into flame until it reaches the edge of the tile, b, over the whole edge of
which it is made to run fairly well by a distributing arrangement. A rapid
combustion takes place here, but some unconsumed tar falls on to the bed
below. About one-third of the grate area is filled up by a fire tile, and
on this the tar coke falls. The tile, c, is moved away from time to time,
and the tar coke that accumulates in front of it is pushed back on to the
fire bars, e, at the back of the furnace, to be there consumed. Air is thus
admitted, by three narrow slot-like openings, to the front of the furnace
between the tiles, a, b, and c, and under c and through the fire bars, e.
The air openings below are about three times the area of the openings in
the front of the furnace; but as the openings between the fire bars and the
tiles are always more or less covered by tar coke, it is impossible to say
what the effective openings are. This disposition answers admirably, and
requires little attention. Three minutes per hour per fire seems to be the
average, and the labor is of a very light kind, consisting of clearing the
passages between the tiles, and occasionally pushing back the coke on to
the fire bars. These latter are not interfered with, and will not require
cleaning unless any bricks in the furnace have been melted, when a bed of
slag will be found on them.


THE AMOUNT OF DRAUGHT

required for these fires is very small, and less than with coke firing. I
find that 0.08 in. vacuum is sufficient with tar fires, and 0.25 in. for
coke fires. The fires would require less attention with more draught and
larger tar supply, as the apertures do not so easily close with a sharp
draught, and the tar is better carried forward into the furnace. A regular
feed of tar is required, and considerable difficulty seems to have been
experienced in obtaining this. So long as we employed ordinary forms of