of Edinburgh, Session 1884 - 85 . 
151 
Magnesium B was prepared from a solution to which ammonium 
was added till a very slight turbidity was produced, and then 
peroxide of hydrogen added. 
In the case of magnesium C, the quantity of ammonia was some- 
what increased, and in magnesium D to a still greater extent. 
Other portions of the magnesium compound were prepared under 
slightly different conditions, and two portions were dried over sul- 
phuric acid in vacuo, but without showing any decided difference. 
The dry powder obtained from these precipitates, by drying them 
at 100° C., possessed properties similar to those of the zinc and 
cadmium compounds, viz., they set free iodine from an acidified 
solution of iodide of potassium, &c. 
Magnesium B was selected for further analysis, as it contained 
the greatest percentage of available oxygen. On ignition it lost 
36 T 5 per cent., and was found to contain 2T7 per cent, of C0 2 , 
30-9 per cent, of water, and to leave a residue of 60 -85 per cent, 
of magnesia. 
The percentage composition was therefore — 
MgO 
63-85 
li 2 0 . . . 
30-9 
co 2 ... 
2-17 
O (by titration) 
3-92 
100-84 
Deducting carbonate of magnesium represented by 2*17 per cent. 
C0 2 and the water the percentage of O in the anhydrous oxide = 6’03. 
In the case of aluminium our methods had to be somewhat 
altered, owing to the insolubility of alumina in ammonia. We 
tried a solution of alumina in caustic soda, containing no more 
alkali than just sufficient to dissolve the alumina. 100 grs. of 
ammonia alum were dissolved in the smallest quantity of water, 
and caustic soda added until the hydrate at first precipitated was 
nearly all, but not quite, redissolved. When this solution was 
filtered and peroxide of hydrogen added, a copious white precipi- 
tate was obtained, which was filtered off, washed, and dried — one 
portion at 100° C., and another over sulphuric acid in vacuo. A 
third portion was allowed to stand in contact with some of the 
