iv GENEBAL PHYSIOLOGY OF NEKVOUS SYSTEM 197 



united into bundles, and that the true ramifications of the con- 

 ducting elements probably exist only in the terminal and peripheral 

 organs, where distinction and localisation of the physiological 

 effects of the excitation transmitted along the conducting filaments 

 is no longer necessary. 



Another fundamental fact of nerve conduction is what William 

 James, the psychologist, termed thQ/orward direction. Conduction 

 is normally centripetal, i.e. from the periphery to the centre in 

 sensory fibres and afferent fibres in general, and centrifugal, from 

 centre to periphery, in the motor fibres and efferent fibres in 

 general. Again, when the nerves are artificially stimulated along 

 their course, the effect is expressed in movement for motor nerves, 

 in sensation for the sensory. We shall see, in fact, in discussing 

 the physiology of the special nerve roots, that on stimulating the 

 central stump of a root that contains motor fibres only all sensory 

 reaction fails, and on stimulating the peripheral stump of a root 

 containing only sensory fibres no motor reactions are obtained. 



This fact at first sight justifies the conjecture that sensory 

 nerves can only conduct the excitation in an afferent direction 

 when excited along their course, and motor nerves only in an 

 efferent direction, as though there were some valvular mechanism 

 which allows the transmission of the impulse in one direction and 

 blocks it in the other. Certain experimental facts, however, show 

 this hypothesis to be untenable, and indicate that nerves in 

 general, when artificially excited at any point of their course, are 

 capable of conducting in both directions, but the effect is manifested 

 only at the centre for sensory nerves, and at the periphery for 

 motor nerves. 



The best argument for double conduction appears from the 

 study of the electrical phenomena that accompany the excitation 

 of nerve. This will be discussed in a separate section. When a 

 nerve is stimulated midway, while the two ends are joined up to 

 two galvanometers, the so-called negative variation is seen on 

 both. This occurs not only with a mixed nerve, which contains 

 both sensory and motor fibres, but also, as Du Bois Eeymond 

 pointed out, with a nerve which contains only motor (efferent) 

 fibres, e.g., the ventral spinal roots. 



Gotch and Horsley repeated and varied this experiment, both 

 with efferent and afferent nerves. They divided a ventral root of 

 the sciatic plexus in the cat ; connected it with a highly sensitive 

 galvanometer, and then excited the trunk of the sciatic. A double 

 reaction followed of the muscles of the limb, which proved 

 centrifugal conduction in the motor fibres, and of the galvano- 

 meter, which showed centripetal conduction in the same motor 

 fibres. Similar effects were obtained with sensory nerves. On 

 exciting a dorsal root and connecting the central end of the 

 divided sciatic with the galvanometer, the negative variation 



