120 Journal of Comparative Neurology. 



after gently heating to stiften the balsam. The latter process, 

 recommended by Huber, seems non-essential if the paraffin im- 

 bedding is used. We have not experimented a sufficiently long 

 time to be sure that no change will be affected, but sections 

 prepared without any more heating than that necessary to 

 coagulate the fixative have remained unchanged for over three 

 years. The results obtained compare favorably with anything 

 we have seen and it is quite possible to cut serially or by a rib- 

 bon process if desired. Naturally the specimen must be small 

 if thorough permeation is to take place. 



In the olfactoty tuber a fair number of the cells are perme- 

 ated. There seems to be no essential difference between the 

 more superficial and the deeper ganglion cells. The peripheral 

 processes pass to the glomerules and there subdivide into a 

 close brush which is in close contact with similar fibres from 

 the olfactory fibres. From the base a long practically un- 

 branched fibre (presumably the axis cylinder) can be traced 

 toward the cerebrum. The protoplasmic processes of the 

 deeper cells pass between the peripheral cells and subdivide 

 similarly. 



The relations described are illustrated by Figs, i-^ of Plate 

 XIX. Fig. I. is a section from the ventricle to the glomerule 

 layer showing the general distribution of cells and processes. 

 Fig. 2. shows two of the peripheral ganglion cells on a larger 

 scale with their protoplasmic processes entering a glomerule. 

 Fig. J. shows the minute subdivision of the protoplasmic pro- 

 cesses in the glomerules. 



The structure of the cerebrum is not greatly simplified by 

 a resort to the Golgi method. Everywhere evidence is afforded 

 that the cells are abundantly supplied with processes but the 

 clear and unambiguous distinction claimed by many writers be- 

 tween protoplasmic and axis cylinders is not obvious to the 

 writer in all or even the majority of cases. The appearances 

 seem to favor the interpretation of an obliquely peripheral pro- 

 cess as the axis cylinder in the mid-dorsal cortex {Fig. 4. Plate 

 XIX.) This is certainly not the case in the hippocampal region, 

 however, where the unambiguous evidence of the haematoxylin 



