methods of lighting the mines without producing its explosion. 17 
half in circumference, and i t 7 <j of an inch deep, communicated 
explosion, but four concentric canals, of the same depth and 
diameter, and of which the smallest was two inches in dia- 
meter, and separated from each other only by their sides, 
which were of brass, and about of an incn in thickness, did 
not suffer the explosion to act through them. 
It would appear then, that the smaller the circumference 
of the canal, that is the nearer it approaches to a tube, the 
greater must be its depth, or the less its diameter to render 
it safe. 
I did not perceive any difference in these experiments, when 
the metals of the apertures were warmed by repeated explo- 
sions; it is probable, however, that considerable elevation of 
temperature would increase the power of the aperture to pass 
the explosion; but the difference between the temperature 
of flame, and that marked on our common mercurial scale, 
is so great that the addition of a few degrees of heat pro- 
bably does not diminish perceptibly the cooling power of a 
metallic surface, with regard to flame. 
By diminishing the diameter of the air canals, their power 
of passing the explosion is so much diminished that their depth 
and circumference may be brought extremely low. I found 
that flame would not pass through a canal of the ~ of an 
inch in diameter, when it was ^ of an inch deep, and forming 
a cylinder of only ~ of an inch in circumference; and a num- 
ber of apertures of °f an inch are sa ^ e when their depth 
is equal to their diameter. It is evident from these facts, 
that metallic doors, or joinings in lamps, may be easily made 
safe by causing them to project upon and fit closely to paral- 
lel metallic surfaces. 
MDCCCXVL 
D 
