PROPAGATION ALONG NERVE FIBRES 285 



transmit impulses in either direction is shown further by the experi- 

 ment known as Kiihne's gracilis experiment. The gracilis muscle of the 

 frog is separated into two portions by a tendinous intersection, so 

 that there is no muscular continuity between the two halves. The 

 nerve to the muscle divides into two branches, one to each half, and at 

 the point of junction there is division of the axis cylinders themselves. 

 If the section a in the diagram (Fig. 105), which is quite isolated 

 from the rest of the muscle, be stimulated, as by snipping it with 

 scissors, the whole muscle contracts. If the portion of the muscle 

 which is free from nerve fibres be stimulated in the 

 same way, the contraction is limited to the fibres 

 directly stimulated, showing that in the first case the 

 stimulus excited nerve fibres w T hich transmitted the 

 impulse up the nerve to the point of division and 

 then down again to the other half of the muscle. 



Since nerves have this power of conduction in both 

 directions, it might be thought that a single set of 

 nerve fibres might very well subserve both afferent 

 and efferent functions, at one time conducting sensory 

 impulses from, periphery to cord, at another time 



motor impulses from cord to muscles. But this is . F IG. 105. 



-, -. Kiihne s graciiis 



not the case. As a matter of tact we find in the experiment, 

 body a marked differentiation of function between 

 various nerve fibres. Thus Bell and Majendie showed that the 

 spinal roots might be divided into afferent and efferent, the anterior 

 root carrying only impulses from spinal cord to periphery, while the 

 posterior roots carried impulses from periphery to central nervous 

 system. The law known by the name of these observers states indeed 

 that a nerve fibre cannot be both motor and sensory. We may find 

 both kinds of fibres joined together into a single nerve-trunk, but 

 the fibres in each case are isolated and conduct impulses only in one or 

 other direction. Under normal conditions the afferent fibres are 

 excited only at their endings on the surface of the body, while the 

 efferent fibres are excited only at their origin from the spinal cord. The 

 difference in the function of different nerve fibres depends therefore 

 not so much on the structure of the nerve fibre itself as on the con- 

 nections of the fibre. We can show this experimentally by grafting 

 one set of nerve fibres on to another. If the cervical sympathetic be 

 united to the lingual nerve, stimulation of the sympathetic, instead of 

 causing, as usual, constriction of the vessels of the head and neck, will 

 cause dilatation of the vessels of the tongue and secretion of watery 

 saliva. In the same way the finer functional differences between the 

 various forms of sensory nerves seem to be determined by their 



