OF THE ATOMIC WEIGHT OF ALUMINUM. 
1017 
a second time, it seemed to be possible to remove this source of error, as out of several 
specimens thus treated only one afforded a barely detectable trace of sulphate. 
Details of method adopted .—The actual experiments on the ignition of the alum 
were carried out as follows. To render uniform the amount of atmospheric conden¬ 
sation on the surface of the vessel weighed, a very light glass bottle was specially 
made, with a delicately blown stopper, the latter carefully ground in and fitting 
quite air-tight; the bottle of a size to contain the platinum crucible in which the 
ignition was to be effected. The crucible was heated to bright redness, and while 
still quite hot was placed in a desiccator at some distance above a surface of recently 
distilled sulphuric acid. When cooled down to the temperature of the balance-room 
the crucible was as quickly as possible transferred to the weighing bottle, which was 
at once closed, and the combined weight of bottle and crucible was taken. The 
stopper was then removed for an instant, the cover of the crucible raised, and the 
quantity of alum desired, which had been roughly weighed off 1 in a tube, having been 
poured in, occupying in no instance more than one-third depth of the vessel, cover 
Fig. 1. 
and stopper were replaced, and a second weighing gave, by the gain upon the first, the 
exact amount of alum used. Attached to the inner side of the crucible cover was a 
piece of rather stout platinum wire, which, when the cover was in place, ran down 
into the crucible in the line of its axis of figure, carrying two little diaphragms of 
platinum foil perforated with small holes (see fig. 1) ; such a diaphragm having been 
suggested and used by Dumas' 1 ' in his researches on atomic weights as the means of 
preventing any loss of solid particles which might otherwise be carried off mechanically 
from the substance ignited. To avoid inconvenience from the fusion of the alum in its 
water of crystallisation, and the swelling up of the salt to a bulky, porous mass, the 
heating was conducted very gradually. The platinum crucible was kept for an hour 
at 90° C., then for an hour at 100°, for an hour at 110°, another hour at 120°, a like 
time at 140°, and was then gradually brought to ignition over an argand alcohol lamp. 
It was then placed inside a larger platinum crucible, resting on a flat bit of unglazed 
porcelain at the bottom of the latter, and exposed to a gradually increased, and at 
MDCCCLXXX. 
* ‘ Atmales de Chimie et de Physique,’ loc. cit. 
G r 
