284 KXPLOSIVE COMPOUND. 



scene! as the air enters, excepting what is retained in the cap- 

 sule, and which covers the globule of the compound j but as 

 this small quantity very quickly filters through the blotting 

 paper, and falls in drops through the hole in the capsule, the 

 compound is left exposed to the gas, and the effects of this 

 exposure immediately appear. 

 When the In our former communication we mentioned that the com- 



ke'TwJth wa- P ounc ^ ma )' De preserved for any length of time, in small tubes 

 ter in a sealed hermetically sealed, provided that the quantity of water, or of 



tube, it be- j i nc i u ded with it, did not exceed seventeen times its bulk: 



comes dissolv- ' .,.,,,. . , , 



ed in process we have since found that this is strictly true, only when the 



ot time, link ss % quantity of water in the tube is very inconsiderable compared 

 the quantity ot ,y ..,,,/• , , i . • i 



water be hv to that of the air included j for that when the tube is nearly 



considerable. fi]i e( i with water, the compound, after some months, disappears 

 and is dissolved in it. 



Remarks on j n tne same communication we described an analysis of the 

 si urees of - , . . . . . . . 



eirour in the compound, remarking, at the same time, that not having re- 

 former analy- peated it, we could not place any confidence in its results, and 

 that our principal object in giving an account of it was, to show 

 an easy and practicable mode of analysing the compound. In 

 the interval, sinc,e that was published, we have paid particular 

 attention to that analysis, and have found that there were two 

 very material sources of error in it 5 the first was owing to the 

 imperfect means which we then possessed of obtaining two 

 The globules globules of equal weights - } and the second, to a circumstance of 

 of the com- which we were not then aware (but which our subsequent ex- 

 unequaT^and periments have proved to have a considerable influence) 

 too much wa- v jz. that the quantity of water with which the explosive com- 

 flent* VaS Pl °" P oun d was ,n contact when it was decomposed by potash, was 

 much too large, and occasioned less azotic gas to be given out, 

 than would otherwise have been collected. To obviate these 

 two sources of error has been the object of our recent labours, 

 and we have fortunately succeeded in removing both. 

 Remedv. The The mode by which we have succeeded irt always operating 

 exact measure vvith known weights of the compound, is by using a glass sy- 

 poimd waaas- r,n g e * ot the f° rra represented in fig. 4. pi. VI. the lower part 

 curtained by a of which terminates in a tttbe of small bore j such as is used 

 capillary sy- ^ or thermometers. This tube is graduated on the outside into 

 inches and decimal parts, and when the point is placed in a 

 globule of the explosive compound under water, and the piston 



raised? 



