OF THE FLAME IN THE EXPI.OSION OF OASES. 
wave and inclined at a small angle to it. The velocity of the detonation-wave being 
2320, the mean velocity of the following compression-wave is given by measurements 
of three photographs as 1635. The difference between this number and the velocity 
of the compression-wave measured in stationary gas (a difference of 500 metres per 
second) is due to the movements, and probably also to the higher temperature of the 
o'us in tlie immediate wake of the detonation-wave. 
PART lY. 
Ox THE CoLTjsiox OF Two Detoxatiox-Waves axd the Effect of Juxctioxs 
IX THE Tubes. 
[In conjunction with R. H. Joxes and J. Bower.) 
In 1897 we began the investigation (by means of the moving film) of the 
phenomena marking the collision of two explosion-waves meeting end-on in a glass 
tube. To make the two flames meet in the field of the camera, the glass firing piece 
A (fig. 24) was fastened to a lead pipe bifurcating at B into two arms of equal length, 
which were bent round at their extremities and opened out to hold the two ends of 
the glpss tube. The junctions at C and D were made gas-tight by slipping pieces of 
rubber tubing ovei‘ the ends of the glass tulie and pushing them into the leaden caps. 
By shutting the tap B"^^ and opening E, the whole apparatus could lie filled tlirough A 
witli the explosive mixture; tlien the taps at A and E were shut and B was opened. 
On passing the spark the flame followed the two arms of the tube, and the two 
detonation-waves generated met in the centre of the glass tube C, D. 
We were greatly puzzled by tlie photographs obtained with this apparatus. The 
first few experiments showed tlie rebound waves (after the collision) to be much 
lirighter and to be travelling (backwards) much faster than the two detonation-waves 
themselves before the collision. Other pliotographs showed that the two flames were 
~ The bore of the tap B was the same size as the lead tube; it is shown in fig. 18. 
