1010 
PROFESSOR J. W. MALLET ON A REVISION 
The air was first expelled from the apparatus by a current of dry carbon dioxide, 
and not until the gas passing through was capable of being completely absorbed by a 
solution of potash was the hydrochloric acid introduced, this latter itself freed from 
atmospheric air. The gas escaping from the tube was now collected in a graduated 
jar, and the temperature of the tube containing the aluminum was raised to a red 
heat. As soon as hydrogen ceased to come over, the gas in the jar was shaken up with 
potash to absorb any carbon dioxide which it might contain, and the volume was 
measured, and reduced by calculation to its equivalent under normal temperature and 
pressure. The aluminum chloride left in the tube was pulverulent and snow white. 
No details are given of the method by which pure metallic aluminum was prepared, 
although this has been the great difficulty in the way of obtaining accurate results 
from experiments made with the metal as the starting point, nor is there any record 
of the tests applied to prove the purity of the metal used. The gas was collected 
over water, in which hydrogen is not altogether insoluble, and from which more or less 
of the gases of atmospheric air would be given off into the hydrogen. Nothing is 
said of the vapour of water, mixed with the hydrogen in proportion depending upon 
the temperature, having been removed, or its amount calculated and allowed for; 
though, as it is not likely that so obvious a precaution was neglected, it may be 
supposed that the potash spoken of as used to absorb any carbon dioxide left was 
either solid hydrate or so strong a solution as to have also removed most if not all the 
aqueous vapour. 
The results of the single experiment reported were— 
Weight of aluminum.. ‘410 grin. 
Volume of hydrogen collected at 11° and 768 m.m. 530 c.c. 
Corresponding volume at normal temperature and pressure . . 508'2 c.c. 
Weight of this hydrogen. .. '0455 grm. 
from which the author calculates 
*0455 : *4100=1 : 9*01, 
giving the atomic weight 13'515 (for ATClg) or 27'03 (for A1C1 3 ). 
In verifying the above calculation I have found as the result of reducing the volume 
of hydrogen from the given to normal temperature and pressure 514*85 c.c. instead of 
508*2 c.c., but this, I am satisfied, arises from the number representing the pressure at 
the time of experiment being, doubtless by a printer’s error, wrongly given as 768 
instead of 758 m.m. With the latter figures the result is as recorded in the paper, 
and of course such atmospheric pressure is more frequently observed than that which 
appears in the above table. 
It is pretty plain that from the researches which have been quoted we may reject 
those of Davy, Thomson, and Mather as incapable of giving exact results, this being 
