1 8 Journal of Comparative Neurology. 



length with a more or less dense moss-Hke covering which pre- 

 sents an uncommonly delicate aspect. They people the mole- 

 cular layer in great numbers and form, when many are impreg- 

 nated, an almost impenetrable thicket. These elements have 

 in fact a great resemblance to ependyma fibers. One can 

 imagine, that these fibers have retained their original connection 

 with the limitans externa and that with the progressive thick- 

 ening of the cerebellar plate the cell body, released from the 

 limitans interna, is gradually withdrawn into the interior of the 

 plate. I 'am more inclined, however, to the view that these 

 fibers are genetically dissimilar to the ependyma cells, that they 

 owe their origin to a part of the derivates of the ^^ germinal 

 cells " and are accordingly to be regarded as secondary support- 

 ing elements or as glm elemcjits in a narrower sense. The same 

 obtains for the Bergmann's fibers in general. However, as far 

 as I know, a positive proof for this conception is not yet 

 brought forward. Further investigations are necessary here. 

 Should they actually prove to be true glia elements, the fibers 

 described would represent, according to my observations at all 

 events, the only ones of their kind in the cerebellum of 

 selachians, inasmuch as I have not succeeded in demonstrating 

 any elements comparable to "astrocytes" or ^^ mossy cells'' in 

 my preparations. 



It emerges from the foregoing considerations that the fun- 

 damental structure of the cerebellar cortex of the selachians as 

 a whole shows already the typical features of that of the higher 

 vertebrates. A lower stage of development, however, can be 

 established in regard to the individual cellular elements, which 

 is expressed both in the less complexity of the dendrites and 

 axis-cylinder and in the prevailing ependymal character of the 

 supporting structure. 



Harvard Medical School, Boston, Mass., Nov. ij, i8gy. 



