ScHAPER, Structure of Selachian Cerebellum. 13 



the axis-cylinders are obliged to pursue their destination along 

 the membrana Imiitans iiiternia. 



If we further consider the granular layer, we find here 

 in general a dense tangle of fibers proceeding in all directions 

 (Plate III, fig. 17); in the upper part near the layer of Purkinje 

 cells a perpendicular direction of the fibers does indeed prevail, 

 in the deeper layers, however, they cross each other without 

 any order. In figure 17 (Plate III), which illustrates these re- 

 lations, the delicate axis-cylinders of the granule cells are 

 omitted ; the larger fibers present come in part from the Pur- 

 kinje cells, another part, however, undoubtedly belongs to the 

 ^^ ascending fibers.'' As mentioned above, there do not appear 

 to be any characteristic marks attached to these two groups of 

 fibers, so that they can only be distinguished by following them 

 to their origins or to their terminal arborizations. This is, nat- 

 urally, only practicable in a few fibers and in the cerebellum of 

 selachians is especially difficult from reasons mentioned below. 



The course of the fibers within the granular layer is some- 

 what differently arranged in those parts of the cerebellar plate 

 where the former does not border directly upon the mc^nbrana 

 liniitans interna, but is separated from it by a compact layer of 

 medullated fibers (the first rudiment of a central white matter) 

 and where, besides, externally the Purkinje cells are as a rule 

 entirely lacking so that the granular layer borders immediately 

 upon the molecular layer, the latter being usually very thin in 

 such places (Plate III, fig. 18). Here one sees, instead of an 

 irregular network, well marked fiber-hindles at definite intervals 

 ascending in a vertical direction within the granular layer. 

 They stand in connection below with the basal fiber layer and 

 extend upward to the boundary between the granular and 

 molecular layers or also somewhat into the latter. Here they 

 end, as though cut off, usually with a small hook-like bend. 

 The fiber-bundles are so compact that they traverse the gran- 

 ular layer as completely closed masses. The axis-cylinders 

 of the granule cells do not participate in their formation but 

 proceed upwards between the bundles in the usual manner. 

 They are not drawn in figure 18 (Plate III). I have unfortu- 



