in contact with the rim of the burner. This is best seen when the gas
is turned low--with a batswing burner, for instance--turned so low
that only a small non-luminous flame is left, the space between burner
and flame will appear as great as the flame itself, while, if the gas
is mixed with an inert diluent like carbon dioxide, the space can be
very much increased.

Several theories have been brought forward to explain this phenomenon,
but the true one is that the burner abstracts so much heat from the
flame at that point that it is unable to burn there, and this can be
proved by the fact that where a cold object touches the flame, a
dividing space, similar to that noticed between flame and burner, will
always be observed, and the colder the object and the more diluted the
gas the greater is the observed space. If a cold metal wire or rod is
held in a non-luminous flame, it causes an extinction of the gas for
some considerable space around itself; but as the temperature of the
rod rises, this space becomes smaller and smaller until the rod is
heated to redness, and then the flame comes in contact with the rod.

In the same way, if the burner from which the gas is issuing be heated
to redness, the space between burner and flame disappears. It has
already been shown that cooling the flame by an inert diluent reduces
the illuminating value, and finally renders it more luminous; and we
are now in a position to discuss the points which should be aimed at
in the construction of a good gas burner.

In the first place, a sensible diminution in light takes place when a
metal burner is employed, and the larger the surface and thickness of
the metal the worse will be its action on the illuminating power of
the flame; but this cooling action is only influencing the bottom of