S78 Mr Deuctiar 07i t/te Nature qf ' Flame. 
four pieces of paper the flame did not pass through. When pa- 
per and flannel were used at the same time, it was found that 
the interior one was always a little scorched. 
From several of these results it would appear, that when the 
flame proceeds with great velocity, its power of acting upon in- 
flammable or other substances is so far prevented ; but when- 
ever we, by any means, retard that motion, we facilitate its ac- 
tion. In the 8th experiment, when one piece only of flannel 
was used, the flame passed through without scorching it ; but 
it was found in the 9th experiment, that as the resistance was 
increased by additional pieces of flannel, the more of the scorch- 
ed effect was observed. The same was illustrated in the 10th 
experiment, when a piece of paper was put below one piece of 
flannel, and when the flannel was put below the paper. 
III. Independent of the proposed application of the appara- 
tus, it appeared of considerable importance to ascertain how far 
the flame might be thus propelled, and the greatest distance at 
which it could explode the gunpowder. 
Exp. 11. — To determine the distance to which the unretard- 
ed flame could be forced by the explosion, there was fixed to 
the apparatus a tube inches long. After numerous trials with 
this, and with tubes of decreasing lengths, it was found that no 
flame appeared at the bottom, till the tube was shortened to 23^ 
inches, and then it was very feeble, and of a pale blue colour. 
Exp. 12 . — The next circumstance for inquiry was the dis- 
tance at which the flame could explode the gunpowder. When 
the gunpowder was put immediately in contact with the lower 
aperture of a tube 23 1 inches long, it was fired ; but when a 
piece of flannel intervened it was not acted upon. Thus it ap- 
peared, that although the flame, at 23| inches distance from the 
source of its production, could inflame a quantity of powder, 
yet it had not force enough to pierce even one piece 'of flanneL 
The tube was now gradually shortened, and trials made at each 
change, till it was reduced to between 19 and 20 inches, when 
it fired the gunpowder through the flannel. 
IV. The next experiments were with the view of ascertaining 
how the results, obtained with this apparatus, might stand rela- 
