ORGANIZATION OF PRIMITIVE CENTRAL NERVOUS SYSTEM 401 



tinguish between accelerating and inhibiting afferent pathways where these 

 both are observed in the same species. However, there must be some form of 

 addressing within the ganghon in certain species of jellyfish, e.g. Cyanea, 

 Cassiopea, which have a second efferent path from the ganghon; this runs 

 only locally to muscles of the bell and brings about the asymmetrical compo- 

 nent of the beat as the animal rights itself when tilted. This efferent pathway 

 is excited principally by the receptors of the direction of gravity, situated in 

 the region of the ganglion, and the same receptors can also modulate the fre- 

 quency of the pacemaker in the giant fibre net. There is no a priori reason for 

 rejecting either anatomically or chemically addressed connexions for the 

 transmission of excitation to these two efferent pathways within the gangUon. 

 The existence of the neuropile at all, with an enormous number of apparently 

 redundant and haphazardly arranged arborizing neurons in each marginal 

 ganghon, is perhaps more compatible with the chemically addressed system. 

 The point is that evolution from one non-addressed nerve net to two inter- 

 acting nets and then to a simple ganglion which integrates several types of 

 sensory excitation could have occurred by a differentiation of classes of 

 sensory neurons distinguished by their specific transmitters, and present 

 techniques do not distinguish this from an anatomically addressed system. 

 On the efferent side the differentiation has been very small, in this instance 

 to only two distinct classes of neurons. 



There is a limitation of technique which influences any consideration of the 

 modulation of ongoing neuron activity, even at the simple level of the coel- 

 enterate pacemaker. The variation between individuals, between similar 

 ganglia where there are several in one animal, and in repeated performances 

 of one preparation, means that the results must be treated statistically. When 

 working with a ganghon with several distinguishable efferent neurons the 

 same consideration apphes for each neuron. Therefore the electrophysio- 

 logical data on which a theory of the ganghonic activity is based, is the 

 influence of various factors on the "probabifity of motor firing", a term which 

 will recur. Input-output relations of this type cannot, by their nature, dis- 

 criminate between the suggested mechanisms of preservation of information 

 during transmission throughaganglion. The mechanism withinmust be studied. 



A more complicated example of a widespread effect is the inhibition of 

 crawhng in the earthworm (Collier, 1938). Stimulation of the head end puts 

 a stop to forward crawhng, causing the movement to "freeze", but the tonus 

 of all the muscles along the body is maintained. Here a pathway, probably a 

 single neuron, runs the length of the worm, and its effect when active is to 

 lower the probability of firing of motoneurons along the whole length of the 

 worm. On account of the distribution of the peristaltic wave, the inhibitory 

 impulses catch adjacent segments at different stages in the sequences of im- 

 pulses from corresponding motoneurons in different segments. The whole 

 motor output is evidently not stopped, because control of tone remains. The 



