56 DURATION OF EXPLOSIVE REACTIONS. 



into long leaden tubes. Further on it will be seen that the 

 author, together with M. Vieille, has measured velocities of 

 several thousands of metres per second in explosive gaseous 

 compounds. 



(b) Progressive combustion transmitting itself, step by step, 

 under the conditions in which the cooling due to conductivity 

 by contact with inert substances, etc., lowers the temperature to 

 the lowest degree compatible with the continuation of the re- 

 action ; all heat is thus dissipated with the exception of a very 

 small fraction necessary to propagate the reaction in the 

 adjacent parts. The velocity of combustion in explosive gases 

 measured by Bunsen (p. 49) is attributable to this mode of 

 inflammation. 



In the case of solid or liquid explosives the propagation of 

 simple inflammation is rendered more difficult by the movement 

 of the gases, which distribute themselves throughout a large 

 space around the point inflamed, instead of acting in a volume 

 equal to, or slightly different from that of the original bodies ; 

 they thus share their temperature with a large mass of the sub- 

 stance up to such a point that the latter cannot be raised to the 

 desired degree for it to commence decomposition. Thus we 

 often see the substance dispersed by the gases without ex- 

 periencing total combustion, and even without undergoing any 

 change. This happens particularly with explosive substances 

 not confined within an envelope which concentrates the action 

 of the gases and gives to it a common resultant (p. 40). 



This is the case with nitroglycerin, which is found unaltered 

 in the vicinity of progressive deflagrations, and it also occurs 

 with dynamite placed on the ground in a thin layer. Damp 

 gun-cotton, which is not inflammable, has also furnished 

 numerous instances of this dispersion resulting from the use 

 of an insufficient detonator. It is by reason of this special 

 action of gases that a simple inflammation of a dynamite 

 cartridge, owing to the use of a badly placed fuse or of ful- 

 minate not in sufficiently close contact, should be avoided, 

 inflammation thus preceding the direct action, which ought to 

 be produced in immediate contact with the fulminate. 



2. Between these two limits an entire series of intermediate 

 states are observed, and they are unlimited in number as the 

 various modes of inflaming dynamite demonstrate. This is 

 proved by the influence of a sufficiently strong tamping (p. 40), 

 which transforms an inflammation into a true detonation. 



Finally, we might here cite the inequality of the effects 

 produced by successive explosions of charges of the same 

 agent detonating by influence at limiting distances beyond 

 which the explosion would no longer propagate itself (see 

 further on). 



This variety in the phenomena is due to two orders of causes, 



