146 "Journal of Comparative Neurology and Psychology. 



density of the jDerijjheral ring that has resisted solution during 

 the earlier phases of chromatolysis. The larger part, however, 

 is distributed in the form of very fine granules through the re- 

 mainder of the protoplasm. At this stage the majority of the 

 large cells correspond quite well to Lugaro's description of the 

 cells of his second type. They are large and medium-sized clear 

 cells with very fine tigroid masses which are larger at the periph- 

 ery of the cell. They do not, however, long conform to this 

 type, since the chromatic substance is rapidly increasing in quan- 

 tity and is laid down in the form of large granules scattered uni- 

 formly throughout the protoplasm. By the seventeenth day the 

 central portions of the cells contain Nissl-bodies of considerable 

 size, and by the twentieth day the distinction between the coarsely 

 granular peripheral ring and the rest of the cell has disajDpeared. 

 The cells present a uniformly coarsely granular appearance. 

 In this way we are able to trace the large cells from the height 

 of chromatolysis through the stages of gradually increasing chromatic 

 substance, until they present the pylcnomorphous appearance charac- 

 teristic of the large cells 20 days after the operation (Figs. 2 and 4). 



The importance of thus following the transformation is two-fold. 

 In the first place, it shows that the large cells undergo repair; and 

 in the second place it shows that the large cells present in the 20- 

 day preparations and counted as such in the enumerations given 

 in Table II are in reality the same large cells as were originally 

 present in the ganglion, and cannot l)e regarded as hypertrophied 

 small cells. In this way we can be sure that the conclusions derived 

 . from Table II are not misleading, namely, that it is the small rather 

 than the large cells which undergo complete degeneration. 



Conclusions. 



The more important results of the investigation may be sum- 

 marized in the form of answers to the problems suggested in the 

 introduction. 



1. It is ]n'obable that the medium sized coarsely granular cells 

 which fail to react to a lesion of the nerve close to the dorsal root 

 ganglion, belong to neurones which send no axon into the nerve 

 (Dogiel's Type II, or new types III, IV, VIII and XI). 



