PHYLOGENETIC ORIGIN OF NERVOUS SYSTEM 147 



circular and dorso-ventral muscles, while that of the posterior nerve 

 causes the antagonistic muscles, the longitudinal muscles, to 

 contract. 



Muscles and their antagonists are then very apt to receive 

 their motor supply from the central nervous system by way of 

 nerves which run separately ; but a muscle and its antagonist are 

 necessarily bound closely together in their innervation, which is 

 of a reciprocal character, so that the contraction of the one muscle 

 is accompanied by the inhibition of the other. If then it is found 

 that the axon of a nerve cell in the central nervous system splits 

 to form two nerve fibres, one of which passes out into one nerve 

 and the other into a quite separate nerve, it seems impossible that 

 both those fibres should be motor to muscles of the same group, 

 and much more probable that they supply antagonistic muscles 

 with efferent nerve fibres of opposite character. 



The origin of reciprocal innervation is on this scheme to be 

 found in a single excitor neuron, which innervates a muscle and 

 its antagonist in opposite directions by means of the splitting of 

 its axon. Such cells may well reciprocally innervate such muscles 

 as the voluntary circular and longitudinal muscles of the leech, 

 the motor fibres of which run respectively in the anterior and pos- 

 terior nerves. 



My son has also observed the cells with splitting processes, 

 which Retzius described, and finds that the smaller lateral cells 

 with such processes do not coincide with the lateral adrenaline 

 cells. Preparations clearly demonstrating these splitting fibres 

 are only occasionally obtained, and the number of such cells is 

 therefore difficult to determine. Retzius only described two, 

 situated laterally ; and undoubtedly two cells, staining exception- 

 ally easily, whose axons split in the ganglion, can sometimes be 

 seen in a single preparation. There are however indications that 

 other cells also have fibres which split in the same manner, 

 whose identification is much more difficult, as the staining of 

 more and more cells renders the picture more confused. The 

 lateral adrenaline cells stain badly and very late, so that identi- 

 fication of any splitting of their axons in the ganglion is a matter 

 of extreme difficulty. The chief fact that emerges is that two or 

 more cells exist, which do not contain adrenaline but supply both 

 anterior and posterior nerves with efferent fibres. 



A similar reciprocal innervation for the musculature of the 



10 * 



