ON GASEOUS COMBUSTION. 475 



reaching the near end of the tube, is reflected back again as G H. The 

 ' detonation wave ' E P passes onwards through the mixture with its 

 characteristic uniform high velocity and intense luminosity. Fig. 3 

 (Plate VII.) is a similar photograph showing the development of 'detona- 

 tion ' and the phenomenon of ' retonation ' in the case of a mixture of 

 cyanogen and oxygen (C 2 N 2 + 2 ) fired near the closed end of a tube. 



Except in special circumstances (e.g., when it is reinforced by 

 another reflected wave) the velocity of the retonation wave is always 

 inferior to that of detonation; thus Le Chatelier gives 2,990 metres per 

 second for the ' detonation wave ' and 2,330 metres per second for the 

 ' retonation wave ' in an equimolecular mixture of acetylene and oxygen. 

 When, however, the ' retonation wave ' is developed just at the closed 

 end of a tube (e.g., when the explosive mixture is fired at such a distance 

 from the closed end that ' detonation ' is set up just as the flame arrives 

 at the end) it may be reinforced by a reflected wave, in which case its 

 velocity cannot be distinguished from that of a true ' detonation. ' 



The explanation of the intense luminosity of the retonation wave, 

 and its higher velocity than sound through the exploded gases, is to be 

 found in the fact that the combustion during the initial stages of an 

 explosion is very much slower than when detonation is set up. Under 

 the extreme conditions of ' detonation ' the temperature of each succes- 

 sive layer of the explosive mixture is suddenly raised to the ignition 

 point by adiabatic compression, and it is probable that a large proportion 

 of collisions between chemically opposite molecules are fruitful of 

 change. The whole combustion is probably completed in an immeasur- 

 ably short interval of time, as the result of a comparatively limited 

 number of successive molecular collisions. But during the initial period 

 of the explosion (' inflammation ') not only is the flame propagated with a 

 much slower velocity, but also the actual process of combustion is much 

 more prolonged than in detonation, and at the moment when detonation 

 is set up combustion is still proceeding in the layers of gas for some 

 distance behind the flame-front. The ' retonation ' wave, in passing 

 backwards through these layers, quickens this residual combustion, and 

 is itself thereby rendered highly luminous. This interpretation is sup- 

 ported by the repeated observations of Dixon that the collision of two 

 flames, in neither of which ' detonation ' has been determined, will 

 frequently give rise to reflected waves more rapid and more luminous 

 than the incident waves. There can be little doubt as to the important 

 part played by reflected waves in determining the violent shattering 

 effects associated with gaseous explosions on a large scale. 



Section II. — The Explosion Wave. 



Berthelot and Vieille, in announcing their discovery of the develop- 

 ment of denotation (' l'onde explosive ') in gaseous explosions, 1 described 

 it as ' une certaine surface r&gulaire, oil se ddveloppe la transformation, 

 et qui realise un m&me Mat de combinaison, de temperature , de pressure, 

 etc. Cette surface, une fois produite, se propage ensuite, de couche en 



1 Ann. Chim. Phys., 1881 [5], 28, 289. 



