they 

 give to tungsten an atomic weight of about ten units too low. We may, 

 therefore, neglect this series and go on to combine the others : 



^ Proc. Amer. Phil. Soc, 43, 

 ^Journ. prakt. Chem., 69, 10. 



123. 3904. 

 ]857. 



