The brain of Acipenser. 193 
must perforce become a great coördinating center for the efferent 
impulses in response to cutaneous and special senses. That it shall 
serve this function there must be a great increase in the number 
of associative elements. These elements are the granule cells and 
the cells with short neurites primarily, and secondarily the newly 
developed PURKINJE cells. The structure of the adult acusticum 
and cerebellum in fishes shows that the PURKINJE cells are not 
primitive in their character, but are among the youngest cells in 
the cerebellum. Their development was due to the increase in as- 
sociative elements. As the granule cells increased in number their 
neurites came to form a definite layer on the external surface of 
the cerebellum, the molecular layer. There was then a continuous 
tract of comparatively short fine fibres connecting the grey matter 
of the dorsal region of successive segments throughout the cord, 
medulla and cerebellum (cf. page 174). In the cerebellum this tract 
is much more voluminous than in the cord, and forms a definite 
layer which continues back over the acusticum as the cerebellar 
crest. Now those large cells in the acusticum and cerebellum whose 
dendrites happened to be imbedded in this layer began to take on 
a peculiar character by the development on their dendrites: of fine 
spines. At first only a part of the dendrites of a given cell developed 
these spines because only a part of the dendrites were surrounded 
by the fine fibres. This condition is still present in the acusticum 
of Acipenser. Gradually the spiny dendrites became more prominent 
and in the cerebellum where the dendrites of any given cell were 
entirely surrounded by fine fibres the typical PURKINJE cells were 
developed. It is worthy of notice that in the acusticum of Aci- 
penser the PURKINJE and transitional cells may have a double 
function. They may receive impulses from afferent sensory fibres 
and also from the neurites of granule cells. In the valvula also 
are cells which hold similar relations to mid brain tracts and fibres 
of the molecular layer. The tendency has been for the granule cells 
and the cells of the II type to receive the in-coming impulses and 
transfer them, either directly or by way of the Korbzellen, to the 
PuRKINJE cells which continue to give rise to descending or motor 
stimuli. The need of coördinating apparatus is greatest in the cere- 
bellum, and the granule cells, molecular layer and PuURKINJE cells 
have come to be restricted to the cerebellum. 
The anterior crus of the cerebelium. — By this I mean 
all those tracts of fibres which connect the cerebellum with parts 
Zool. Jahrb, XV. Abth. f, Morph. 13 
