leave it to itself, to go where it will.

As soon as it arrives in a current of heated air, it will ascend,
passing along with the current, and descending or rising as the current
is either warm or cold. The effect of the cold fresh air from windows or
doors, as well as the effect of the radiant heat from the fire, can be
thus thoroughly studied. Some of our pet theories may receive a cruel
shock from this experiment; but, in the end, the ventilation of the room
will doubtless be benefited, if we apply the information obtained. It
will be discovered that the wide-throated chimney is the cause of the
little black arrows turning their backs on the right path and our
theoretical outlets for vitiated air becoming inlets. The chimney flue
must have an enormous supply of air, and it simply draws it from the
most easily accessible places. From 1,000 to 2,000 cubic feet of air per
hour is a large "order" for a small room. Therefore, until we have made
ample provision for the air supply to the fire, it is quite useless to
attempt to ventilate the upper part of the room, either by ventilating
gas lights or one of the cheap ventilators with little talc flappers,
opening into the chimney when there is an up draught, and shutting
themselves up when there is any tendency to down draught. The success of
these and all other ventilators depends upon there being a good supply
of air from under the door or through the spaces round the window
frames. These fresh air supplies are, of course, unendurable; but if one
of the spaces between the joists of the floor is utilized to serve as an
air conduit, and made to discharge itself under the fender (raised about
two inches for the purpose), quite another state of things will be set
up. Then the supply of air thus arranged for will satisfy the fire,
without drawing from the doors and windows, and at the same time supply
a small quantity of fresh air into the room. But the important fact that
the radiant heat from the fire will pass through the cold air without
warming it all must not be lost sight of. In reality, radiant heat only
warms the furniture and walls of the room or whatever intercepts its
rays. The air of the room is warmed by passing over these more or less
heated surfaces; and as it is warmed, it rises away to the ceiling.
Therefore, if we desire to warm any of this fresh air supplied to the
fire, it must be made to pass over a heated surface. The fender may be
used for this purpose b