PouioK—On the Extraction of Glucinum from Beryl. 143 
and water added, and the whole boiled up by steam, allowed to settle, the solution 
siphoned off, and the silica thrown on a linen filter and well washed. The 
solution was now precipitated by ammonia, and the very bulky precipitate of 
iron, alumina, and glucina, filtered off and dissolved in hydrochloric acid, and 
the solution saturated with hydrochloric acid gas, which precipitates nearly all the 
aluminium as chloride, but leaves the whole of the iron and glucina in solution. 
Any soluble silica present comes down as a gelatinous precipitate before the pre- 
cipitation of the aluminium chloride begins. ‘The precipitated aluminium chloride 
was filtered off through glass wool, by a vacuum pump, and the precipitate well 
washed with a saturated solution of hydrochloric acid. The precipitated alumi- 
nium chloride was pure white, crystalline, and easily filtered. On boiling off the 
hydrochloric acid from the solution and evaporating down, a deep reddish-yellow 
syrupy solution was obtained, containing the iron and glucinum as chlorides. This 
solution was added in quantities of 100 cc. at a time to two litres of saturated 
solution of ammonium carbonate, and well shaken up in Winchester quarts. At 
first a heavy curdy precipitate came down, which, on shaking, redissolved com- 
pletely to a perfectly clear solution; in about five minutes this became cloudy, 
and in ten minutes quite a heavy precipitate came down; this very curious 
phenomenon being exactly repeated in all the Winchesters used. The time of 
appearance of the first cloudiness was very regularly five minutes, the precipitation 
being apparently complete in ten minutes. 
To ascertain the effect of varying the proportion of carbonate of ammonia, 
and its degree of dilution, a series of experiments were made, using measured 
quantities of saturated solution of glucinum chloride, carbonate of ammonia, and 
water. ‘The mixture was well shaken up until it became perfectly clear, the time 
then noted, and the liquid left at rest until the first appearance of cloudiness, 
when the time was again noted; then, so far as possible, the time of apparent 
complete precipitation was noted ; and the following results were obtained :— 
ue aca Water On shaking, aiscen | mean 
GlCl, &e. | Am. Card. aeldted col in— in— 
2 ce. 20 ce. 0 Milky — 
Ih ©@ | 2X0) Ge, 0 Clear 6/ 
1 ce. | 20) ee, 20 cc. Clear 8/ 
Il @@, | 60 cc. 0 Clear | 20’ 
In the first experiment there was not sufficient carbonate of ammonia for 
complete solution, between the second and third experiments there was little 
