EXPLOSI VENESS OF NITRE. 13 



the persistence of the floors. That, prior to the last explosion, the nitre must 

 have been collected in the cellar, may be assumed from the fact, that the temper- 

 ature being inevitably far above its fusing point, the salt must have been all lique- 

 fied, and occupying the lowest accessible cavity, on account of its superior specific 

 gravity. This assumption is moreover justified by the circumstance that the force 

 of the explosion appears to have been especially exerted upon the parietes of the 

 cellar, the walls and surrounding earth having given way in a manner which crea- 

 ted astonishment. 



19. In order to amplify the practical basis upon which the preceding inferences 

 had been founded, I made some experiments on the combustion of gunpowder in 

 an exhausted receiver, so as to secure the gaseous products evolved. A cylindri 

 cal glass receiver, such as is usually employed as a candle shade, was ground upon 

 a lap-wheel, so as to fit air-tight between two disks of sheet brass. The disk for 

 closing the upper opening of the receiver was furnished with two cocks severally, 

 for communicating with an air-pump and barometer gauge. The disk for closing 

 the lower opening of the receiver, so as to form the bottom of the space included, 

 was furnished with an arch of platinum wire soldered to two stouter brass wires, 

 of which one was soldered to the disk, the other secured and insulated in passing 

 through it by a collet of leather, compressed about it by an appropriate screw. 

 These preparations being made, a portion of gunpowder, weighing about 25 grains, 

 was so supported on a tray, as to include the middle portion of the platinum wire. 

 The receiver being put into its place, so as to be duly supported by the lower 

 disk and covered by the other, the air was withdrawn as far as practicable, with 

 a good air-pump. In the next place, the wire was ignited to incandescence. To 

 my surprise, the gunpowde;* only smoked at first, and did not flash until a percep- 

 tible interval had elapsed. When this result ensued, it appeared to be owing to 

 the radiant heat, as the early volatilization of a portion of sulphur had driven the 

 granules away from the wire, so that it did not touch any of them. Subsequently, 

 on allowing the air to enter, and removing the receiver, it appeared that the gun- 

 powder was only partially burned. Thus it became evident that in this way a 

 complete combustion could not be effected. The feebleness of the flash in vacuo 

 shows how much confinement is essential to give energy to the explosion of this 

 powerful agent, and its not being forthwith ignited by an incandescent wire, 

 demonstrates that, as in the case of the apparent quiescence of a globule of vola- 

 tile matter in an intensely heated cavity, a capability of volatilization delays this 

 process, by preventing the contiguity requisite to a communication of heat. 



20. This leads to a discrimination which has not, to the best of my knowledge, 

 been made heretofore. I allude to the difference existing between fulminating 

 combinations and fulminating mixtures. As an example of the latter, we have gun- 

 powder and other pulverulent mixtures consisting partially of nitre, or chlorate of 

 potash, while, as an exemplification of the former, we may advert to aurum or 

 argentum fulminans, or to the fulminates of mercury or silver ; also to the chlo- 

 ride or iodide of nitrogen, or perchloric ether. Compounds of the last-mentioned 

 kind, without confinement, break the vessel on which they are exploded. They 

 cannot be used in gunnery, because the force in their immediate vicinity, in pro- 

 portion to its durability, is too great, so that they burst the chamber before the 



