77 



Dutlina from /»hich many processss r-iiiate. These processes 

 raraly branch. Some of them may proceed to a distance equal 

 to several times the greater axis of the cell, but most of the 

 branches are far shorter. Thsy are placed so closely along 

 the margin of the cell that the whole has something the effect 

 of a halo. These cells are clearly homologous with the astro- 

 cytes of higher vertebrates. 



The other type of neuroglia cell (Pig.l9,bg. f . ) is refer- 

 able to the category of Bergmann's fibres of the mammalian 

 cerebellira. The cell-body has fewer processes than the astro- 

 cyte, but it gives origin to one stout fibre which takes a 

 peripheral course, without branching, through the entire thick- 

 ness of the molecular layer. The path of the fibre is not one 

 directly toward the surface, for it runs almost parallel with 

 the layer of Purkinje neurones for a short distance; it then 

 turns upward, terminating at the limitans externa in a conic- 

 al expansion. During the proximal part of its length, it bears 

 fine processes similar to those emerging from the cell-body. 

 The distal part of the fibre has a remarkably vast number of 

 fine processes. These may remain separate from each other, 

 but at intervals they become matted together so closely that 

 the whole has the appearance of felt or even that of a solii. 



The account given by Schaper ('93) of what he took to be 

 Bergmann's fibres does not correspond with my findings in sev- 

 eral particulars. This author does not mention the processes 

 of the cell-body, nor do his figures show them. It may be that 



