NEAL : NEKVOUS SYSTEM IX SQUALUS ACANTHIAS. 225 



to determine that other processes, instead of leaving the medullary wall, 

 extend posteriorly in the wall and parallel with it to form the ventral 

 fibre tract. The nerve process (Figure G) shows a differentiation in its 

 distal portion into tsvo deeply staining fibrils surrounded by more faintly 

 staining plasma, the two fibrils dividing distally into three, which enter 

 the fine processes with which the nerve ends. The nerve process on the 

 other side of the same embryo (Figure H) does not, however, show this 

 same evidence of histological differentiation. Here the nerve appears as 

 a highly refractive fibril, and, while having a greater extent than that of 

 its mate of the opposite side, is composed, except at its root, of a single 

 undivided fibril. The connection of this fibril with the axis-cylinder 

 process from a neuroblast cell in the ventral horn seems indisputable, 

 since this passes directly through the limiting membrane at the base of 

 the brain wall, and projects into the shrinkage space directly opposite 

 the chief root of the nerve, as is shown at ax-cyl. I have no evidence 

 to offer, such as that stated by His ('88, '89), for Mammals and other 

 Vertebrates, of a migration of the neuroblasts from the " inner layer " 

 of medullary cells, nor do I find any evidence of migration of cells from 

 the neural tube, as stated by Dohrn ('01). I find at this stage neither 

 nuclei connected with the roots of the nerve outside the neural tube, 

 nor such as are half in and half out of the tube. 



* The connection of the oculomotorius with cells of the mesocephalic 

 ganglion is attained very quickly, and in embryos of 54 or 55 somites has 

 already taken place. At this stage of development, as seen in embryos 

 fixed with the corrosive sublimate-acetic mixture (Davidofifs fluid), the 

 nerve appears (Plate 8, Fig. 58) as a cellular strand, which extends from 

 the inner side of the mesocephalic ganglion toward the ventral wall of 

 the midbrain, with which the nerve unites by at least two main roots. 

 To detect the proximal roots as well as the relations of these with 

 medullary cells, sagittal sections are much more favorable than frontal, 

 since the nerve roots are situated one behind the other.^ The fact that 

 the nerve is several cells in thickness near the ganglion, while its calibre 

 diminishes as it passes toward the brain wall, Avould naturally, if one 

 were unacquainted with the conditions shown in the embryo of 52 

 somites, lead to the inference that the growth of this nerve takes place 

 from the ganglion toward the brain (vide Miss Piatt, '91, Mitrophanow, 



^ Also, for the reason already stated by His ('88*, p. 344) for spinal ventral 

 nerves, that "die Wurzelbiindel treten in grijsseren Abstanden aus dem Riicken- 

 raark hervor. Jedes Biindel bezieht seine Fasern aus einem entsprechend brelten 

 Bezirk des Riickenmarks. Die Sammlung derselben erfolgt zum Theil noch inner- 

 halb des Markes, zum Theil erst in der Leibeswand." 

 VOL. XXXI. — NO. 7. 6 



