206 Capt. Sir A. Noble. [June 21, 



and to Professor Dewar's advice and assistance I am indebted, I 

 can hardly say to what extent. 



The heat developed by explosion, and the quantity of permanent 

 gases generated were also determined as described in our researches, 

 but the amount of water formed plays so important a part in the 

 transformation that special means were adopted in order to obtain 

 this product with exactness. 



The arrangement employed was as follows : 



After explosion the gases formed were allowed to escape through 

 two U -tubes filled with pumice stone and concentrated sulphuric 

 acid ; when the gases had all escaped the explosion cylinder was 

 opened, and the water deposited at the bottom of the cylinder was 

 collected in a sponge, placed in a closed glass vessel and weighed. 

 The cylinder was then nearly closed and heated, and a measured 

 quantity of air was, by means of an aspirator, drawn slowly through 

 the U -tubes till the cylinder was perfectly dry. This was easily 

 ascertained by observing when moisture was no longer deposited 

 on a cooled glass tube through which the air passed. 



The (J -tubes were then carefully weighed, the amount of moisture 

 absorbed determined, and added to the quantity of water directly 

 collected. The aqueous vapour in the air employed for drying was, 

 for each experiment, determined and deducted from the gross 

 amount. 



Numerous experiments were made to ascertain the relation of the 

 tension of the various explosives employed, to the gravimetric 

 density of the charge when fired in a close vessel, but I do not pro- 

 pose here to pursue this part of our enquiry, both because the sub- 

 ject is too large to be treated of in a preliminary note and because 

 approximate values have already been published* for several of the 

 explosives with which we have experimented. 



With certain explosives, the possibility or probability of detona - 

 tion was very carefully investigated. In some cases the explosive 

 was merely placed in the explosion vessel in close proximity to a 

 charge of mercuric fulminate by which it was fired, but I found that 

 the most satisfactory method of experiment was to place the charge 

 to be experimented with in a small shell packed as tightly as possible, 

 the shell then being placed in a large explosion vessel and fired by 

 means of mercuric fulminate. The tension in the small shell at the 

 moment of fracture and the tension in the large explosion vessel were 

 in each experiment, carefully measured. 



It may be desirable here to explain that I do not consider the 

 presence of a high pressure with any explosive as necessarily denot- 

 ing detonation. With both cordite and gun-cotton I have developed 

 enormous pressures, close upon 100 tons per square inch (about 

 * Noble, ' Internal Ballistics/ 1892, p. 33 ; ' Eoy. Soc. Proc.,' vol. 52, p. 128. 



