1 6 Journal of Comparative Neurology. 



only in certain parts of the cerebellar plate and, furthermore, 

 that other neuroglia elements also are present, which are prob- 

 ably not derived from the ependyma cells. 



The most primitive forms of ependyma cells with an entirely 

 embryonic habit we find along and in the immediate neighbor- 

 hood of the median line of the cerebellar plate, where the me- 

 dian furrow nearly reaches the surface. Cells of this kind are 

 shown in figure 19 (plate III). We see here one or two proces- 

 ses arise from a round-oval or triangular cell body, which pro- 

 cesses may again divide and extend in a fairly straight course to 

 the surface where they lie against the limitans externa with a 

 conical expansion. These processes are entirely smootJi. Usually 

 a number of processes arising from different cells are united into 

 a thick bundle and form in this way bulky column-like structures 

 which come out very clearly even in simply stained preparations. 

 Not infrequently there arises also from the under pole of the 

 cell body a short frequently branched process. Purkinje cells 

 are not present in this portion of the cerebellar plate. 



All other ependyma cells are to be essentially distinguished 

 from those here described, above all because most of thejn have com- 

 pletely lost their connection with the mcmbrana limitans externa. 

 The few which retain a permanent connection with the surface 

 I have found, in my preparations, almost exclusively in the 

 thinner regions of the cerebellar plate only and especially imme- 

 diately cephalad of the transition of the latter into the velum 

 meduUare posterius. Such an ependyma fiber is shown in fig- 

 ure 20 (plate IV). We see here the fiber arise in the typical way 

 from a pyramidal cell body lying close against the membrana 

 limitans interna, pursue an irregular course through granular 

 and molecular layers and attach itself with a conical expansion 

 to the limitans externa. On its way the fiber gives off several 

 lateral branches which partially fall into numerous terminal 

 twigs. 



Those supporting fibers of ependymal origin which have 

 lost their connection with the surface of the cerebellum termin- 

 ate in various planes of the granular layer or reach the Purkinje 

 cells. These elements are impregnated with extraordinary ease 



