336 
rROFESSOR H. B. DTXOX OX THE HOVE]\!EXTS 
The knowledge gained iroin these experiments with junctions made it now possible 
to investigate the phenomena of collisions, but the experiments had also showm that 
it wmuld be of interest to explore photographically the region of the exj^losion prior 
to the initiation of the detonation-wave. 
When two detonation-^vaves come into collision, the tube remains brightly luminous 
at the j)oint of contact for some time,"'^ and two reflected waves are sent backwards 
with velocities w'hich increase at first, owing to the movement of the gas through 
which they are propagated. 
The photoglyphs of two detonation-waves meeting in collision are shown in figs. 40 
and 41, the explosive mixture being cyanogen with two volumes of oxygen, and in 
fig. 42 (Plate 14), the mixture being two volumes of hydrogen with one volume 
of oxvgen. 
A comparison of all the photographs show's that the gases are more luminous after 
a collision than wdien the explosion-w’'ave strikes a flat surface of metal fastened at 
the end of the tube. The reflected waves in the twm cases are similar in character, 
but the reflexion generated by collision wdth another detonation-wave seems ahvays 
to ti'avel slightly fastei'. If w'e w'ere dealing only with waves produced mechanically, 
the reflected w'aves w'ould be exact copies of the incident-w'aves with velocities 
reversed in both cases. But in the detonation-wave we have chemical as w'ell as 
mechanical action, wBile the reflected wave is mainly mechanical. We should expect 
therefore the reflected waves to travel more slowdy than the incident-w'aves, but Ave 
should also expect the reflected Avaves to traA'el Avith the same A'elocity whether they 
Avere produced by collision Avith a rigid diaphragm or AAnth a similar and equal AA'ave 
traA'elling in the opposite direction, unless there Avas some chemical difference invoh'ed 
in the two kinds of collisions. Noav in the hypothesisf I haA'e advanced for the mode 
of propagation of the detonation-AAmve I have assumed that the exj)losion is 
propagated not only by the forAAmrd moA^ements of the molecules produced by the 
chemical change, but also partly by the moA'ement of the yet unburnt molecules. 
For instance, in the explosion of hydrogen and chlorine, molecules of hydrogen 
chloride just formed in the AvaA^e-front may inoA'e forAA-ard until they come into 
collision Avith molecules of unburnt hA'drogen or molecules of unburnt chlorine. These 
molecules (by exchange) noAv moA'e forAA'ard Avith increased A'elocity, and in turn meet 
molecules of the opposite kind, Avith AA'hich they combine. The combination therefore 
does not pioceed betAveen cold molecules entirely, nor betAveen heated molecules 
entirely, l)ut mainly betAveen molecules half of AA'hich are at the ordinary temperature 
and half are heated by collision Avith the products of combustion. If this roughly 
represents the state of the AAmA'e-front, there Avould be a chemical difference betAA'een 
the collision Avith a diaphragm and Avith another explosion-AA’av’e. For in the latter case 
* General Hess, of the Austrian artillery, has photographed a luminous band at the point of collision 
of the tAA'o compression-AAmves produced by exploding tAvo cartridges suspended in the air a short distance 
apart. See ‘ Bulletin Soc. de ITndustrie Minerale,’ Saint Etienne, 1900, a'oI. 14, 3, p. 116. 
t ‘ Phil. Trans.,’ A, vol. 184, p. 131 (1893). 
