OF THE ATOMIC WEIGHT OF ALUMINUM. 
1029 
surface, while the last one-fifth of its length remained at first unoccupied, to be filled 
later as will be described ; the vapour of water being finally collected beyond this 
combustion tube in a single light drying tube of calcium chloride, one of sulphuric 
acid on pumice, and one of phosphorus pentoxide. 
The whole of the apparatus having been put together, with the exception of the 
three final drying tubes, for which at first a single unweighed calcium chloride tube 
was substituted, with sodium hydrate solution in the bulb, and metallic aluminum in 
the tube ( b ), a slow stream of carefully purified and well dried air was passed through 
from ( x ) for some time, while the combustion tube was heated to low redness. Having 
allowed it to cool down to the temperature of the room, the stream of air from ( x) was 
replaced by one of dry nitrogen ; a plug of soft copper turnings with bright unoxidised 
surface was introduced into the further end of the combustion tube, so as to fill up the 
unoccupied fifth of its length, and it was again brought to and kept at a red heat. 
After the nitrogen had passed through at this temperature for about twenty minutes, 
the three drying tubes for the collection of the water to be formed, having been 
standing for some time in the balance case, were accurately weighed, and connected 
with the further end of the combustion tube, the previously used and unweighed 
calcium chloride tube being removed. The tube ( x ) having been closed, the tube (b) 
was now tilted so as to make the aluminum slip down into the alkaline liquid, and as 
in the experiments already described, the rate at which the hydrogen was evolved was 
controlled by inclining the bulb so as to vary the extent of surface of the metal 
attacked, and by cooling the outside of the bulb with water. 
As soon as the metal was all dissolved, nitrogen was again introduced by (,xj to sweep 
out the remaining hydrogen, limiting the quantity of the former gas to such an amount 
as was thought necessary for this purpose. This nitrogen was then in turn replaced 
by pure and dry air, which was passed through the apparatus until the surface of the 
copper which had been reduced was reoxidized, this being done to avoid any risk of 
occluded hydrogen being retained, while the nitrogen had served to obviate the danger 
of explosion. The drying tubes were then finally removed from the further end of 
the combustion tube, and weighed after exposure to the atmosphere of the balance 
case long enough to permit the surface of the glass attaining a constant condition. 
In both weighings the reduction to equivalent weights in vacuo was duly attended 
to, the weights and densities of all the materials making up the drying tubes and 
their contents having been previously ascertained. The last tube of the set was 
weighed separately, as was the last of the drying tubes connected with the reaction 
bulb, and it was found that, there being no increase of weight on the part of either of 
them, the absorption of aqueous vapour was sensibly complete. 
Although all the precautions I could think of were taken in these experiments, the 
well known difficulty of absolutely excluding moisture, of which every joint to the 
apparatus becomes a possible source, so that in ordinary organic analysis the amount 
of hydrogen found may be expected to come out rather above than below the truth, 
