228 Mr. Thompson’s Experiments , &c. 
impurities of all kinds are fo very detrimental to gun-powder. 
It is not that they deftroy or alter the properties of any of the 
bodies of which the powder is. compofed, but Amply, that by 
obftrufting the progrefs of the inflammation, they leflen its 
force, and render it of little or no value. Too much care, 
therefore, cannot be taken in manufacturing powder to free 
the materials from ail heterogeneous matter. 
Of an attempt to foot flame inflead of bullets . 
Having often obferved paper and other light bodies to come 
out of great guns and fmall arms inflamed, I was led to try if 
other inflammable bodies might not be fet on fire in like man- 
ner, and particularly inflammable fluids j and I thought if this 
could be effected, it might be poffible to projedt fuch ignited 
bodies by the force of the exploiion, and by that means 
communicate the fire to other bodies at fome confiderable 
diftance ; but in this attempt I failed totally. I never could 
fet dry tow on fire at the diftance of five yards from the 
muzzle of my barrel. I repeatedly difcharged large wads 
of tow and paper, thoroughly foaked in the moft inflammable 
fluids, fuch as alkobol , retherial fpirit of turpentine , balfam of 
fulphur , &c. ; but none of them were ever fet on fire by the 
explofion. Sometimes I difcharged three or four fpoonfuls of 
the inflammable fluid, by interpoling a very thin wad of cork 
over the powder, and another over the fluid ; but ftill with no 
better fuccefs. The fluid was projected againft the wall as be- 
fore, and left a mark where it hit ; but it never could be made 
to take fire ; fo 1 gave up the attempt. If it had fucceeded, 
probably it would have turned out one of the moft important 
difcoveriss in the art of war that have been ’ made fi nee the 
invention of gun-powder. 
