170* PHILOSOPHICAL TRANSACTIONS. [ANNO 1797. 



following experiment. 1 160 grains Troy of apparently dry gunpowder, taken from 

 the middle of a cask, on being exposed 15 minutes in dry air, heated to the tem- 

 perature of about 200°, was found to have lost 1 1 grains of its weight. This 

 shows that each cubic inch of this gunpowder actually gave out 2^ grains of 

 water on being exposed to this heat; and there is no doubt but that at the end of 

 the experiment it still retained much more water than it had parted with. 



If now we compute the quantity of water which would be sufficient, when re- 

 duced to steam under the mean pressure of the atmosphere, to fill a space equal in 

 capacity to 1 cubic inch, we shall find that either that contained in the nitre which 

 enters into the composition of 1 cubic inch of gunpowder as water of crystalliza- 

 tion, or even that small quantity which exists in the powder in the state of mois- 

 ture, will be much more than sufficient for that purpose. 



Though the density of steam has not been determined with that degree of pre- 

 cision that could be wished, yet it is quite certain that it cannot be less than 2000 

 times rarer than water, when both are at the temperature of 212°. Some have 

 supposed it to be more than 10,000 times rarer than water, and experiments have 

 been made which seem to render this opinion not improbable; but we will take its 

 density at the highest possible estimation, and suppose it to be only 2000 times 

 rarer than water. As 1 cubic inch of water weighs 253.175 grains, the water 

 contained in 1 cubic inch of steam at the temperature of 212° will be 4 ' 00 part of 

 253.175 grains, or 0.1 2659 of a grain. But we have seen that 1 cubic in,ch of 

 gunpowder contains 10. 927 grains of water of crystallization, and 4.873 grains in 

 a state of moisture. Consequently the quantity of water of crystallization in gun- 

 powder is 86 times greater, and the quantity which exists in it in a state of mois- 

 ture is 38 times greater, than that which would be required to form a quantity of 

 steam sufficient to fill completely the space occupied by the powder. 



Hence we may venture to conclude, that the quantity of water actually existing 

 in gunpowder, is much more than sufficient to generate all the steam that would 

 be necessary to account for the force displayed in the combustion of gunpowder, 

 supposing that force to depend solely on the action of steam, even though no 

 water should be generated in the combustion of the gunpowder. It is even very 

 probable that there is more of it than is wanted, and that the force of gunpowder 

 would be still greater, could the quantity of water it contains be diminished. From 

 this computation it v*ould appear, that the difficulty is not to account for the force 

 actually exerted by fired gunpowder, but to explain the reason why it does not 

 exert a much greater force. But I shall leave these investigations to those who 

 have more leisure than I now have to prosecute them.* 



* We have bestowed more than ordinary remarks on this paper, as the experiments are exceedingly 

 curious, the observations ingenious, and the subject of great importance. The author has greatly 

 merited the praise of the philosophical world for his ingenuity and uncommon exertions, and the mi- 

 nuteness with which he has related all the circumstances. But as we have often heard it remarked that 

 Count R. did not always understand his own experiments, particularly in the case of the present paper \ 



