15/11 ADAPTATION IN ITERATED AND SERIAL SYSTEMS 



to further step-function changes, its field will not alter, and it 

 will at all times react in its characteristic way. So the whole 

 system is equivalent to an ultrastable system B interacting with 

 an ' environment ' A. It would also be equivalent to an ultra- 

 stable system interacting with an inborn reflex, as in S. 3/12. B 

 will therefore change its step-mechanism values until the whole 

 has a field which is stable and which holds within limits the 

 variable (blood-glucose concentration) whose extreme deviations 

 cause the step-mechanisms to change. We know from S. 8/11 

 that, whatever the peculiarities of A, B's terminal field will be 

 adapted to them. 



It should be noticed that the seven sets of variables (Figure 

 15/10/2) are grouped in one way when viewed anatomically and 

 in a very different way when viewed functionally. The anato- 

 mical point of view sees five sets in the animal's body and two sets 

 in the outside world. The functional point of view sees the whole 

 as composed of two parts : an ' adapting ' part B, to which A 

 is ' environment '. 



It is now possible to predict how the system will behave after 

 the above processes have occurred. Because part A, the ' avoid- 

 ing ' system, is unchanged, the behaviour of the whole will still 

 be such that collisions do not occur; and the reactions to the 

 food supply will maintain the blood-glucose within normal limits. 

 But, in addition, because B became adapted to A, the getting of 

 food will be modified so that it does not involve collisions, for all 

 such variations will have been eliminated. 



15/11. What of the time required for adaptation of all the essen- 

 tial variables when the environment is so joined in a chain ? 



The dominating subsystem A will, of course, proceed to adapta- 

 tion in the ordinary way. B, however, even when A is adapted 

 may still be disturbed to some degree by changes coming to it 

 from A, changes that come ultimately from the disturbances to 

 A that A must adapt against. C -also may get upset by some of 

 these disturbances, transmitted through B, and so on. Thus each 

 subsystem down the chain is likely to be disturbed by all the 

 disturbances that come to the subsystems that dominate it, and 

 also by the reactive, adaptive changes made by the same domin- 

 ating subsystems. 



It is now clear how important is the channel-capacity of the 



203 



