ORGANIZATION OF PRIMITIVE CENTRAL NERVOUS SYSTEM 397 



such blanks can convey information to the effectors. Such a system, with 

 the pattern of excitation in time as its only parameter, is non-addressed. Far 

 from being necessarily random, all degrees of anatomical pattern may be 

 found in such a system, but the physiologically significant connexions cannot 

 be divided into recognizable classes. Neurons in a coelenterate nerve net may 

 meet with a characteristic number of others, or have a regular pattern, but 

 there is no discrimination between one individual neuron and another. Each 

 can act vicariously for any of the others in the net and we have no anatomical 

 or physiological means of distinguishing an individual neuron when it recurs 

 in different animals. As conceived here, in a simple nerve net, the neurons 

 themselves do not distinguish. 



There is a third possible type of system which, until ruled out by experi- 

 ment, must be considered as a possibiHty in the interpretation of neuron-to- 

 neuron transmission in a system which is more complex than a single nerve 

 net. This depends on the differential sensitivity of the efferent neurons to 

 different transmitter substances produced from thereby distinguishable afferent 

 neurons and interneurons. If we imagine a primitive ganglion with different 

 classes of afferent neurons having distinct transmitter substances and only 

 a few classes of efferent neurons, then it can be readily seen that differential 

 sensitivity of the efferents, to the transmitters which have access to them, can 

 account for their differential responses to various types and combinations of 

 afferent excitation. From points of observation on the efferent side there is 

 no way of distinguishing such a system from one which is fully anatomically 

 addressed although it need contain no anatomically addressed connexions 

 at all. Such a system, called chemically addressed, is one of a class of systems 

 which are physiologically addressed. Other physiologically addressed systems 

 can be imagined; they could, for example, have codes based upon the propor- 

 tion of afferents which are active, or the rate of change or other temporal 

 pattern of the total afferent excitation. A purely physiologically addressed 

 system has the character that the morphogenetic specifications define no 

 particular anatomical connexions although afferent excitation can be clas- 

 sified into different kinds, as judged solely by the differential sensitivity and 

 responses of the output; only a quantity of branching and plenty of con- 

 nexions are required. 



In terms of the same analogy with defivery of mail, a chemically addressed 

 system is Uke an indiscriminate delivery of more than one kind of leaflet. 

 Many recipients receive a copy of many kinds of leaflet and some may receive 

 all kinds but their different sensitivities ensure that only the intended class of 

 recipients respond to the contents of certain classes of leaflet, and they do so 

 in their characteristic manner. It should be noted that both the anatomically 

 and the chemically addressed neurons are distinctively labelled lines. These 

 are physiological pathways known to be labelled because they carry a parti- 

 cular identifiable class of excitation, which the next stage into which these 



