540 EVENING DISCOURSES. 
some constant velocity (which, however, varied between 25 and 50 metres per second 
in different experiments), the explosion tube being placed at such a distance from 
the camera that the size of the image was about one-thirtieth of the flame. In this 
way they analysed the progress of an explosion from its origin up to the final attain- 
ment of its maximum force and velocity in ‘detonation.’ They also discovered 
(i) the wave of ‘retonation,’ which is thrown back through the still burning or 
chemically active medium from the point where detonation starts (a phenomenon 
also independently discovered by Le Chatelier in 1900), (ii) the effects of collisions 
between two explosion waves, as well as of the passage of ‘ reflected waves’ through 
the hot products of combustion behind the flame front. 
It is hardly possible to give any adequate idea of the wealth of new information 
derivable from the beautiful series of over seventy photographs of explosion flames 
included in the memoir ultimately published by Dixon in 1903 ;? but a selection of 
them will be shown on the screen, from which their principal features may be judged. 
Within recent years the experimental method has been further developed and 
improved in our laboratories at the Imperial College, South Kensington, chiefly by 
means of the new high-speed photographic machines designed by Mr. R. P. Fraser, 
which have so increased its analysing power that (as will be shown later) it is now 
possible to photograph and measure movements in explosion flames occurring 
periodically with frequencies up to 250,000 per second. And with a new type of 
camera embodying the principle of a mirror revolving in vacuo at high constant 
velocity (30,000 r.p.m.) and projecting the image of the explosion flame on to a 
stationary film, we hope soon further to increase the analysing power four- or five-fold, 
so that in the near future we hope to be able to photograph and measure periodic 
flame-movements occurring with frequencies of a million per second. In reviewing 
some of these developments in Nature two years ago, the late H. B. Dixon, who had 
followed them with the closest interest, said that he envied us the luxury of our feelings. 
A selection of some thirty typical photographs, illustrative of our results,* will 
be shown on the screen. They have already modified and corrected many of our 
former ideas concerning the nature of the initial phase of slow uniform flame-movement 
in gaseous explosions, and (inter alia) completely disproved the supposed ‘ law of flame 
speeds.’ Indeed, it now seems probable that what Mallard and Le Chatelier visualised 
as flame propagation ‘ by conduction’ is an ideal condition perhaps realisable only 
when a stagnant explosive mixture is ignited without impulse at the centre of a 
spherical vessel of infinite radius. 
Important new information has been obtained regarding the influences of ‘ com- 
pression ’ and ‘shock’ waves upon the speeding up of combustion and flame move- 
ments during explosions ; and it is now proved that the speed can be abruptly raised 
from a lower to a higher uniform value when a flame is overtaken by a ‘ shock’ wave 
travelling in the same direction. Indeed, it has been shown that in such manner the 
flame-speed can be successively raised ‘ per saltum’ many times, and that it may 
assume any uniform value between the limits of that theoretically corresponding with 
propagation ‘ by conduction’ and that due to propagation ‘ by adiabatic compres- 
sion’; examples of this will be shown on the screen. 
For purposes of exposition it is convenient to distinguish between three successive 
phases of a gaseous explosion as developed in a tubular enclosure, namely: (1) the 
initial period, when the flame is travelling at a speed less than those of any shock 
waves overtaking it in the medium, (ii) the pre-detonation stage, when the accelerated 
flame is travelling at a greater speed than any ‘ shock’ waves, and therefore catching 
up with, and overtaking, such as may have already passed through it, and (iii) the 
final phase ‘ detonation.’ 
Much new light has been thrown on what may be termed the ‘ pre-detonation’ 
phase of explosion, when the flame is advancing at a speed greater than that of any 
‘shock’ wave through the unburnt medium, and is therefore overtaking any such 
shock waves which are ahead of it. 
The photographs also show how ‘ detonation ’ is ultimately set up when a rapidly 
moving flame on the verge thereof is just about to overtake a shock wave immediately 
in front of it. Thereupon an ‘ignition ahead’ of the flame-front occurs, and 
2 Phil. Trans. A 200 (1903), p. 315. 
8 Some of these have already appeared in the Proceedings and Philosophical 
Transactions of the Royal Society during the past five years, and another instalment 
is now in the course of publication therein. 
seen ad 
omc 
o,. Seeds 
Bite it 
——— re 
