OF THE ATOMIC WEIGHT OF ALUMINUM. 
1021 
was, in separate portions, heated in the Sprengel vacuum as long as any gas was 
expelled, and, having been treated with pure hydrochloric acid to remove any possibly 
adhering particles of lime, the granules were finally washed with pure water, dried, 
and kept for use in a glass stoppered bottle. An approximate calculation having been 
made of the quantity of silver which would be required to precipitate each of the 
specimens of aluminum bromide, an amount less than this by something under a 
decigramme was dissolved in nitric acid in a strong flask closed with a stopper, which 
was carefully opened when cold, and the contents were somewhat further diluted with 
water. The amount of specially purified nitric acid used was apportioned so as to 
leave the smallest possible excess after solution of the metal. 
Precipitation of silver bromide—Details of method used .—To avoid the danger of 
losing aluminum bromide when it was brought in contact with water, the action being 
quite violent and attended with dispersion of white fumes, each one of the sealed thin 
glass tubes containing the bromide was, when the time came for using it, cautiously 
marked with transverse scratches at intervals of about half-an-inch by means of a 
writing diamond, and a strong glass bottle with a very well ground stopper having 
had a sufficient quantity of pure wafer placed at the bottom, the tube was broken 
across at the uppermost scratch, above the surface of the bromide, the empty point 
was dropped into the bottle, and the rest of the tube carefully lowered by means of a 
loose fitting spiral of platinum wire held sideways, so as to rest by the closed end on 
the bottom of the bottle without allowing 1 the water to reach the bromide until the 
stopper had been inserted and tied down. By now gently inclining the bottle the 
water was brought very gradually into contact with the aluminum bromide, without 
dangerously violent action and without possibility of loss. 
As soon as solution was complete and the bottle had cooled down the stopper was 
removed, any liquid adhering to it washed back into the bottle, and a stout glass rod 
with square end was used to gently crush the tube to small fragments, as otherwise 
its contents could not have been brought fairly into contact with the silver solution, 
since the interior of the tube would have become plugged up with silver bromide. 
The scratches previously made upon the glass rendered it easy to break it up without 
any splashing, and the rod was then well washed with pure water allowed to run 
directly into the bottle. The silver nitrate solution destined for this particular 
specimen was now washed out of the flask in which it had been shortly before 
prepared into the bottle containing the bromide, the stopper was again inserted, and 
the bottle was vigorously shaken as in the usual Gay-Lussac silver assay. The 
precipitation of the bromine was completed with a very carefully adjusted solution of 
silver nitrate, containing 1 milligramme of silver per cubic centimetre, and delivered 
from a burette'" reading clearly to c.c. The correspondence in capacity of the 
burette and measuring flask used was well ascertained. It was at first intended to 
* This burette was simply drawn down to proper bore at the bottom, and the flow of the liquid was 
regulated by the admission of air through a well ground stop-cock at the top. 
