1902.] 



of the Flame in the Explosion of Gases. 



481 



Part VI. — On the Initial Phases of the Explosion. 

 [In conjunction with R. H. Jones and J. Bower.] 



As several of our photographs had shown that the flame in the 

 period of acceleration was overtaken by a more rapidly moving wave, 

 we photographed the beginning of an explosion of cyanogen with 

 oxygen, and found that the bright flame which overtook the primary 

 flame came from the end of the tube near the firing wires. The " kick- 

 off" which the explosion gets from the closed end, apparently enables 

 the reflected wave to overtake the original wave of combustion. 

 When the gases are fired by means of wires just penetrating the 

 stopper closing the end of the tube, the explosion proceeds more 

 slowly, and without disturbance by any sudden changes of accelera- 

 tion until at last the point of detonation is reached. The photo- 

 graphs of these explosions are in marked contrast to those taken when 

 the firing wires were 3 inches from the end of the tube. In the 

 less rapid explosions it is seen that the flame does not travel direct to 

 the near end of the tube, but while still a short distance from it 

 recedes and again approaches with an oscillatory motion, which is 

 repeated before the flame finally reaches the end of the tube. From 

 the point where the flame is first checked a luminous wave is seen 

 running back and overtaking the main flame, which at this point 

 acquires greater brightness and velocity. From the point of collision 

 where the faster overtakes the slower flame, a reflection is thrown 

 back to the near end of the tube where it is again reflected. 



Now when an explosive mixture is fired by a spark, the suddenly 

 ignited gases must expand and transmit a compression-wave in both 

 directions. This travels with the velocity of sound in the unburnt 

 gas, and will be reflected from the end of the tube. The propagation 

 of the flame from the firing point is, in most gaseous mixtures, less 

 rapid than the velocity of sound in the unburnt gas, but the rate of 

 propagation of the flame augments much more rapidly in some 

 mixtures than in others. If the tube is a long one the flame will 

 overtake the sound-wave after a more or less prolonged chase, according 

 to' the nature of the mixture. But if the tube is short the sound- 

 wave may reach the end of the tube and return as a reflected wave 

 to meet the flame, which is still advancing. Our measurements show 

 this to be the origin of the " return- wave " from the end of the tube. 



When gases are fired in the centre of a long tube the phenomena 

 of explosion are simple ; when they are fired in the centre of a short 

 tube, the sound-waves generated reach the end of the tube before the 

 flames, and the impact of their reflections with the flames produces 

 cross-waves of great intricacy and beauty. 



