OXY-ETHER AND HYDRO-CARBON LIGHTS 97
oxygen at the jet than vapour from the other generator, and
this test being conclusive. A tank can also be made to con¬
tain ether enough for any reasonable time, while it is
difficult with woollen to saturate the vapour sufficiently for
more than an hour and three quarters.
58. Management of the Ether Light.—I consider either
generator, as now made, to be practically safe with experienced
operators, and especially with such as understand the inherent
risks of using volatile fluids. But it is clear that these risks
are a serious addition to the ordinary risks of the mixed jet ;
and for this reason it is not to be desired that the oxy-ether
light should be widely used by the ordinary class of operators.
I cannot condemn it wholesale, as Sir Henry Roscoe has done,
especially in the porous form ; for I should use it without hesi¬
tation if I wanted high power where I could not get gas. But
since the oxy-spirit jet supplies light for most purposes in
country rooms, and compressed gas can be procured for others,
the ordinary exhibitor would do well not to run the extra risks
of a process, which he will not probably thoroughly under¬
stand all the dangers of. As the oxy-ether light will, how¬
ever, probably continue to be used to some extent by skilled
operators, in certain circumstances, a few words are necessary
regarding its practical manipulation.
In the first place, the oxygen pressure must be good. The
oxygen has to supply two branches of tubing, and in addition,
each current has to be driven through the resistance of the
pumice chamber, which should always be employed. Where
this light is habitually used, the pumice chamber must be
examined, and if necessary replenished from a store of sifted
material kept always at hand, at short intervals. If a bag is
used, at least cwt., and 2 cwt. to finish, will be needed for
a good light ; if a cylinder and regulator, from 9 to 12 inches
pressure.
In the second place, the ether used must be of a certain
purity, known by its specific gravity, which should not exceed
* H