300 ^Ir. Crum on the Acetates »f Alumina. 



thickness. When the liquid is poured off, and the crust allowed to dry, 

 it separates readily from the vessel in hard plates like porcelain. If the 

 solution be not left strictly at rest, it becomes turbid after some days, 

 and the crust is produced in a more friable state. In cold weather the 

 solution remains unaltered for a much longer time. Heated to redness, 

 the salt becomes black from the decomposition of part of its acetic acid, 

 and in this state it very slightly affects the colour of moistened litmus 

 paper, showing that it retains but little alkali. Dissolved in nitric acid 

 and tested with nitrate of silver, it showed a trace of chlorine, and it 

 contained also a little iron. It was fi'eed from these impurities by 

 processes to be afterwards detailed. 



To ascertain the quantity of alumina in one of these salts, it is enough 

 to moisten it with rectified and concentrated sulphuric acid in a platinum 

 crucible, to evaporate the excess of acid over a spirit lamp, and then to 

 subject it to a white heat half-an-hour in a furnace. The sulphuric 

 acid expels acetic acid, and thereby secures the absence of charcoal. It 

 facilitates the disengagement of water ; and the powder, which, without 

 it, becomes so pulverulent by calcination as to be with difficulty pre- 

 vented from escaping, is thereby rendered coherent. 



Attempts were first made to determine the acetic acid by the alka- 

 limeter, taking potash alum of known purity for a standard, and for a 

 comparison of results. The substance being kept in water at a boiling 

 heat during the addition of the alkali, (soda,) and by a careful compari- 

 son of the shades of the litmus paper, appeared at first to redden that 

 test, so long as any portion of the salt remained uudeconiposed. But it 

 was found, after some practice, that at least 5 per cent, of the insoluble 

 salt remained untouched by the alkali after all trace of acid reaction had 

 disappeared in the process, and that within a range of 3 or 4 per cent, 

 the method was not to be depended upon. It gave valuable assistance, 

 however, in this investigation, in the testing of solutions of alumina, with 

 which it gave results much more accurate than with the insoluble salts. 



33'24 grains of the substance, dried first in the air, then pulverized 

 and dried further in a stove twenty hours at 100" Fahr., were moistened 

 with 80 grains of rectified sulphuric acid in a platinum crucible, and cal- 

 cined as above. There remained 8*64 grains. When this residue was 

 treated with water and filtered, the solution showed no trace of sulphuric 

 acid, and was alumina nearly pure ^ 25'99 per cent, of the original 

 substance. 



In another experiment, 32-64 grains, dried four hours in the stove, 

 and treated as above, left 8'4 grains of alumina =. 25'74 per cent. The 

 mean is 25"86 per cent, of alumina. 



Burned with oxide of copper, 2'985 grains produced 5'526 cub. inches 

 of dry carbonic acid = 2*612 grains, and corresponding to 50'69 per cent, 

 of acetic acid. In another experiment TGSS grains indicated 50'58 of 

 acetic acid. The mean is 50"63. It was assumed that the remaining 

 23 "SI per cent, was water. The salt was therefore composed of — 



