162 PROFESSOR MARSHALL. 



are to be found in any of the cranial nerves roots strictly comparable 

 to the anterior roots of the spinal nerves. 



In dealing with this question, it is first necessary to establish cer- 

 tain general conclusions concerning the cranio-spinal nerves. As was 

 first pointed out by Balfour, the posterior roots of the spinal nerves 

 must be regarded as of a more primitive nature than the anterior 

 roots, the grounds on which this conclusion is based being the fol- 

 lowing : — 



1. The actual mode of development of the two kinds of roots in the 

 spinal nerves. As noticed in a previous page, the posterior roots 

 appear before the anterior ones, and are also in their mode of develop- 

 ment of a more primitive character than these latter, the posterior 

 roots consisting at first entirely of undifferentiated spherical or poly- 

 gonal cells, while the anterior roots are almost from their first ap- 

 pearance fibrillar. 



2. The condition of the nervous system in Amphioxus, in which, 

 according to Balfour, all the nerves arise by single roots, which 

 roots correspond to the dorsal or posterior roots of other vertebrates, 

 and must clearly in Amphioxus be of mixed motor and sensory 

 function. 



From these facts the further conclusion is drawn " that primitively 

 the cranio-spinal nerves of vertebrates were nerves of mixed function 

 with one root only, and that root a dorsal one ; and that the present 

 anterior or ventral root is a secondary acquisition." 1 



Concerning the several steps by which these anterior roots have 

 been acquired, the evidence at our disposal is of an imperfect, and in 

 great part merely conjectural character. Still I think that, although 

 we may not be able to solve the problem completely, we can at any 

 rate define its limits fairly accurately, and perhaps indicate the path 

 along which the solution will ultimately be found. 



The problem is how, from animals resembling AmpMoxus, in pos- 

 sessing only dorsal roots to the nerves, and these dorsal roots con- 

 sequently of mixed function, has the type of spinal nerve met with 

 among existing vertebrates, with two distinct roots, dorsal or sensory 

 and ventral or motor, been derived 1 



It appears to me that there are two ways in which we can conceive 

 this change as having come about : — 



1 Balfour, Elasmobranch Fishes, p. 193. 



