324 
rrjOFEssoE n. b. dixox ox the moyemexts 
compoiincls) burns first to cai'bonic oxide. In the earlier of these experiments we 
were j)uzzled l)y a re-diijjlication of the image which constantly ap^^eared in the 
photographs, until at last this was found to be due to an unsuspected constriction in 
the lead pipe, which had sent back a reflected wave sufficiently intense to re-illuminate 
the window. This led us to examine more precisely the detonation-wave and its 
reflexions l)y means of the rotating film. The camera and film being arranged as 
l3efore, so that the image was thrown on the film as the latter moved vertically 
downwards, and the explosion-tube being fixed horizontally, the photograph showed 
an inclined line of light compounded of the horizontal movement of the image of the 
flame and tlie vertical movement of the film. AVhen the explosion-tube was placed 
at such a distance from the camera that the length of the image was that of 
the tube, a velocity of the flame of 3000 metres j^er second corresponded with a 
horizontal velocity of the image of 100 metres per second. When the wheel 
was rotated twenty-five times per second this caused the film to move vertically 
at a rate of 25 metres per second, the circumference of the wheel being 1 metre. 
The line described bv the iniao;e on the film thus made an ancle with the 
horizontal wliose tangent was ^ (nearly 14°); or when the wheel-was rotated at 
twice this velocity (50 metres per second) the angle was nearly 27°. For most 
experiments a rate of rotation between these limits was used, although the wheel 
could be rotated 80 to 100 times per second."^ On account of the slip of the catgut 
cord it was not found possible to determine the rate of the wheel from the gear and 
the rate of the motor. The true velocity of the wheel was determined by the trace 
of a tuning-fork, which sliowed that several revolutions were made at a sensiblv 
uniform speed, and therefore that the speed of the film might be taken as rigorouslv 
uniform during its movement through the small arc affected by the photograph. The 
films were slowly developed in a long trough, wliicli was set rocking by suitable gear. 
For tlie brigliter photographs an hour was generally required for proper development, 
for the fainter images two or three hours were recpiired. After fixing and washing, 
the films were soaked in glycerine and water, which prevented them from curling up 
on drying. 
It was found that the detonation-wave of all the undiluted mixtures we tried could 
l)e photographed on the moving films, but the intensity of the images varied very 
considerably. Cyanogen, carbon disidphide, and acetylene fired with oxygen gave 
the brightest flames, hydrogen with chlorine and carbonic oxide with nitrous oxide 
gave the least bright flames. Thin strips of black paper were fastened round the 
explosion-tube to give reference marks on the photograph. These strips appear as 
black vertical bands on the prints, and are useful for measuring the angles made by 
the detonation or reflected waves. 
Figs. 7, 8 and 9 (Plate 11) show the detonation-wave in a mixture of cyanogen with 
* Our method of photogivaphing on a rotating film has given most interesting results in I’rofessor A. 
Schuster’s research on the constitution of the electric spark. See ‘ Phil. Trans.,’ A, vol. 193,1900, p. 1S9. 
