DIFFERENTIATIONS OF ECTODERM IN NECTURUS. 9305 
the Selachii, is equally applicable to Necturus. Nevertheless 
we frequently hear that nerves are not primarily cellular in 
their structure, but arise always as fibrous outgrowths from 
the peripheral ganglion or central nervous system. In reply 
to Dohrn, His tells us (20, p. 342) that no thoughtful investi- 
gator will assume that sharks can differ from other Vertebrates 
fundamentally in the manner in which their nerves are formed ; 
and that when it can be shown that in higher Vertebrates, such 
as Amphibia and bony fishes, the motor nerve-fibre arises as a 
thread-like outgrowth from particular spinal cells, the state- 
ment must (?) be valid for ray and shark. 
It may, therefore, not be out of place to add a few words in 
regard to the formation of the nerves in a primitive Amphibian ; 
for if the Cyclostomata, Selachii, and Amphibia are shown 
to agree in the cellular structure of their nerves, motor as 
weli as sensory, the “ thoughtful investigator” might perhaps 
conclude that the development of the human nervous system 
differs from that of the shark and tadpole. 
All of the nerves which I have described in Necturus, 
whether motor or sensory, cranial or spinal, are formed by 
the continued migration of modified ectoderm cells, either 
from the neural crest, the closed medullary tube, the peripheral 
ganglia, or the superficial ectoderm. The manner of their 
formation, however, is not always the same, and nerves of 
different orders vary greatly in the number of cellular elements 
they contain. 
In the development of the embryo we may distinguish two 
factors, known to philosophy as efficient cause and final cause. 
Efficient cause leads to that unfolding of the embryo by which 
each stage grows out of the preceding one in the sequence of 
its original development—its phylogeny modified by the present 
changed surroundings. Final cause, however, stamps mysteri- 
ously on the younger stage the image of a later form; and 
thus we find throughout development rudiments of later 
structures, which can be neither of immediate use to the de- 
veloping embryo nor yet associated with the ancestral type its 
present form repeats. To disentangle these factors is the 
