404 G. ADRIAN HORRIDGE 



Another preparation where the non-specific inhibition plays a different 

 role is in the last abdominal ganglion of the Mantis (Milburn et al., 1960). 

 When the last connexion of the de-afferentated terminal gangUon is severed 

 from the rest of the nervous system, patterned bursts of impulses soon began 

 to appear in the phaUic nerve, similar to those causing copulatory movements 

 if the nerve is intact. Apparently, descending inhibitory impulses are cut off 

 by severing the cord. 



Widespread connexions are also a feature of many second-order sensory 

 neurons or interneurons of arthropods, but in these examples it is on the 

 input side that the neuron shows various degrees of specificity. Examples are 

 accumulating which show that one interneuron may collect from a large 

 number of primary sensory axons, not only of many segments of the body 

 but even of differing modahty. In his detailed analysis of the crayfish cord, 

 Wiersma (1958) finds examples of differing sizes of the sensory field covered 

 by an interneuron, with examples of overlap and inclusiveness of the fields 

 of different interneurons. For example, one interneuron may respond to 

 sensory field A, another to sensory field B, and a third to A and B. In another 

 type of pattern, an interneuron can be found which responds to any of a 

 number of similar sensory fields repeated in adjacent segments, whereas each 

 of these are also represented individually by other interneurons. Occasionally 

 a very small group of hairs have their own interneuron. It is evident that there 

 are all degrees of specificity in the connexions between sensory axons and 

 interneurons. It is tempting to think that the more specific ones have evolved 

 later. The interneurons are clearly individually differentiated to a high degree 

 and some are unique in that only one is found in each side of the ventral cord. 

 Perhaps here is an ideal situation in which it could be tested whether the 

 various degrees of specificity of the input connexions lie in the exact ana- 

 tomical addressing of sensory endings terminating on them, with a single 

 transmitter, or on the varying sensitivity of different interneurons to a range 

 of transmitters with different ranges of effect, formed by the differentiated 

 peripheral sensory cells. 



When we turn to the detailed structure of the neuropile as seen in single 

 sections examined with the electron microscope we find no data which is 

 relevant to the present question. Synapses are identified as contacts between 

 axons having synaptic vesicles at one side of the bounding membranes, but 

 there is no way of telhng whether they are anatomically addressed, or whether 

 widespread branches are selective in their contacts. The special methods for 

 single neurons, using Golgi or intravitam methylene blue methods, have 

 shown that the neuropile of many invertebrates consists of branched afferent 

 axon arborizations and branched dendrites of efferent neurons. However, 

 for the finer endings, half a century of detailed work by the best methods on 

 many favourable preparations has failed to demonstrate any pattern or circuit 

 diagram which is evidently anatomically addressed. The finer afferent and 



