VOL. LXXXVII.] PHILOSOPHICAL TRANSACTIONS. 175 



But even here the heat is still much below that which is most undoubtedly gene- 

 rated in the combustion of gunpowder. The temperature which is indicated by 

 722° of Fahrenheit's scale, (which is only 122 degrees higher than that of boiling 

 quicksilver, or boiling linseed oil), falls short of the heat of iron which is visibly 

 red-hot in day-light by 355 degrees: but the flame of gunpowder has been found 

 to melt brass, when this metal, in very small particles, has been mixed with the 

 powder; and it is well known that to melt brass a heat is required equal to that of 

 3807 degrees of Fahrenheit's scale; 2730 degrees above the heat of red-hot iron 

 or 3085 degrees higher than the temperature which gives to steam an elasticity 

 equal to the pressure of 131072 atmospheres. That the elasticity of steam would 

 actually be increased by heat in the ratio here assumed, can hardly be doubted. It 

 has absolutely been found to increase in this ratio in all the changes of temperature 

 between the point of boiling water, I may even say of freezing water, and that of 

 280° of Fahrenheit's scale; and there does not appear to be any reason why the 

 same law should not hold in higher temperatures. 



A doubt might possibly arise with respect to the existence of a sufficient quantity 

 of water in gunpowder, to fill the space in which the powder is fired, with steam, 

 at the moment of the explosion; but this doubt may easily be removed. The best 

 gunpowder, such as was used in my experiments, is composed of 70 parts in weight 

 of nitre, 18 parts of sulphur, and ]6 parts of charcoal; hence 100 parts of this 

 powder contain 67 -ro P arts °f nitre, 17-jV parts of sulphur, and of charcoal 15-^ 

 parts. Mr. Kirwan has shown that in 100 parts of nitre there are 7 parts of water 

 of crystallization; consequently, in 100 parts of gunpowder, as it contains 67-jV 

 parts of nitre, there must be 4 i ' \ parts of water. 



Now as 1 cubic inch of gunpowder, when the powder is well shaken together, 

 weighs exactly as much as 1 cubic inch of water at the temperature of 55° p. 

 namely 253.175 grs. Troy, a cubic inch of gunpowder in its driest state must con- 

 tain at least lO-^Vograins of water; for it is 100 to 4.711, as 253.175 to IO.927. 

 But besides the water of crystallization which exists in the nitre, there is always a 

 considerable quantity of water in gunpowder, in that state in which it makes bodies 

 damp or moist. Charcoal exposed to the air has been found to absorb nearly -f of 

 its weight of water; and by experiments I have made on gunpowder, by ascertain- 

 ing its loss of weight on being much dried, and its acquiring this lost weight again 

 on being exposed to the air, I have reason to think that the power of the charcoal, 

 which enters into the composition of gunpowder, to absorb water remains unim- 

 paired, and that it actually retains as much water in that state, as it would retain 

 were it not mixed with the nitre and the sulphur. 



As there are 15vV parts of charcoal in 100 parts of gunpowder, in 1 cubic inch of 

 gunpowder = 253.175 grains Troy, there must be 38.989 grains of charcoal; and 

 if we suppose -f of the apparent weight of this charcoal to be water, this will give 

 4.873 grains in weight for the water which exists in the form of moisture in 1 cubic 

 inch of gunpowder. That this estimation is not too high, is evident from the 



