THE DEVELOPMENT OF THE VENTRAL NERVES IN SELACHII. 295 



observations of the early stages of development are necessary to prove the truth of 

 the assumption, which is by no means self-evident. The discovery by von Lenhossek 

 through the Golgi method of a motor fibre showing the terminal swelling charac- 

 teristic of a growing fibre in a 20-millimetre selachian embryo does not exclude the 

 possibility that the "cell-chain" mode of nerve development may also occur in 

 ontogeny. It is but just that those who differ from Dohrn, Beard, and Raffaele 

 should, before seeking to discredit the results of these investigators on the ground 

 that the methods used are unsuitable for the special purpose of nerve study, make 

 a more extended investigation of nerve histogenesis in the Anamnia than they have 

 done. Beard's request ('96, p. 395) for an extended work upon selachian-nerve his- 

 togenesis, with at least a single plate of figures in disproof of the cell-chain hypothesis, 

 has not yet been granted. 



Instead of criticism of technical methods and a priori reasoning, renewed 

 investigations by the best methods applicable, covering all stages of nerve develop- 

 ment, from first appearance until adult conditions are attained, are needed. In this 

 conviction the present investigation was undertaken. While not all the problems 

 of nerve histogenesis come within the scope of the present paper, which covers only 

 enough of the general problem to afford a sufficient basis for comparison with the 

 histogenesis of the eye-muscle nerves, positive conclusions concerning some of the 

 more important questions have been reached. These are: 



1. The neuraxones of spinal ventral nerves of selachians develop like those of 

 Amniota as processes of neuroblast cells. In their growth they are secondarily surrounded 

 by sheath cells. 



2. Medullary cells, but not those which form the neuraxones, migrate into the ventral 

 nerves in early stages of development. Their presence in the embryonic ventral nerves 

 gives rise to the "cellular structure" in these early stages of development and thus 

 obscure, in preparations made by the conventional embryological methods, their 

 fibrillar structure. Suitable methods show that neuraxones are present in the earliest 

 as in all later stages. 



3. The migrant medullary cells form the neurilemma sheaths, but take no part in 

 the formation of the neuraxones or ganglia of ventral nerves. 



4. The epineurium and perineurium sheaths are in chief part added to the embryonic 

 nerve from the adjacent mesenchyme. This addition of mesenchyme cells, however, 

 occurs in somewhat advanced stages of development. 



These results confirm generally accepted opinions except as regards the question 

 of the migration of medullary cells and the participation of these in the formation 

 of nerve sheaths, which have usually been regarded as of mesenchymatous origin. 



