78 



his prspar-itions were insufficiently impregnated. Weii^ht is 

 given to sach a possibility from the fact thnt he dii not find 

 astrocytes at all. It is certainly of some phylogenstic inter- 

 est that Bergmann's fibres appear to be derived from astrocytes. 

 I have found numerous instances where transitional forms are 

 recognizable, linking the extremes of the simple astrocyte 

 flith the one provided with a long process, the fibre of Berg- 

 mann. 



The position occupied by the cell-body of a Bergmann's 

 fibre is also significant. Kolliker ('9^,0.3^-3) states that 

 at birth Bergmann's fibres in mammals lie at the boundary of 

 the granular and molecular layers, and that during growth they 

 normally migrate into the granular layer. Now the permanent 

 position of Bergmann's fibre in Mustelus corresponds to the 

 embryonic state in the mammal. An additional comparison may 

 be instituted regarding the form of the fibre, the adult ele- 

 ment of Mustelus remaining simple, while the mammalian fibre 

 becomes much branched. That the condition in Vustelus corres- 

 ponds to the embryonic stage of development in higher verte- 

 brates is, of course, no more than should be expected. 



The physiological interpretation of the cerebellar neu- 

 roglia is not difficult. The numerous processes from both the 

 astrocytes and Bergmann's cells provide a delicate suspensory 

 apparatus for the neurones of Purkinje. The great Bergmann's 

 fibres, reaching upward as they do through the molecular layer, 

 may be likened to so many telegraph poles; and the matted tan- 



