228 Count Rumford’s Experiments to determine 
density ; but if this supposition should prove false, all those 
computations, with all the practical rules founded on them, 
must necessarily be erroneous ; and the influence of these 
errors must be as extensive as the uses to which gunpowder 
is applied. 
Having found by experience how difficult it is to confine 
the elastic vapour generated in the combustion of gunpowder, 
when the smallest opening is left by which any part of it can 
escape, it occurred to me, that I might perhaps succeed better 
by closing up the powder entirely, in such a manner as to 
leave no opening whatever, by which it could communicate 
with the external air ; and by setting the powder on fire, by 
causing the heat employed for that purpose to pass through 
the solid substance of the iron barrel used for confining it. In 
order to make this experiment, I caused a new barrel to be con- 
structed for that purpose : its length was 3.4,5 inches, and the 
diameter of its bore of an inch ; its ends were closed up by 
two screws, each one inch in length, which were firmly and 
immoveably fixed in their places by solder ; a vacuity being 
left between them in the barrel 1.45 inch in length, which 
constituted the chamber of the piece; and whose capacity 
was nearly ^ of a cubic inch. An hole, 0.37 of an inch in 
diameter, being bored through both sides of the barrel, 
through the centre of the chamber, and at right angles to its 
axis, two tubes of iron, 0.37 of an inch in diameter, the dia- 
meter of whose bore was of an inch, were firmly fixed in 
this hole with solder, in such a manner that while their in- 
ternal openings were exactly opposite to each other, and on 
opposite sides of the chamber, the axes of their bores were in 
the same right line. The shortest of these tubes, which pro- 
