Many experiments have been made with the view of determining the
percentage of carbon monoxide in air which is fatal to human or,
rather, animal life, and the most reliable as well as the latest
results are those obtained by Dr. Stevenson, of Guy's Hospital, in
consequence of the two deaths which took place at the Leeds forge from
inhaling uncarbureted water gas containing 40 per cent. of carbon
monoxide. He found that one per cent. visibly affected a mouse in one
and a half minutes, and in one hour and three quarters killed it,
while one-tenth of a per cent. was highly injurious. Let us, for the
sake of argument, take this last figure 0.1 per cent. as being a fatal
quantity, so as to be well within the mark.

In ordinary carbureted water gas as supplied by the superheater
processes, such as the Lowe, Springer, etc., the usual percentage of
carbon monoxide is 26 per cent., but in the Van Steenbergh gas--for
certain chemical reasons to be discussed later on--it is generally
about 18 per cent., and rarely rises to 20 per cent. An ordinary
bedroom will be say 12 ft. X 15 ft. X 10 ft., and will therefore
contain 1,800 cubic feet of air, and such a room would be lighted by a
single bats-wing burner consuming not more than four cubic feet of gas
per hour. Suppose now the inmate of that room retires to bed in such a
condition of mental aberration that he prefers to blow out the gas
rather than take the ordinary course of turning it off--a process, by
the way, of putting out gas which is decidedly easier in theory than
in practice, especially in his presumed mental condition--you would
have in one hour the 1,800 cubic feet of gas in the room mixed with
four fifths of a cubic foot of carbon monoxide--the carbureted water
gas being supposed to contain 20 per cent.--or 0.04 per cent. In such
a room, however, if the doors and windows were absolutely air tight,