130 Scientific Intelligence. 
jum and sodium may be advantageously employed in place of chlorid of 
aluminium in the preparation of this remarkable metal. The author 
recommends small crucibles of cast iron; into these the Cryolite powder 
is pressed with thin layers of sodium; the whole is then covered with 
a layer of chlorid of potassium and heated. The proportions used 
were 5 parts of Cryolite, 2 of sodium and five of chlorid of potassium, 
and the crucible was kept at a red ree for half an hour. The fused 
mass is to be treated with water in a platinum or silver vessel ; grains of 
aluminium weighing half a gramme are thus obtained. These are to 
be cleaned and melted under a layer of fused chlorid of eee or 
better of the double chlorid of Senne — potassium. The largest ‘ 
yield which the author obtained was 0-8 gramme of aluminium for 10 . 
eee of Cryolite instead - ; > er whit the pois contains. In 
many cases however only 0°3 g r less was obtained. Rose con- 
siders ae results as proving “ gree ae it will be well worth while 
to make further experiments with Cryolite as a cee for the much 
more expensive chlorid of aluminium employed by Deville. After 
pao his investigation he found that large quantities of ees ite 
ere to be had in Berlin at the very low price of 3 tha 
—per hundred weight. It appears that the mineral was brought from 
Greenland to Stettn via Copenhagen and sold to the soap boilers under 
the name of mineral soda. By boiling with caustic lime a caustic lye 
was obtained whic as Rose observes, on account of the alumina held 
in solution is well adapted to the manufacture of some kinds of soap. 
The thes proved to be of great purity.—Pogg. Ann., xcvi, 152. 
{Nore.—It may be worth while to remark that kryolite would be of 
great vabie’ ini dhevtaboen ratory as a means of obtaining pure fluobydric 
acid for mineral analyses. On distillation with sulphuric acid, a residue 
of sulphate of soda and sulphate of alumina would remain whic 1 
of fluorid of silicon, in which case the quantity of silica in the 
eral could readily be calculated from the loss of weight. This pro 
to furnish an expeditious and accurate mode of determining silica. 
lime and magnesia could be estimated, if necessary, in the resi 
ignited mass.—w. G.] 
On the di 
methods of determining the strong or w 
pon this subject H. Rose has published « af 
vestigation which promises to yield resu! 
mistry. In the paper before us ss 
ifferent toward the salts of 
— treats of the behavior c 
monium and particularly toward 
first place, to the alkalies, aes fc 
