081 
Mr Deuchar on the Nature of Flame. 
was irregular in its effect, and formed a cake of crocus at the? 
top. In firing this mixture, however, it occasionally happen- 
ed, that the flame passed through at B, without inflaming the 
gunpowder ; this at first was supposed to be owing to the burst- 
ing of the paper containing the powder, but I afterwards began 
to suspect, that it might proceed from some peculiar character 
of the flame. Probably it might require more air than was 
present, to erfable it to display its full energy. Under these 
circumstances, gunpowder placed in the joinings of the tube^, 
would not be acted upon by the flame passing through it ; and 
this is so far verified by the following experiment. 
Eocp. 18. — At the bottom of one of the divisions (a, h or c)^ 
of the tube, was put a small piece of flannel, so as to cover the 
hole there ; and upon this were poured two, three, or four, and 
in some trials even five grains of gunpowder. Upon exploding, 
the fulminating powder at A, it was found in some of the expe- 
riments, that the flame had passed through the gunpowder 
without firing it ; but at other times this did not take place. In a 
few experiments, gunpowder was put at two of the divisions (a, and 
5), when it was found that the flame sometimes went through 
both, without firing either of them ; at other times one portion, 
was inflamed, and one left unaltered. The result, however, is 
by no means regular in its occurrence ; sometimes I found it to 
succeed in three successive trials ; at other times I failed four or 
five times before obtaining this curious effect. In all these 
cases, the powder was made to cover the whole surface of the 
piece of flannel upon which it was poured. This variety of effect 
seems to take place from some accidental circumstances, which 
have as yet escaped my notice. At some times I have succeeded 
best when gunpowder was used, the grains of which were large, 
and at other times, again, the small-grained gunpowder was most 
uniform. 
This, at first, appeared to be an objection to the proposed ap- 
plication of the apparatus. But after many trials, it was found 
that the above curious result only took place when the stroke 
with the hammer was slight ; for when a smart blow was given, 
inflammation always took place. 
The principal cause we may adduce for the powder remain- 
ing uninflamed, under the above circumstances, is the rapid mo- 
