THE CEREBELLUM. 515 



The thick body of these elements, measuring from 60 to 70 p 

 in length, and from 20 to 30 /j. in breadth, is most frequently 

 vesiculaiiy expanded towards the granule layers internally, and 

 is there without processes, whilst the large processes directed 

 outwards are always visible. According to 'Kb'lliker, Deiters, 

 KoschewnikofF,and Hadlich, each cell gives off on its inner aspect 

 only a single process, which the three last observers maintain 

 is niedullated and unbranched, whilst Stilling considers there 

 are several branches which immediately divide in a plexiform 

 manner in the granule layer, and, according to Gerlach, unite 

 with the granules. Hadlich attributes the connection with 

 the granules to the finest ramifications of the thick finely 

 striated processes which run into the pure grey layer, the 

 minutest filaments of which curving back, as already stated 

 by Obersteiner, near the surface of the cerebellar cortex, join a 

 plexus which is in direct connection with the granules. He 

 adduces also, as a ground of probability for this, the broad fact 

 that at the level ,of the lamellae of convolution the closest 

 approximation of the cells of Purkinje coincides with the 

 greatest breadth of the granule layer ; and again, that around 

 the sulci the most sparing distribution of the former occurs, 

 with the greatest attenuation of the latter. According to 

 Obersteiner, the cells of Purkinje spread forth their numerous 

 antler-like branches in a single plane, so that they do not each 

 form a brush or cone of fibrils, but resemble the ribs of a leaf, 

 whence the laminar structure of the medulla of the cerebellum 

 would also be expressed in the structure of the cortex. Lastly, 

 Obersteiner and Hadlich have found that the remarkable 

 transverse divergence of the large primary processes of the 

 cells of Purkinje, by virtue of which they extend parallel to 

 the surface of the cortex in opposition to the similar processes 

 of other cells which run radially to the surface, represent a 

 mode of branching that occurs only in those portions of the 

 cortex that surround the sulci. 



The large cells appear to me to be invested by a loose hyaline 

 sheath which is extended for a little distance along the larger 

 processes. Obersteiner regards it as a plexus of supporting 

 fibres loosely investing the cell. The internal layer, which was 

 already recognized by Purkinje, and has been subjected to a 



L L2 



