The brain of Acipenser. 189 
he finds the “ascending” fibres ending in the molecular layer. He 
does not say on what he bases the identification of these fibres, but 
apparently it is upon the appearance of the fibre without tracing 
it to is origin. As I know from my own study, it would be very 
difficult in this way to be sure of the character of any fibre which 
had an ascending course in a given section. In work on the fish 
cerebellum it is not too much to expect that the source of fibres 
should be made out in order to determine their character. In Aci- 
penser I have described the endings of many different tracts of 
fibres in the cerebellum. In every case the fibres have been traced 
with perfect certainty from their origin, or at least from without 
the cerebellum, into the cerebellum and their endings have been 
studied in one and the same continuous and complete series of sections. 
In all cases, too, the descriptions rest not on the findings in one 
series, but in several series. In no case have I found any con- 
siderable number of fibres ending in the molecular layer. The only 
fibres which I have succeeded in tracing to the molecular layer are 
those of the tractus tecto-cerebellaris I, a few of whose fibres reach 
the molecular substance of the keel in the caudal end of the body. 
Here, however, the sharp limit between the molecular and granular 
layers is lost. In the case of the other tracts from the mid and 
tween brain, all fibres break up in the granular layer. In the case 
of the VIII, lateral line and V root fibres the branching is simple 
and never passes beyond the layer of PURKINJE cells. This is what is 
to be expected from the origin of the cerebellum and its adult re- 
lations to the acusticum. As I have pointed out (98b, p. 600) 
the acusticum with the cerebellar crest is strictly comparable with the 
cerebellum. That VIII or lateral line root fibres should end in the 
molecular layer of the cerebellum is as little to be expected as that 
they should end in the cerebellar crest. If it be true that the as- 
cending fibres in Teleosts end in the molecular layer, there must be 
a great difference in the functioning of the several nerve elements 
in the cerebellum of Teleosts and Ganoids. In Acipenser, as I interpret 
the structure, in-coming impulses arrive in the granular layer, where 
they are received by the granule cells and by the cells of the II type. 
The latter serve to spread the impulses more widely among the granule 
cells. These in turn send the impulses along their very slender neurites 
to the molecular layer, where the dendrites of the PURKINJE cells are 
especially developed to receive impressions from the myriads of fine 
fibres as they course among these dendrites. In the molecular layer 
