98 ON CERTAIN TUNGSTEN COMPOUNDS. 



Hence we have : — 



grms. 



Original weight of oxide 7*8840 



Oxide after first reoxidation 7*8806 



,, second reoxidation 7*8792 



Weight of metal, first reduction 6 -243 8 



„ second reduction 6*248 1 



„ third reduction 6*2488 



It is evident from these numbers that the second and 

 third weights of oxide and the second and third weights of 

 metal are the only ones which can be relied on as being 

 perfectly pure. Taking the means of these two series, we 

 have 7*8799 grms. of oxide giving 6*24845 grms. metal, or 

 79*296 per cent. This corresponds to the atomic weight 

 1 83*84. In order to have obtained the number 184, 

 7*8799 grms. of oxide must have yielded 6*24960 grms. of 

 metal, differing by 0*00115 grm. from the experimental 

 observed weight. 



2. From analysis of the Hexachloride. — Perfectly pure 

 hexachloride was prepared from the pure metal (obtained 

 from the oxychloride), the first product being distilled in a 

 current of chlorine into a clean part of the tube. 



No traces of red oxychloride could be detected in this 

 product ; and it yielded a perfectly pure canary-yellow tri- 

 oxide on treatment with water, proving absence of any 

 pentachloride. For the determination of the chlorine, the 

 hexachloride employed was weighed in the piece of drawn- 

 out combustion- tube, in which it was afterwards reduced in 

 hydrogen, the hydrochloric acid being collected and pre- 

 cipitated as silver chloride. 



The determination of metal was made in a porcelain boat, 

 in which the weighed quantity of hexachloride was first 

 carefully converted into trioxide by exposure for two days 

 to a moist atmosphere, and afterwards reduced in a current 

 of hydrogen. 



The following are the experimental data : — 



