ature of a yellow stone till the yellow
tint turns to a pink of the colour desired. The topaz is harder than
quartz, as will be seen on reference to the "Hardness" table, and is
composed of a silicate of aluminium, fluorine taking the place of some
of the oxygen. Its composition averages 16.25 per cent. of silica, 55.75
per cent. of alumina, or oxide of aluminium, and fluoride of silicium,
28 per cent. Its formula is [Al(F,OH)]_{2} SiO_{4}, or (AlF)_{2}SiO_{4}.
From this it will be understood that the fluorine will be evolved when
the stone is fused. It is, however, very difficult to fuse, and alone it
is infusible under the blowpipe, but with microcosmic salt it fuses and
evolves fluorine, and the glass of the tube in the open end of which the
stone is fixed is bitten with the gas.

Such experiments with the topaz are highly interesting, and if we take a
little of the powdered stone and mix with it a small portion of the
microcosmic salt, we may apply the usual test for analysing and proving
aluminium, thus: a strongly brilliant mass is seen when hot, and if we
moisten the powder with nitrate of cobalt and heat again, this time in
the inner flame, the mass becomes blue. Other phenomena are seen during
the influence of heat. Some stones, as stated, become pink on heating,
but if the heating is continued too long, or too strongly, the stone is
decoloured. Others, again, suffer no change, and this has led to a
slight difference of opinion amongst chemists as to whether the colour
is due to inorganic or organic matter. Heating also produces
electricity, and the stone, and even splinters of it, will give out a
curious phosphorescent light, which is sometimes yellow, sometimes blue,
or green. Friction or pressure produces strong electrification; thus the
stones may be electrified by shaking a few together in a bag, or by the
tumbling of the powdered stone-grains over each other as they roll down
a short inclined plane. The stones are usually found in the primitive
rocks, varying somewhat in different localities in their colour; many of
the Brazilian stones, when cut as diamonds, are not unlike them.

In testing, besides those qualities already enumerated, the crystalline
structure is specially perfect and unmistakable. It is doubly
refractive, whereas spinel and the diamond, which two it closely
resembles, are singly refractive. Topaz is readily electrified, and, if
perfect at terminals, becomes polarised; also the commercia