282 Drs. Mott and Halliburton and Mr. Edmunds. [June 12, 



are growths from the central ends of divided nerve trunks. The experimental 

 facts recorded by those who, like Bethe and Kennedy, hold the opposite view 

 are susceptible of easy explanation, mainly on the lines emphasised by Langley 

 and Anderson of accidental and unnoticed connection of the peripheral 

 segments with the central nervous system by means of other nerves cut 

 through in the operation. If such connection is effectually prevented, real 

 regeneration of structure and restoration of function never occurs. 



Moreover, the regenerated fibres always degenerate in a peripheral direction, 

 and in a peripheral direction only, when the link that binds them to the 

 central nervous system is again severed. Perhaps the most striking of 

 the facts brought out in our own work is in reference to the development of 

 the medullary sheath ; this appendage of the axis cylinder appears earliest at 

 situations near the point where the ends of a nerve have been joined together, 

 and reaches the distal portions later. 



What takes place in the peripheral segment of a divided nerve is a mul- 

 tiplication, elongation and union into long chains of the neurilemmal cells. 

 The same change is even more vigorous at the central termination of the cut 

 nerve; and the view of the phagocytic and nutritive function which we 

 attribute to this sheath has been supported independently by some striking 

 observations of Graham Kerr which we have fully referred to. At the 

 central end this nutritive function is effective and provides for the nourish- 

 ment of the actively lengthening axis cylinders. At the peripheral end, 

 unless the axons reach it, it is ineffective in so far as any real new formation 

 of nerve-fibres is concerned. If, however, the axons reach the peripheral 

 segment, the work of the neurilemmal cells has not been useless, for they 

 provide the supporting and nutritive elements necessary for its continued 

 and successful growth. The neurilemmal activity appears to be essential, for 

 without it, as in the central nervous system, regeneration does not take place. 



According to Graham Kerr the formation of neuro-fibrillae may possibly 

 take place in the protoplasmic residue of the degenerated axis cylinder ; 

 according to Marinesco this property is assigned to the neurilemmal elements 

 themselves, a proposition we regard as extremely improbable, seeing that 

 these elements are mesoblastic. In either case these two observers consider 

 that the neuro-fibrillae, however formed, are ineffective until they are 

 activated by union with those of the central axons. Our own observations 

 do not entirely exclude this view, but on the other hand (as we have never 

 seen these fibrillar in the peripheral segment unless union with the central 

 end has occurred) they lend it no support. All our facts are readily explic- 

 able, however, on the theory that the nerve-fibres are growths from the 

 central ends of divided nerves. 



