Theory of Explosions. 179 



free paths, we get for oxygen w=5x 10 6 . Here the discre- 

 pancy is a thousand times as great ; so that if longitudinal 

 vibrations are to be considered as likely to produce any 

 effect, they will certainly be considerably more likely to do so 

 if transmitted through solids or liquids than through gases. 



We have still, however, got the ether to fall back on, and 

 there we are safe, for there is no reason why vibration of the 

 right period should not be transmitted through it. 



The experiments with tubes, however, seem to point exclu- 

 sively to the air as the medium through which vibrations are 

 to be transmitted, and that may, I fancy, be fairly regarded 

 as unlikely, If the theory of synchronous vibration can be 

 disproved at all by experiment, then Abel has, at all events, 

 made the most telling experiment against it ; there may of 

 course be other experiments with which I am unacquainted, 

 though not for want of endeavour on my part to discover 

 them ; and these may point in the opposite direction. Still, 

 in the light of what has been published on the subject, there 

 is little doubt that our natural hesitation to accept a theory 

 of vibrations is justified by a consideration of the facts. On 

 the other hand, if we allow that vortex motion may exist, it 

 will account for some of the effects observed in the neighbour- 

 hood of violent explosions. The most important effect to be 

 accounted for is the capriciousness of explosions. Instances 

 are so numerous that it is hardly worth while to dwell on 

 them in detail ; I happen, however, to have an account of 

 the famous explosion at Bremer Haven from an eye-witness, 

 and was especially struck by the way in which the bystanders 

 seemed actually selected for injury, and that not always from 

 flying debris. Such effects as these are difficult to account 

 for, on any theory of uniform propagation of wave-motion. 

 Again, any of the observed phenomena of propagation of 

 explosion are as well explained by vortex propagation as by 

 wave-motion. There is no reason why the two states of pro- 

 pagation should not exist together, varying in their relative 

 importance according as the explosion is of long or short 

 duration. In ordinary cases of detonation I imagine the 

 shocks are given chiefly by the explosion-gases before any 

 considerable break-down has taken place. In conclusion, I 

 wish to say that, though I have given prominence to the 

 vortex-ring method of propagation, I do not intend to offer 

 it as necessarily the most important of the phenomena in all 

 cases, but only to point out that in some circumstances, 

 chiefly those which have been inadequately explained, a con- 

 sideration of its possible influence helps us to understand the 

 facts. 



