THE DEVELOPMENT OF THE VENTRAL NERVES IN SELACHII. 303 



larity in the development of dorsal and ventral nerves, in as much as both were com- 

 posed of cells ectodermal in origin. Hoffmann ('97) and Harrison (:01)* confirmed 

 Dohrn's first opinion that medullary cells wander from the neural tube into spinal 

 ventral nerves. Thus while observers agree that embryonic selachian nerves contain 

 cells, the source of these cells is in dispute. Similarly in doubt is the question of 

 the part that these cells play in nerve histogenesis. 



c. Are the ventral nerves ab initio connected with their terminal organ and how 

 intimate is this connection? All observers are agreed that from their first appearance 

 spinal ventral nerves are connected with the myotomes. The only question under 

 discussion is that of the intimacy of this connection. Hensen ('64) and Sedgwick 

 ('94) have maintained that nerve fibres are in immediate protoplasmic continuity 

 with their terminal organ. A similar intimate cell to cell continuity seems involved 

 in the "fibrillar theory" of Apdthy. Bardeen ('98), on the other hand, states that it 

 is only comparatively late in embryonic development that terminal fibres enter into 

 intimate relations with the tissue elements of the part they are to supply. Barker 

 ('99, p. 193) also implies that this is true, for, according to him, peripheral nerves, 

 like blood-vessels, follow in their growth the channels of least resistance. 



B. HISTOGENESIS OF NETJRAXON AND NEURILEMMA IN SQUALUS. a. Is the 

 neuraxon multicellular in origin or the process of a single medullary cell in 

 selachians? The following evidence seems to me to favor the process theory and to 

 confirm the conclusions of His and von Lenhoss6k: 



First, in their earliest stages of development (PI. XXII, Figs. 1-6) spinal ventral 

 nerves are made up of processes from medullary cells, and are wholly devoid of 

 nuclei. 



Second, in all later stages of development after nuclei have made their appear- 

 ance in the nerve, the processes of the medullary cells can be traced into the fibrillar 

 portion of the forming nerve (PI. XXII, Figs. 9-11; PI. XXIII, Figs. 12-16). 



Third, the number of neuroblasts whose neuraxones can be traced from each 

 neuromere into the forming nerve corresponds closely with the estimated number of 

 the neuraxones in the nerve. As the number of neuraxones increases in the 

 nerve the number of neuroblasts whose neuraxones can be traced to the nerve 

 increases. 



These facts, however, do not prove that all of the neuraxones in the nerve are 

 formed as processes from medullary cells. They only make it probable that most 

 of the neuraxones are such processes. The possibility still remains that some of the 



* Harrison (:01) finds that, in the salmon, ventral nerves are formed as processes of neuroblast cells in the 

 ventr al half of the neural tube. In much later stages of development, nuclei migrate out of the neural tube 

 into the nerve to form motor elements in the sympathetic ganglia. 



