Conduction in the Central Nervous System. 



245 



idiolateral limb it is extremely difficult, often impossible, to then 

 obtain by this facile long spinal path any discharge at all on the 

 side of the transected afferent roots, although that side is usually 

 peculiarly accessible. T find that similarly severance of the dorsal 

 (afferent) roots in their extraspinal course greatly impairs the re- 

 action from the /. graciles and /. cuneati. Thus, if when flexion of 

 right knee or right hallux is being regularly evoked by excitation of 

 /. gracilis at the top of the cord, the extraspinal dorsal (afferent) roots 

 of the right pelvic limb be severed, the reaction, until then regularly 

 obtained, disappears or almost disappears. The section of right band 

 roots annuls the right hand reaction, but not the left hand, and con- 

 versely. On the other hand, the flexion of knee, or of hallux, or of 

 elbow obtained by excitation of the Rolandic cortex or of the lateral 

 column (pyramidal tract fibres) is, as has been shown in a previous 

 number of these ' Proceedings '* by Dr. Mott and myself, not im- 

 paired after the root severance, indeed often appears, on the contrary, 

 to be facilitated. In this respect, therefore, the reaction obtainable 

 by direct excitation of /. graciles and /. cuneati is shown to be curiously 

 different from that- obtainable from the pjTamidal tract fibres and 

 Rolandic cortex. On the other hand, it is seen to resemble in this 

 respect to a remarkable degree the " long spinal reflexes " as defined 

 above. 



What, then, is tbe nature of this reaction obtainable from the 

 /. graciles and cuneati ? The reaction is evidently one which 

 involves each dorsal column of the cord as a conducting path, in 

 many cases as a " long " — in not a few as a remarkably long — conduct- 

 ing path, even employing its whole length. In light of the evidence 

 given above, I infer that although certainly, as has been long estab- 

 lished, the dorsal column is, with the single exception of its short, 

 scanty, and deeply placed ground-bundle, a functionally purely 

 upward path, consisting of nothing else than sensory root fibres, 

 the vast majority of which fibres — and all the longest of which — 

 are ascendant ; the conduction along it in these experiments is 

 downward, even extending its whole length. That is to say, the 

 conduction must be downward and cellulipetal along ascending axons 

 which function in a. cellulifugal direction ; that is to say, the propa- 

 gation of the impulses artificially started in my observations must 

 have been antidrome instead of orthodrome. The motor discharges 

 evoked I refer to the spread of the excited condition into the collaterals 

 of the axons excited to antidrome conduction, their collaterals 

 impinging upon motor neurons. 



The direction of propagation occurs therefore in opposition to the 



* " Experiments upon the Influence of Sensory Nerves upon Movement and 

 Nutrition of the Limbs," ' Roy. Soc. Proc.,' vol. 57, March 7, 1895. 



