136 'Journal of Comparative Neurology and Psychology. 



not react. Making allowance for these differences, we may saj that 

 it is a fact, accepted by all who have studied this question, that 

 somewhere from 85 to 100 per cent of the spinal ganglion cells 

 show chromatolvsis after the section of the peripheral nerve close 

 to the ganglion. The preparations of the ganglia of the rat also 

 show the greater number of cells in the various stages of reaction. 

 The extensive chromatolvsis, which seems to be established beyond 

 any possibility of doubt, is very difficult to harmonize with the 

 results obtained by the numerical investigation of the architecture 

 of the ganglia, results which are also beyond cavil. 



It has been clearly demonstrated by many observers, Hodge, 

 Biihler, Lewin, Hardesty, Ilatai, and Eanson, that there are several 

 times as many cells in the spinal ganglion as there are medullated 

 afferent fibers. This well established fact has been discussed in 

 detail in a previous j^aper and we do not need to consider it here. 

 It is very interesting, however, to note that we have here two well 

 established facts which seem contradictory to each other. Whereas 

 at least 85 per cent of the cells show what has been interpreted by 

 all observers as a typical axonal reaction after section of the per- 

 ipheral nerve, the numerical results show that on the average only 

 33 per cent of the cells are connected with the medullated fibers 

 in the nerve. In ^vhat way are we to explain the axonal reaction 

 in the other 52 per cent of the cells ? Lugaro and Koster, who had 

 not themselves worked with the numerical method, felt justified in 

 doubting the correctness of the numerical results, because they con- 

 flicted with facts that they knew to l)e correct. But having worked 

 with both methods and being convinced of the correctness of both 

 sets of results, I cannot so lightly set aside either in order to avoid 

 the dilemma. There seem to be but two alternatives : either the 

 reaction in this 52 per cent of the cells is not an axonal reaction 

 at all but is secondary to some intra-ganglionic disturbance, or 

 there must be a very large number of non-medullated fibers in the 

 nerve. It has been shown (Hanson '08) that the histology of the 

 ganglion would not exclude the possibility of the first alternative. 

 ''The spinal ganglion is not to be regarded as an aggregation of 

 more or less spherical cells each independent of the others, and 



