The brain of Acipenser. 231 
end freely without forming a network. Each mitral cell supplies 
only one glomerulus, occasionally two. The neurites traverse the 
granular layer and turn backward. Collaterals are present on their 
whole course, or at least on the horizontal part. Usually each 
glomerulus receives in addition to the dendrite from one mitral cell 
several dendrites from larger or smaller Pinselzellen. These are 
situated either superficial to the glomeruli or among and beneath 
them. Their neurites take a centripetal course and give off collaterals 
just beneath the mitral cells. Beneath the mitral cells is a layer of 
fibres, granules, and cells of the II type. The granule cells are 
small, oval, pyramidal, or fusiform, and have their long axis with 
their chief processes placed vertically. Their character is in dispute. 
S. Ramon (90) considers them nervous, the peripheral process serving 
as neurite. VAN GEHUCHTEN holds them in doubt, denying that 
the peripheral process is a neurite, and being unable to find a 
neurite. KOLLIKER considers them certainly neuroglia. The peri- 
pheral processes are covered with spines, and according to KOLLIKER 
break up beneath the mitral cells, but according to S. Ramon and 
VAN GEHUCHTEN & MARTIN (91) among the small Pinselzellen just 
beneath the glomeruli. The granules have a few small central pro- 
cesses. The stellate cells of the II type have neurites ending out- 
side of the mitral cells. 
Comparing these results it is to be noted that in all important 
respects the relations of the olfactory fibres, glomeruli, and large 
and small mitral cells are identical in all Vertebrates. The presence 
of II type cells in the deep layers has been noted only in Mammals, 
and it is not very probable that the cells of the II type in Aci- 
penser are the same as these. This is, however, a minor point and 
the main question in the interpretation of the olfactory lobe is con- 
cerning the character and homology of the numerous cells lying 
beneath the mitral cell layer, present in all Vertebrates in great 
numbers, and presenting constant anatomical characters at least from 
the Amphibia onward. P. Ramon has very properly pointed out 
that in spite of the apparent absence of neurites on these cells, they 
must be important factors in the constitution of the olfactory lobe. 
In my preliminary paper (98a) I made a strong point of the 
evidence for the nervous character of the granule cells in Acipenser, 
being at that time inclined to homologize them alone with the 
granule cells in Mammals. This would leave no means of accounting 
for the large number of other cells in the granular zone, a difficulty 
