ON LIGHT. 303 



case of our coloured rings, it must be mentioned that in 

 reckoning the number of undulations or semi-undulations 

 by which the second reflected ray actually is in arrear of 

 the first on emergence, we have to consider the different 

 modes in which the reflexion of a wave is accomplished 

 at the surface of a medium denser or rarer than that in 

 which it moves and is reflected. To present this clearly, 

 we will take the most familiar illustration that of the 

 propagation of motion by the collision of elastic balls. 

 Imagine a great number of equal ivory balls (supposed 

 perfectly elastic) in contact, but connected only by an elastic 

 string passing centrally through each and along the com- 

 mon axis of all; and pinned or fastened to each at its centre 

 so that the separation of any two shall stretch only that 

 part of the string between their centres. Suppose now 

 that a shock is given to the extreme ball at one end in 

 the direction of the common axis, by another similar and 

 equal detached ball driven against it. By the received 

 laws of elastic collision it will give up its whole motion to 

 that which it first strikes, and be itself reduced to rest. 

 In like manner the motion so communicated to the first 

 will be handed on undiminished to the second ; itself rest- 

 ing and therefore remaining in contact with the striking 

 ball, and so on. Thus what may be termed a "wave of 

 compression," will run along the series till it reaches the 

 ball at the other end. This, having none in front to 

 communicate its motion to, will start off; and, were it 

 free, would quit the series. But this it cannot do, by 

 reason of the elastic thread ; which however it will stretch 

 in its effort to do so, and be ultimately brought back 



