DISASTROUS EFFECTS OF COLLISION ON RAILWAYS. 307 



difference of velocities will be generated: this will occasion the first jolt, 

 which being applied at the back of the pilot engine considerably lower 

 than its centre of gravity, will cause it to lift up its fore wheels. By 

 the elasticity of the buffers, (which, when attached to engines, are not 

 usually furnished with springs like those attached to the carriages), the 

 velocity of approach will be immediately converted into a velocity 

 of separation, according to the laws of elasticity and impact, and, 

 for the moment, the front engine will be again driven ahead of the 

 other; but by the continued excess of the retardation of the first 

 engine, the velocity of separation will at length be reconverted into a 

 velocity of approach, giving rise to a second jolt, and occasioning the 

 front engine a second time to lift up its fore wheels; and the same 

 process might be repeated a great number of times in succession. 



In the above explanation, for the sake of simplicity, I have omitted 

 the consideration of the effect of the train following the second engine ; 

 but since the carriages are exempt from one great cause of retardation 

 to which the engines are subject, namely, the friction of internal 

 machinery, it is clear that they would, if left to themselves, be retarded 

 less rapidly than the engines, whence it is easy to see, that the effect 

 of the train of carriages, will tend to push on the second engine, and in- 

 crease the effect which has been described. 



I have supposed, moreover, that the engine has time to resume its 

 natural position after each jolt before the succeeding jolt is commu- 

 nicated, but if the jolts succeed one another more rapidly than the 

 natural time of an ascent and descent of the front wheels, it is easy 

 to see how, under favourable circumstances, they might conjoin their 

 effects so as to increase the angle of elevation very considerably. The 

 most favourable case for producing this effect, would be, when the 

 second jolt is communicated at the precise moment when the front 

 wheels have attained their highest elevation due to the motion im- 

 pressed by the first, and so on for the third and following jolts. 



These successive ascents of the fore wheels were exactly the phe- 

 nomena which it was our object to account for, being such as were 

 observed to precede the overthrow of the engine, and such as no doubt 



