89 
ancl reduced in hydrogen and oxidised in a current of air. 
After each reduction the boat was found to be partially 
coated inside with a thin black film having a metallic ap- 
pearance which oxidised completely when heated in air. 
A second boat was placed in the tube beyond that containing 
the substance for the purpose of ascertaining whether any 
metal was volatilized, but this boat was not found to become 
the least discoloured. The results of the determinations 
were as follows : — 
1. Original weight of Oxide 7 '8840 grams. 
2. Oxide after 1st Oxidation 7 "8806 „ 
3. 2nd 7-8792 „ 
4. Weight of Metal, 1st reduction. 6’2438 „ 
5. 2nd 6-2481 „ 
6. 3rd 6 -2488 ,, 
It is evident from these numbers that the 2nd and 3rd 
weights of oxide and the 2nd and 3rd weights of metal are 
the only ones which can be relied on as being perfectly pure. 
Taking the mean of these two series, we have 7'8799 grams 
of oxide, giving 6-24845 grams of metal, or 79 - 296 per cent. 
This corresponds to the atomic weight 183-84. In order to 
have obtained the number 184 - 00 the weight 7'8799 grams 
of oxide must have yielded 6-24960 grams of metal, differing 
by 0-00115 grams from the experimental number. 
2. By Analysis of the Hexachloride. 
Perfectly pure hexachloride was prepared from the pure 
metal (itself obtained from oxychloride). No traces of oxy- 
chloride could be detected in the hexachloride employed, 
and it yielded a perfectly canary yellow trioxide on treat- 
ment with water, showing absence of any pentachloride. 
In the determination of the chlorine, the substance was 
weighed in the piece of drawn-out combustion tubing, in 
which it was afterwards reduced in hydrogen, the hydro- 
chloric acid being collected and estimated as silver salt. 
The determination of metal was made in a porcelain boat in 
which the weighed hexachloride was first carefully converted 
into trioxide by exposure for two days to a moist atmo- 
sphere, and afterwards reduced in hydrogen. Analysis 
gave : — 
