1901.] on Memory. 597 



departure, and again subject it to the action of the weight. We find 

 that the application of the same weight produces no addition to the 

 permanent set. We may apply the weight as many times as we 

 please, or we may leave it hanging for a thousand years, and still, 

 when it is removed, the wire will return to the same position, the set 

 will not be increased. Besides the structural memory or change of 

 shape, the wire has undergone another change, a change of the 

 stresses among its particles, such that further modification in the 

 same direction has become more difficult. The same force no longer 

 suffices to produce a set. We may call this a dynamical memory of 

 the experience that the wire has undergone. 



If, instead of twisting the wire, we distort its shape in any other 

 way, the result, mutatis mutandis, is the same. If the change of 

 shape is within the limit of elasticity, the wire springs back to its 

 former shape the moment the disturbing influence ceases to act. If 

 the change of shape is beyond the limit of elasticity, the metal retains 

 a permanent set, which is a statical or structural memory of the ex- 

 perience that it has undergone; and at the same time, with the 

 permanent set, it acquires a change in molecular disposition, an 

 alteration of the stresses existing among the molecules, such that in 

 future it reacts differently to incident motion. It retains a dynamical 

 memory of the experience. 



What is true of iron is true in various degree of other metals, 

 and what is true of metals is true in various degree of other solids, 

 so long as these solids are unorganised. All possess some elasticity, 

 though the limit of perfect elasticity varies very widely. All take 

 permanent set, or entertain structural memories, and all react 

 differently thereafter to incident motion, though the degrees in 

 which they admit of these modifications are very various. 



When we pass from unorganised to organised solids, we find that 

 their behaviour under experiences of incident motion is in some 

 respects similar, and in others is widely different. We find that they 

 too are distorted by incident motion ; that they too possess elasticity ; 

 that they too recover their exact shape after distortion within the 

 limit of elasticity ; that they too take a set when distorted beyond the 

 limit of elasticity ; and that they too react differently thereafter to 

 motion incident upon them ; but there are important differences. 



Test the behaviour of a walking-cane or a whip-stock which is 

 bent, and which we may straighten with the hands. By holding the 

 stick at the ends of the bend, and placing the camber across the 

 knee, the stick may easily be straightened. But as soon as we release 

 it, it springs back again into its curve. It is elastic. In order to 

 allow for this elastic recoil, we must over-bend it. We must bend it 

 at least as far in the opposite direction as the bend that we are trying 

 to obliterate. When this is done, and the stick is released, it springs 

 back by its elasticity to a position intermediate between that im- 

 pressed upon it and that in which we found it. So far the stick 

 behaves in the same way as an iron rod would behave under the same 



