OF THE ATOMIC WEIGHT OF ALUMINUM. 
1009 
account is given of the means taken to prepare a carbon crucible free from iron and 
silicon; and to prevent destruction of the crucible by its material burning away during 
the fusion at high temperature of the aluminum salt with sodium. The metal was 
tested for iron by dissolving it in nitrodiydrochloric acid, evaporating to dryness with 
a large excess of nitric acid, and igniting the residue of alumina, which was observed 
to be of brilliant whiteness, while the addition of a solution containing a few thou¬ 
sandths of iron sufficed “ to colour it very strongly red.” Why the much more 
delicate tests available for iron in the original solution were not used does not appear. 
As regards silicon, it is stated that “the solution of the metal by means of hydro¬ 
chloric acid left no trace of silicon no mention is made of the solution having been 
evaporated to dryness, the residue remoistened with strong hydrochloric acid and 
dissolved in water in order to see whether silica was left. • A portion of the solution 
obtained with nitro-hydrochloric acid was evaporated to dryness, the residue ignited, 
and the alumina so left was digested with a strong and boiling solution of ammonium 
nitrate. This solution was evaporated to dryness, and left a residue of sodium nitrate 
representing 0'135 per cent, of sodium in the metallic aluminum. 
1 ‘935 grin, of this aluminum was dissolved in hydrochloric acid, the solution evapor¬ 
ated with an excess of nitric acid until all chlorine was completely driven off, and the 
residue heated until the nitric acid was also completely removed and alumina only was 
left. This alumina weighed 3'645 grms. In the paper recording this single experiment 
the resulting atomic weight is not calculated, but the author simply points out that 
the number 14, which he says many chemists adopt as representing aluminum, must 
be too high, while 13'75, the number assigned by Dumas, is in all probability accurate. 
In support of this view it is calculated that, if Al=14, the alumina obtained in the 
above described experiment should have weighed 3 - 590 r ' grms.; whereas, with Al= 13 - 75, 
its weight should have been 3‘624 grms. In getting these figures 0 is taken =8. 
But, if the minute quantity of sodium stated to have existed in the metal used be 
deducted, and allowed for as sodium oxide (aluminate) in the last weighed residue, 
and if the results obtained be calculated, for alumina = ALO :i , with Stas’ number for 
oxygen (15*96), the atomic weight of aluminum will be represented by 27'068, a 
number much nearer to 27 than to 27'5 (13'75x2), the value assumed as most 
probably correct by Tissier. 
8. Experiments of Terreil, 1879.—Lastly, about a year ago Terreil+ made a 
determination of the constant in question by passing hydrochloric acid gas over 
metallic aluminum, collecting and measuring the hydrogen evolved. He placed a 
known weight of aluminum in a tube of hard glass, the tube wrapped with foil so as 
to allow of its being made red hot. By one end a stream of well dried gaseous hydro¬ 
chloric acid could be introduced, while a smaller tube extended from the other end and 
dipped into a vessel of water. 
* This ought to read 3'594. 
f ‘Bulletin de la Societe Chimique de Paris,’ xxxi. (20 Fev., 1879), p. 153. 
G o 
M DCCCLXXX. 
