OF THE FLA^IE IN THE EXPLOSION OF GASES. 
343 
with C.iN, + O.^, and hg. 57 with + 2 O 3 . In all cases the flame begins to travel 
right and left from the wires with equal velocity in both directions. In fig. 5G tlie 
flame develops the retonation-wave at the near end of the tube (on the left) as is 
shown by the intense-wave running nearly parallel to the detonation, which is started 
at about the same distance on the other side of the firing point. 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 Anally reaches the end of the 
tube. From the point where the flame is first checked, a luminous wave is seen 
(in flg. 55 ) running hack and overtaking the main flame, which at this point acquires 
greater brightness and velocity. How does this new “ return-wave ” arise ? 
Now, when an explosive mixture in a tube is fired by a sj)ark, 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. This seems to be the 
orig-in of the “ return-wave.” 
o 
Fig. 57a gives in outline the movements of the flame in Ag. 57. The flame, starting 
at A, moves to the left, tracing the curve A to C. At C the detonation is set up, 
and the retonation-wave C D is thrown back. Now the velocity of sound in the 
unburnt mixture C 3 N 0 + SOg is about 312 metres per second. If a sound-wave of 
this velocity had started towards the left-hand from A at the same time as the 
flame, it would have been overtaken by the flame at the point B; but the sound¬ 
wave starting to the right-hand would reach the end of the tube at E and been 
reflected to G before meeting the flame. On constructing the figure from the 
approximately-known velocities of the film and the flame, the point G is seen to be 
the spot where the movement of the flame is retarded and the “ return’’-wave is 
visible—within the limits of experimental error. 
The photograph 58 permits this point to be determined with considerable accuracy. 
The gases CgNg + O 2 were fired close* to one end of a tube 8 inches long 
(205 millims.). In a second tube, parallel wdth the first, the detonation-wave in the 
same mixture was set up and the two explosions photographed together. The 
detonation is showm at the top of the picture. Fig. 58a gives an outline of the 
explosion of the first tube. Starting at A, the flame moves to the left, but is checked 
* The spark was passed between wires which just penetrated the stopper. 
