10/3 DESIGN FOR A BRAIN 



If the cell is sufficiently sensitive to be affected by changes of 

 atomic size, then such changes would usually be of step-function 

 form, for they could change only by a quantum jump. But this 

 source of step-functions is probably unavailable, for changes of 

 this size may be too indeterminate for the production of the 

 regular and reproducible behaviour considered here (S. 1/10). 



Round the neuron, and especially round its dendrons and axons, 

 there is a sensitive membrane that might provide step-functions, 

 though the membrane is probably wholly employed in the trans- 

 mission of the action potential. Nerve ' fibrils ' have been des- 

 cribed for many years, though the possibility that they are an arte- 

 fact cannot yet be excluded. If they are real their extreme delicacy 

 of structure suggests that they might behave as step-functions. 



The delicacy everywhere evident in the nervous system has 

 often been remarked. This delicacy must surely imply the 

 existence of step-functions ; for the property of being w delicate ' 

 can mean little other than 4 easily broken ' ; and it was observed 

 in S. 7/6 that the phenomenon of something ' breaking ' is the 

 expression of a step-function changing value. Though the argu- 

 ment is largely verbal, it gives some justification for the opinion 

 that step-functions are by no means unlikely in the nervous system. 



' The idea of a steady, continuous development ', said 

 Jacques Loeb, ' is inconsistent with the general physical 

 qualities of protoplasm or colloidal material. The colloidal 

 substances in our protoplasm possess critical points. . . . 

 The colloids change their state very easily, and a number of 

 conditions . . . are able to bring about a change in their 

 state. Such material lends itself very readily to a discon- 

 tinuous series of changes.' 



10/3. Another source of step-functions would be provided if 

 neurons were amoeboid, so that their processes could make or 

 break contact with other cells. 



That nerve-cells are amoeboid in tissue-culture has been known 

 since the first observations of Harrison. When nerve-tissue 

 from chick-embryo is grown in clotted plasma, filaments grow 

 outwards at about 0-05 mm. per hour. The filament terminates 

 in an expanded end, about 15 x 25 fx in size, which is actively 

 amoeboid, continually throwing out processes as though explor- 

 ing the medium around. Levi studied tissue-cultures by micro- 

 dissection, so that individual cells could be stimulated. He found 



126 



