DESIGN FOR A BRAIN 15/6 



take in, by means of the tentacles on that side, one fragment 

 of food after another. If now bits of food be alternated with 

 bits of filter paper soaked in meat juice, the two materials 

 will be accepted indiscriminately for some eight or ten trials, 

 after which only the meat will be taken and the filter paper 

 will be discharged into the sea water without being brought 

 to the mouth. If, after having developed this state of affairs 

 on one side of the mouth, the experiment is now transferred 

 to the opposite side, both the filter paper and the meat will 

 again be taken in till this side has also been brought to a state 

 of discriminating.' 



15/6. What of the time taken by the iterated set to become 

 adapted ? T 3 (of S. 11/5) is applicable here; so the extremeness 

 of T x is not to be feared. Thus, however large the whole, if it 

 should actually consist of iterated subsystems, then the time it 

 takes to get adapted may be expected to be of the same order as 

 that taken by one of its subsystems. If this time is fairly short, 

 the whole may be very large and yet become adapted in a fairly 

 short time. 



15/7. If Figure 15/4/2 is re-drawn so as to show explicitly its 

 relation to the system of Figure 7/5/1 the result is that shown 

 in Figure 15/7/1 (where the subsystems have been reduced to 

 three for simplicity in the diagram). 



At once the reader may be struck by the fact that the three 

 reacting parts in the organism (in its brain usually) are represented 

 as having no connexion between them: is this not a fatal flaw ? 



The subject is discussed more thoroughly in S. 17/2; here a 

 partial answer can be given. Let us compare the course of 

 adaptation as it would proceed (1) with the two left-hand sub- 

 systems wholly unconnected as shown, and (2) with the reacting 

 part of subsystem A having some immediate effect on subsystem B. 



The first case is straightforward: each subsystem is a little 

 ultrastable system, homologous with that of S. 7/5/1, and each 

 would proceed to adaptation in the usual way. 



When B is joined so as to be affected by A, however, the whole 

 course is somewhat changed. A is unaffected, so it will proceed 

 to adaptation as before; but B, previously isolated, is now affected 

 by one or more parameters that need no longer be constant. The 

 effect on B will depend on whether the effect comes to B from A's 

 reacting part or from A's step-mechanisms. If from the step- 



198 



