directed downward. This reverses the direction of the
current of gas, which then escapes through the pin-holes downward into a
chamber, then turns upward along its sides to the tip, on entering which
it again turns. Each burner is able to consume economically a flow of
gas peculiar to itself, which can be ascertained by a minute's
experiment, and then regulated by the tap in the pipe. But this requires
much care, and is apt to be neglected. A very small tap in the burner
(as in the Wood and Ellis burners), which can be adjusted so as to
require no further attention, seems the best method of effecting this
graduation."

The chemist now pulled a manuscript from his pocket and read from it as
follows: "The quantity of light decreases with disproportionate rapidity
by reduced consumption; for, as experiments have shown, when consuming
only two feet per hour, eighty-five per cent. of the gas is lost; with
two and a half feet the loss is sixty per cent.; and with three and a
half feet it is thirty-four per cent. of that derived from the gas when
burning the full quantity for which the burner is constructed. In some
experiments made upon this matter under the direction of referees
appointed by the London Board of Trade the loss at the other extreme is
given. They report: 'Instead of the gas giving increased light as the
rate of consumption is increased, it will be seen that _in every case_
there is a point beyond which the _light decreases_ relatively to the
proportion of gas consumed. In every case, too, this point lies far
below the maximum of gas-consumption, observing the turning-points in
the case of the different burners.' Again, every burner has a certain
amount of gas which it will consume to the greatest advantage as to both
light and economy; which in a completely-regulated burner is quickly
found, and the delivery fixed by the small tap. When the gas is issuing
from the burner at so low a pressure that the flame is just on the point
of smoking, the maximum effect for the quantity of gas consumed in that
particular burner is attained, because in that case the quantity and
intensity of the light are most advantageously balanced. For the same
reason, the burner best suited for light is one in which the
jet-openings are proportionately large, so as to prevent as much as
possible too great contact with the air in the lower part of the flame.
In case the air-currents disturb the light, it is necessary to turn on a
strong