night or on a Monday morning--no doubt because a large body of fire
had formed before it was detected. A certain number of accidents
occur in summer in private houses from persons on hot nights opening
the window behind the toilet glass in their bedrooms, when the draught
blows the blind against the candle. Swallows do not more certainly
appear in June, than such mishaps are found reported at the sultry
season.

If we watch still more narrowly the habits of fires, we find that they
are active or dormant according to the time of the day. Thus, during
a period of nine years, the percentage regularly increased from 1.96
at 9 o'clock A.M., the hour at which all households might be
considered to be about, to 3.34 at 1 P.M., 3.55 at 5 P.M., and 8.15
per cent at 10 P.M., which is just the time at which a fire left to
itself by the departure of the workmen, would have had swing enough to
become visible.

The origin of fires is now so narrowly inquired into by the officers
of the Brigade, an by means of inquests, that we have been made
acquainted with a vast number of curious causes, which would never
have been suspected. From an analysis of fires which have occurred
since the establishment of the Brigade, we have constructed the
following Tables:--

Curtains 2,511
Candle 1,178
Flues 1,555
Stoves 494
Gas 932
Light dropped down Area 13