No. 4.] 



AMPULLAE OF THE S EL AC H II. 



171 



upward for a short distance, then divide more or less dichoto- 

 mously, sending out branches at right angles to their former 

 course (Fig. 6, n.ax.cyl.}. By the interlacing of these axis 

 cylinders a plexus is formed, occupying nearly the whole area 

 of the centrum cap. All the best results of previous workers 

 were obtained by the use of osmic acid. Since this reagent 

 differentiates the medullary sheath only, their observations 

 could be carried only to the point where this sheath disappears. 

 The fibers were then pursuing a perpendicular course, and the 

 assertion was but natural that this course was continued for a 



a.c.c. 



71. 77?. 



FIG. 6. Nerve fibers i.i lotrjitudinal section just below centrum cap : a.c.c., ampulla centrum cap ; 

 n.ax.cyl., axis cylinder of nerve ; ti.i.sli., medullary sheath of nerve, x 465. 



very short distance, the fibers then coming into connection with 

 the cells at the top of the centrum. As we have seen, this is 

 not at all true. From the plexus already described bundles of 

 axis cylinders collect at the periphery of the centrum cap, and 

 pass out into the connective tissue of the partitions which sep- 

 arate the ampullary pockets. Thus far the nerves have, so to 

 speak, avoided all relations of contact or of continuity with 

 cells. 



Having passed into the region where the double layer of cells 

 is found, the nerve fibrils begin to divide, sending out branches 

 which play over the bases of the deeper layer of cells (Fig. 8, 

 n.ax.cyl.}. The manner of giving off branches is interesting. 

 In well-stained specimens (Fig. 7) the larger axis cylinders are 



