FULLY CONNECTED SYSTEMS 11/1 



excessively and causes no deviation from the routine of chewing and 

 swallowing. 



An even more dramatic example, showing how defenceless is 

 the living organism if pain has not its normal effect of causing 

 behaviour to change, is given by those children who congenitally 

 lack the normal self-protective reflexes. Boyd and Nie have 

 recently described such a case : a girl, aged 7, who seemed healthy 

 and normal in all respects except that she was quite insensitive 

 to pain. Even before she was a year old her parents noticed 

 that she did not cry when injured. At one year of age her arm was 

 noticed to be crooked : X-rays showed a recent fracture-disloca- 

 tion. The child had made no complaint, nor did she show any 

 sign of pain when the fragments were re-set without an anaesthetic. 

 Three months later the same injury occurred to her right elbow. 

 At the seaside she crawled on the rocks until her hands and knees 

 were torn and denuded of skin. At home her mother on several 

 occasions smelt burning flesh and found the child leaning uncon- 

 cernedly against the hot stove. 



It seems, then, that if an imperfectly formed ultrastable system 

 is, under certain conditions, defenceless, so may be an imperfectly 

 formed living organism. 



(2) Even if the ultrastable system is suitably arranged — if the 

 critical states are encountered before the essential variables reach 

 their extreme limits — it usually cannot adapt to an environment 

 that behaves with sudden discontinuities. In the earlier examples 

 of the homeostat's successful adaptations the actions were always 

 arranged to be continuous ; but suppose the homeostat had con- 

 trolled a relay which was usually unchanging but which, if the 

 homeostat passed through some arbitrarily selected state, would 

 suddenly release a powerful spring that would drag the magnets 

 away from their ' optimal ' central positions : the homeostat, if 

 it happened to approach the special state, would take no step 

 to avoid it and would blindly evoke the 'lethal' action. The 

 homeostat's method for achieving adaptation is thus essentially 

 useless when its environment contains such ' lethal ' discontinuities. 



The living organism, however, is also apt to fail with just the 

 same type of environment. The pike that collided with the 

 glass plate while chasing minnows failed at first to avoid collision 

 precisely because of the suddenness of the transition from not 

 seeing clear glass to feeling the impact on its nose. This flaw 



131 



