192 Journal of Comparative Neurology. 



hold strictly, sometimes several successive cells sending their 

 neurites to the same side. 



Entering the lateral bundle the neurite may pass either 

 cephalad or caudad (Figs, i and 2). Dahlgren '97 finds that 

 in the order Heterosomata the neurites all run caudad. In exam- 

 ining upward of three hundred cells in which the neurites were 

 followed into the bundle, approximately one-third were found 

 to send the neurite through the bundle caudad, the other two- 

 thirds cephalad. This harmonizes with results obtained by 

 counting the number of fibers in the bundle at different parts 

 of its course, which shows that the majority of the neurites run 

 cephalad. 



In approximately two-thirds of the number of cells exam- 

 ined the neurite was found to divide into axis cylinders of equal 

 diameter. In the other one-third no such branching could be 

 seen. This may sometimes have been due to the imperfection 

 of the preparations, but in a fev instances at least it would seem 

 that the neurites do not divide. 



There is a remarkable and interesting variation in the man- 

 ner of this division. In the most common type (Fig. 4 b) the 

 neurite passes ventrad and laterad nearly to the level of the 

 bundle and then splits into two equal axis cylinders at least one 

 of which enters the bundle and passes through its entire course. 

 The division may take place higher up near to the cell, the two 

 branches diverging as they pass downward (Fig. 4,<: and e). 

 Or the division may occur so high up that the two processes 

 come directly from the cell (Fig. 4,^). In a few cases the di- 

 vision was observed to take place after the neurite had entered 

 the bundle, the two resulting processes continuing parallel for 

 some distance. 



The axis cylinder stains deeply with iron hematoxylin, 

 Kenyon's or Sahli's method, and is uniformly stained through- 

 out its length. Frequently however the initial part of the fiber 

 immediately adjacent to the cell takes the stain but lightly, the 

 protoplasm of that part of the neurite staining precisely like 

 the cytoplasm of the cell, with which it seems to be continuous 

 and identical. The axis cylinders are throughout their course 



