ON THE MAMMALIAN NERVOUS SYSTEM. 485 



in which there is from physiological evidence reason to believe considerable delay 

 occurs in the passage of the impulses through the centre. 



Such histological evidence as has been referred to in Section 1 suggests that the 

 structure is of such nature as not to facilitate the passage of impulses through this 

 region, but when we performed the following series of experiments we were surprised 

 to find the degree to which such hypothetical obstruction really prevailed. It 

 occurred to us that the obstruction might be more marked to excitatory impulses 

 passing in the reverse direction, i.e., from efferent to afferent mechanisms. We were 

 not at all prepared, however, to discover, as we have done, that that obstruction was 

 actually an absolute block ; nevertheless, such appears to be the case. 



To ascertain this we arranged the experiment as follows : 



Having divided the cord at the level of the 10th dorsal vertebra (see fig. 19), we 

 raised its peripheral end for observation as described in previous pages and connected 

 it with the galvanometer by its cross section and surface. The cauda equina having 

 been exposed to an adequate degree, a pair of nerve roots was selected (usually the 

 7th lumbar), divided just above the intervertebral foramen and their central end 

 raised in the air by ligatures and separated. The exciting electrodes were then 

 applied respectively to the central ends of the posterior and anterior roots (see fig. 1!, 

 Ex. p. and Ex. a.). 



The excitation in the first case invariably evoked a large deflection of the galvanometer, 

 but in the second case absolutely nothing even when the secondary coil was brought 

 to 12500 of the Kronecker scale, thus completely covering the primary, i.e., the zero 

 point of the centimetre scale. In other words, the excitatory condition, arriving 

 by the afferent channel or posterior root, not only passed up the few direct fibres of 

 the postero- external, and later of the posterior median column, but probably aroused 

 the afferent portion of the centre from which additional effects might ascend the cord 

 by the internuncial fibres, thus producing a very large deflection in the galvanometer. 



On the other hand, the excitatory condition which arrives at the nerve centre by 

 passing up the anterior root or efferent channel is totally unable to reach the afferent 

 side of the centre, and so ascend the cord. 



The importance of the negative result of this experiment caused us to vary it 

 as follows : 



In two Cats we divided the posterior roots of the 4th, 5th, Gth, and 7th lumbar 

 nerves and the 1st sacral on the left side. We then divided the cord in the lower 

 dorsal region (10th dorsal), and connected the upper end of the lower fragment with 

 the galvanometer. On exciting the left (mixed) sciatic nerve, which was in connec 

 tion with the cord by the anterior roots only, no electrical change could be evoked in 

 the observed region of cord however intense the stimulus used, whereas, excitation 

 of only one opposite posterior root gave deflections of 350, 320, and with a stimulus 

 of only 500, 130. This method avoided any chance escape of stimulating current, 



