222 



substance. The neuntes of the anterior bilaterally placed cells at 

 first run caudad near the surface of the cord for some distance, then 

 curving ventrad, laterad, and finally cephalad pass forward through 

 the fibre bundle. The neurite (unpaired ones) having passed ventrad 

 through one -half or two -thirds the distance to the caualis centralis 

 curves gradually either to the right or to the left, sometimes dividing 

 in its course, and finally enters, or sends one of its two branches into, 

 the lateral nerve bundle, which is made up of similar fibres. The 

 neurites in general pass alternately to the right and left, but this does 

 not hold strictly, for sometimes several successive cells send their 

 neurites to the same side. 



Entering the lateral bundle the neurite may pass either cephalad 

 <>r caudad 1 ) (Figs. 1,2). In examining upward of three hundred cells 

 from Ctenolabrus in which the neurite was 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 fibres in the bundle at 

 different parts of its course, which show that the majority of the 

 neurites run cephalad. 



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

 neurite was found to give rise to two axis cylinders of equal diameter. 

 In the other third no such branching could be seen. The failure to 

 observe branching may sometimes have been due to the imperfection 

 of the preparations, but in a few instances at least it was evident 

 that the neurites did not divide. 



There are interesting variations in the manner of this division. 

 In the most common type (Figs. 3, 4, a) the neurite passes ventrad 

 and laterad nearly to the level of the bundle and then parts into two 

 equal axis cylinders, of which at least one enters the bundle and 

 passes through its entire course. But the division may take place 

 more dorsally, near the cell, the two branches diverging as they pass 

 ventrad (Figs 4 c, 9), or it may occur still more dorsally, so that the 

 two processes come directly from the cell (Figs. 4, d, 10). In a few 

 cases the division was observed to take place after the neurite had 

 entered the bundle, the two resulting processes continuing to run par- 

 allel for some distance. 



1) Dahlgken ('97) finds that in the order Heterosomata the neu- 

 rites all run caudad. 



