680 COMPARATIVE ELECTRO-PHYSIOLOGY 



appear at most unexpected moments. Thus in a given case 

 of the present experiment, performed in the afternoon, the 

 subject perceived this recurrence for some time, after which 

 the alternating impressions seemed to disappear, and were 

 completely forgotten. On retiring at night, however, these 

 recurrent images suddenly reappeared. Thinking the matter 

 to be an effect of light, the observer hastily extinguished his 

 lamp. But the recurrent images now became only the more 

 intense. 



In another case the recurrence was observed in a dream, 

 about three weeks after the original impression was made, 

 and in this case it was seen as the crossing and recrossing 

 of bright swords. These instances of the revival at night of 

 impressions made in the daytime, when the interference o 

 distracting influences is withdrawn, is significant. Since an 

 intense stimulation of nerve is -liable to recur spontaneously, 

 without the action of the will, or even in spite of it, it follows 

 that any single impression, when very intense, may become 

 dominant, and persist in recurring automatically. Examples 

 of this are only too familiar. 



We have hitherto dealt with that aspect of memory in 

 which it is a more or less immediate after-effect of sensory 

 stimulation. But we encounter a much more difficult pro- 

 blem when we come to the question of the revival of an 

 image long after it has apparently faded. It has been sug- 

 gested that this process of revival depends upon the existence 

 of some * scar,' or fixed impression, in the brain, or on a 

 certain persistent disposition or tendency to movement 

 created there. It is perhaps worth while to point out here, 

 however, that though when a blow is recent, the smarting 

 effect will persist for some time, yet, when once healed, no 

 scar could of itself reproduce the original excitation. It is of 

 course recognised that such expressions are merely figurative, 

 and that the entire process is not clearly understood. We 

 are more likely, however, to arrive at a true explanation of 

 the phenomenon if we recognise in it two distinct factors, 

 first, that of molecular change, with concomitant change of 



