set fire to her friend's
furniture, &c., ten or eleven times in the course of one or two days.
In neither case could anything like disagreement or harshness be
elicited, but the reverse. In other instances, it has been strongly
suspected that this disease was the cause of repeated fires, but there
was no positive proof. In all these cases, known or suspected, the
parties were generally from fourteen to twenty years of age.

FIRE-PROOF STRUCTURES.

What is "Fire-proof Construction?" is a question which has given rise
to a great deal of discussion, simply, as it appears to me, because
the size of the buildings, and the quantity and description of the
contents, have not always been taken into account. That which may be
perfectly fireproof in a dwelling house, may be the weakest in a large
warehouse. Suppose an average-sized dwelling-house 20 x 40 x 50 =
40,000 cubic feet, built with brick partitions, stone or slate stairs,
wrought-iron joists filled in with concrete, and the whole well
plastered. Such a house will be practically fire-proof, because there
is no probability that the furniture and flooring in any one room,
would make fire enough to communicate to another. But suppose a
warehouse equal to twenty such houses, with floors completely open,
supported by cast-iron pillars, and each floor communicating with the
others by open staircases and wells; suppose, further, that it is half
filled with combustible goods, and perhaps the walls and ceilings
lined with timber. Now, if a fire takes place below, the moment it
bursts through the upper windows or skylights, the whole place becomes
an immense blast furnace; the iron is melted, and in a comparatively
short time the building is in ruins, and, it may be, the half of the
neighbourhood destroyed. The real fire-proof construction for such
buildings is groined brick arches, supported on brick pillars only.
This mode of building, however, involves so much expense, and occupies
so much space, that it cannot be used with advantage. The next best
plan is to build the warehouses in compartments of moderate size,
divided by party-walls and double wrought-iron doors, so that if one
of these compartments takes fire, there may be a reasonable prospect
of confining the fire to that compartment only. Again, cast iron gives
way from so many different causes, that it is impossible to calculate
when it will give way. The castings may have flaws in them; or they
may be too weak for t